SCREENING
Criteria for a screening program to benefit society (WHO Monograph. 1968. Wilson JMG and Junger G)
The condition should be an important health problem.
There should be an accepted treatment for patients with recognized disease.
Facilities for diagnosis and treatment should be available.
There should be a recognizable latent or early symptomatic stage.
There should be a suitable test or examination (i.e. safe, rapid, inexpensive, relatively easy to apply, reproducible, and has a high sensitivity, some specificity).
The test should be acceptable to the population.
The natural history of the condition, including development from latent to declared disease, should be adequately understood.
There should be an agreed policy on whom to treat as patients (and treatment for early stage disease should reduce morbidity and mortality).
The cost of case-finding should be economically balanced in relation to possible expenditure as a whole.
Case-finding should be a continuing process and not a "once and for all" project.
Screening in prevention
Primary prevention of disease involves elimination of risk factors (i.e. smoking cessation)
Secondary prevention involves early detection of disease - most screening falls into this category
Tertiary prevention involves a reduction in complications of an already established disease
Potential harms of screening
Cost, time, discomfort
Complications of the screening test
Complications of the intervention
False positives
Psychological harm
Complications of additional testing
Screening guidelines
Many guidelines for screening issued by the American Cancer Society (ACS) and subspecialty societies are not based on evidence that screening improves morbidity or mortality.
The U.S. Preventative Services Task Force (USPSTF) was established by the U.S. Public Health Service in 1984 for the purpose of developing evidence-based clinical policies for preventative care.
Created in 1984
This task force was modeled after the Canadian Task Force.
JAMA published 12 background papers from 1987-1990.
Work of the first USPSTF culminated in 1989 with publication of Guide to Clinical Preventive Services, first edition.
The second task force was created in 1990, and published its first paper in JAMA in 1993.
Work of the second USPSTF culminated in 1996 with publication of Guide to Clinical Preventive Services, second edition.
The third task force was created in 1998 by the Agency for Health Care Research and Quality (AHRQ). The systematic reviews that that support Task Force recommendations are now conducted by AHRQ supported Evidence-based Practice Centers (EPC's) at Oregon Health Sciences University in Portland, Oregon and at Research Triangle Institute inResearch Triangle Park, North Carolina. USPSTF then crafts final recommendations based on the quality of the evidence and the relative balance of benefits and harms. The task force includes representatives from family medicine, internal medicine, pediatrics, obstetrics and gynecology, geriatrics, preventive medicine, public health, behavioral medicine, and nursing. Recommendations will be posted at www.ahrq.gov/clinic/uspstfix.htm and available in print through the AHRQ Clearinghouse at 800-358-9295.
USPSTF does not directly examine the cost of implementing guidelines.
U.S. Public Health Service is responding to this limitation by creating a cost-effectiveness panel on clinical preventative services.
2010 Affordable Care Legislation (ACA) included waivers of Medicare copayments and deductibles for screening services with a USPSTF Grade A or B recommendation
Estimated cost of screening (based on scientific modeling)
FOBT screen: $20,000 - $40,000/QALY (quality adjusted life year).
Combination of FOBT and sigmoidoscopy screening: $50,000/QALY.
Mammography: $15,000/QALY in women age 50-70; $105,000/QALY in women age 40-49
Hypertension in females: $44,000/QALY at age 20; $12,000/QALY at age 60
Hypertension in males: $29,000/QALY at age 20; $8,000/QALY at age 60
Smoking cessation in males aged 35-69: $1,000/QALY with brief physician counseling, $6,000/QALY with counseling and nicorette gum.
Screening for asymptomatic carotid artery disease: $120,000/QALY
Drug treatment of hypercholesterolemia in 20-30 year olds based on NCEP I guidelines: $1,000,000/QALY.
Drug treatment of hypercholesterolemia: $34,640/year of life saved in the West of Scotland study for primary prevention in middle aged men with increased risk of coronary artery disease.
NOTE: screening may be a good way to use public resources, but screening generally does not save money.
Breast cancer screening
268,600 cases/year; 41,760 deaths/year [2019 data] 10% lifetime risk; 3% mortality
Modalities of screening
Breast self-exam (BSE)
In 2009, the USPSTF recommends against clinicians teaching women how to perform breast self-examination (Ann Intern Med. 2009. 151. 716-726).
In 2002, the USPSTF found insufficient information to recommend for or against teaching or performing breast self-exam.
A randomized trial of breast self-exam in Shanghai, China in which 132,979 female factory workers were randomized to a group in which careful BSE instruction was given at baseline and reinforced 1 and 3 years later, and another 133,085 women were randomized to a control group, showed no difference in mortality from breast cancer at 10 years (J Natl Cancer Inst. 2002. 94. 1445-1457).
Breast awareness (i.e. awareness of changes in a breast and bringing this to the attention of a doctor) is VERY LIKELY important in reducing mortality from breast cancer.
Clinical breast exam:
In 2009, the USPSTF found that the current information is insufficient to assess the additional benefits and harms of clinical breast examination beyond screening mammography in women 40 years or older (Ann Intern Med. 2009. 151. 716-726).
In 2015, the American Cancer Society (ACS) does not recommend clinical breast exam for screening amongst average risk women; this is a qualified recommendation, based on very low quality of evidence (JAMA. 2015. 314. 1599-1614 and editorial 1569-1571).
A Canadian study (NBSS2) which randomized 40,000 women showed that mortality reduction with clinical breast exam alone is equal to the mortality reduction with mammography at 7 years!
False positives are the main downside to screening clinical breast exam. A 10 year retrospective study in which 10,905 screening breast exams were done, with an average of 5 clinical breast exams per decade in women aged 40-69, showed a 13.4% false positive rate per decade for screening clinical breast exam (N Engl J Med. 1998. 338. 1089-1096).
Mammography:
Screening guidelines for mammography are detailed in the section below of this outline (entitled “Recommendations for screening mammography”), along with risks and benefits of mammography screening, by age cohort of 40-49 years old, 50-69 years old, and > 70 years old.
Benefit: in order to prevent one death from breast cancer at 10 years, one must screen 1250 women aged 50-59 years, 476 women aged 60-69 years, and 769 women aged 70-74 years (USPSTF recommendation statement. Ann Intern Med. 2016. 164. 279-296).
Historically it was estimated that mammography had a 64-93% sensitivity and 90-95% specificity for detection of breast cancer in 50-69 year old women – more recent data indicates a lower sensitivity.
In a study of 42,760 average risk women who participated in the Digital Mammographic Imaging Screening trial, sensitivity was 55% (N Engl J Med. 2005. 353. 1773-1783).
In a study of 2725 high risk women (dense breast tissue, half previously treated for breast cancer and the other half with more than one family member diagnosed with breast cancer), the sensitivity of mammography was only 49% (JAMA. 2008. 299. 2151-2163 and editorial 2203-2205).
In BRCA1 mutation carriers, sensitivity can be as low as 25% (JAMA. 2004. 292. 1317-1325; Lancet. 2005. 365. 1769-1778).
Annual vs. biennial mammography screening – virtually no direct data (i.e. randomized trial comparing annual with biennial screening); a paper reporting results using 6 separate models to predict risk versus benefits of annual versus biennial mammography screening concluded that screening biennially maintained an average of 81% of the benefit of annual screening (range across 6 models of 67-99%) with almost half the number of false positive results (Ann Intern Med. 2009. 151. 738-747).
One vs. two view mammography screening - even though standard of care is most often two view mammography, one view mammography screening may be adequate; one versus two view mammography has not been adequately studied.
Radiation exposure from screen film mammography with modern dedicated mammography equipment is only 0.05 - 0.2 rad (exposure from a PA and lateral CXR is 0.04 rad, and yearly environmental exposure to radiation is approximately 0.1 rad).
It is estimated that the radiation exposure from 10 annual mammographic screenings in 10,000 women will cause only one additional breast cancer (Cancer. 1996. 77. 818-822).
Turmeric (Curcumae longae) may protect against radiation-induced damage caused by mammography – suggested dose is 500 mg three times a day of a 5:1 aqueous extract for a course of 21 days, beginning 7 days before imaging is scheduled (Alt Med Alert. 2008. 11. 64-68). The study cited (Adv Exp Med Biol. 2007. 595. 301-320) reports that this herb paradoxically protects normal cells from radiation while sensitizing cancer cells to the radiation. Ginkgo may also offer protection against radiation.
Discomfort associated with mammography can be reduced by applying 4% lidocaine gel to the skin of the breasts and chest wall one hour before the exam (Radiology. 2008. 248. 765-772).
Women with mammography screening have approximately 25% fewer metastases to regional nodes at time of diagnosis (provocative data).
Some of apparent benefit of mammography may be due to lead time bias, estimated to be as much as 3.5 years.
False positives associated with mammography are significant. See more info on false positives in the section below on Recommendations for screening mammography
Nationally, an average of 11% of mammograms are read as abnormal, and necessitate further diagnostic workup.
Breast cancer is found in only approximately 3% of women with an abnormal mammogram (Am J Roentgenol. 1995. 165. 1373-1377).
A 10 year retrospective study in which 9762 screening mammograms were performed, with a median of 4 screening mammograms per decade in women aged 40-69, showed a 23.8% false positive rate per decade for screening mammography. This translates to a 49% risk of a false positive mammogram after 10 mammograms. This same study estimated that after 10 mammograms, there is a 19% risk of a needle or open biopsy (N Engl J Med. 1998. 338. 1089-1096).
A table (N Engl J Med. 2003. 348. 1677) provides a very useful visual regarding the chance of a false positive mammogram, need for biopsy, and development of breast cancer among 1000 women aged 40-49, 50-59, and 60-69 who undergo annual screening mammography for 10 years.
Hormone replacement therapy increases breast density in some women, and decreases both the sensitivity and the specificity of mammography screening for breast cancer (J NatlCancer Inst. 1996. 88. 643-649; Ann Intern Med. 2003. 138. 168-175).
Analysis of 100,622 screening mammograms found that obese women had a 20% increased risk having a false positive result, compared with normal weight and underweight women. Sensitivity was unchanged in obese women (Arch Intern Med. 2004. 164. 1140-1148).
Mammography results in overdiagnosis of breast cancers (i.e. diagnosis in women who would have lived their lives and died of another cause without ever knowing they had breast cancer). See more info on false positives in the section below on Recommendations for screening mammography
Analysis of data from the Malmo mammographic screening trial (42,283 women enrolled over 10 years, with follow-up for another 15 years) showed that even though during the follow-up period women in both groups had mammography screening, there continued to be more diagnoses of breast cancer in the screened group. The failure of women initially in the non-screened group to ‘catch up’ with regard to new diagnoses of breast cancer implies overdiagnosis of breast cancer. It is estimated from this study that 10% of women diagnosed with breast cancer by mammography were overdiagnosed (BMJ. 2006. 332. 689-692).
Analysis of data from the Norwegian Breast Cancer Screening Program, in which the incidence of invasive breast cancer in 1996-2005 in counties in which the screening program was implemented was compared with the incidence in which the program was not yet implemented, concluded that 15-25% of cases of cancer are overdiagnosed. The nationwide Cancer Registry of Norway is nearly 100% complete, and 39,888 patients with invasive breast cancer were included in this analysis (Ann Intern Med. 2012. 156. 491-499). Overdiagnosis may occur more often in the US than in Norway, as US radiologists are more likely to read a mammogram as abnormal, and most women in the US start screening at a younger age, and are screened more often (Editorial. Ann Intern Med. 2012. 156 536-537).
Mammography has led to the diagnosis of many more cases of ductal carcinoma in situ, but it is unclear whether or not finding this cancer saves lives, because while on the one hand we know that ductal carcinoma in situ can progress to invasive cancer, on the other hand autopsy studies in women who have died from other causes show a substantial number of ductal carcinoma in situ.
Mammography with computer-aided detection
On the business side, startup cost is higher, but Medicare reimbursement is approximately $20 higher per mammogram interpreted (Editorial. N Engl J Med. 2007. 356. 1464-1466).
Data on 222,135 women and 429,345 mammograms at 43 centers found that computer-aided detection is associated with reduced accuracy of detection of breast cancer – there was no significant increase in the cancer detection rate, but there was an increase in false positives (N Engl J Med. 2007. 356. 1399-1409).
A retrospective cohort study of SEER data in 409,459 mammograms in 163,099 women showed that computer-aided detection in Medicare enrollees aged 67-89 “is associated with increased DCIS incidence, the diagnosis of invasive breast cancer at earlier stages, and increased diagnostic testing in women without breast cancer.” The effect on breast cancer mortality could not be assessed (Ann Intern Med. 2013. 158. 580-587).
Digital mammography
In 2009, the USPSTF found that the current information is insufficient to assess the additional benefits and harms of screening digital mammography instead of film mammography (Ann Intern Med. 2009. 151. 716-726).
Overall diagnostic accuracy of digital and film mammography are similar (N Engl J Med. 2005. 353. 1773-1783), but the diagnostic accuracy of digital mammography is superior in women younger than age 50, as well as for those with dense breasts (and less accurate in women over age 65), based on data gathered in 42,760 women who had both digital and film mammography in random order at the same screening encounter, in the DMIST trial (Radiology. 2008. 246. 376-383).
In the Oslo II trial in which women aged 45-69 were randomized to undergo digital versus film-screen mammography, the cancer detection rate was higher with digital mammography, but specificity was lower (Radiology. 2007. 244. 708-717).
In a population based screening program in Spain, digital mammography was associated with a higher recall rate and similar cancer detection rates, as compared with film-screen mammography (Radiology. 2009. 252. 31-39).
In the UK breast cancer screening program, there was no difference in recall rates or cancer detection between digital mammography and film-screen mammography (Radiology. 2009. 251. 347-358).
Prospective observational data gathered on 329, 261 women (231,034 digital mammograms and 638,252 film-screen mammograms) aged 40-79 in mammography facilities in the NCI-funded Breast Cancer Surveillance Consortium show that (1) in women age 50-79, cancer detection is similar with film-screen screening as compared with digital screening (2) digital screening may be superior in women age 40-49, presumably related to increased breast density in these younger women, and also a higher prevalence of estrogen receptor negative cancer in younger women, and (3) digital screening may be superior at detecting estrogen receptor negative cancer screening (Ann Intern Med. 2011. 155. 493-502). Based on greater number of false positive digital mammograms, greater detection of in situ tumors of unknown clinical significance, and no evidence of breast cancer death with digital mammography, conclude digital mammography may actually reduce the efficacy of breast cancer (Editorial. Ann Intern Med. 2011. 155. 554-555).
Economic modeling shows that screening by using all-digital mammography is not cost effective; the cost is $331,000 per QALY gained relative to all-film mammography. Targeted digital mammography for women less than age 50 costs $26,500 per QALY gained relative to film mammography; targeted age and breast density digital mammography costs $84,500 per QALY gained relative to film mammography (Ann Intern Med. 2008. 148. 1-10).
On the business side, startup costs are higher with a digital mammography unit ($500,000 vs. $150,000 for a conventional unit), but Medicare reimbursement is approximately $50 higher per mammogram interpreted (Editorial. N Engl J Med. 2007. 356. 1464-1466).
Digital breast tomosynthesis (DBT)
This is a newer mammographic technique (approved by the FDA in 2011) which was developed to improve upon limitations of traditional digital mammography. Specifically, as compared with digital mammography, DBT improves sensitivity and reduces the number of call backs (especially in those with very dense breast tissue) for what are ultimately determined to be false positive mammograms. As per an editorial (JAMA. 2022. 327. 2198-2200), multiple studies have confirmed the benefit of improved sensitivity of DBT (citations provided in the editorial).
Prevalence in 2022, based on FDA data - DBT is offered by 83% of breast imaging facilities in the US, and offered in 45% of all mammography units
Acquisition time (to obtain all images) and radiation exposure - if synthetic 2D mammogram images are generated from the DBT image data, the acquisition time plus radiation exposure from DBT + synthetic mammography is similar to that of digital mammography alone, but if the facility performs a combination of digital mammography and DBT, the acquisition time plus radiation exposure is approximately twice that associated with DBT alone (Editorial. JAMA. 2022. 327. 2198-2200).
Cost - in the vast majority of facilities in the US, in 2022 there is no difference in cost to the patient between digital mammography with DBT versus DBT alone (Editorial. JAMA. 2022. 327. 2198-2200).
The USPSTF concludes in 2016 that current evidence is insufficient to assess the benefits and harms of digital breast tomosynthesis as a primary screening method for breast cancer (Ann Intern Med. 2016. 164. 279-296).
Benefit pertaining to risk of interval invasive cancer (defined as cancer detected in between screening studies) and risk of advanced cancer - in a cohort of 504,427 women at 44 US Breast Cancer Surveillance Consortium (BCSC) facilities, at 1 year of follow up post mammogram, in the entire cohort, there were NOT statistically significant differences in rates of invasive cancer or advanced cancer (but there was a trend toward lower rates); in the 3.6% of women in the cohort with extremely dense breasts and a high risk of breast cancer (BCSC 5-year breast cancer risk), advanced cancer rates per 1000 examinations were significantly lower (0.27 vs. 0.80, 95 %CI -0.97 to -0..10). [JAMA. 2022. 327. 2220-2230].
The USPSTF concludes in 2024 that DBT is superior to digital mammography, based on simulation modelling data, because DBT is associated with fewer false positives; benefits were similar with DBT as compared with digital mammography (JAMA. 2024. 331. 1947-1960).
Thermography is inferior to mammography for screening. A systematic review of published literature showed that compared to mammography, sensitivity of thermography is 25% and specificity 74%; compared to histology, sensitivity of thermography ranged from 25% (specificity 85%) to 97% (specificity 12%) [N Z Med J. 2012. 125. 80-91].
Ultrasound
Inferior to mammography when used alone for screening, but can be complementary to mammography in high risk women, based on a study in 2637 participants – sensitivity of breast cancer detection was increased from 49% to 77.5%, but there were many false positives, such that the positive predictive value was only 8.6%. In this study, the ultrasound screening was performed by a physician, and took an average of 19 minutes of physician time to perform the exam, so this becomes a very expensive approach when physician time is factored into the mix (JAMA. 2008. 299. 2151-2163 and editorial 2203-2205).
