Search

This article summarizes the phenomenon of cancer overdiagnosis-the diagnosis of a "cancer" that would otherwise not go on to cause symptoms or death. We describe the two prerequisites for cancer overdiagnosis to occur: the existence of a silent disease reservoir and activities leading to its detection (particularly cancer screening). We estimated the magnitude of overdiagnosis from randomized trials: about 25% of mammographically detected breast cancers, 50% of chest x-ray and/or sputum-detected lung cancers, and 60% of prostate-specific antigen-detected prostate cancers. We also review data from observational studies and population-based cancer statistics suggesting overdiagnosis in computed tomography-detected lung cancer, neuroblastoma, thyroid cancer, melanoma, and kidney cancer. To address the problem, patients must be adequately informed of the nature and the magnitude of the trade-off involved with early cancer detection. Equally important, researchers need to work to develop better estimates of the magnitude of overdiagnosis and develop clinical strategies to help minimize it.

Background: Although there is uncertainty about the effect of prostate-specific antigen (PSA) screening on the rate of prostate cancer death, there is little uncertainty about its effect on the rate of prostate cancer diagnosis. Systematic estimates of the number of men affected, however, to our knowledge, do not exist., Methods: We obtained data on age-specific incidence and initial course of therapy from the National Cancer Institute's Surveillance, Epidemiology, and End Results program. We then used age-specific male population estimates from the US Census to determine the excess (or deficit) in the number of men diagnosed and treated in each year after 1986-the year before PSA screening was introduced., Results: Overall incidence of prostate cancer rose rapidly after 1986, peaked in 1992, and then declined, albeit to levels considerably higher than those in 1986. Overall incidence, however, obscured distinct age-specific patterns: The relative incidence rate (2005 relative to 1986) was 0.56 in men aged 80 years and older, 1.09 in men aged 70-79 years, 1.91 in men aged 60-69 years, 3.64 in men aged 50-59 years, and 7.23 in men younger than 50 years. Since 1986, an estimated additional 1 305 600 men were diagnosed with prostate cancer, 1 004 800 of whom were definitively treated for the disease. Using the most optimistic assumption about the benefit of screening-that the entire decline in prostate cancer mortality observed during this period is attributable to this additional diagnosis-we estimated that, for each man who experienced the presumed benefit, more than 20 had to be diagnosed with prostate cancer., Conclusions: The introduction of PSA screening has resulted in more than 1 million additional men being diagnosed and treated for prostate cancer in the United States. The growth is particularly dramatic for younger men. Given the considerable time that has passed since PSA screening began, most of this excess incidence must represent overdiagnosis.

Background: To make sense of the disease risks they face, people need basic facts about the magnitude of a particular risk and how one risk compares with other risks. Unfortunately, this fundamental information is not readily available to patients or physicians. We created simple one-page charts that present the 10-year chance of dying from various causes according to age, sex, and smoking status., Methods: We used the National Center for Health Statistics Multiple Cause of Death Public Use File for 2004 and data from the 2004 US Census to calculate age- and sex-specific death rates for various causes of death. We then combined data on smoking prevalence (from the National Health Interview Survey) and the relative risks of death from various causes for smokers vs never smokers (from the American Cancer Society's Cancer Prevention Study-II) to determine age-, sex-, and smoking-specific death rates. Finally, we accumulated these risks for various starting ages in a series of 10-year life tables. The charts present the 10-year risks of dying from heart disease; stroke; lung, colon, breast, cervical, ovarian, and prostate cancer; pneumonia; influenza; AIDS; chronic obstructive pulmonary disease; accidents; and all causes., Results: At all ages, the 10-year risk of death from all causes combined is higher for men than women. The effect of smoking on the chance of dying is similar to the effect of adding 5 to 10 years of age: for example, a 55-year-old man who smokes has about the same 10-year risk of death from all causes as a 65-year-old man who never smoked (ie, 178 vs 176 of 1000 men, respectively). For men who never smoked, heart disease death represents the single largest cause of death from age 50 on and the chance of dying from heart disease exceeds the chances of dying from lung, colon, and prostate cancers combined at every age. For men who currently smoke, the chance of dying from lung cancer is of the same order of magnitude as the chance dying from heart disease and after age 50 it is about 10 times greater than the chance of dying from prostate or colon cancer. For women who have never smoked, the magnitudes of the 10-year risks of death from breast cancer and heart disease are similar until age 60; from this age on, heart disease represents the single largest cause of death. For women who currently smoke, the chance of dying from heart disease or lung cancer exceeds the chance of dying from breast cancer from age 40 on (and does so by at least a factor of 5 after age 55)., Conclusion: The availability of simple charts with consistent data presentations of important causes of death may facilitate discussion about disease risk between physicians and their patients and help highlight the dangers of smoking.

