Estimating the cumulative risk of false positive cancer screenings

Modeling the mean time to interval cancer after negative results of periodic cancer screening

Interval cancers are cancers detected symptomatically between screens or after the last screen. A mathematical model for the development of interval cancers can provide useful information for evaluating cancer screening. In this regard a useful quantity is MIC, the mean duration in years of progressive preclinical cancer (PPC) that leads to interval cancers. Estimation of MIC involved extending a previous model to include three negative screens, invoking the multinomial-Poisson transformation to avoid estimating background cancer trends, and varying screening test sensitivity. Simulations show that when the true MIC is 0.5, the method yields a reasonably narrow range of estimated MICs over the range of screening test sensitivities from 0.5 to 1.0. If the lower bound on the screening test sensitivity is 0.7, the method performs considerably better even for larger MICs. The application of the method involved annual lung cancer screening in the Prostate, Lung, Colorectal, and Ovarian trial. Assuming a normal distribution for PPC duration, the estimated MIC (95% confidence interval) ranged from 0.00 (0.00 to 0.34) at a screening test sensitivity of 1.0 to 0.54 (0.03, 1.00) at a screening test sensitivity of 0.5 Assuming an exponential distribution for PPC duration, which did not fit as well, the estimated MIC ranged from 0.27 (0.08, 0.49) at a screening test sensitivity of 0.5 to 0.73 (0.32, 1.26) at a screen test sensitivity of 1.0 Based on these results, investigators may wish to investigate more frequent lung cancer screening.

Breast cancer overdiagnosis in stop-screen trials: More uncertainty than previously reported

OBJECTIVE

According to the Independent UK Panel on Breast Cancer Screening, the most reliable estimates of overdiagnosis for breast cancer screening come from stop-screen trials Canada 1, Canada 2, and Malmo. The screen-interval overdiagnosis fraction is the fraction of cancers in a screening program that are overdiagnosed. We used the cumulative incidence method to estimate screen-interval overdiagnosis fraction. Our goal was to derive confidence intervals for estimated screen-interval overdiagnosis fraction and adjust for refusers in these trials.

METHODS

We first show that the UK Panel's use of a 95% binomial confidence interval for estimated screen-interval overdiagnosis fraction was incorrect. We then derive a correct 95% binomial-Poisson confidence interval. We also use the method of latent-class instrumental variables to adjust for refusers.

RESULTS

For the Canada 1 trial, the estimated screen-interval overdiagnosis fraction was 0.23 with a 95% binomial confidence interval of (0.18, 0.27) and a 95% binomial-Poisson confidence interval of (0.04, 0.41). For the Canada 2 trial, the estimated screen-interval overdiagnosis fraction was 0.16 with a 95% binomial confidence interval of (0.12, 0.19) and a 95% binomial-Poisson confidence interval of (-0.01, 0.32). For the Malmo trial, the estimated screen-interval overdiagnosis fraction was 0.19 with a 95% binomial confidence interval of (0.15, 0.22). Adjusting for refusers, the estimated screen-interval overdiagnosis fraction was 0.26 with a 95% binomial-Poisson confidence interval of (0.03, 0.50).

CONCLUSION

The correct 95% binomial-Poisson confidence interval s for the estimated screen-interval overdiagnosis fraction based on the Canada 1, Canada 2, and Malmo stop-screen trials are much wider than the previously reported incorrect 95% binomial confidence intervals. The 95% binomial-Poisson confidence intervals widen as follow-up time increases, an unappreciated downside of longer follow-up in stop-screen trials.

Maximum likelihood estimation with missing outcomes: From simplicity to complexity

Many clinical or prevention studies involve missing or censored outcomes. Maximum likelihood (ML) methods provide a conceptually straightforward approach to estimation when the outcome is partially missing. Methods of implementing ML methods range from the simple to the complex, depending on the type of data and the missing-data mechanism. Simple ML methods for ignorable missing-data mechanisms (when data are missing at random) include complete-case analysis, complete-case analysis with covariate adjustment, survival analysis with covariate adjustment, and analysis via propensity-to-be-missing scores. More complex ML methods for ignorable missing-data mechanisms include the analysis of longitudinal dropouts via a marginal model for continuous data or a conditional model for categorical data. A moderately complex ML method for categorical data with a saturated model and either ignorable or nonignorable missing-data mechanisms is a perfect fit analysis, an algebraic method involving closed-form estimates and variances. A complex and flexible ML method with categorical data and either ignorable or nonignorable missing-data mechanisms is the method of composite linear models, a matrix method requiring specialized software. Except for the method of composite linear models, which can involve challenging matrix specifications, the implementation of these ML methods ranges in difficulty from easy to moderate.