Sometimes recommended as supplemental screening for women with dense breasts (and sometimes mandated by state laws that physicians discuss ultrasound as an option for women with dense breasts
A study SEER data in women aged 50-74 and 3 validated simulation models, examining cost and of screening and quality adjusted life-years reported (in the Conclusion section of the abstract) “Supplemental ultrasound screening for women with dense breasts would substantially increase costs while producing relatively small benefits” (Ann Intern Med. 2015. 162. 157-166).
The USPSTF concludes in 2016 that current evidence is insufficient to assess the benefits and harms of ultrasound as adjunctive screening for women with dense breasts and a negative mammogram (Ann Intern Med. 2016. 164. 279-296).
MRI – this is emerging in the 2000’s as a more sensitive screening test in women with a statistically high risk of breast cancer (N Engl J Med. 2007. 356. 1362-1364. Editorial).
In 2009, the USPSTF found that the current information is insufficient to assess the additional benefits and harms of screening MRI instead of film mammography (Ann Intern Med. 2009. 151. 716-726).
The American Cancer Society in the 2003 update to its guideline for breast cancer screening stated that women at increased risk for breast cancer might benefit from the earlier initiation of screening, shorter screening intervals, or the addition of screening methods such as breast ultrasound or MRI (CA Cancer J Clin. 2003. 53. 141-169).
The American Cancer Society in 2007 recommends annual breast-cancer screening by means of MRI for women with approximately 20% or greater lifetime risk of breast cancer, according to risk models that are largely dependent on a strong family history of breast or ovarian cancer (CA Cancer J Clin. 2007. 57. 90-104). A “risk calculator” is available at http://bcra.nci.nih.gov/brc/questions.htm.
There is also a role for MRI in “secondary screening” for breast cancer in the contralateral breast in women with recently diagnosed breast cancer (N Engl J Med. 2007. 356. 1295-1303).
The USPSTF concludes in 2016 that current evidence is insufficient to assess the benefits and harms of MRI as adjunctive screening for women with dense breasts and a negative mammogram (Ann Intern Med. 2016. 164. 279-296).
USPSTF recommendations for women with dense breasts (JAMA. 2024. 331. 1918-1930)
Current evidence is insufficient to assess the balances and harms of supplemental screening for breast cancer using breast ultrasonography or breast MRI, in those women identified with dense breasts and an otherwise negative mammogram.
NOTE approximately half of US women are categorized on mammography as having dense breasts; the increase in breast cancer risk associated with dense breasts on mammography screening is small in magnitude, similar to the increased risk associated with having an aunt with breast cancer (Ann Intern Med. 2018. 169. 474-484).
Effective in September 2024, the FDA will mandate all US screening facilities to inform women of their breast density in conjunction with informing women of their mammography results (Editorial. JAMA. 2024. 331. 1896-1897).
Recommendations for screening mammography
Mammography post autologous breast reconstruction – not beneficial for screening, based on data in 541 patients, as physical exam is sensitive for detection of local recurrence (Ann Surg Oncol. 2014. 21. 3256-3260).
Age 50-74:
In 2009, 2016, and 2024 the USPSTF recommends biennial screening mammography for women between the ages of 50 and 74 [Grade B recommendation] (Ann Intern Med. 2009. 151. 716-726; Ann Intern Med. 2016. 164. 279-296; JAMA. 2024. 331. 1918-1930).
Historically mammography was recommended every 1-2 years by almost all authorities in the US, based on results from the individual large trials and a meta-analysis of the 5 Swedish trials which show a 30% reduction in mortality at 10-12 years (Lancet. 1993. 341. 973-978).
HOWEVER, a re-analysis of these large trials by two Danish researchers, Olsen and Gotzsche, published in Lancet in 2000 suggests that the randomization process failed to create similar groups in 6 of the 8 trials, and that there were imbalances at baseline in 4 of the 5 trials included in the meta-analysis of Swedish trials.
Furthermore, the results of the two studies with adequate randomization, the Malmo study (BMJ. 1988. 297. 943-948) and the Canadian study (Can Med Assoc J. 1992. 147. 1477-1488) showed no decrease in the risk of death from breast cancer in women screened with mammography. The authors conclude that "screening for breast cancer with mammography is unjustified" (Lancet. 2000. 355. 129-134). This conclusion generated much debate. It was followedby publication in Lancet in 2001 of a research letter reporting that a Cochrane review confirmed the findings of the analysis published in Lancet in 2000 (Lancet. 2001. 358.1340-1342). An editorial accompanying the 2001 research letter concludes that "At present, there is no reliable evidence from large, randomised trials to support screening mammography programmes" (Lancet. 2001. 358. 1284-1285). Analysis of surgery rates (mastectomies and lumpectomies) posted on the Lancet web site determines that there are approximately 40 extra surgeries per every 10,000 women screened - this can be considered a significant risk of screening mammography (Ann Intern Med. 2002. 137. 363-365)!
Nonetheless the United States Preventive Services Task Force (USPSTF) disagreed with the authors of the Cochrane review and disagreed with the conclusions published in Lancet and found that only one of the eight trials was flawed and that the combined evidence from the seven acceptable trials showed reductions in breast cancer mortality of at least 25% in women aged 50-70, and continued in 2002 to recommend mammography screening every one to two years, categorizing this as a Grade B recommendation, meaning that the evidence is only fair in quality (Ann Intern Med. 2002. 137. 344-346).
Additionally, in 2002 the National Cancer Institute (NCI) rejected its advisory board's position and continued to recommend mammography screening every one to two years.
Reanalysis of the 4 Swedish breast cancer screening mammography trials to extend the follow-up showed that the benefit of mammography screening was increased until 12 years after randomization and was preserved thereafter (Lancet. 2002. 359. 909-919).
The number needed to screen to prevent one breast cancer death after 14 years is 838 (Harris R. Screening for breast cancer: what to do with the evidence [editorial]. Am Fam Physician. 2007. 75. 1623-1626).
In aggregate, the data shows that screening biennially from ages 50-69 achieved a median 16.5% reduction in breast cancer deaths, compared to no screening (Ann Intern Med. 2009. 151. 738-747); this translates into one death averted for every 1300 women age 50-59 screened for 10 years, and one death averted for every 400 women age 60-69 screened for 10 years (Ann Intern Med. 2009. 151. 727-737).
However, In the Canadian NBSS, in which 89,835 women aged 40-59 were randomized to usual care or 5 years of annual mammograms, at a mean follow up of 22 years, there was no reduction in breast cancer mortality in the mammography group (BMJ. 2014. 348. g366).
Risks versus benefits - risks of mammography screening include false positive mammograms and overdiagnosis of breast cancer. Overdiagnosis is defined as the identification of tissue that has all of the pathologic indications of cancer on biopsy, but if not treated will not progress to cause symptoms or death, and may even regress.
In the Canadian NBSS, in which 89,835 women aged 40-59 were randomized to usual care or 5 years of annual mammograms, at a mean follow up of 22 years, excess numbers of invasive cancer in the mammography group represented 22% of all screen-detected cancer and 50% of nonpalpable cancer. Since breast cancer mortality at 22 years was the same in the screening group as in the usual care group, these statistics seem to reflect overdiagnosis [i.e. an invasive cancer that would not have become clinically apparent during life] (BMJ. 2014. 348. g366).
SEER data on tumor-size distribution in 1975-1979 (prior to implementation of widespread mammography) as compared with 2000-2002 showed 30 fewer cases of cancer per 100,000 women in 2000-2002, but 162 more cases of cancer per 100,000 women in 2000-2002. Assuming the prevalence of cancer did not change over the 30 years from 1970's to 2000's, this data suggests that 132 cases per 100,000 women represent overdiagnosis (N Engl J Med. 2016. 375. 1438-1447 and editorial 1483-1486).
Data in a racially, ethnically, and geographically diverse cohort of 35,986 women (Breast Cancer Surveillance Consortium), and based on using a natural history likelihood model concludes that among women biennially screened from age 50 to 74, approximately 1 in 7 screen-detected breast cancers represents overdiagnosis (Ann Intern Med. 2022. 175. 471-478 and editorial 598-599).
An overview and re-analysis of systematic reviews concludes that overdiagnosis occurs in approximately 25% of those diagnosed with cancer (Cancer Epidemiol. 2023. 84. 102352).
Aggregate date from RCTs shows that if 1000 50-year old women have annual mammograms for the next 10 years, about 600 will have at least one false positive mammogram, as compared with 25 diagnosed with invasive breast cancer (a benefit, except in those who are overdiagnosed), and 3 lives will be saved by mammography screening (Commentary by Suzanne Fletcher in ACP Journal Club. 2014. 160. 10. JC7, in reference to BMJ. 2014. 348. g366).
Age 40-49: Historically, recommendations have fluctuated (Editorial. Ann Intern Med. 2007. 146. 529-531; Editorial. Ann Intern Med. 2024. 177. 389-392)
Guidelines
In 2024, the USPSTF recommends biennial screening mammography for women 40-49 years old [Grade B recommendation] (JAMA. 2024. 331. 1918-1930). Note the draft of the 2024 recommendations was published on May 9, 2023, and this prompted published critiques of the change from the 2009 and 2016 USPSTF recommendations for women in this age group.
As per a “Dissenting View” (N Engl J Med. 2023. 389. 1061-1064), these guidelines are not based on new data from RCTs but rather the new guidelines are based on complex statistical modelling, and estimated effects derived from this type of modelling are inherently extremely sensitive to the modelling assumptions. The dissenting view promotes continuing to recommend screening to commence in average risk women at age 50, rather than age 40 (N Engl J Med. 2023. 389. 1061-1064).
As per an editorial (Ann Intern Med. 2024. 177. 389-392) by a USPSTF reviewer of 20 years and USPSTF member for 5 years, the 2023 DRAFT recommendations are not evidence-based.
Prior USPSTF recommendations
In 2009 and in 2016, the USPSTF recommends against routine screening mammography in women aged 40-49, stating that “the decision to start regular, biennial screening before the age of 50 years should be an individual one and take into account patient context, including the patient’s values regarding specific benefits and harms [Grade C recommendation] (Ann Intern Med. 2009. 151. 716-726; Ann Intern Med. 2016. 164. 279-296).
In 2002 the USPSTF did recommend mammography screening every 1-2 years [Grade B recommendation], qualifying this recommendation by writing “the precise age at which the benefits from screening mammography justify the potential harms is a subjective judgement and should take into account patient preferences“ (Ann Intern Med. 2002. 137. 344-346).
In 1989 and 1996 made no recommendation for this age group, based on uncertainty of the evidence.
Historically, the NIH Consensus Conference in 1997 recommended neither for nor against screening in this age group (NIH Consensus Statement. 1997. 15. 1-35). NOTE the U.S. Senate responded to these guidelines by voting 98 to 0 in favor of screening women in their 40s (Editorial. N Engl J Med. 1997. 336. 1180-1183).ACP in 2007 recommended individualizing assessment for risk of breast cancer to help guide decisions about screening mammography, as the magnitude of benefit for the average risk woman is small, and the risks are significant (Ann Intern Med. 2007. 146. 511-515).
Historically, the Canadian Task Force in 2001 recommended discussion of potential risks versus benefits with each woman (CMAJ. 2001. 164. 469-476).
Historically, ACOG in 2003 recommended mammograms every 1-2 years (Obstet Gynecol. 203. 101. 821-831).
Historically, American Cancer Society in 2006 recommended annual mammograms starting at the age of 40 (CA Cancer J Clin. 2006. 56. 11-25).
Data on benefits
A meta-analysis of 11 RCTs with 13 years of follow up, conducted by the Independent UK Panel on Breast Cancer Screening concluded that screening is associated with a 20% reduction in breast cancer mortality. For women between the ages of 50-75, this translates into preventing 1 breast cancer death for every 235 women invited to screening and 1 breast cancer death for every 180 women actually screened (Lancet. 2012. 380. 1778-1786 as cited in JAMA. 2013. 309. 2555-2556).
A Cochrane review of the same studies meta-analyzed by the UK Panel estimated a 15% reduction in breast cancer mortality, because of concerns about quality in some of the trials (Cochrane Database Syst Rev. 2011. CD001877).
USPSTF updated analysis in 2009 reports that mortality reduction with screening in this age group is 15%, and this translates into a need to screen 1900 women aged 40-49 years for 10 years to prevent one death. Stated another way, 3.5 out of 1000 women in their 40’s will die of breast cancer over the next 10 years; screening reduces this to 3 of 1000 women (Ann Intern Med. 2009. 151. 727-737).
A systematic review of the 8 original mammography trials and an additional 117 studies which met criteria (of 873 studies reviewed) concluded that screening mammography in women age 40-49 is associated with a 7-23% reduction in breast cancer mortality (Ann Intern Med. 2007. 146. 516-526). The number needed to screen to prevent one breast cancer death after 14 years is 1792 (Harris R. Screening for breast cancer: what to do with the evidence [editorial]. Am Fam Physician. 2007. 75. 1623-1626).
The Age trial, which enrolled 160,921 women and was published after completion of the analysis for the above systematic review, found that at a mean follow up of 10.7 years, the relative risk of death due to breast cancer in the screened group was 0.83 in screened women in their 40’s (p=0.11). These findings on relative risk reduction (17%, but not statistically significant) are consistent with the findings of the above systematic review. This translates though to an absolute risk reduction in mortality of only 0.40 per 1000 women (Lancet. 2006. 368. 2053-2060).
A meta-analysis indicates an 18% mortality reduction with mammography screening every 1-2 years (J Natl Cancer Inst Monogr. 1997. 87-92). In terms of absolute numbers, it is estimated that over the course of a decade of screening, 16 of every 1000 women in their 40's will be diagnosed with breast cancer, but mammography screening will benefit only 2 of these 16.
The Canadian National Breast Screening Study fails to show any benefit from mammography screening, as compared with annual clinical breast exams (standardized, lasting 5-10 minutes).
In the NBSS-1, in which 50,430 volunteers aged 40-49 years were randomized from 1/80-3/85 to 4 or 5 annual screenings with mammography along with breast physical exam and instruction in breast self-exam or usual community care, at 11-16 years of follow up, there was no reduction in breast cancer mortality in the mammography group (Ann Intern Med. 2002. 137. 305-312).
In the NBSS, in which 89,835 women aged 40-59 were randomized to usual care or 5 years of annual mammograms, at a mean follow up of 22 years, there was no reduction in breast cancer mortality in the mammography group (BMJ. 2014. 348. g366).
NOTE, RCTs might underestimate the benefit of screening because a women randomized in a study to be screened who dies from breast cancer despite never going for screening counts as a mortality in the screening group, and a woman in the control group who has a mammogram outside the trial is counted as not screened. Two observational studies, which include millions of person years of follow up, show mortality reduction of 30-40% (Int J Cancer. 2007. 120. 1076-1080: Lancet. 2003. 361. 1405-1410), compared with 15% in RCTs.
Data on risks/harms (Armstrong K et al. Screening mammography in women 40 to 49 years of age: a systematic review for the ACP. Ann Intern Med. 2007. 146. 516-526).
The authors of a systematic assessment of benefits and risks of breast cancer screening conclude that the false positive rate over the course of a decade for a 40-year-old woman undergoing 10 years annual mammography screening is about 61%, and that the overdiagnosis rate is approximately 19% (meaning 19% of cancers diagnosed would not have become clinically apparent without screening). The net benefit for an individual depends upon baseline cancer risk for that individual (Review. JAMA. 2014. 311. 1327-1335 and editorial 1298-1299).
The Independent UK Panel on Breast Cancer Screening estimated an overdiagnosis rate of 11%, which translates into 129 overdiagnosed case per 10,000 women screened (Lancet. 2012. 380. 1778-1786 as cited in JAMA. 2013. 309. 2555-2556).
After 10 years of annual screening, more than half of women age 40-59 at the age of the first screen will receive at least one false positive recall, and 7-9% will receive a false positive biopsy recommendation, based on prospective data on 169,456 women gathered from 7 mammography registries in the NCI-funded Breast Cancer Surveillance Consortium. “Biennial screening appears to reduce the cumulative probability of false positive results after 10 years, but may be associated with a small absolute increase in the probability of late-stage cancer diagnosis (Ann Intern Med. 2011. 155. 481-492).
Increased diagnosis of ductal carcinoma in situ (DCIS), the natural history of which is uncertain. Based on 15 year follow up from the Malmo trial, 2 women are overdiagnosed for every breast cancer death avoided (BMJ. 2006.332. 689-692); a Cochrane review concludes that 10 women are overdiagnosed for every breast cancer death avoided (Cochrane Database Syst Rev. 2009. CD001877).
A systematic analysis of 23 eligible studies (n=313,967) found that a greater percentage of women with false positive mammograms returned for routine screening, and that the women with false positive mammograms conducted more frequent breast self- examinations, but did not have significantly higher levels of pathologic distress or anxiety in the long run (Ann Intern Med. 2007. 146. 502-510).
Risk from radiation exposure is small – death secondary to radiation exposure from a screening mammogram estimated at 30-200 deaths per 100,000 women screened. Nonetheless, a systematic review of 20 studies showed that the risk of radiation-induced breast cancer was increased in 14 studies (RR 1.3 to 11), with no increased risk seen in 6 studies (Ann Intern Med. 2007. 146. 516-526).
Women screened yearly from age 40-50 will have a rate of false positive mammograms over this decade of 50%, and a biopsy rate of 20% (N Engl J Med. 1998. 338. 1089-1096). It is estimated that about 85% of abnormal mammograms turn out to be false positives. False positive mammograms are associated with increased anxiety during the time of a workup. HOWEVER, in a survey of 479 women, 63% felt that 500 or more false positive mammograms per life saved was a reasonable tradeoff (BMJ. 2000. 320. 1635-1640).
False reassurance if the result is normal.
Discomfort with the procedure.
Cost effectiveness - the estimated cost per year of life saved is $105,000 for women in their forties, compared with $21,400 for women aged 50-59. In order to decrease the incremental cost for women in their forties to $50,000 per year of life saved, mammography cost would have to be reduced to $45 (Ann Intern Med. 1997. 127. 955).