Background: Despite the considerable attention given to the prostate-specific antigen (PSA) as a screening test for prostate cancer, it is needle biopsy--and not the PSA test result--that actually establishes the diagnosis of prostate cancer. We sought national estimates on the proportion of men found to have prostate cancer after a needle biopsy of the prostate and the risk of subsequent biopsies among those not found to have prostate cancer., Methods: We linked Medicare claims data to Surveillance, Epidemiology, and End Results (SEER) data to analyze outcomes after 10,429 needle biopsies performed in 1993 through 2001 in 8273 men aged 65 years and older enrolled in Medicare Part B who resided in a SEER area. We determined the proportion of needle biopsies that were followed by a diagnosis of prostate cancer, the cumulative risk of prostate cancer following multiple biopsies, and the risk of subsequent biopsy among men not found to have prostate cancer in the previous biopsy. All statistical tests were two-sided., Results: The overall proportion of needle biopsies found to contain prostate cancer was 32% (95% confidence interval [CI] = 31% to 33%). The yield increased with age (26% for men aged 65-69 years, 31% for men aged 70-74 years, 35% for men aged 75-79 years, and 41% for men aged 80 years and older; P(trend)<.001). The cumulative risk of prostate cancer diagnosis increased with repeated biopsy, with 50% of men receiving a prostate cancer diagnosis after two biopsies, 62% after three biopsies, and 68% after four biopsies. Among men whose first recorded biopsy did not detect prostate cancer, the risk of having a subsequent biopsy was 11.6% (95% CI = 11% to 12%) at 1 year and 38% (95% CI = 36% to 40%) at 5 years., Conclusions: About one-third of prostate biopsies identified prostate cancer in this population. Men not found to have prostate cancer on a first biopsy frequently undergo repeat biopsies, which raise the cumulative risk of prostate cancer diagnosis.

Background: The finding that some men with a normal prostate-specific antigen (PSA) level (i.e., less than 4 ng/mL) nonetheless have microscopic evidence of prostate cancer has led to some suggestions that the threshold defining abnormal should be lowered to 2.5 ng/mL. We examined the effect of this lower threshold on the number of American men who would be labeled abnormal by a single PSA test., Methods: We obtained PSA data on a nationally representative sample of American men 40 years of age and older with no history of prostate cancer and no current inflammation or infection of the prostate gland (n = 1308) from the 2001-2002 National Health and Nutrition Examination Survey. We obtained data on the 10-year risk of prostate cancer death in the pre-PSA era from DevCan, the National Cancer Institute's software to calculate the probability of dying of cancer., Results: Based on NHANES data, approximately 1.5 million American men aged 40 to 69 years have a PSA level over 4.0 ng/mL. Lowering the threshold to 2.5 ng/mL would label an additional 1.8 million men as abnormal, if all men were screened. For men aged 70 years or older, the corresponding numbers are 1.5 and 1.2 million. The proportion of the population affected by different thresholds would vary with age. Among men in their 60s, for example, 17% have a PSA level over 2.5 ng/mL, 5.7% have a PSA level over 4.0 ng/mL, and 1.7% have a PSA level over 10.0 ng/mL. For context, only 0.9% of men in their 60s are expected to die from prostate cancer in the next 10 years., Conclusion: Lowering the PSA threshold to 2.5 ng/mL would double the number of men defined as abnormal, to up to 6 million. Until there is evidence that screening is effective, increasing the number of men recommended for prostate biopsy--and the number potentially diagnosed and treated unnecessarily--would be a mistake.