Cumulative probability of false-positive recall or biopsy recommendation after 10 years of screening mammography: a cohort study

BACKGROUND

False-positive mammography results are common. Biennial screening may decrease the cumulative probability of false-positive results across many years of repeated screening but could also delay cancer diagnosis.

OBJECTIVE

To compare the cumulative probability of false-positive results and the stage distribution of incident breast cancer after 10 years of annual or biennial screening mammography.

DESIGN

Prospective cohort study.

SETTING

7 mammography registries in the National Cancer Institute-funded Breast Cancer Surveillance Consortium.

PARTICIPANTS

169,456 women who underwent first screening mammography at age 40 to 59 years between 1994 and 2006 and 4492 women with incident invasive breast cancer diagnosed between 1996 and 2006.

MEASUREMENTS

False-positive recalls and biopsy recommendations stage distribution of incident breast cancer.

RESULTS

False-positive recall probability was 16.3% at first and 9.6% at subsequent mammography. Probability of false-positive biopsy recommendation was 2.5% at first and 1.0% at subsequent examinations. Availability of comparison mammograms halved the odds of a false-positive recall (adjusted odds ratio, 0.50 [95% CI, 0.45 to 0.56]). When screening began at age 40 years, the cumulative probability of a woman receiving at least 1 false-positive recall after 10 years was 61.3% (CI, 59.4% to 63.1%) with annual and 41.6% (CI, 40.6% to 42.5%) with biennial screening. Cumulative probability of false-positive biopsy recommendation was 7.0% (CI, 6.1% to 7.8%) with annual and 4.8% (CI, 4.4% to 5.2%) with biennial screening. Estimates were similar when screening began at age 50 years. A non-statistically significant increase in the proportion of late-stage cancers was observed with biennial compared with annual screening (absolute increases, 3.3 percentage points [CI, -1.1 to 7.8 percentage points] for women age 40 to 49 years and 2.3 percentage points [CI, -1.0 to 5.7 percentage points] for women age 50 to 59 years) among women with incident breast cancer.

LIMITATIONS

Few women underwent screening over the entire 10-year period. Radiologist characteristics influence recall rates and were unavailable. Most mammograms were film rather than digital. Incident cancer was analyzed in a small sample of women who developed cancer.

CONCLUSION

After 10 years of annual screening, more than half of women will receive at least 1 false-positive recall, and 7% to 9% will receive a false-positive biopsy recommendation. 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.

PRIMARY FUNDING SOURCE

National Cancer Institute.

Cumulative probability of false-positive recall or biopsy recommendation after 10 years of screening mammography: a cohort study

BACKGROUND

False-positive mammography results are common. Biennial screening may decrease the cumulative probability of false-positive results across many years of repeated screening but could also delay cancer diagnosis.

OBJECTIVE

To compare the cumulative probability of false-positive results and the stage distribution of incident breast cancer after 10 years of annual or biennial screening mammography.

DESIGN

Prospective cohort study.

SETTING

7 mammography registries in the National Cancer Institute-funded Breast Cancer Surveillance Consortium.

PARTICIPANTS

169,456 women who underwent first screening mammography at age 40 to 59 years between 1994 and 2006 and 4492 women with incident invasive breast cancer diagnosed between 1996 and 2006.

MEASUREMENTS

False-positive recalls and biopsy recommendations stage distribution of incident breast cancer.