Age over 74: Consider life expectancy and probability of risk versus benefit
In 2009, 2016, and 2024, the USPSTF concludes that the current information is insufficient to assess the additional benefits and harms of screening mammography in women 75 years or older [I statement] (Ann Intern Med. 2009. 151. 716-726; Ann Intern Med. 2016. 164. 279-296; JAMA. 2024. 331. 1918-1930).
A Markov model suggests that mammography screening in women aged 70-79 is moderately expensive and adds little to life expectancy. In this population, biennial screening of a population of 10,000 women with bone mineral density in the top 3 quartiles would prevent 9.4 deaths and increase life expectancy 2.1 days at a cost of $66,773 per year of life saved; biennial screening of all women would add 7.2 hours to life expectancy at an added cost of $117,689 per year of life saved (JAMA. 1999. 282. 2156-2163).
A study that compared outcomes among women aged 70-75 and those aged 50-69 concluded based on data from 187,207 women aged 70-75 who had mammograms performed (65.6% of the 315,103 women offered mammography) that the duration of the preclinical detectable phase of breast tumors increases steadily beyond age 69, leading to an unfavorable balance between benefits and harms in women older than age 75. The conclusion was that age 75 is an appropriate upper limit for screening mammography in women (Int J Cancer. 2006. 118. 2020-2025).
The authors of a review entitled “Screening Mammography in Older Women” states that modelling studies estimate 2 fewer breast cancer deaths per 1000 women who in their 70s continue biennial screening for 10 years instead of stopping screening at age 69. Potential harms of continued screening include false positive mammograms in approximately 200 per 1000 women screened, and overdiagnosis approximately 13 per 1000 women screened (JAMA. 2014. 311. 1336-1347).
HOWEVER, observational data in 1,058,013 Medicare beneficiaries aged 70 to 84 years shows benefit of mammography screening in women aged 70-74 years but no benefit in women aged 75-84 years (Ann Intern Med. 2020. 172. 381-389). An accompanying editorial comments that this data precedes widespread use of digital or 3-D mammography as the cohort was from 2000-2008 (Ann Intern Med. 2020. 172. 427-428).
HOWEVER, a Markov model using data from SEER concludes that concludes that “Although annual mammography (beyond age 75 years) is not cost effective, biennial screening mammography to age 80 years is; however the absolute number of deaths averted is small, especially for women with comorbidities.” The cost per QALY associated with biennial screening from ages 75-79 was $54,000 for those with zero comorbidities, $65,000 for those with one comorbidity, and $85,000 for those with two comorbidities (Ann Intern Med. 2022.. 175.. 11-19).
Randomized trials of screening mammography (N Engl J Med. 2003. 348. 1672-1680)
The USPSTF in 2009 updated the evidence from randomized controlled trials (Ann Intern Med. 2009. 151. 727-737).
8 major trials have randomized 500,000 women (J NCI. 1993. 20. 1644-1650; Lancet. 1993. 341. 973-978; Ann Intern Med. 1989. 111. 389-399).
4 studies in Sweden comparing mammography with no screening.
3 studies (one in Edinburgh Scotland, one in Canada, and one in New York) comparing mammography and clinical breast exam with no screening.
1 study in Canada comparing addition of mammography to standardized 10-15 minute clinical breast exam.
The most recent RCT reported the effect of mammographic screening from age 40 years on breast cancer mortality at 10 years of follow up (Lancet. 2006. 368. 2053-2060).
These studies vary in periodicity of screening (12-36 months) and also in the number of views (1 vs. 2) of each breast.
NOTE: other than the Canadian study (NBSS2) which also included a head to head trial of mammography vs. clinical breast exam, none of these studies have head to head randomized comparisons of 12 vs. 24 months mammography screening, 1 vs. 2 view mammography or mammography vs. clinical breast exam.
One possible limitation of the data available from the above studies is that major improvements were made in mammography technique around 1985.
Effect of screening mammography on breast cancer mortality
Historically, RCTs showed an approximately 30% relative reduction in breast cancer mortality.
A systematic assessment of benefits and risks of mammography concludes that screening is associated with a 19% overall reduction in breast cancer mortality, with a reduction approximately 15% for women in their 40s, and 32% for women in their 60s (Review. JAMA. 2014. 311. 1327-1335 and editorial 1298-1299).
Observational data in 40,075 women with breast cancer in Norway, using two historical control comparison groups, showed that screening mammography (every other year) “was associated with a reduction in the rate of death from breast cancer, but the screening itself accounted for only about a third of the total reduction.” The death rate was reduced by 7.2 deaths per 100,000 person-years in the screening group as compared with the historical screening group, and 4.8 deaths per 100,000 person-years in the nonscreening group, as compared with the historical nonscreening group. Thus, the difference between current and historical groups attributable to screening alone was 2.42 deaths per 100,000 person-years, a relative reduction of 10% [p=0.13] (N Engl J Med. 2010. 363. 1203-1210). An accompanying editorial addresses possible explanations for the discrepancy between the 10% relative reduction in this study, as compared with 30% relative reduction in historical RCTs of mammography screening, and concludes that while the discrepancy might be due to the fast that the current data is observational, it might be due to an increase in breast cancer awareness amongst women (and thus the possibility of increased self-detection) as well as improvements in outcomes in unscreened women due to advances in treatment. The editorial concludes “But no one can argue that screening mammography is one of the most important services we provide in medicine. The time has come for it to stop being used as an indicator of the quality of our health care system.” (Welch HG. N Engl J Med. 2010. 363. 1276-1278).
No reduction in all-cause mortality (Cochrane Database Syst Rev. 2001. 4:CD001877). NOTE a Markov model using Australian data provides data on differences in the absolute risk of death from breast cancer and all causes in women screened biennially compared with women not screened, including a table (BMJ. 2005. 330. 936-938).
However, the USPSTF in 2016 identified 10 prospective randomized trials, and concluded that collectively these studies suggest that routine mammography reduces relative risk for cancer death in women aged 50-70 by 20-33% (Ann Intern Med. 2016. 164. 279-296).
Chances of development of and death from breast cancer within next 10 years (N Engl J Med. 2003. 348. 1672-1680)
NOTE the statistic of a 1 in 8 lifetime risk of breast cancer is for a newborn who lives for 90 years.
Age 10 year risk 10 year risk 10 year risk
of invasive breast cancer death from breast cancer death from any cause
Age 40 1.5% 0.2% 2.1%
Age 50 2.8% 0.5% 5.5%
Age 60 3.7% 0.7% 12.6%
Age 70 4.3% 0.9% 30.9%
Age 80 3.5% 1.1% 67.0%
Chemoprophylaxis of breast cancer
USPSTF III recommends against the routine use of tamoxifen or raloxifene for the primary prevention of breast cancer in women at low or average risk of breast cancer.
A systematic review reports that tamoxifen, raloxifene, and tibolone (not yet FDA approved in the US in 2009) reduce the risk for primary breast cancer, but tamoxifen and raloxifene increase the risk for thromboembolic events, tamoxifen also increases the risk for endometrial cancer, and tibolone reduces the risk of stroke. Limitations of the data analyzed in this systematic review include bias, trial heterogeneity, lack of head to head trials of the agents, lack of data on doses, duration, and timing of the medications, lack of data on long term effects, and lack of data for nonwhite and premenopausal women (Ann Intern Med. 2009. 151. 703-715).
A decision about chemoprophylaxis in women at high risk of breast cancer involves a consideration of risks versus benefits. A risk calculator can be accessed at http://bcra.nci.nih.gov/brc/ or at www.cancer.gov/bcrisktool/ or 800-4-CANCER (800-422-6237).
The largest RCT of tamoxifen, the Breast Cancer Prevention Trial (BCPT), which enrolled 13,388 women, found a risk reduction of invasive cancer of 49% among women at high risk for breast cancer. There were 175 cases of invasive breast cancer in the placebo group and 89 cases in the tamoxifen group (J Natl Cancer Inst. 1998. 90. 1371-1388). Two other RCT's of tamoxifen did not show similar benefit. The benefit was seen only against estrogen-receptor positive tumors.
One RCT of raloxifene in postmenopausal women showed a 76% risk reduction in the development of invasive breast cancer.
Both tamoxifen and raloxifene increased the risk of thromboembolic events, but the increased risk was statistically significant only in women over age 50. The increase in risk was comparable to the increase in risk associated with oral contraceptives or hormone replacement therapy. In the BCPT trial, after a median of 55 months of use of tamoxifen, there were 2.2 strokes per 1000 women in the treatment group compared with 1.3 strokes per 1000 women in the placebo group, 1.0 cases of pulmonary embolism in the treatment group compared with 0.3 cases per 1000 women in the placebo group, and 1.5 cases of DVT per 1000 women in the treatment group compared with 0.9 cases per 1000 women in the placebo group.
In women over age 50 in the BCPT trial, the risk of developing endometrial cancer was 3.1 cancers per 1000 women in the treatment group compared with 0.8 cancers per 1000 women in the placebo group. In women under age 50, there was no significant increase in risk of endometrial cancer. Raloxifene has not been associated with an increased risk of endometrial cancer.
In the BCPT, 45.7 % of women in the tamoxifen group reported bothersome hot flashes compared to 28.7 % in the placebo group, and 12.4 % of women reported bothersome vaginal discharge versus 4.5 % in the placebo group. In the MORE trial of raloxifene, 10.7 % of treatment patients reported hot flashes, compared to 6.4% in the placebo group (JAMA. 1999. 281. 2189-2197).
In general, the balance of benefits versus risks is more favorable in women in their 40s at increased risk for breast cancer and with no predisposition to thromboembolic events.
Evaluation of palpable breast mass (N Engl J Med. 1992. 327. 937-942)
Mammogram is recommended to search for clinically occult lesions, not primarily to characterize the mass, because mammograms are normal in approximately 20-30% of women with palpable breast cancer.
Beware that dense breasts (i.e. younger women) limit the utility of mammography.
Initial workup should be needle aspiration (do not rely on physical exam to rule out cancer).
Non-bloody fluid and resolution of mass confirms diagnosis benign disease.
Submit only bloody fluid for cytologic exam.
Use a 22 gauge needle, small syringe, no anesthesia.
Fine needle aspiration can be useful with an experienced cytopathologist.
Cervical cancer screening
13,170 cases/year; 4250 deaths/year [2019 data] 3% lifetime risk unscreened women; 0.8% lifetime risk screened women
Modalities of screening (Ann Intern Med. 2000. 133. 1021-1024; Editorial. N Engl J Med. 2007. 357. 1650-1653)
Cytology screening - Pap test (first published data on this modality was a monograph in 1943 by Papanicolaou and Traut) which has decreased the incidence and mortality of cervical cancer by 40% between 1973 and 2000, and receives an "A" recommendation from the USPSTF, even though a randomized, controlled trial has never been done.
Herpes papillomavirus (HPV) testing.
Rationale:
High sensitivity - in a RCT in 10,154 women in Montreal, Canada (Canadian Cervical Cancer Screening Trial) in which both HPV testing and Pap testing were performed, and results on colposcopy were used as the gold standard, the sensitivity of HPV testing for cervical intraepithelial neoplasia grade 2 or 3 was 94.6% (95% CI 84.2% to 100%), whereas the sensitivity of Pap testing was 55.4% (95% CI 33.6% to 77.2%). Specificity was 94.1% for HPV testing and 96.8% for Pap testing. The sensitivity of both tests used together was 100% with a specificity of 92.5% (N Engl J Med. 2007. 357. 1579-1588).
Highly reproducible, easily monitored, and provides an objective outcome.
Pap test has a low sensitivity, estimated at 51% in a systematic review (Ann Intern Med. 2000. 132. 810-819). 1/2 to 2/3 of false negatives are estimated to be caused by poor specimen collection, the other 1/3 to 1/2 by screening errors in the laboratory.
The overall prevalence of HPV among cervical cancers in a large international study is 99% (J Pathol. 1999. 189. 12-19). Of the more than 70 types of HPV, 4 types (16, 18, 31, and 45) are associated with most cases of invasive cervical cancer.
Disadvantages
Low specificity
Most HPV infections resolve spontaneously.
Due to high cost, and limited specificity, testing is currently not recommended for routine screening.
May have a role in the management of minor Pap abnormalities (JAMA. 2000. 283. 87-93).
In a simulation model of neoplasia natural history, screening every two years with combined Pap and HPV testing until age 75 was associated with a cost of $70,347 per QALY when compared with Pap testing every 2 years (JAMA. 2002. 287. 2372-2381).
In a RCT in 12,257 women aged 32 to 38 in Sweden, the addition of HPV testing to community screening with Pap testing reduces the incidence of cervical intraepithelial neoplasia grade 2 or 3 detected by subsequent screening exams (N Engl J Med. 2007. 357. 1589-1597). The method of HPV detection used in this study is not readily available (Editorial. N EnglJ Med. 2007. 357. 1650-1653).
USPSTF in 2011 recommends against HPV testing alone or in combination with cytology, in women younger than age 30 (D recommendation) [Ann Intern Med. 2012. 156. 880-891]. USPSTF in 2011 concluded that data is insufficient to recommend HPV-enhanced primary screening, with the high sensitivity of this test offset by a lower specificity, as compared with cytology screening (Ann Intern Med. 2011. 155. 687-697).
Guidelines for screening
Age of onset for screening
USPSTF in 2012 and again in 2018 recommends against screening prior to the age of 21 (D recommendation), because cancer is rare in women under age 21, and false positives are common. The data is insufficient to recommend deferring on initiation of screening until age 25 (Ann Intern Med. 2012. 156. 880-891; JAMA. 2018. 320. 674-686).
American College of Obstetrics and Gynecology (ACOG) in 2009 recommend screening should start at age 21, and should be avoided prior to age 21 (because cervical cancer is rare under age 21, but self-limited cytologic abnormalities are common under age 21).
American Cancer Society (ACS) guidelines in 2012 recommend initiation of screening at age 21 (CA Cancer J Clin. 2012. 62. 147-172).
Historically, guidelines recommended starting screening at age 18, or at the onset of sexual activity.
Frequency of screening
USPSTF in 2012 and again in 2018 recommends Pap smears every 3 years in women aged 21-29. In women aged 30-65, options include Pap smears every 3 years, Pap smears every 5 years in combination with HPV testing, or HPV testing every 5 years without Pap smears (Ann Intern Med. 2012. 156. 880-891; JAMA. 2018. 320. 674-686).
American Cancer Society (ACS) guidelines in 2012 recommend screening fromage 21 to 29 with cytology alone every 3 years, and screening from age 30 to 65 with cytology and HPV testing every 5 years (preferred) or cytology alone every 3 years (acceptable) [CA Cancer J Clin. 2012. 62. 147-172].
American College of Obstetrics and Gynecology (ACOG) in 2009 - women at average risk should be screened only every other year from age 21-29, and thereafter every 3 years for women with 3 consecutive negative Pap smears.
Screening over age 65 is controversial (Controversies in Internal Medicine. Arch Intern Med. 2004. 164. 243-248)
On the one hand 25% of new cases of cervical cancer and 41% of deaths from this cancer occur in the 13% of the female population over age 65 (National Cancer Institute. Surveillance Epidemiology and End Results at http://cancer.gov/statistics).
On the other hand cervical neoplasia has a pre-invasive period of 10-30 years and it is quite unusual for a woman who has been screened regularly up to age 65 to develop a new onset abnormal Pap smear and progress to invasive cervical cancer.
Prospective study of over 800 elderly women (JAMA. 1986. 256. 367-371) found a prevalence of 16 abnormal Pap smears/1000 women.
In this study, 25% of women over age 65 never hada previous Pap smear, 50% had infrequent Pap smears, and only 25% had Pap smears every one to three years.
Beware: In this study, 33% of women giving a past medical history of hysterectomy had an intact cervix.
Mathematical modeling with a deterministic semi-Markov model shows that screening every 2 years with HPV and Pap smears would capture 97.8% of the benefits of lifetime screening at a cost of $70,347 per QALY (JAMA. 2002. 287. 2372-2381).
Guidelines:
United Services Preventive Services Task Force in 2002, 2012 and again in 2018 recommends recommends stopping Pap screening after age 65 in women who have had regular previous screenings, with no history of cervical cancer or CIN on previous screenings (Guide to Clinical Preventive Services, second edition, 1996, pp. 105-107; Ann Intern Med. 2012. 156. 880-891; JAMA. 2018. 320. 674-686).
American Cancer Society (ACS) guidelines in 2012 recommend no screening over age 65 for women with evidence of adequate negative prior screening and no history of CIN2+ within the last 20 years (CA Cancer J Clin. 2012. 62. 147-172).
American Geriatrics Society recommends stopping Pap screening after age 70 in women with regular previous screenings (J Am Geriatr Soc. 2001. 49. 655-657).
American College of Obstetrics and Gynecology (ACOG) in 2009 recommends discontinuation of screening in women between 65 and 70 years of age who have had 3 or more consecutive normal tests and no abnormal results within the previous 10 years.
If no regular Pap smears up to age 65, then at the least do two Pap smears one year apart before discontinuing Pap smears
There is little value in screening women who have never been sexually active; the concern with this approach is whether the sexual history is accurate.
American College of Obstetrics and Gynecology (ACOG) in 2009 recommend that if colposcopy is performed, restraint is recommended in the treatment of most biopsy confirmed cytologic abnormalities (because spontaneous regression of CIN 1 and CIN 2 are common)
Pap smear post hysterectomy (with known removal of cervix)
Up to 37% of women in the U.S. have had their cervix removed by age 65 (J Womens Health Gender Based Med. 2002. 11.103-111).
It can be justified for women who had a hysterectomy for premalignant disease of cervix, because these women have risk of 0.9%- 6.8% for premalignant disease of vagina.
It is not cost effective in the general population because the incidence of vaginal cancer is only 0.6/100,000.
USPSTF III and ACS in 2002 recommend against screening in women who have had a total hysterectomy for benign disease. This recommendation is reiterated in USPSTF 2012 and 2018 updates (Ann Intern Med. 2012. 156. 880-891; JAMA. 2018. 320. 674-686).
Digital rectal exam during routine pelvic exam
This is clinically useful to only 1 of every 1000 women under age 40 (Journal of Family Practice. 1998. 46. 165-167).