Celebrities often promote cancer screening by relating personal anecdotes about their own diagnosis or that of a loved one. We used data obtained from a random-digit dialing survey conducted in the United States from December 2001 through July 2002 to examine the extent to which adults of screening age without a history of cancer had seen or heard or been influenced by celebrity endorsements of screening mammography, prostate-specific antigen (PSA) testing, or sigmoidoscopy or colonoscopy. The survey response rate was 72% among those known to be eligible and 51% among potentially eligible people accounting for those who could not be contacted. A total of 360 women aged 40 years or older and 140 men aged 50 years or older participated in the survey. Most respondents reported they "had seen or heard a celebrity talk about" mammography (73% of women aged 40 years or older), PSA testing (63% of men aged 50 years or older), or sigmoidoscopy or colonoscopy (52% of adults aged 50 years or older). At least one-fourth of respondents who had seen or heard a celebrity endorsement said that the endorsement made them more likely to undergo mammography (25%), PSA testing (31%), or sigmoidoscopy or colonoscopy (37%).

Background: The incidence of esophageal adenocarcinoma is rising dramatically. This increase may reflect increased disease burden, reclassification of related cancers, or overdiagnosis resulting from increased diagnostic intensity, particularly upper endoscopy for patients with gastroesophageal reflux disease or Barrett esophagus., Methods: We used the National Cancer Institute's Surveillance, Epidemiology, and End Results database to extract information on incidence, stage distribution, and disease-specific mortality for esophageal adenocarcinoma as well as information on related cancers., Results: From 1975 to 2001, the incidence of esophageal adenocarcinoma rose approximately sixfold in the United States (from 4 to 23 cases per million), a relative increase greater than that for melanoma, breast, or prostate cancer. Reclassification of squamous cell carcinoma is an unlikely explanation for the rise in incidence, because the anatomic distribution of esophageal cancer in general has changed. The only location with increased incidence is the lower third of the esophagus-the site where adenocarcinoma typically arises. Reclassification of adjacent gastric cancer is also unlikely because its incidence has also increased. Because there has been little change in the proportion of patients found with in situ or localized disease at diagnosis since 1975 (from 25% to 31%) and because esophageal adenocarcinoma mortality has increased more than sevenfold (from 2 to 15 deaths per million), overdiagnosis can be excluded as an explanation for the rise in incidence., Conclusion: The rising incidence of esophageal adenocarcinoma represents a real increase in disease burden.

Background: Cancer mortality should include not only deaths from cancer but also deaths from cancer treatment. By convention, deaths within 30 days of a surgical procedure are considered treatment-related deaths in the calculation of operative mortality-that is, the chance of dying from surgery. How cause of death is attributed in patients who die within 1 month of cancer-directed surgery is unknown., Methods: The National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) program data from 1994 through 1998 were used to examine the cause of death in patients diagnosed with one of 19 common solid tumors who had died within 1 month of diagnosis and had also received cancer-directed surgery. We determined the proportion of deaths not attributed to the cancer and the magnitude of the undercount in cancer-specific mortality., Results: Among 4135 patients with only one cancer who died within 1 month of diagnosis and cancer-directed surgery, the proportion of deaths not attributed to the coded cancer was 41% (1714/4135), ranging from 13% (1/8) for cervical cancer to 81% (13/16) for laryngeal cancer. Selected intermediate values include 25% (14/56) for esophageal cancer, 34% (177/525) for lung cancer, 42% (719/1695) for colorectal cancer, 59% (110/186) for breast cancer, and 75% (80/106) for prostate cancer. Restricting the analysis to deaths following specific major procedures (e.g., esophagectomy, pneumonectomy, colectomy) had little effect on the findings. If all deaths within 1 month of cancer-directed surgery were attributed to cancer, cancer mortality would rise about 1%., Conclusion: Some deaths that are conventionally attributed to surgery are not being attributed to the cancer for which the surgery was performed. Although the estimated effect of this misclassification on overall cancer mortality is modest, it may be indicative of more widespread confusion about how to code treatment-related deaths of patients with cancer.