RESULTS

False-positive recall probability was 16.3% at first and 9.6% at subsequent mammography. Probability of false-positive biopsy recommendation was 2.5% at first and 1.0% at subsequent examinations. Availability of comparison mammograms halved the odds of a false-positive recall (adjusted odds ratio, 0.50 [95% CI, 0.45 to 0.56]). When screening began at age 40 years, the cumulative probability of a woman receiving at least 1 false-positive recall after 10 years was 61.3% (CI, 59.4% to 63.1%) with annual and 41.6% (CI, 40.6% to 42.5%) with biennial screening. Cumulative probability of false-positive biopsy recommendation was 7.0% (CI, 6.1% to 7.8%) with annual and 4.8% (CI, 4.4% to 5.2%) with biennial screening. Estimates were similar when screening began at age 50 years. A non-statistically significant increase in the proportion of late-stage cancers was observed with biennial compared with annual screening (absolute increases, 3.3 percentage points [CI, -1.1 to 7.8 percentage points] for women age 40 to 49 years and 2.3 percentage points [CI, -1.0 to 5.7 percentage points] for women age 50 to 59 years) among women with incident breast cancer.

LIMITATIONS

Few women underwent screening over the entire 10-year period. Radiologist characteristics influence recall rates and were unavailable. Most mammograms were film rather than digital. Incident cancer was analyzed in a small sample of women who developed cancer.

CONCLUSION

After 10 years of annual screening, more than half of women will receive at least 1 false-positive recall, and 7% to 9% will receive a false-positive biopsy recommendation. 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.

PRIMARY FUNDING SOURCE

National Cancer Institute.

Modelling the cumulative risk of a false-positive screening test

The goal of a screening test is to reduce morbidity and mortality through the early detection of disease; but the benefits of screening must be weighed against potential harms, such as false-positive (FP) results, which may lead to increased healthcare costs, patient anxiety, and other adverse outcomes associated with diagnostic follow-up procedures. Accurate estimation of the cumulative risk of an FP test after multiple screening rounds is important for program evaluation and goal setting, as well as informing individuals undergoing screening what they should expect from testing over time. Estimation of the cumulative FP risk is complicated by the existence of censoring and possible dependence of the censoring time on the event history. Current statistical methods for estimating the cumulative FP risk from censored data follow two distinct approaches, either conditioning on the number of screening tests observed or marginalizing over this random variable. We review these current methods, identify their limitations and possibly unrealistic assumptions, and propose simple extensions to address some of these limitations. We discuss areas where additional extensions may be useful. We illustrate methods for estimating the cumulative FP recall risk of screening mammography and investigate the appropriateness of modelling assumptions using 13 years of data collected by the Breast Cancer Surveillance Consortium (BCSC). In the BCSC data we found evidence of violations of modelling assumptions of both classes of statistical methods. The estimated risk of an FP recall after 10 screening mammograms varied between 58% and 77% depending on the approach used, with an estimate of 63% based on what we feel are the most reasonable modelling assumptions.

Improving the biomarker pipeline to develop and evaluate cancer screening tests

The biomarker pipeline to develop and evaluate cancer screening tests has three stages: identification of promising biomarkers for the early detection of cancer, initial evaluation of biomarkers for cancer screening, and definitive evaluation of biomarkers for cancer screening. Statistical and biological issues to improve this pipeline are discussed. Although various recommendations, such as identifying cases based on clinical symptoms, keeping biomarker tests simple, and adjusting for postscreening noise, have been made previously, they are not widely known. New recommendations include more frequent specimen collection to help identify promising biomarkers and the use of the paired availability design with interval cases (symptomatic cancers detected in the interval after screening) for initial evaluation of biomarkers for cancer screening.

Proportional Hazards Regression for Cancer Studies

There has been some recent work in the statistical literature for modeling the relationship between the size of cancers and probability of detecting metastasis, i.e., aggressive disease. Methods for assessing covariate effects in these studies are limited. In this article, we formulate the problem as assessing covariate effects on a right-censored variable subject to two types of sampling bias. The first is the length-biased sampling that is inherent in screening studies; the second is the two-phase design in which a fraction of tumors are measured. We construct estimation procedures for the proportional hazards model that account for these two sampling issues. In addition, a Nelson-Aalen type estimator is proposed as a summary statistic. Asymptotic results for the regression methodology are provided. The methods are illustrated by application to data from an observational cancer study as well as to simulated data.