The disadvantage is that it takes time, is uncomfortable, and may cause embarrassment.
Liquid-based cervical cytology
This is a technique for preserving and preparing cells for cytologic study – it involves suspending the sample of cells in a vial of liquid preservative rather than spreading on a glass slide.
In 2011, the USPSTF concluded that liquid-based cytology has equivalent sensitivity and specificity to conventional cytology, based on data from 4 fair to good quality studies of 141,566 participants (Ann Intern Med. 2011. 155. 687-697). Historically, 2003, the USPSTF found the evidence insufficient to make a recommendation about using liquid-based cervical cytology.
HPV Vaccine for Prevention of cancer of the cervix
The CDC ACIP voted 6/29/06 to recommend routine vaccination of females within certain age groups.
This is a quadrivalent vaccine (protective against HPV types 6, 11, 16, and 18), and is given as a series of 3 intramuscular injections.
Colorectal cancer screening
145,600 cases/year; 51,020 deaths/year [2019 data]
5% lifetime risk; 2.5% mortality
Incidence and mortality have declined approximately 3% between 2001-2016
10.5% of new cases occur in persons under age 50 (CA Cancer J Clin. 2017. 67. 177-193).
Incidence of adenocarcinoma of colon has increased by almost 15% in those aged 40-49 years, when comparing the incidence in 2000-2002 with the incidence in 2014-2016 (Ann Intern Med. 2021. 174. 157-166).
Rationale for screening
Colon cancer evolves slowly from premalignant adenomatous polyps.
Technological advances in the early 1970's with regard to guaiac cards with a stable reagent, sigmoidoscopy and colonoscopy made screening feasible.
Guidelines for screening
USPSTF 2021 (JAMA. 2021. 325. 1965-1977)
Screening is recommended for those 50-75 (A recommendation), those age 45-49 (B recommendation), and with consideration of screening on an individualized basis in those age 76-85 (C recommendation). NOTE the absolute risk of colon cancer in those ages 45-49 is 35 cases per 100,000 people.
USPSTF 2016 (JAMA. 2016. 315. 2564-2575) - update of 2008 guideline
"The USPSTF recommends screening for colorectal cancer starting at age 50 and continuing until age 75 (A recommendation). The decision to screen for colorectal cancer in adults aged 76 to 85 years old should be an individual one (C recommendation)."
"... colonoscopy every 10 years, annual FIT, sigmoidoscopy every 10 years with annual FIT, CTC every 5 years ... provided similar LYG (life-years gained) and a comparable balance of benefit and screening burden" in a microsimulation modeling study (JAMA. 2016. 315. 2595-2609). Shared decision making between patient and practitioner is recommended regarding which modality of screening to choose (Editorial. JAMA. 2016. 315. 2529-2531).
US Multi-Society Task Force on Colorectal Cancer 2016 (Robertson DJ et al. Gastroenterology. 2016. ePub) - updated guideline
Recommends screening for all individuals over age 50, using any one of a variety of methods.
Consensus statement summarizes the advantages of annual FIT testing over annual gFOTB testing.
Offers useful recommendations regarding implementation of FIT testing, including one stool test per screening round, screening once/year, and use of a quantitative immunochemical assay with a cutoff of < 20 mcg blood per gram of stool.
Evidence supporting these particular recommendations is recognized as weak.
In the US in 2016, the FDA has approved only qualitative FIT results reporting.
American Cancer Society 2018
Start screening in average risk individuals at age 45 instead of age 50.
The rationale was that the SEER database showed a 22% relative increase in colorectal cancer incidence from 2000 to 2013 in persons aged 40 - 49 years old. NOTE some of this increase is due to detection of carcinoid tumors which are likely incidental (Ann Intern Med. 2021. 174. 157-166).
NOTE that because the absolute risk of colorectal cancer is low in the age 45 - 49 age group, the cost is $6.1 million per one averted death from colorectal cancer (Ann Intern Med. 2018. 169. 405-416).
NOTE the SEER database lumps carcinoid tumors of the colon with adenocarcinoma of the colon and categorizes both as colorectal cancer, and data would suggest that most carcinoid tumors of the colon are incidental findings and do not require treatment.
Analysis of the SEER database shows that 4 - 20% of tumors categorized as colorectal cancer are carcinoid tumors (Ann Intern Med. 2021. 174. 157-166 and editorial 263-264).
USPSTF 2021 (JAMA. 2021. 325. 1965-1977) - recommends initiating colorectal cancer screening at age 45
American College of Physicians (ACP) 2023 (Ann Intern Med. 2023. 176. 1092-1100)
Start screening in asymptomatic average-risk adults at age 50
Consider not screening asymptomatic average-risk adults between the ages of 45-49, based on uncertainty pertaining to benefits versus harms of screening in this population
Stop screening in asymptomatic average-risk adults at age 75, or in those with a life expectancy of 10 years or less
Select a screening test in consultation with the patient based on discussion of benefits, harms, costs, availability, frequency, and patient values and preferences. Select amongst fecal immunochemical or high-sensitivity guaiac fecal occult blood tests every 2 years, colonoscopy every 10 years, or flexible sigmoidoscopy every 10 years plus fecal immunochemical test every 2 years as screening options (recommends against stool DNA, CT colonography, capsule endoscopy, urine, or serum screening tests)
Surveillance – US Multi-Society Task Force on Colorectal Cancer recommendations (Gastroenterology. 2020. 158. 1131-1153):
Patients with hyperplastic polyps should be considered to have a normal colonoscopy, with subsequent colonoscopy recommended at 10 years.
Patients with 1-2 adenomas < 10 mm in size should have their next colonoscopy in 7-10 years (prior recommendation was next colonoscopy in 5-10 years). Note that 70% of individuals with adenomas on colonoscopy have only 1-2 adenomas less that 10 mm in size.
Patients with 3-4 adenomas < 10 mm in size should have their next colonoscopy in 3-5 years (prior recommendation was next colonoscopy in 3 years).
Patients with > 11 adenomas should have their next colonoscopy in 1 year (prior recommendation was next colonoscopy in < 3 years).
Patients with advanced adenomas (> 10 mm in size, with villous features, or with high grade dysplasia) should have their next colonoscopy 3 years (same as prior recommendation).
Following curative resection of colorectal cancer, repeat colonoscopy should be done at 1 year.
FOBT screening
Guaiac-based test; requires the addition of hydrogen peroxide to oxidize substances such as blood in the stool.
Specificity is higher if the person institutes specific dietary restrictions such as eliminating meat from the diet for a few days prior to the test.
The ACS and USPSTF recommend annual screening in patients age 50-75 (USPSTF) as one of several options for screening.
Covered by most health insurance, including Medicare 1/1/98 (Level 1 evidence).
Recommendations from USPSTF updated to recommend against routine screening in adults age 76-85 and recommends against screening in adults older than age 85 (Ann Intern Med. 2008. 149. 627-637).
Data from the RCT's below indicate that annual screening is significantly more sensitive than biennial screening, but less specific, and that rehydration of the cards is also more sensitive but less specific.
USPSTF III states that whether patients need to restrict their diet or avoid certain medication is not established.
Screening is defined as use of guaiac-based test cards prepared at home on 3 consecutive stool samples; USPSTF III states that testing a single stool specimen obtained at the time of DRE is not recommended as an adequate screening strategy.
Cost effectiveness: data suggests a cost of $10,000 - $25,000 per year of life saved (Ann Intern Med. 2002. 137. 132-141).
Recommended in conjunction with flexible sigmoidoscopy.
NOTE that the recommendation is for testing 6 samples in 3 bowel movements. A study in 3121 asymptomatic men at 13 VA medical centers showed that hemocult testing a single sample obtained at the time of rectal exam at an office visit had a sensitivity of only 4.9% for advanced neoplasia compared to 23.9% sensitivity in the same study when 6 samples were obtained (Ann Intern Med. 2005. 142. 81-85).
The case for screening:
Biologic rationale: colon cancer and adenomatous polyps may bleed.
Four randomized, controlled, prospective trials show that it (hemocult II) saves lives. A systematic review of these 4 trials including nearly 330,000 participants followed for an average of about 9 years shows a 16% lower mortality from colorectal cancer in the screened group. In the subgroup of individuals assigned to screening who actually followed through (i.e. complied with completion of the stool cards), mortality from colorectal cancer was reduced 23% (BMJ. 1998. 317. 559-565).
N Engl J Med. 1993. 328. 1365-1371.
46,551 subjects (volunteers), age 50-80, 13 years follow up (83% of cards rehydrated). Minnesota. Screening positive in 10% of cases/year.
FOBT yearly: 6 of 1,000 died of colon cancer.
No screening: 9 of 1,000 died of colon cancer.
Conclude FOBT lowers colon cancer death risk by 1/3.
Extrapolation to the 60 million Americans over age 50; save 20,000 lives/year.
Note: every other year FOBT was studied; not as good as yearly in this study.
In a published 18 year follow up of the above study, compared with controls, the incidence of colorectal cancer with annual screening was reduced 20% and incidence with biannual screening was reduced 17% (N Engl J Med. 2000. 343. 1603-1607).
Benefit persists at 30 years – risk of death from colon cancer was 32% lower in the annual screening group and 22% lower in the biennial screening group. Note that 32% relative risk reduction translates into a relatively small absolute risk reduction – 6 deaths prevented per 1000 persons screened for 30 years (N Engl J Med. 2013. 369. 1106-1114 and editorial 1164-1166).
Lancet. 1996. 348. 1472-1477.
Randomized allocation of 150,000 subjects in the United Kingdom who received either FOBT by mail every other year or no screening.
No rehydration - screening was positive in 2% of cases. Approximately 60% of patients will accept FOBT.
In patients with positive FOBT, 40-50% have a positive colonoscopy and 10% have colorectal cancer
At 8 years, there was a 15% reduction in cumulative colorectal cancer mortality in the screening group, but with no difference in all-cause mortality.
Lancet. 1996. 348. 1467-1471.
Randomization of 140,000 subjects age 45-75 in Denmark, in which screening group patients were sent FOBT cards every other year.
67% completed the first screening round, and screening was positive in 4.3% of cases (no rehydration).
At 10 years, there were 205 deaths from colorectal cancer in the screened group and 249 deaths from colon cancer in the controls, which translates into an 18% reduction in mortality in the screened group.
Scand J Gastroenterol. 1994. 29. 468-473.
Randomization of 68,308 subjects aged 60-64.
Two screens 16-22 months apart.
Most cards rehydrated.
Unpublished results show a 12% reduction in mortality from colorectal cancer with biennial screening.
The case against screening:
Low sensitivity:
Sensitivity of hemocult and hemoquant for detecting asymptomatic cancer is approximately 30% and for detecting large polyps (1-2 cm) is only about 10% (JAMA. 1993. 269. 1262-1267).
Sensitivity for detection of advanced neoplasia (defined as tubular adenoma 10 mm or greater, adenomas with villous histology or high-grade neoplasia, or invasive cancer) was 23.9% in a prospective cohort study in 3121 asymptomatic men at 13 VA medical centers (Ann Intern Med. 2005. 142. 81-85).
Note: average blood loss from a polyp is only 1.4 cc/day, but 3 cc/day is necessary for a positive hemocult test.
Note: the citation of 80-90% sensitivity in the N Engl J Med study is an erroneous statistic based on the number of patients in whom cancer was detected in one year of the test, rather than the percentage of patients with cancer in whom the test was positive.
Low specificity: In the above N Engl J Med study with rehydration, 38% of subjects had a positive result over 13 years, but only 2.5% of the colonoscopies done were positive for cancer.
Low yield: Meta-analysis of the four randomized, controlled trials shows that screening reduced mortality 16% for those allocated to screening and 23% for those actually screened (BMJ. 1998. 317. 559-565). If 10,000 people over age 40 were offered biennial screening, 8.5 deaths from colorectal cancer would be prevented over 10 years. Stated differently, the number needed to screen to prevent one death from colorectal cancer over 10 years is 1173 people (BMJ. 1998. 317. 559-565).
No reduction in all-cause mortality:
A meta-analysis of 3 large published randomized studies in which a total of 245,217 people were followed for a mean of 12 years showed a 1.9% relative increase in noncolorectal cancer deaths in the non-screened group and no difference in overall mortality between screened and unscreened groups (Am J Gastroenterol. 2006. 101. 380-384).
Thirty year follow up data for the participants in the study initially reported in N Engl J Med. 1993. 328. 1365-1371 showed no decrease in all-cause mortality associated with screening (N Engl J Med. 2013. 369. 1106-1114 and editorial 1164-1166).
High cost:
Direct cost: $5.00/FOBT x 60 million adults over age 50 = $300 million/year.
Downstream cost: $500-$1000/colonoscopy x 6 million colonoscopies per year if stool cards are rehydrated = $3-$6 billion/year.
Added cost of ongoing surveillance for positive colonoscopies.
FIT (Immunochemical FOBT) and Fecal DNA – see “novel methods of screening” below.
Sigmoidoscopy
Recommended by the ACS and USPSTF as an option for screening.
Covered by most health insurance, including Medicare, effective Jan 1, 1998.
Recommendations from USPSTF updated to recommend against routine screening in adults age 76-85 and recommends against screening in adults older than age 85 (Ann Intern Med. 2008. 149. 627-637).
The interval of screening is not clear.
In 2002, the ACS recommendations are for every 5 years instead of every 3 years.
USPSTF recommends a screening interval of every 5 years (Ann Intern Med. 2008. 149. 659-669).
In the PLCO trial, a community based RCT of cancer screening conducted in 10 screening centers in the United States, 9317 (80%) of 11,583 eligible participants for repeat flexible sigmoidoscopy screening 3 years after an initial negative exam returned. 1292 of these 9317 returning participants (13.9%) had a polyp or mass detected by flexible sigmoidoscopy screening 3 years after the initial exam. 292 of these 9317 (3.1%) were found to have an adenoma or cancer. A total of 6 individuals had cancer and another 72 had advanced adenomas. The yield for cancer detection on the initial screen was 27 cancers detected per 10,000 individuals screened; the yield for cancer detection on the repeat screen was 6.4 cancers detected per 10,000 individuals screened. The yield for advanced distal adenoma detection on the initial screen was 3.1%; the yield for advanced distal adenoma detection on the repeat screen was 0.8%. Medical record review determined that 80% of the advanced adenomas and cancers had arisen since the previous screening just 3 years ago whereas 20% might have been present and missed due to inadequate preparation or decreased depth of scope insertion on the initial exam (JAMA. 2003. 290. 41-48).
An issue of relevance in determining the optimal interval of screening is data on the sensitivity of a single screen. Studies of tandem colonoscopy in which two practitioners perform colonoscopy in tandem on the same day indicate miss rates of 13% for adenomas smaller than 1 cm (Gastrointest Endosc. 1991. 37. 125-127), 27% for adenomas smaller than 5 mm, and 6% for adenomas of at least 1 cm (Gastroenterology. 1997. 112, 24-28).
The case for screening:
Biologic rationale:
Adenomatous polyps are precursors to colon cancer, and the progression from polyp to cancer requires 5-10 years in an average risk population (Cancer. 1975. 36. 2241-2270; Gastroenterology. 1987. 93. 1009-1013).
50% colon cancers arise in distal 60 cm colon.
50% persons with adenomatous polyps in distal colon will have more proximal lesions.
Prospective observational data from participants in the Nurses’ Health Study and the Health Professionals Follow-up Study, in 88,902 participants followed over 22 years – relative risk in those screened with sigmoidoscopy was 0.59 (N Engl J Med. 2013. 369. 1095-1105 and editorial 1164-1166).
High quality case-control study (N Engl J Med. 1992. 326. 653-657).
Kaiser Permanente record review. 261 case patients, 868 control patients.
70% reduction in colon cancer and 60% reduction in colon cancer mortality in the part of the colon visualized by 20 cm rigid sigmoidoscopy.
Other published reports also refer to a 50% reduction in colon cancer and a 60% reduction in colon cancer deaths (J Natl Cancer Inst. 1992. 84. 1572-1575; Ann Intern Med. 1995. 123. 904-910; Arch Intern Med. 1995. 155. 1741-1748; Cancer Causes Control. 1998. 9. 455-462).
RCT data
A large multicenter RCT in the UK(the UKFSST) reported a 33% reduction in colorectal cancer incidence and a 43% reduction in colorectal cancer mortality with once-only flexible sigmoidoscopy screening (Lancet. 2010. 375. 1624-1633).
A RCT in Italy (SCORE trial) reported a reduction in colorectal cancer incidence, but no reduction in colorectal cancer mortality, with once-only flexible sigmoidoscopy screening (J Natl Cancer Inst. 2011. 103. 1310-1322). A study examining long term follow up of this cohort concluded that “the strong protective benefit of a single flexible sigmoidoscopy screening for colorectal cancer incidence and mortality was maintained up to 15 and 19 years” (Ann Intern Med. 2022, 175. 36-45 and editorial 129-130).
In an arm of the PLCO trial, in which 154,900 men and women aged 55-74 were randomized to flexible sigmoidoscopy screening versus usual care, with repeat screening at 3-5 years, in those randomized to screening, colorectal cancer incidence was reduced 21% (with benefit observed in both the proximal and the distal colon), and colorectal cancer mortality was reduced 26% (mortality benefit restricted to the distal colon, with a 50% reduction in colorectal cancer mortality in the distal colon). Median follow-up was 11.9 years (N Engl J Med. 2012. 366. 2345-2357). Despite positive results, several factors in this trial might have masked the magnitude of true benefit: (1) 47% of participants in the usual care group underwent flexible sigmoidoscopy, as compared with 83.5% in the flex-sig screening group, (2) only 54% of the flex-sig screening group underwent repeat screening at 3-5 years, and (3) colonoscopy was not performed in 20% of patients who had a polyp detected by flexible sigmoidoscopy (Editorial. N Engl J Med. 2012. 366. 2421-2422).