Background: The most widely accepted end point in randomized cancer screening trials is disease-specific mortality. The validity of this end point, however, rests on the assumption that cause of death can be determined accurately. An alternative end point is all-cause mortality, which depends only on the accurate ascertainment of deaths and when they occur. We compared disease-specific and all-cause mortality in published randomized cancer-screening trials to indirectly assess the validity of the disease-specific mortality end point., Methods: We examined all 12 published randomized trials of cancer screening for which both end points were available (seven of mammography, three of fecal occult blood detection, and two of chest x-ray screening for lung cancer). For each randomized trial, we subtracted disease-specific mortality observed in the screened group from that observed in the control group and all-cause mortality in the screened group from that in the control group. We then compared the differences in these two mortality measures., Results: In five of the 12 trials, differences in the two mortality rates went in opposite directions, suggesting opposite effects of screening. In four of these five trials, disease-specific mortality was lower in the screened group than in the control group, whereas all-cause mortality was the same or higher. In two of the remaining seven trials, the mortality rate differences were in the same direction but their magnitudes were inconsistent; i.e., the difference in all-cause mortality exceeded the disease-specific mortality in the control group. Thus, results of seven of the 12 trials were inconsistent in their direction or magnitude., Conclusion: Major inconsistencies were identified in disease-specific and all-cause mortality end points in randomized cancer screening trials. Because all-cause mortality is not affected by bias in classifying the cause of death, it should be examined when interpreting the results of randomized cancer-screening trials.

Background: Screening with a fecal occult blood test (FOBT) has been shown to reduce colorectal cancer mortality in controlled trials. Recently, Medicare approved payment for FOBT screening. We evaluated the pattern of diagnostic testing following the initial FOBT in elderly Medicare beneficiaries. Such follow-up testing would in the long run influence both the cost and the benefit of widespread use of FOBT., Methods: Using Medicare's National Claims History System, we identified 24 246 Americans 65 years old or older who received FOBT at physician visits between January 1 and April 30, 1995. Prior to FOBT, these people had no evidence of any conditions for which FOBT might be used diagnostically. We examined relevant diagnostic testing in this cohort during the subsequent 8 months and determined what proportion of those received an evaluation recommended by the American College of Physicians., Results: For every 1000 Medicare beneficiaries who received FOBT, 93 (95% confidence interval = 89-96 per 1000) had positive findings and relevant testing in the subsequent 8 months. Of these, 34% had the recommended evaluation of either colonoscopy or flexible sigmoidoscopy with an air-contrast barium enema. Another 34% received a partial colonic evaluation with either flexible sigmoidoscopy or a barium enema. The remaining 32% received other gastrointestinal (GI) testing without evaluation of the colonic lumen: computed tomography or magnetic resonance imaging of the abdomen (15%), upper GI series (10%), carcinoembryonic antigen (7%), and upper endoscopy (2%). Restricting the analysis to testing performed within 2 months of the initial FOBT yielded similar results., Conclusion: Following FOBT, many Medicare beneficiaries get further diagnostic testing, but only a small proportion receives the recommended evaluation. With this pattern of practice, population screening is likely to be more costly and less effective than estimated from controlled trials.

Background: To provide some sense of the general frequency and timing of diagnostic testing following screening mammography in the United States, we investigated the experience of women screened in the Medicare population., Methods: By use of Medicare's National Claims History System, we identified a cohort (n=23172) of women 65 years old or older screened during the period from January 1, 1995, through April 30, 1995, and tracked each woman over the subsequent 8 months for the performance of additional breast imaging and biopsy procedures. Using two claims-based definitions for newly detected breast cancer, we also estimated the positive predictive value of screening mammography., Results: For every 1000 women aged 65-69 years who underwent screening, 85 (95% confidence interval [CI]=79-91) had follow-up testing in the subsequent 8 months; 76 (95% CI=71-82) had additional breast imaging, and 23 (95% CI=20-26) had biopsy procedures. Corresponding numbers for women aged 70 years or more were similar. Some women underwent repeated examinations; 13% of those receiving diagnostic mammograms had more than one; 11% of those undergoing biopsy procedures had more than one. About half of the women who underwent a biopsy had the procedure more than 3 weeks after the imaging test upon which the decision to perform a biopsy was presumably made. The estimated positive predictive value of an abnormal screening mammogram (defined as a mammogram that engendered additional testing) was 0.08 (95% CI=0.06-0.10) for women aged 65-69 years and 0.14 (95% CI=0.12-0.16) for women aged 70 years or more., Conclusion: Additional testing is a frequent consequence of screening mammography and may require a considerable period of time to come to closure. The need for additional testing, however, is weakly predictive of cancer.