A RCT in Norway (NORCCAP Trial) in which 100,201 individuals aged 50-64 years old were randomized either to control, once-only flexible sigmoidoscopy screening or once-only flexible sigmoidoscopy screening and FOBT testing showed a reduced incidence of and mortality from colorectal cancer in both treatment groups. There was 20% relative reduction in colorectal cancer incidence and a 27% relative reduction in colorectal cancer mortality, at an average of 11 years of follow up. This translates to one less colorectal cancer mortality per 1000 persons screened. The mortality curves in treatment groups versus control group did not diverge until the 9th year, suggesting that there may be additional benefit with longer follow up. Screening was effective in the 50-54 year old age group, as well as the 55-64 year old age group (JAMA. 2014. 312. 606-615 and editorial 601-602).
A pooled analysis of the above four RCTs (n = 358,204) of individuals aged 55 to 64 concludes that “sigmoidoscopy screening demonstrates a significant and sustained effect of sigmoidoscopy on CRC incidence and mortality for 15 years.” The individual trials were done in Norway (NORCCAP), USA (PLCO), UK, (UKFSST) and Italy (SCORE). Attendance rates were 58-84% in these trials (Ann Intern Med. 2022. 175. 1525-1533).
Note a prior pooled analysis of 3 trails with 10-12 years of follow up (n = 288,000) also showed benefit of screening, but a lesser magnitude of benefit in older women (BMJ. 2017. 356. i6673).
Cost effectiveness: data suggests a cost of $10,000 - $25,000 per year of life saved (Ann Intern Med. 2002. 137. 132-141). USPSTF III states that cost is likely less than $30,000 per year of life gained.
The case against screening:
It is imperfect - approximately 70% of patients with proximal colon cancer do not have a distal "sentinel" lesion (Arch Intern Med. 1994. 154. 185-856). It is estimated in terms of absolute percentages in the population that 1% - 3% of adults have advanced proximal lesions that would not be detected by sigmoidoscopy alone (Gastrointest Endosc Clin N Am. 2002. 12. 41-51).
It is not risk free - bowel perforations occur in 1 per 25,000-50,000 exams in centers of excellence (J Natl Cancer Inst. 2003. 95. 230-236), and 3.2% (40 of 1235 patients) reported post procedure bleeding (Gut. 1998. 42. 560-565). Serious complications estimated to occur in 3.4 per 10,000 procedures, perforation in 4.6 per 100,000 procedures (Ann Intern Med. 2008. 149. 627-637).
Cost:
$100/scope for 50 million Americans 50-75 = $1 billion/year if screening is done every 5 years.
Downstream costs: 10% will have adenomatous polyps on sigmoidoscopy. 1 million colonoscopies at $1000 = $1 billion/year.
Surveillance colonoscopy adds more cost.
Colonoscopy
The ACS and USPSTF state that colonoscopy every 10 years is an option for screening for colorectal cancer.
The initial data supporting screening colonoscopy came from the National Polyp Study (NPS), in which 9112 patients underwent colonoscopy, and 3778 underwent polypectomy. This study demonstrated that patients who had polypectomies developed colorectal cancer up to 90% less than untreated historical controls (N Engl J Med. 1993. 329. 1977-1981). The unanswered question was whether or not the lower incidence of cancer translated into a mortality benefit. Follow up data (median of 15.8 years of follow up) of patients who underwent polypectomy in the NPS show a 53% reduction in death rate (95% CI 0.26 – 0.80. absolute risk of colorectal cancer of 0.8% as compared with 1.5%) in this cohort, as compared with the death rate from colorectal cancer in the SEER Program, with the latter considered a representative sample of the general population (N Engl J Med. 2012. 366. 687-696). NOTE that this data is based on a cohort in which 100% of the population complied with screening, which is not a real-life scenario. NOTE also that the SEER cohort had a higher all-cause mortality than the NPS cohort, and this may bias the results (Editorial. N Engl J Med. 2012. 366. 759-760).
Prospective observational data from participants in the Nurses’ Health Study and the Health Professionals Follow-up Study, in 88,902 participants followed over 22 years – relative risk in those screened with colonoscopy was 0.32 [as compared with a RR of 0.59 in those screened with sigmoidoscopy (N Engl J Med. 2013. 369. 1095-1105 and editorial 1164-1166).
Population-based case-control studies suggest that colonoscopy markedly reduces the risk of colorectal cancer (Gastroenterology. 2010. 138. 870-876; Ann Intern Med. 2011. 154. 22-30).
In a population-based case control study in Germany, with 1688 case patients with colorectal cancer and 1932 control participants age 50 or older, colonoscopy in the preceding 10 years was associated with a 77% lower risk of colorectal cancer, with an 84% lower risk of left-sided cancer, and a 56% lower risk of right sided cancer (Ann Intern Med. 2011. 154. 22-30 and editorial 68-69).
In a population-based RCT (NordICC; conducted in Norway, Poland and Sweden; n = 85,179) of persons aged 55 to 64, a single colonoscopy screening invitation versus no invitation reduced CRC incidence at 10 years, but did not reduce mortality at 10 years. Participation rate was only 42% in this trial (N Engl J Med. 2022. 387. 1547-1556).
Cost-effectiveness: In a hypothetical cohort of 100,000 people, a Markov model indicates that colonoscopy every 10 years starting at age 50 prevents 4428 colorectal cancers and saves 7951 life years at a cost of $10,983 per year of life saved (Gastroenterology. 2002. 122. 78-84).
Screening interval
In a study in 2436 persons with no adenomas on screening colonoscopy, 51.6% of these individuals were re-screened a mean of 5.34 years later; none were found to have cancer, and only 1.3% were found to have an advanced adenoma (50% of advanced adenomas on re-screening were distal to the splenic flexure). Even though the percentage of individuals who were re-screened was low at 51.6%, sensitivity analysis performed by the authors showed that the risk of advanced adenoma at most would be 1.9% at 5 years (N Engl J Med. 2008. 359. 1218-1224 and editorial 1285-1287).
According to a validated microsimulation model, “Compared with the currently recommended strategy of continuing colonoscopy every 10 years after an initial negative examination, rescreening at age 60 years with annual HSFOBT (highly sensitive guaiac fecal occult blood testing), annual FIT (fecal immunochemical testing), or CTC (computed tomographic colonography) every 5 years provides approximately the same benefit in life-years with fewer complications at a lower cost” (Ann Intern Med. 2012. 157. 611-620 and editorial 673-674).
Prospective observational data from participants in the Nurses’ Health Study and the Health Professionals Follow-up Study, in 88,902 participants followed over 22 years – the decreased risk of colon cancer in those screened was stable for up to 10 years, except among those with a first degree relative with colon cancer (N Engl J Med. 2013. 369. 1095-1105 and editorial 1164-1166).
Downstream cost is significant
It is estimated that by age 50 1/3 of US population has adenomatous polyps, by age 70 1/2 of US population has adenomatous polyps.
A trial examining the effectiveness of screening at age 40-49 found that tubular adenomas were seen in 8.7% of the 906 participants, for a number needed to screen of 12; 3.5% of the subjects had an advanced adenoma for a number needed to screen of 29. There is currently no data on cost effectiveness of screening average risk individuals under age 50 (N Engl J Med. 2002. 346. 1781-1785).
Risk is not negligible
A study of patients in VA Medical Centers found that 10 of 3121 patients (0.3%) had major complications during or immediately after the procedure (N Engl J Med. 2000. 343. 162-168).
Perforation occurs in an estimated 3.8 per 10,000 procedures, major bleeding in 12.3 per 10,000 procedures, all ‘serious complications’ in 25 per 10,000 procedures (Ann Intern Med. 2008. 149. 627-637).
The 30-day procedure-related mortality is 3 per 100,000 (Gastrointest Endosc. 2019. 90. 863-876).
Hot snare polypectomy using electrocautery can cause delayed bleeding, post polypectomy syndrome, or perforation, with the risk ranging from 0.07% to 1. 26% in healthy individuals, and as high as 14% in participants receiving anticoagulants (Endoscopy. 2022. 54. 290-298; Gastrointest Endosc. 2014. 79. 417-423). In an open label single blinded trial of 40-70 participants, cold snare polypectomy was associated with reduced risk, as compared with hot snare polypectomy (Ann Intern Med. 2023. 176. 311-319).
Sensitivity is far from perfect
Studies of tandem colonoscopy in which two practitioners perform colonoscopy in tandem on the same day indicate miss rates of 13% for adenomas smaller than 1 cm (GastrointestEndosc. 1991. 37. 125-127), 27% for adenomas smaller than 5 mm, and 6% for adenomas of at least 1 cm (Gastroenterology. 1997. 112. 24-28; Gastroenterology. 2006. 101. 343-350).
Chemoprevention trials show that in 0.3 – 0.9% who had all identified polyps removed at baseline colonoscopy develop invasive colorectal cancer within three years (Gastroenterology. 2005. 129. 34-41).
In a population based study, 6% of patients with newly discovered right sided colon cancer had undergone colonoscopy within 6 months to 3 years prior to the diagnosis, suggesting a substantial miss rate in the community setting (Gastroenterology. 2005. 132. 96-102).
Higher sensitivity is associated with slower endoscopy withdrawal rates (N Engl J Med. 2006. 355. 2533-2541).
A population based case-control study in Ontario in which administrative claims data was used found that colonoscopy was associated with fewer deaths from colon cancer, but the benefit was primarily limited to deaths from cancer developing on the right side of the colon (Ann Intern Med. 2009. 150. 1-8). There are a number of nuances which are pertinent to the findings of this study, as listed below (Editorial. Ann Intern Med. 2009. 150. 50-52). Nonetheless, additional population-based studies reported similar results, with little or no protection against right sided cancer (J Natl Cancer Inst. 2010. 102. 89-95; Gastroenterology. 2010. 139. 1128-1137).
The database does not distinguish whether colonoscopy was done for screening or for diagnosis, based on symptoms. However, the criteria used of colon cancer diagnosis > 6 months after colonoscopy should capture screening colonoscopies rather than diagnostic colonoscopies. The discrepancy between findings of benefit for left sided colon lesions and lack of benefit with right sided colon lesions, and the fact that the magnitude of the benefit for left sided lesions is similar to that seen in previous case control studies of sigmoidoscopy tend to indicate that the results of this study are valid.
It may be that a greater percentage of right sided lesions are flat rather than pedunculated, making detection more difficult.
It is conceivable that a greater percentage of right sided lesions grow quickly, thus causing disease before the next screening colonoscopy is performed.
70% of colonoscopies were done in this study by internists and surgeons, who might not be as skilled at detecting right sided lesions as gastroenterologists or colorectal surgeons.
An exam recorded as ‘complete’ might not have truly reached the cecum.
Colon preparation might have been poor, obscuring lesions on the right side of the colon.
A systematic review and meta-analysis of tandem colonoscopies concludes that there is 15 - 30% miss rate for neoplasia (Gastroenterology. 2019. 156. 1661-1674).
Factors associated with lower sensitivity of screening
Performance characteristics – indicators of a quality endoscopy include > 6 minute withdrawal time, adenoma detection rate of at least 25% for average risk women and men, cecal intubation rate > 95% (N Engl J Med. 2010. 362. 1795-1803; ESGE guideline - update 2020. Endoscopy. 2020. 52. 687-700).
Practitioner factors – fewer missed lesions when exam conducted by a gastroenterologist rather than a GP or surgeon, when exam conducted in a hospital setting rather than an office (Gastroenterology. 2007. 132. 96-102).
Patient factors – increasing age, female sex, presence of diverticulosis, comorbidities, quality of the prep (Singh H et al. Am J Gastroenterol. 2010; Gastroenterology. 2007. 132. 96-102).
USPSTF III (2002) stated it was unclear if the procedure's increased accuracy offsets its additional complications, inconvenience, and costs.
Recommendations from USPSTF updated to recommend against routine screening in adults age 76-85 and recommends against screening in adults older than age 85 (Ann Intern Med. 2008. 149. 627-637).
CT Colonography (CTC)
This is recommended as an option for screening as of 2008, based on a joint guideline issued by the ACS, US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology (Gastroenterology. 2008. 134. 1570-1595). Previous consensus guidelines had suggested performing every 5 years if negative, performing in 3 years if 6-9 mm polyp identified, and not reporting polyps 0-5 mm to the patient (Radiology. 2005. 236. 3-9). However, the USPSTF concludes that evidence is insufficient to assess the benefits and harms (Ann Intern Med. 2008. 149. 627-637).
This is also referred to as virtual colonoscopy, and involves imaging by CT scan after a bowel prep (Editorials. Mayo Clin Proc. 2007. 82. 659-661 and 662-664).
First described in an abstract published in 1994.
Patient must be able to lie prone and supine, and hold breath for 15-20 seconds.
No sedation required, entire exam takes 20 minutes.
Complications are lower than with colonoscopy, and patient acceptance is higher, BUT there is radiation exposure, with one report concluding that a single CT colonographicexamination in a 50 year old producing a 1 in 714 risk of a radiation-induced tumor (Gastroenterology. 2005. 129. 328-337), and colonoscopy is required if a > 5 mm lesion is identified.
Much less sensitive than colonoscopy for detection of adenomas less than 0-5 mm.
Specificity is low – in the multisite study of 2600 asymptomatic patients referred to just below, 17% had a finding considered positive for a polyp > 1 cm, but only 25% actually had a polyp > 1 cm on subsequent colonoscopy.
Extracolonic findings are common - in the multisite study of 2600 asymptomatic patients referred to just below, 16% had an extracolonic finding worrisome enough to trigger additional testing; effective treatment was available for very few of these incidental findings.
Cost effectiveness – 2005 data indicates that this costs $3600 more per life-year saved than conventional colonoscopy (ACP Journal Club. 2005. 143. 78).
Evidence supporting this modality of screening
A multisite study of 2600 asymptomatic patients in which colonoscopy was used as the gold standard showed that CT colonography had a 90% sensitivity for detection of polyps > 1 cm, and a 78% sensitivity for detecting polyps > 6 mm (N Engl J Med. 2008. 359. 1207-1217 and editorial 1285-1287).
A prior meta-analysis of 33 studies found a high specificity but an unacceptably low sensitivity for detection of adenomatous polyps (Ann Intern Med. 2005. 142. 635-650).
Novel methods of screening
Immunochemical fecal occult blood testing (i-FOBT or FIT) – recommended annually in 2012 by some authorities as preferable to FOBT testing for colorectal cancer screening.
Developed to improve specificity and eliminate the need for dietary restriction.
Detects human hemoglobin in fecal samples (via monoclonal or polyclonal antibodies that bind to intact human hemoglobin).
Some FIT’s have poor stability at room temperature and need more rapid transport to the laboratory (or refrigeration), whereas FOBT’s are relatively stable and can be analyzed up to 21 days after collection (Int J Cancer. 2009. 125. 646-750).
Medicare reimburses $22 for the test in 2010.
This is recommended as an option for screening as of 2008, based on a joint guideline issued by the ACS, US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology (Gastroenterology. 2008. 134. 1570-1595).
Sensitivity of 66% for detecting colon cancer, but only 27% for polyps > 1 cm, based on data in 21,805 asymptomatic adults who underwent one-time immunochemical FOBT and simultaneous colonoscopy. Specificity 95% (Gastroenterology. 2005. 129. 422-428).
Sensitivity of 88% for detecting colon cancer, with a specificity of 90%, and sensitivity of 62% for detecting significant neoplasia, with a specificity of 93%, at the manufacturer-recommended threshold of 100 ng/ml. If a threshold of 50 ng/ml used instead of 100 ng/ml, 100% sensitivity for detecting cancer, with an 84% specificity. Data gathered in 1000 consecutive ambulatory patients who were scheduled to undergo colonoscopy (either for symptoms or screening) and agreed to participate in this study. Despite the nature of the study group, the prevalence of cancer was similar to that in screening colonoscopy studies, and 16 of the 17 cases of cancer discovered by i-FOBT screening were in an early stage (Ann Intern Med. 2007. 146. 244—255 and editorial 309-311).
A prospective study in which 1319 participants at average risk for colon cancer who were to undergo colonoscopy for cancer screening were evaluated prior to colonoscopy with one of six different qualitative immunochemical FOBT tests showed that performance characteristics of i-FOBT were better than guaiac-based FOBT, but there was tremendous variability amongst the i-FOBT tests (Ann Intern Med. 2009. 150. 162-169).
Additional studies show that semi-quantitative FIT is more accurate than guaiac testing for detection of colorectal cancer and advanced adenomas (Int J Cancer. 2011. 128. 2415-2424; J Gastroenterol. 2010. 45. 703-710; Br J Cancer. 2009. 100. 1103-1110; Gastroenterology. 2008. 135. 82-90; Eur J Cancer. 2008. 44. 2254-2258).
A blinded comparison of FIT with screening colonoscopy in 1256 asymptomatic persons showed that using a quantitative cutoff of 50 ng/ml, a single FIT had a 35% sensitivity and 93% specificity for detecting advanced adenoma, and 38% sensitivity and 93% specificity for detecting advanced neoplasia (Am J Gastroenterol. 2012. 107. 1570-1578 and ACP Journal Club. 2012. 157. JC5-10).
A systematic review and meta-analysis of 19 eligible studies concluded “Fecal immunochemical tests are moderately sensitive, are highly specific, and have high overall diagnostic accuracy for detecting CRC. Diagnostic performance of FITs depends on the cutoff value for a positive test result (Ann Intern Med. 2014. 160. 171-181).
A retrospective cohort study of 323,349 members of Kaiser Permanente, ages 50-70 concluded "Annual FIT screening was associated with a high sensitivity for CRC with high adherence to annual follow up screening among initial participants" (Ann Intern Med. 2016. 164. 456-463).
FIT versus gFOBT (Ideas and Opinions. Ann Intern Med. 2017. 166. 297-298 and this references data in the 2016 US Multi-Society Task Force on Colorectal Cancer consensus statement).
In pooled analyses, sensitivity of FIT is 79% as compared with approximately 35% for gFOBT.
In a meta-analysis, FIT detected twice as many colorectal cancers and advanced adenomas as gFOBT.
Patients prefer FIT to gFOBT.
Fecal DNA analysis (requires collecting and shipping an entire bowel movement)
First generation test marginally more sensitive than ordinary fecal OB testing, but much more expensive; second generation test significantly more sensitive than fecal OB testing, but with poor specificity (Ann Intern Med. 2008. 149. 441-450 and editorial 509-510).
USPSTF concludes that evidence is insufficient to assess the benefits and harms (Ann Intern Med. 2008. 149. 627-637).
This is recommended as an option for screening as of 2008, based on a joint guideline issued by the ACS, US Multi-Society Task Force on Colorectal Cancer, and the American Collegeof Radiology (Gastroenterology. 2008. 134. 1570-1595).
In a cross sectional study of 9989 participants at 90 sites, comparing one time second generation DNA fecal analysis with one time FIT testing, DNA testing was more sensitive, but less specific. Using colonoscopy as the gold standard, sensitivity for detection of colorectal cancer 92.3%, sensitivity for detection of advanced adenoma 42.4%, specificity 90% (N Engl J Med. 2014. 370. 1287-1297 and editorial 1350-1351).
Multitarget stool DNA testing every 3 years in conjunction with FIT testing annually
Increases sensitivity, but decreases specificity, as compared with FIT testing alone (N Engl J Med. 2014. 370. 1287-1297).
Next-generation testing, consisting of updated DNA biomarkers, shows preservation of sensitivity, with improved specificity, based on indirect comparison with the first-generation test, in a prospective study of 20,176 participants (N Engl J Med. 2024. 390. 984-993 and editorial 1045-1046).
Epi proColon blood- based screening - approved by FDA 4/13/16. Relies upon quantitative detection of the methylated septin 9 gene (SEPT9) [Viewpoint. JAMA. 2016. 315. 2519-2520].
Rectal exam with an optical probe to assess nanoscale architectural manifestations of adenomas via “light scattering signatures.” Accurate in experimental models, but no clinical studies yet.
Cell-free DNA (cfDNA) from whole blood ("liquid biopsy") to detect genomic alterations, aberrant methylation, and DNA fragment changes.
Cell-free DNA is a composite of extracellular DNA molecules found in bodily fluids. Short fragments of DNA molecules make up the majority of cfDNA. The fragmentation patterns have been found to carry diagnostic information - the study of these fragments is named fragmentomics. The epigenetic signature of the tissue of origin is carried by cfDNA (Lu YMD et al. Science. 2021. 372. eaaw3616 and editorial. N Engl J Med. 2024. 390. 1047-1049).
In the ECLIPSE trial of 7861 patients subjected to one-time testing, this test (commercially available assay - Shield test) had a 83% sensitivity for colorectal cancer, 90% specificity for advanced neoplasia, and 13% sensitivity for advanced precancerous lesions. Note, the sensitivities in the study are lower and the sensitivities reported for next-generation multi target DNA stool assay (N Engl J Med. 2024. 390. 973-983 and editorial 1045-1046).
Primary prevention
Fruits and vegetables, folate, adequate calcium and vitamin D, cereal fiber, and consider selenium supplements and zinc supplements, limit alcohol intake, quit smoking.
Aspirin
Daily intake of aspirin has been demonstrated to decrease the risk of colorectal cancer in multiple cohort and case control studies. Data would suggest that at least a decade of use is necessary to statistically significantly reduce the risk of colorectal cancer.
Aspirin reduces polyp numbers in patients with familial polyposis.
Daily aspirin starting at age 50, in a hypothetical cohort of 100,000 people, would prevent 2952 colorectal cancers and save 5301 life years at a cost of $47,249 per year of life saved, using a Markov statistical model. Using this model, colonoscopy screening every 10 years is more cost effective, preventing 4428 colorectal cancers and saving 7951 life years at a cost of $10,983 per year of life saved. Combining colonoscopy with aspirin would prevent an additional 50% of cancers over colonoscopy alone, but at a cost of more than $200,000 per year of life saved (Gastroenterology. 2002. 122. 78-84).
A prospective cohort study including 27,077 women in the Nurses' Health Study showed that regular short term use of aspirin (< 5 years) is inversely associated with risk of colorectal adenoma, with the greatest protective effect present at high doses (>14 tablets per week). The relative risk of adenoma in users of one aspirin per week was 0.80 compared to non-users, and the relative risk in users of >14 tablets per week was 0.49 compared to non-users. (Ann Intern Med. 2004. 140. 157-166).
In a RCT of 272 patients S/P removal of an adenomatous polyp by colonoscopy, fewer of those on aspirin had adenomatous polyp recurrence by colonoscopy at one year (30% versus 41%), and furthermore fewer of those on 300 mg per day of aspirin had recurrences than those on 160 mg of aspirin (Gastroenterology. 2003. 125. 328-336).
Colon cancer screening in those with a positive family history
If first degree relative developed cancer after age 60, USPSTF III states that it is reasonable to initiate screening at an earlier age.
Expert opinion is to initiate screening at an age 10 years younger than the diagnosis of colorectal cancer in the first degree relative, and to screen with colonoscopy every 5-10 years.
Prospective observational data from participants in the Nurses’ Health Study and the Health Professionals Follow-up Study (88,902 participants followed over 22 years) supports a recommendation to repeat colonoscopy every 5 years in those with a first degree relative with colon cancer – the decreased risk of colon cancer in those screened was stable for up to 10 years, except among those with a first degree relative with colon cancer (N Engl J Med. 2013. 369. 1095-1105 and editorial 1164-1166).
Colorectal Cancer 10 year risk Lifetime risk
Age 40 - 50 0.2%
Age 50 - 60 0.7% 6.0%
Age 60 - 70 1.5%
Colon cancer screening in those over age 75 and with no prior screening (Ann Intern Med. 2014. 160. 750-759 and editorial 804-805)
In persons age 76 or older without previous screening, prevalence of colorectal cancer is nearly 10 times that of older persons with negative screening colonoscopy at ages 50, 60, and 70.
Effects of screening in older individuals are as follows, based on meticulously developed model, and with a threshold of $100,000 per QUALY as a cost effectiveness threshold
In persons with no comorbid conditions, colonoscopy is cost effective until age 83, sigmoidoscopy until age 84, and fecal immunochemical testing until age 85-86.
In persons with moderate and severe comorbid conditions, colonoscopy is cost effective until age 80 and 77, sigmoidoscopy until age 81 and 78, and fecal immunochemical testing until age 82 and 79.
Lung cancer
228,150 cases/year; 142,670 deaths/year [2019 data]
Current recommendation – the USPSTF in 2021 recommends annual low dose helical CT of the chest for smokers aged 50-80 who have a 20 pack-year history of smoking and who currently smoke or who have quit within the past 15 years; Grade B recommendation (JAMA. 2021. 325. 962-970, 971-987, 988-997, Viewpoint 933-934 and Editorial 939-941).
Historical: the USPSTF in 2013 recommended annual low dose helical CT of the chest for smokers aged 55-80 who have a 30 pack-year history of smoking and who currently smoke or who have quit within the past 15 years (Ann Intern Med. 2014. 160. 330-338 and editorial 363-364). This Grade B recommendation is based on a systematic review which concludes “Strong evidence shows that LDCT screening can reduce lung cancer and all-cause mortality. The harms associated with screening must be balanced with the benefits” (Ann Intern Med. 2013. 159. 411-420). This recommendation is supported by a comparative modeling study for the USPSTF which examined 5 independent models and concluded “Annual CT screening for lung cancer has a favorable benefit-harm ratio for individuals aged 55 through 80 with 30 or more pack-years’ exposure to smoking” (Ann Intern Med. 2014. 160. 311-320).
HISTORICALLY, a prospective RCT at Johns Hopkins (Chest. 1986. 89S. 324S-325S) and another prospective RCT at Memorial Sloan-Kettering (Chest. 1984. 86. 44-53) comparing patients receiving annual CXR and pooled sputum cytologic testing every four months with a control group receiving only annual CXR showed no difference in 5 year mortality rates.
HISTORICALLY, a prospective RCT at Mayo Clinic (J Occup Med. 1986. 28. 746-750) comparing CXR and sputum cytologic testing every 4 months for 6 years in 4618 men older than 45 and smoking at least 1 ppd with routine care in 4593 subjects showed no difference in lung cancer mortality between the two groups despite detection of more cancers in the screened group [206] than in the control group [160]. The chief flaw in this study is that approximately 50% of men in the control group did receive a CXR during the study period (Cancer. 1991. 67. 1164-1191).
Helical (spiral) CT allows acquisition of data from the entire thorax within 15-20 seconds, and motion artifact is minimized because the entire scan is typically completed during a single breath hold. Radiation exposure is 1/6 that of conventional CT of the chest, and only 10 times that of a CXR. No contrast administration is necessary.
In the ELCAP study (Lancet. 1999. 354. 99-105), 90% of the cancers detected by low dose spiral CT were resectable and 10 year survival approached 90% in those screened with helical CT (N Engl J Med. 2006. 355. 1763-1771).
Limitations of this trial include a positive predictive value of only about 10% due to many false positives, lack of a control group (i.e. it was not a RCT), the lack of an unbiased outcome measure, and the lack of quantification of the harms of screening (Arch Intern Med. 2007. 167. 2289-2295).
In the NLST study, a U.S. randomized, controlled study in 33 clinical centers of 53,454 patients, 55-74 years of age (59% men), who had a cigarette smoking history of > 30 pack years (and, if former smokers, had quit within the past 15 years), 3 annual screenings with low-dose CT, using multidetector scanners, was associated with a 20% relative reduction in lung cancer mortality, accrued over an average of 6 years of follow-up (N Engl J Med. 2011. 365. 395-409).
The absolute reduction in risk of death from lung cancer was 0.3%, from 1.7% over 6 years to 1.4 % over 6 years. Thus for every 1000 participants screened, 3 lung cancer deaths were averted and 14 lung cancer deaths were not averted. For an individual, the magnitude of benefit will depend on age, sex, smoking history, and exposure to asbestos (Ideas and Opinions. Ann Intern Med. 2012. 157. 571-573).
In a 62 year old man with a 52 pack-year smoking history, who is a current smoker, screening will reduce his reduce his risk of lung cancer from 1.95% over 6 years to 1.56% over 6 years, for an absolute risk reduction of 0.39%, and meaning that 256 individuals would need to be screened to prevent 1 lung cancer death.
In a 70 year old current smoker with a 110 pack-year smoking history, screening will reduce his reduce the risk of lung cancer from 6.09% over 6 years to 4.87% over 6 years, for an absolute risk reduction of 1.22%, and meaning that 82 individuals would need to be screened to prevent 1 lung cancer death.
In a 40 year old female smoker who smoked 1 ppd for 10 years and quit 15 years ago, screening will reduce her lung cancer risk from 0.01% to 0.008%, meaning that 35,186 individuals would need to be screened to prevent 1 lung cancer death.
False positives were common – if 308 patients were screened, extrapolation of data from this trial implies that these individuals would undergo 985 CT scans, 18 PET scans, 8 bronchoscopies, and 9 surgical procedures to yield 8 diagnoses of lung cancer and prevent 1 additional lung-cancer-related death (Commentary. ACP Journal Club. 2011. 155. JC5-6).
Overdiagnosis is a possibility – overdiagnosis is defined as the identification of tissue that has all of the pathologic indications of cancer on biopsy, but if not treated will not progress to cause symptoms or death, and may even regress.
A cohort study of 175 patients who underwent low-dose CT screening annually for 5 years showed that slow growing or indolent cancer comprised about 25% of incident cases (Ann Intern Med. 2012. 157. 776-784).
In another paper, overdiagnosis was estimated at 29% (JAMA Intern Med. 2014. 174. 1198).
In the NELSON study, a European randomized, controlled study of 15,792 participants (84% men) between the ages of 50-74, those randomized to four screenings (baseline, 1 year, 3 years, 5.5 years) had a lower mortality rate at 10 years (24% lower in the men in the trial and 33% lower in subgroup analysis of the women in the trial). Overdiagnosis in this trial was estimated at 10% - fewer cases of overdiagnosis than of lives saved from screening (N Engl J Med 2020. 382. 503-513 and editorial 572-573).
Ovarian cancer
22,530 cases/year; 13,980 deaths/year [2019 data] 1.5% lifetime risk
The inverse relationship between survival and stage at diagnosis (80-90% 5 year survival for Stage I, 10% survival for Stage IV) combined with the lack of early symptoms calls for a screening strategy.
Any screening strategy must have a high specificity (i.e. at least 99.6%) because the condition of ovarian cancer is rare and a positive screening test necessitates a surgical procedure for definitive diagnosis).
USPSTF III (2004) recommends against routine screening for ovarian cancer (D recommendation). Fair evidence that screening with CA-125 or transvaginal ultrasound can detect ovarian cancer at an earlier stage than it can be detected in the absence of screening, but also fair evidence that effect of screening on mortality would be small at best, and fair evidence too that screening could lead to important harms, related to the low incidence of ovarian cancer and false positive results which would lead to unnecessary surgical procedures.
The D recommendation (against screening) was reaffirmed in 2012
The D recommendation (against screening) was reaffirmed in 2018
Pancreatic cancer
56,770 case/year; 45,750 deaths/year[2019 data]
Rationale for screening is that this is the third leading cause of death in the United States, due to a 5-year survival rate of 9.3% based on SEER data from 2009 - 2015 (but 5-year survival rate for localized pancreatic cancer is 37.4%, 5-year survival rate of 12.4% when regional disease is present and 5-year survival rate of 2.9% when distant metastatic disease is present (JAMA. 2019. 322. 438-444).
USPSTF in 2004 and again in 2019 recommends against screening in asymptomatic adults, primarily based on the low incidence of disease in the population of 12.9 cases per 100,000 person years (JAMA. 2019. 322. 438-444).
HOWEVER, the USPSTF guideline does not address individuals at high risk for pancreatic cancer and individuals with greater than a 5% lifetime risk or a 5-fold increased relative risk may benefit from screening (Editorial. JAMA. 2019. 322. 407-408).
Individuals from familial pancreatic cancer kindrids are at increased risk
Individuals with genetic mutations such as BRCA1 and BRCA2 are at increased risk
Individuals with new onset diabetes have up to an 8-fold increased risk of pancreatic cancer
Prostate cancer
174,650 cases/year; 31,620 deaths/year [2019 data] 15% lifetime risk of clinically evident prostate cancer; 3.4% mortality
Rationale for screening
Prostate cancer is slow growing.
Treatment of pathologically organ-confined prostate cancer is successful; retrospective analysis of John's Hopkins data from 1909-1963 shows mortality in men treated with radical prostatectomy is comparable to age matched controls.
Guidelines for screening
American Cancer Society (ACS) and American Urologic Association (AUA) - digital rectal exam (DRE) and prostate specific antigen (PSA) yearly for all men age 50 with a life expectancy of at least 10 years; screen black men and men with positive family history yearly starting at age 40, because prostate cancer is 50% more common in blacks and 100% more common in men with a first degree relative diagnosed with prostate cancer.
Follow up abnormal DRE with transrectal ultrasound (TRUS) guided biopsy.
Follow up abnormal PSA with TRUS.
If positive, TRUS guided biopsy.
If negative, Calculate PSA density (PSA/prostate volume).
If PSA density >0.15, TRUS guided sextant biopsy.
Do not use TRUS for routine screening because it has a high cost, is operator dependent, and there is no data that it is beneficial independent of DRE and PSA.
American College of Physicians (ACP), American College of Preventive Medicine (ACPM), and Canadian Task Force on Preventive Health Care all recommend against PSA and against DRE based in part on the risk of false positive results, and in part on lack of documented mortality benefit.
The United States Preventive Services Task Force (USPSTF) in 2018 states that "For men aged 55 to 69 years, the decision to undergo periodic PSA-based prostate cancer screening should be an individual one and should include discussion of the potential benefits and harms of screening with their clinician. Screening offers a small potential benefit of reducing the chance of death from prostate cancer in some men. However, many men will experience potential harms of screening, including false positive results ... overdiagnosis and overtreatment ... treatment complications (C recommendation. The USPSTF recommends against PSA-based screening for prostate cancer in men 70 years and older (D recommendation)." (JAMA. 2018. 319. 1901-1913). The change from the D recommendation to the C recommendation is based primarily on longer term follow up from the ERSPC trial which shows that the absolute reduction in prostate cancer mortality has increased from 7 per 10,000 men at 9 years to 13 per 10,000 men at 13 years (Editorial. JAMA. 2018. 319. 1866-1868); mathematical modeling suggests that absolute reduction in mortality will decline further with longer follow up (J Clin Epidemiol. 2011. 64. 1412-1417).
Historically, in 2002, and again in 2008 (Ann Intern Med. 2008. 149. 185-191), recommendation iwa a grade I recommendation (i.e. current evidence insufficient), because although there is "good evidence that PSA screening can detect early-stage prostate cancer, there is mixed and inconclusive evidence that early detection improves health outcomes" (Ann Intern Med. 2002. 137. 917-929 and editorial 1866-1888).
In 2008, the USPSTF recommended against screening in men over age 75 (grade D recommendation).
In 2012, USPSTF recommended against prostate cancer screening (grade D recommendation). The grade D recommendation was criticized in Perspective articles (N Engl J Med. 2011. 365. 1951-1953 and 1953-1955) and in Commentaries (JAMA. 2011. 2715-2716 and 2717-2718 and 2719-2720 and 2721-2722).
Case for screening (Level 3 evidence)
Screening by PSA: 63% of cancer is surgically organ-confined (may be 70%).
Screening by DRE: 48% of cancer is surgically organ-confined (may be only 30%).
In men with palpable clinically localized prostate cancer, development of metastases and death is 50% lower in men undergoing radical prostatectomy (Cancer. 1993. 72. 310-322).
90% of cancers detected by PSA are 0.5 cc, which suggests PSA is not so sensitive that it detects the incidental microscopic cancers found at autopsy.
Case against screening
Overdiagnosis - defined as the identification of tissue that has all of the pathologic indications of cancer on biopsy, but if not treated will not progress to cause symptoms or death, and may even regress.
Autopsy data shows that prostate cancer is very common (30% prevalence in men age 50-59, 40% prevalence in men age 70-79, and 67% prevalence in men > age 80). A 50 year old man with a life expectancy of 25 years faces a 42% of microscopic prostate cancer, but only a 10% risk of clinically apparent prostate cancer and 3% risk of death (Lancet. 1994. 343. 1263-1267).
Estimated at 50% based on 15-year follow up of a population-based cancer screening study (J Urol. 2009. 181. 1615-1621).
Estimated at 23-42% in the US population and as high as 67% in a European randomized screening trial (Eur Urol. 2014. 65. 1046-1055).
Complications - risk of sepsis approaches 4% after biopsy (J Urol. 2013. 189 (1 suppl). 512-517).
Limited magnitude of benefit/expensive from a public health perspective - in analysis using a Markov model, which is an accepted method for evaluating clinical problems not adequately assessed by randomized trials, found that a one-time screening effort with DRE and PSA increased life expectancy of men age 50-70 by only 1-2 days, but caused an estimated 3-13 day decrease in QALY (quality adjusted life expectancy). Estimated cost/year life saved is between $113,000 and $729,000 (JAMA. 1994. 272. 770-780).
RCT data
CAP trial (Cluster Randomized Trial of PSA Testing for Prostate Cancer), This trial included 419,582 men aged 50-69 and was conducted at 573 primary care practices across the United Kingdom. At a median follow up of 10 years, there was no significant difference in mortality between men in practices randomized to a single PSA screening intervention versus those men in practices randomized to no screening (JAMA. 2018. 319. 883-895 and editorial 868-869).
PLCO trial (Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial), sponsored by the NCI. 76,693 men aged 55-74 at 10 US study centers randomized to receive either annual screening (PSA annually for 6 years and digital rectal exam yearly for 4 years) or usual care. Compliance was 85% for PSA testing and 86% for DRE. Usual care sometimes included screening; cross sectional surveys of men in the usual care group indicated that 52% had undergone at least a single PSA screen by year 5 of the study. 34% of men participating in the trial had a PSA test within 3 years prior to enrolling in the trial. Adjudication of causes of death were made by a committee whose members were unaware of study group assignments, but were aware of treatment. In an interim report, no mortality benefit during a median follow up of 11 years (N Engl J Med. 2009. 360. 1310-1319 and editorial 1351-1354). Based on 32% of men with a PSA test in the 3 years prior to enrollment in the trial, and 52%in the control group with at least one PSA test in the first 5 years of the trial, this trial in fact ended up comparing two different screening schedules rather than determining whether screening is effective (Editorial. Cleve Clin J Med. 2009. 76. 446-448). An updated analysis of the data reported no changes in the findings of lack of mortality benefit in the screening group (J Natl Cancer Inst. 2012. 104. 1-8).
ERSPC (European Randomized Study of Screening for Prostate Cancer). 182,000 men identified through registries, and 162,243 men aged 55-69 randomized for the purposes of this study. In the treatment group, PSA was repeated once every 4 years on average; contamination of the control group with screening as a part of usual care is not described in the report. Adjudication of causes of death were made by a committee whose members were unaware of study group assignments, but were aware of treatment. In an interim report, PSA screening without DRE was associated with a 20% reduction in relative risk of death from prostate cancer at a median of 9 years of follow up (p=0.01). This translates into a reduction of about 7 prostate cancer deaths over 9 years per 10,000 men screened. The risk of screening was overdiagnosis - 820 men per 10,000 screened received a diagnosis of prostate cancer, compared to 480 per 10,000 men in the control group. The men in the screening group underwent more than 17,000 biopsies, and 277 per 10,000 underwent radical prostatectomy and 220 underwent radiation therapy, compared to 100 men per 10,000 in the control group undergoing radical prostatectomy and 123 men per 10,000 in the control group undergoing radiation therapy, NNT to prevent one death from prostate cancer is 48 (N Engl J Med. 2009. 360. 1320-1328 and editorial 1351-1354). An updated analysis reported that at a median of 11 years of follow-up (2 additional years of follow-up) the relative risk reduction of death from prostate cancer was 2% (p=0.001), similar to the data reported in 2009. Screening did not alter all-cause mortality (N Engl J Med. 2012. 366. 981-990 and editorial 1047-1048).
Goteberg study – this was a center in the ERSPC trial with 14 years of follow up. 44% relative risk reduction for prostate cancer death despite 24% noncompliance in screening group, for a number needed to diagnose of 12 to prevent one prostate cancer death over 14 years (Lancet Oncol. 2010. 11. 725-732). Note thought that after 22 years of follow-up in the Goteborg screening trial, the rate ratio for prostate cancer death in screening participants was 0.59, with a number needed to invite to screening of 221 and a number needed to treat of 9 to prevent 1 death from prostate cancer (J Urol. 2022. 208. 292-300). This is a significant benefit (Editorial. JAMA. 2024. 331. 1448-1450).
A systematic review and meta-analysis of 6 RCTs (n=387,286) concluded that PSA screening does NOT reduce death from prostate cancer in asymptomatic men (BMJ. 2010. 341. c4543). A commentary critiques this systematic review, and the commentary concludes that the data from the Goteberg study may be more representative of the benefits of PSA screening than the data from this systematic review (ACP Journal Club. 2011. 154. JC1-2).
A review of the evidence for the USPSTF, funded by AHRQ, concludes “Prostate-specific antigen-based screening results in small or no reduction in prostate cancer-specific mortality and is associated with harms related to subsequent evaluation and treatments, some of which may be unnecessary” (Annals Intern Med. 2011. 155. 762-771).
Important issues related to screening
Physicians believe in the benefit of screening even though it is not proven – a national survey conducted by mail in 2000 found that 78% of male primary care physicians and 95% of male urologists over age 50 reported having had a PSA test (J Gen Intern Med. 2006. 21. 257).
Clinically organ-confined disease is often metastatic in the operating room.
The sensitivity of DRE and PSA is unknown.
DRE is probably more sensitive when done by a urologist.
Approximately 20% of clinically significant prostate cancers are missed using a PSA cutoff of 4, and many of these are detected by DRE, so these tests are complementary.
Data from the Prostate Cancer Prevention Trial in which all men had a prostate biopsy at the end of the study, regardless of PSA level or placebo versus finasteride treatment showed that 27% of the 2950 men in the placebo arm had prostate cancer with a PSA of 3.1-4 and 6.6% of men had prostate cancer with a PSA of 0.5 or less. However, only 2.3% of placebo patients with a PSA less than 4 had a high grade cancer (Gleason score of 7 or higher). Conclude that PSA cutoff of 4 is NOT sensitive for detection of prostate cancer but is sensitive for detection of high-grade prostate cancer (New Engl J Med. 2004. 350. 2239-2246).
Approximately 75% of men over age 50 have symptoms related to prostate enlargement; the PSA in this context is used partially for diagnosis, not just for screening.
Once biopsy is done, the Grade of the tumor is the most important prognostic feature, but volume of tumor is also significant.
In men with cancer with a Gleason score of 6 and PSA or 10 or less and clinically localized cancer, reasonable to measure PSA every 3 months for 2 years and then every 6 months, intervene if PSA doubling time is less than 3 years, based on one experts opinion (Internal Medicine News. 5/1/03. 24).
Only 4-8% of cancers with a Gleason score of 6 will progress to high grade cancer after 8 years (Internal Medicine News. 5/1/03. 24).
There is a general consensus that screening is not indicated when life expectancy is less than 10 years.
An isolated elevation in PSA level should be confirmed several weeks later before proceeding with further testing, because a high proportion of men (approximately 30-40% depending on the cutoff used) with an abnormal PSA had a normal PSA level again at subsequent testing during 4 year follow up (JAMA. 2003. 289. 2695-2700).
If screening is beneficial, data would suggest that screening every four years for men with a low normal PSA would be as beneficial as screening every year (J Natl Cancer Inst. 2003. 95. 1462-1466).
A prospective RCT in 5855 men aged 50-66, with a median follow-up of 7.6 years found that in men with PSA levels less than 1 ng/ml, PSA screening every 3 years is sufficient (Arch Intern Med. 2005. 165. 1857-1861).
A 2 -4 year interval of PSA testing, as compared with annual screening might reduce false-positive results and overdiagnosis without substantially sacrificing benefits (Ann Intern Med. 2013. 158. 145-153).
PSA
Serine protease produced by prostatic epithelial cells, purified and characterized in 1979, detected in serum in 1980, approved by FDA in 1995 for early detection.
Non-specific for prostate cancer - also elevated in BPH.
Risk versus benefit:
Benefit: Data from 13-year and 16-year follow up of the ERSPC trial indicate that one must screen 570 men aged 55-69 to prevent one death from prostate cancer at 16 years (Lancet. 2014. 384. 2027-2035; Eur Urol. 2019. 76. 43-51).
The USPSTF prepared an infographic in April 2017 (https://screeningforprostatecancer.org/get-the-facts/) - for every 1000 men offered PSA-based screening over 10-15 years, approximately 240 will have an elevated PSA level and will have a biopsy recommended. 100 will have cancer detected on the biopsy and 80 will choose treatment (65 will choose treatment immediately and 15 will choose treatment after a period of active surveillance). 60 will experience incontinence or impotence secondary to treatment. One-two of the 1000 men will avoid death from prostate cancer (Referenced in ACP Journal Club. 2018. 168. JC5).
In a RCT of 162,388 men (ERSCP Trial), prostate biopsies were performed in 20% of men screened (BJU Int. 2012. 110. 1654-1660), cancer was not detected in 75% of the biopsies (Lancet. 201. 384. 2027-2035), and 70% of the cancers found in the screening group were low-grade (N Engl J Med. 2009. 360. 1320-1328).
Prostate biopsy is associated with a risk of bleeding and/or infection requiring hospitalization in 0.5% to 6.9% of men (Eur Urol. 2017. 71. 353-365).
Overdiagnosis (i.e. detection of cancers that would not have been clinically detected during life in the absence of screening) is estimated at 23-42% in the US population and as high as 67% in a European randomized screening trial (Eur Urol. 2014. 65. 1046-1055).
Baseline PSA at age 40 provides prognostic data regarding risk of prostate cancer.
Retrospective data from the Baltimore Longitudinal Aging Study found that men age 40-49 with a PSA above 0.6 ng/ml were 4 times as likely to develop prostate cancer. However, only 29 cancers were detected in this cohort (Urology. 2001. 58. 411-416 as cited in Hoffman RM. Viewpoint: Limiting Prostate Cancer Screening. Ann Intern Med. 2006. 144. 438-440).
A poster presented at the American Urological Association Meetings in San Antonio in 2005 (poster 258) reported that baseline PSA levels of 0.7 ng/ml to 2.5 ng/ml in men in their 40s were associated with a 10 fold increased risk of prostate cancer and baseline PSA levels of 2.5 ng/ml to 4.0 ng/ml in men in their 40s were associated with a 104 fold increased risk of prostate cancer (cited in Catalona WJ et al. Viewpoint: Expanding Prostate Cancer Screening. Ann Intern Med. 2006. 144. 441-443).
PSA-derived measurements which might increase specificity of PSA screening
PSA velocity – was thought to be a more specific measure than PSA, but compared to the PSA, does not add independent information when using the PCPT risk calculator (see final bullet below).
PSA density may improve specificity, but is not a practical screening tool, as calculation requires a transrectal ultrasound to measure prostate volume.
Percent free PSA - if < 15%, more likely that cancer is the present; if > 25%, more likely that elevated PSA is due to BPH.
PSA and DRE are complementary tests. It is estimated that if a PSA cutoff of 4 is used, 10-20% of early prostate cancer will be missed.
PSA is not significantly altered by DRE or urethral catheterization, but prostate biopsy causes a 57-fold rise in PSA which may persist for 6 weeks, and ejaculation can increase the PSA level for up to 48-72 hours.
If PSA screening reduces mortality, data indicates that biennial screening (i.e. every other year) is probably just as good.
The prescription medication Proscar (finasteride 5 mg) will lower the PSA by about 50%, so the upper level of normal in a man taking Proscar is 2.0 and not 4.0.
The OTC medication Propecia (finasteride 1 mg) will also lower the PSA by about 50%, thus requiring a lower PSA cutoff (Lancet Oncology. 2007. 8. 21-25).
PSA accurately predicts the absence of skeletal metastases; if PSA < 10, there is probably no need for bone scan.
PSA > 10 is 92% specific for prostate cancer, but most of these cancers are already metastatic.
There is controversy regarding the PSA cutoff to use for biopsy - some authorities recommend a cutoff of 2.5 for biopsy
Initial data, using sextant biopsies, reported that the probability of prostate cancer in men with a PSA of 4-9.9 was 22%, and that the probability if PSA was > 10 was 67% (N Engl J Med. 1991. 324. 1156-1161). However, if more than 6 biopsies are taken, then the probability of prostate cancer in men with a PSA of 4-9.9 is 40-50% (J Urol. 2006. 175. 485-488).
When a cutoff of 4 is used, approximately 1/3 of tumors have already spread to the margins of the prostate gland or beyond.
The Prostate Cancer Prevention Trial, which used 6-sector biopsies in 2950 men with normal PSA levels showed a positive biopsy rate of 6.6% in men with PSA of 0.5 ng/ml or less, 10.1% in men with a PSA of 0.6 to 1.0 ng/ml, 17.0% in men with a PSA of 1.1 to 2.0 ng/ml, 23.9% in men with a PSA of 2.1 to 3.0 ng/ml, and 26.9% in men with a PSA of 3.1 to 4.0 ng/ml, for an overall detection rate of prostate cancer of 15.2% in men with a PSA less than 4. 2.3% of the men had a high grade cancer, indicating that 44 men required biopsy to find one high grade cancer (N Engl J Med. 2004. 350. 2239-2246; J Natl Cancer Inst. 2006. 98. 529-534).
Age specific PSA (95th percentile) may improve specificity.
Age 40-49 1 - 2.5
Age 50-59 0 - 3.5
Age 60-69 0 - 4.5
Age 70-79 0 - 6.5
Interpretation of PSA in the context of other risk factors, including age, race, family history, findings on digital rectal exam, prostate size, results of earlier prostate biopsies, %free PSA ratio, and whether or not taking a 5 alpha reductase inhibitor. A more meaningful laboratory report provides risk curves for the risk of prostate cancer, and separate risk curves for the risk of high grade prostate cancer, based on results of DRE, family history, and patient age. A PCPT risk calculator is available at www.compass.fhcrc.org/edrnnci/bin/calculator/main.asp and at http://deb.uthscsa.edu/URORiskCalc/Pages/uroriskcalc.jsp (Cleve Clinic J Med. 2008. 75. 33-34; Cleve Clinic J Med. 2009. 76.439-445).
Novel Screening Tests
PCA3 (prostate cancer antigen 3) - molecular urine assay which may be more specific for prostate cancer, as compared with PSA screening. If score is >35, suggests presence of prostate cancer, but does not distinguish high-grade from low-grade Urology. 2008. 179. 1587-1592).
4K score - incorporates total PSA, free PSA, intact PSA, and human kallikrein-related peptidase - this score provides a percentage risk for high grade (Gleason score greater than or equal to 7) prostate cancer on biopsy (Cancer Epidemiol Biomarkers Prev. 2011. 20. 255-261).
Patient education
Risk assessment tool – www.myprostatecancerrisk.com
Developed in 2006 based on data in 5519 men in the Prostate Cancer Prevention Trial (PCPT), all of whom had undergone a prostate biopsy independent of the level of PSA and other risk factors (J Natl Cancer Inst. 2006. 98. 529-534).
Using individual patient information, this tool predicts the probability of (1) a negative biopsy result, (2) low-grade cancer and (3) high-grade cancer.
The tool has been validated in several populations and was recently updated with contemporary data (Ankerst DP et al. Urology. 2014. 83).
Decision aides on line (ACP Internist. July/August 2011)
Harvard www.harvardprostateknowledge.org
American Cancer Society www.cancer.org/prostatemd
Foundation for Informed Medical Decision Making www.healthdialogue.com
Mayo Clinic www.mayoclinic.com/health/prostate-cancer/HQ01273
Endometrial cancer
61,880 cases/year; 12,160 deaths/year [2019 data]
Current recommendations: Insufficient evidence to make clear recommendations about endometrial biopsies or transvaginal ultrasound (Levels 4, 5 evidence).
Esophageal cancer
17,650 cases/year; 16,080 deaths/year [2019 data]
Current recommendations: Insufficient evidence that screening would decrease mortality (Level 5 evidence).
Gastric cancer
27,510 cases/year; 11,140 deaths/year [2019 data]
Current recommendations: Insufficient evidence that screening would decrease mortality (Level 5 evidence).
Skin cancer
96,480 melanoma/year; 7,230 deaths/year [2019 data]
Current recommendations (USPSTF 2016 and 2023): Insufficient evidence to recommend for or against routine screening (Level 5 evidence).
Testicular cancer
9560 cases/year; 410 deaths/year [2019 data]
Risk factors
Undescended testis (cryptorchidism) - relative risk is 2.9-6.3
Personal history of testicular cancer - risk of developing cancer in the opposite testicle is 12 times that of the general population
Infertility, with mechanism unclear
Trauma to the testicle
NOTE vasectomy is not a risk factor
Guidelines
USPSTF II (1996) found insufficient evidence that screening with clinical exam or by self-exam decrease mortality (Level 5 evidence).
USPSTF III (2005) recommends against routine screening with clinical or self-exam, based on low prevalence and favorable outcomes with current interventions once a diagnosis is made. Conclude harms exceed benefits. Note screening is distinct from case-finding in men with testicular symptoms.
Thyroid cancer
52,070 cases/year; 2170 deaths/year [2019 data]
USPSTF (2017) recommends against screening for thyroid cancer in asymptomatic individuals (D recommendation), as the evidence indicates that the harms outweigh the benefits
Biochemical profiles (Blood tests)
The American College of Physicians (ACP) recommends against routine screening biochemical profiles in asymptomatic adults, with the exception of checking cholesterol, and (in a 1998 recommendation) TSH periodically in women over age 50.
Fasting serum glucose - see section just below - Diabetes
Thyroid disease = see section just below - Thyroid disease
The ACP says a screening serum creatinine may be useful because this may be elevated in 3% of screened patients, and dietary restriction of protein and phosphorous would be appropriate along with possible adjustment of drug dosages.
Some experts recommend screening ALT and AST annually in those who are overweight and with type II diabetes, as the prevalence of fatty liver is high in this group, and lifestyle measures can be effective at reducing liver fat and abnormal lab values.
Some experts recommend screening calcium and uric acid, but there is no data that treating asymptomatic hyperuricemia or hyperparathyroidism changes morbidity or mortality.
Diabetes
The USPSTF in 2021 recommends screening for diabetes and prediabetes in nonpregnant adults age 35 - 70 who are overweight or obese, with referral to those identified with prediabetes to effective preventive interventions (B recommendation). Interval of screening not specified (JAMA. 2021. 326. 736-743 and editorial 701-703).
This is an update of the 2015 recommendation which recommended initiation of screening at 40 years of age.
The American Diabetes Association recommends screening for all adults 45 years or older, every 3 years.
Hemochromatosis
Prevalence
1 in 250 based on large-scale screening studies using genetic testing for C282Y homozygotes, with an incidence of 1 in 200 in those of northern European ancestry.
Initially believed that most but not all of these individuals will show evidence of phenotypic expression (i.e. elevated ferritin or transferring saturation).
In a longitudinal study of 31,192 persons of northern European descent followed for an average of 12 years, 203 were identified as C282Y homozygotes by genotyping. The prevalence of iron overload disease was only 28.4% in males and 1.2% in females. Ferritin levels of >1000 ng/ml were highly predictive of iron overload disease (N Engl J Med. 2008. 358. 221-230 and editorial 291-292).
However, cross sectional screening studies showed that approximately 20% of men and 40% of women with the mutation have normal serum ferritin levels (N Engl J Med. 1999. 341. 718-724; Scand J Gastroenterol. 2001. 36. 1108-1115; N Engl J Med. 2005. 352. 1769-1778), and in one study, less than 1% of the C282Y homozygotes had classic symptoms and signs of hereditary hemochromatosis (Lancet. 2002. 359. 211-218).
Historically, prior to the discovery of the HFE gene in 1996, prevalence estimates based on high % transferrin saturation were variable.
Initial studies showed only a 0.3% incidence but these were done in blood donors repeatedly phlebotomized.
A study of 3012 asymptomatic employees and 3027 asymptomatic patients of nine practitioners in Western Germany found an incidence of hemochromatosis was 1.8% among patients and 1.0% among employees, using values of 60% transferrin saturation in males, and 50% transferrin saturation in females, with positive values repeated to rule out lab error or diurnal variation causing a false positive (Ann Intern Med. 1998. 128. 337-345). In retrospect, this high prevalence rate might have been due to misdiagnosis of individuals with high transferring saturation due to occult chronic liver disease from causes other than hereditary hemochromatosis.
Screening can be done by either phenotypic tests (ferritin, transferrin saturation) or genetic testing (HFE mutation analysis).
Three groups to consider for screening
Family members of an affected person – general consensus is to screen by genetic testing.
Patients with liver disease – approximately 50% of patients with nonalcoholic fatty liver disease (NAFLD), chronic hepatitis C, and alcoholic liver disease will have abnormal iron findings; whether they have co-existent hereditary hemochromatosis cannot be determined unless genetic testing is performed. Data would suggest that these patients should be screened with genetic testing and treated with phlebotomy if positive (editorial – Arch Intern Med. 2006. 166. 269-270).
General population - there is controversy about the utility of screening for hemochromatosis.
There is a lack of clarity regarding the percentage of individuals with the gene mutations (C282Y or H63D mutations of the HFE gene on chromosome 6) or even with high transferrin saturation who actually progress to symptomatic disease (Controversies in Internal Medicine: Hereditary Hemochromatosis and Its Elusive History. Arch Intern Med. 2003. 163. 2421-2427). There is data that cirrhosis as diagnosed by liver biopsy is more likely to develop in C282Y homozygotes with the GSTP1 Val/Val genotype than in those with other genotypes. The authors of this paper hypothesize that this is due to the central role of oxidative stress in the pathogenesis of cirrhosis (Arch Intern Med. 2005. 165. 1835-1840).
A study which assessed disease expression by clinical evaluation and liver biopsy in 672 individuals who were C28Y homozygotes but essentially asymptomatic found that hepatic fibrosis was frequently present in asymptomatic subjects and except when cirrhosis was present, was reversed by phlebotomy. This data provides a rationale for population based screening by % transferrin saturation, followed by liver biopsy (Arch Intern Med. 2006. 166. 294-301 and editorial 269-270).
Finally, while it is a standard of care, the efficacy of phlebotomy for improving survival in patients without cirrhosis is based only on data from uncontrolled case series and a plausible theoretical rationale, not on RCTs (Ann Intern Med. 2005. 143. 517-521).
Longitudinal studies of individuals with hemochromatosis by genotyping show that while transferrin saturation levels tend to rise slowly over time, serum ferritin levels vary up and down but do not increase appreciably over decades (Mayo Clin Proc. 2004. 79. 305-306).
Initial understanding of the natural history of hemochromatosis was based on ill patients who presented to a physician, and we now realize from populations studies and genotyping that most individuals with hemochromatosis live their lives without becoming symptomatic.
The American College of Physicians issued an evidence –based clinical practice guideline (Ann Intern Med. 2005. 143. 517-521)
There is insufficient evidence for or against screening for hereditary hemochromatosis in the general population.
In case-finding for hereditary hemochromatosis, serum ferritin and transferring saturation tests should be performed.
Physicians should discuss the risks, benefits, and limitations of genetic testing in patients with a positive family history or those with elevated serum ferritin or transferring saturation.
Further research is needed to establish better diagnostic, therapeutic, and prognostic criteria.
The USPSTF recommends against routine genetic screening for hereditary hemochromatosis in the asymptomatic general population (Ann Intern Med. 2006. 145. 204-208 and 209-223).
Hepatitis B
Prevalence of chronic hepatitis B estimated at 1.6 million - 2.4 million in the US [2021 data]
ACIP (1988) recommends universal screening of pregnant individuals
The USPSTF (2009) recommends universal screening of pregnant individuals
The USPSTF (2020) recommends screening for HBV infection in adolescents and adults at increased risk for infection (B recommendation). One-time screening is the recommendation unless an adult has ongoing risk. Risk factors include birth in a region with a high prevalence of HBV (countries in Africa, Asia, the Pacific islands and parts of South America), persons who inject drugs, men who have sex with men, persons with HIV infection, as well as sex partners, needle-sharing contacts, and household contacts of persons with chronic HBV infection (JAMA. 2020. 324. 2415-2422).
CDC (2023) recommends onetime screening of all adults aged 18 and older (screening to include HBsAg, HB surface antibody, and HB core antigen (this identifies those with resolved infection, but who may be susceptible to reactivation)
Hepatitis C
The USPSTF (2020) recommends screening for HCV infection in all adults aged 18 to 79 years (B recommendation). One-time screening is the recommendation unless an adult has ongoing risk (injection drug use) [JAMA. 2020. 323. 970-975].
Thyroid Disease
The American Thyroid Association recommends that all adults be screened with a TSH starting at the age of 35 and every 5 years thereafter.
The USPSTF (2004 and 2015) states that the evidence is insufficient to recommend for or against routine screening. There is fair evidence that TSH can detect subclinical thyroid disease but poor evidence that treatment of subclinical thyroid disease improves clinically important outcomes (Ann Intern Med. 2004. 140.125-127; Ann Intern Med. 2015. 162. 641-650).
Urinalysis
Proteinuria - screening for proteinuria to prevent end-stage renal disease is not cost effective unless directed at high risk groups (those with hypertension, diabetes, or over age 60 - screen the latter every 10 years) [JAMA. 2003. 290. 3101-3114).
Hematuria - screening for hematuria to detect cancer of the bladder or cancer of the kidney early is not considered cost effective except possibly in smokers and those with occupational exposures which increase the risk of bladder cancer.
Abdominal aortic aneurysm (by definition, diameter of aorta > 3 cm)
9000 deaths per year Present in 4% to 8% of older men
Rationale for screening (Ann Intern Med. 2003. 139. 516-532)
Prevalence rises steeply after age 50, but deaths are in individuals over age 65.
6 times more common in men than women, and 4 times more common in smokers than non-smokers. 90% of all patients with AAA have smoked. The prevalence in men who have never smoked is the same as in women (Ann Intern Med. 2005. 142. 203-211).
Rupture of aneurysms causes 1.2% of male deaths and 0.6% of female deaths over age 65 - 15th leading cause of death in the U.S. Most AAA-related deaths occur before age 80 in males and after age 80 in females. Only 20% of patients survive rupture.
Small AAA's typically enlarge by 0.2 - 0.3 cm/year, are nearly always asymptomatic until rupture, and rarely rupture before reaching a diameter of 6 cm.
Risk of rupture (most of the data is in men; limited data suggests a slightly greater risk of rupture in women vs. men for an AAA the same size)
5 year survival rate only 6% in those with untreated AAA > 6 cm (Ann Surg. 1966. 164. 678-699).
Data from a population-based study in Rochester MN indicates a risk of rupture of 1.0% per year when aneurysm size is 4.0-4.99 cm and 11% per year when aneurysm size is 5.0-5.99 cm (Arch Intern Med. 1997. 157. 2064-2068).
Risks versus benefits of surgery/endoscopic repair
Operative mortality rate with elective repair by an experienced surgeon is 4% to 6%.
Follow up on 2226 patients initially enrolled in the UKSAT trial and ADM trial showed that at an average of 5.3 years of follow up, outcomes of immediate open repair were no better than outcomes associated with active surveillance of small aneurysms (Mayo Clin Proc. 2013. 88. 910-919).
Surveillance recommendations for a small aneurysm – repair when size reaches 5.5 cm, becomes symptomatic, enlarges by >0.7 cm in 6 months, or > 1.0 cm in in year.
Ultrasonography is nearly 100% sensitive and specific for screening.
A single normal ultrasound in men age 65 and older virtually excludes future risk of AAA-related death (Br J Surg. 2001. 88. 941-944)
Most cost effective strategy for population screening based on the data is a single screen in all men at age 65 - consider screening high risk men at age 60 and high risk women at age 65 too.
A meta-analysis found that aneurysm screening reduced deaths by 43% over 4-5 years in men over age 65 (Ann Intern Med. 2005. 142. 203-211).
Randomized, controlled trials (Ann Intern Med. 2003. 139. 516-532)
A RCT of 67,770 men, mean age 69.2 years, at four centers in the south of England (MASS), 80% of whom accepted screening, showed a reduction in deaths due to aneurysm of 42%, with 65 deaths in the screening group and 113 deaths in the control group. At 4 years, the cost effectiveness was $57,686 per QALY. The predicted cost effectiveness at 10 years would be $12,819 per QALY, with 710 subjects screened to prevent one death. There was also in this trial a statistically significant reduction in deaths from ischemic heart disease in the screened group (BMJ. 2002. 325. 1135-1138; Lancet. 2002. 360. 1531-1539). At 7 years of follow-up, the mortality benefit was maintained, with a reduction in all-cause mortality actually documented. Cost-effectiveness was estimated at $19,500 per life-year gained based on AAA-related mortality, and $7600 per life-year gained based on all-cause mortality (Ann Intern Med. 2007. 146. 699-706 and editorial 749-750). Extrapalating from this study, and assuming a 50% smoking rate and that the study results apply to women, the 5 year number needed to screen (NNS) to prevent one AAA death is 335 for a man who ever smoked, 536 for all men, 1340 for a nonsmoking man, 2011 for a woman who ever smoked, 3217 for all women, and 8044 for a nonsmoking woman. By comparison, the 5-year NNS for mammography in 60-69 year old women is 1251 (ACP Journal Club. 2007. 147. 57).
A RCT of 6058 men and 9342 women aged 65-80 in Chichester, U.K. in which 74% of men and 65% of women accepted screening reported a 41% reduction in deaths due to aneurysm in men at 5 years (BrJ Surg. 1995. 82. 1066-1070) and a 21% reduction in men at 10 years (Br J Surg. 2002. 89. 861-864), but neither of these reductions reached statistical significance. No benefit was seen in the women screened.
A RCT of 12,658 men aged 65-73 in Viborg, Denmark, 76% of whom accepted screening, reported a 68% reduction in inpatient AAA deaths; information on outpatient deaths was not recorded (Eur J Vasc Endovasc Surg. 2002. 23. 55-60).
A RCT of 39,166 men aged 65-83 in western Australia, 62% of whom accepted screening, reported a 28% reduction in aneurysm deaths, which was not statistically significant. This has been published only in abstract form (BrJ Surg. 2003. 90. 492).
A RCT of all 12,639 men in Denmark who were born between 1921 and 1929 in which 76.6% of those randomized to undergo screening in fact did undergo screening (at an average age of 67.7 years) reported 9 deaths from AAA in the screened group versus 27 deaths in the control group over 4.33 years, translating into a number needed to screen of 349 to prevent one AAA-related death over 4.33 years. Overall mortality however was the same in both groups (BMJ. 2005. 330. 750-752).
A RCT of 67,770 men aged 65-74 at 4 centers in the UK (MASS study) found that 1 time screening by ultrasound reduced AAA-related mortality, with a small magnitude of effect (>200 men screened to prevent 1 aneurysm-related death over 13 years) [Br J Surg. 2012. 99. 1649-1656].
A systematic evidence review for USPSTF identified 4 RCTs (n=137,214) and concluded “One-time invitation for AAA screening in men aged 65 years or older was associated with decreased AAA rupture and AAA-related mortality rates but had little or no effect on all-cause mortality” (Ann Intern Med. 2014. 160. 321-329).
USPSTF recommendations for screening (Ann Intern Med. 2005. 142. 198-202; Ann Intern Med. 2014. 161. 281-290; JAMA. 2019. 2211-2218 and editorial 2177-2178)
Recommends one-time screening by ultrasound in men age 65-75 who have ever smoked (more than 100 lifetime cigarettes) – Grade B recommendation. Good evidence that screening for aneurysms and repair of large (>5.5 cm) aneurysms leads to decreased AAA-specific mortality, based on prevalence in this group estimated at 7%. Good evidence that ultrasound is an accurate screening test. There is also good evidence of important harms from screening and early treatment, but benefits outweigh harms.
Recommends clinicians selectively offer screening in men age 65-75 who have never smoked – Grade C recommendation. Good evidence that screening for aneurysms and repair of large (>5.5 cm) aneurysms leads to decreased AAA-specific mortality, but benefit is small due to lower prevalence (estimated at 2%) and harms are significant, with risk of benefit and risk of harm approximately equal.
Recommends against routine screening in women who have never smoked and have no family history – Grade D recommendation. Due to low prevalence of AAA in this group (estimated at 0.03-0.60%) , and significant harms associated with screening and early treatment, risk of harm exceeds risk of benefit.
Concludes evidence is insufficient to assess the balance of benefits versus harms of screening by ultrasound in women age 65-75 who have ever smoked or who have a family history - Grad I recommendation.
NOTE since the 2014 update of the USPSTF recommendation, a review of 4 RCTs concluded that one-time screening of ALL men at age 65 or older for AAA is cost effective (J Vasc Surg. 2016. 64. 1855-1868 as cited in ACP Journal Club. 2017. 166. JC27).
Management of aneurysms found by screening (Ann Intern Med. 2003. 139. 516-532)
Operate electively if the aneurysm is greater than 5.5 cm
Two RCT's have demonstrated that elective repair of AAAs smaller than 5.5 cm does not improve survival (Lancet. 1998. 352. 1649-1655; N Engl J Med. 2002. 346. 1437-1444).
A systematic review concludes that elective repair of AAAs smaller than 5.5 cm does not improve survival (Ann Intern Med. 2007. 146. 735-741). Previous reviews concluded the same (AHRQ Publication #06-E017; Br J Surg. 2005. 92. 937-946).
Cost effectiveness of elective repair is addressed in a separate systematic review published in Int J Technol Assess Health Care. 2007. 23. 205-215.
Endovascular repair reduces the cost of ICU care, hospital days, and blood transfusions, but due to the cost of the graft ($10,000) and the higher rates of re-intervention (14% within 4 years), but it does not improve overall survival or quality of life, and is actually less cost effective than open repair (Ann Intern Med. 2007. 146. 735-741 and editorial 749-750).
Serial ultrasound
Every 6 months for aneurysms 4 - 5.5 cm, and every 2-3 years for aneurysms less than 4 cm. Be aware that variations in measurement up to 0.5 cm are common (J Vasc Surg. 2004. 39. 267-269).
A meta-analysis of individual patient data concluded that in men, to control the risk of rupture to below 1%, screening intervals of 8.5 years for a 3 cm aneurysm and 17 months for a 5 cm aneurysm are sufficient (JAMA. 2013. 309. 806-813).
Research into medical methods to slow enlargement of small aneurysms is preliminary at this time - antibiotics MAY slow growth through inhibition of macrophage proteolytic enzymes.
Time to Rethink Screening for Abdominal Aortic Aneurysm? (Arch Intern Med. 2012. 172. 1462-1463)
This Invited Commentary comments that even though there are no new RCT data since the USPSTF review of 2005, data has emerged that mortality from ruptured AAA has decreased by as much as 50% in the past 10-15 years, paralleling reduction in smoking prevalence and MI prevalence, and this would suggest that the incidence of AAA is dropping, and if so, this would reduce the potential benefit of screening.
In 2011, and estimated 70-80% of AAA repairs are performed using an endovascular technique (EVAR). 30 day mortality is lower with this technique, but 5 year mortality is the same as with an open repair. A consequence though of less invasive surgery is that as many as 41% of patients undergoing EVAR have an AAA diameter of < 5.5 cm (Circulation. 2011. 123. 2848-2855). Benefit is not proven when surgery is done on individual with smaller aneurysms, and small aneurysms are common.
Screening leads to label of AAA in those with small aneurysms, and psychological effects of living with a potentially fatal condition may be significant.
Asymptomatic carotid stenosis
The USPSTF in 2007, 2014, and in 2021 recommends against screening for asymptomatic carotid artery stenosis in the general adult population, grade D recommendation (Ann Intern Med. 2014. 161. 336-346, 356-362 and editorial 370-371; JAMA. 2021. 325. 476-481 and editorial 443-444).
Page Updated October 20, 2024