Cumulative false positive recall rate and association with participant related factors in a population based breast cancer screening programme

STUDY OBJECTIVE

To investigate the cumulative false positive recall rate throughout the period of participation in a population based breast cancer screening programme and to examine its association with women related factors.

DESIGN

Analysis of a database to estimate the cumulative false positive recall rate after 10 biennial mammograms in a cohort of women. Cumulative risk after 10 rounds was calculated by projecting forward the information available on the four rounds. Logistic regression was used to evaluate the association between the cumulative risk of false positive recall and women related factors.

SETTING

Population based breast cancer screening programme in Barcelona City (Spain).

PARTICIPANTS

8502 women aged 50-69 years who participated in four consecutive screening rounds. Eligible women had received a mammogram in the first screening round between 1 December 1995 and 31 December 1996.

MAIN RESULTS

The false positive recall rate in the first screening for women who entered the screening programme at the age of 50-51 years was assessed at 10.6% (95% CI 8.9, 12.3). In the second screening this risk decreased to 3.8% (95% CI 2.7, 4.9) and remained almost constant in subsequent rounds. After 10 mammograms, the cumulative false positive recall rate was estimated at 32.4% (95% CI 29.7, 35.1). The factors associated with a higher cumulative risk of false positive recall were: previous benign breast disease (OR = 8.48; CI 7.39, 9.73), perimenopausal status (OR = 1.62; CI 1.12, 2.34), body mass index above 27.3 (OR = 1.17; CI 1.02, 1.34), and age 50-54 years (OR = 1.15; CI 1.00, 1.31).

CONCLUSIONS

One third of women could have at least one false positive recall over 10 biennial screens. Women participating in screening programmes should be informed about this risk, especially those with associated factors.

Evaluating markers for the early detection of cancer: overview of study designs and methods

BACKGROUND

The field of cancer biomarker development has been evolving rapidly. New developments both in the biologic and statistical realms are providing increasing opportunities for evaluation of markers for both early detection and diagnosis of cancer.

PURPOSE

To review the major conceptual and methodological issues in cancer biomarker evaluation, with an emphasis on recent developments in statistical methods together with practical recommendations.

METHODS

We organized this review by type of study: preliminary performance, retrospective performance, prospective performance and cancer screening evaluation.

RESULTS

For each type of study, we discuss methodologic issues, provide examples and discuss strengths and limitations.

CONCLUSION

Preliminary performance studies are useful for quickly winnowing down the number of candidate markers; however their results may not apply to the ultimate target population, asymptomatic subjects. If stored specimens from cohort studies with clinical cancer endpoints are available, retrospective studies provide a quick and valid way to evaluate performance of the markers or changes in the markers prior to the onset of clinical symptoms. Prospective studies have a restricted role because they require large sample sizes, and, if the endpoint is cancer on biopsy, there may be bias due to overdiagnosis. Cancer screening studies require very large sample sizes and long follow-up, but are necessary for evaluating the marker as a trigger of early intervention.

Breast biopsy and race/ethnicity among women without breast cancer

BACKGROUND

Breast biopsy is essential for definitive breast cancer diagnosis, but may also play a role in determining eligibility for breast cancer preventive measures or clinical trials. In addition, the prevalence of a history of negative breast biopsy can be viewed as an indicator of the adequacy or intensity of health care in a given population. We therefore analyzed the association of a history of breast biopsy with race/ethnicity and other factors in a cohort of women without a cancer diagnosis who completed a risk assessment form for participation in the Study of Tamoxifen and Raloxifene (STAR) and a sociodemographic questionnaire.

METHODS

Subjects were recruited at our large, urban teaching hospital. We developed a logistic regression model with biopsy (ever/never) as the outcome and age, race/ethnicity, educational attainment, and insurance coverage as the independent variables.

RESULTS

Among 805 unaffected predominantly minority subjects, white women were more than three times as likely as black and Hispanic women (OR=3.3, 95% CI 1.9-5.9), and insured women were twice as likely as uninsured women (OR=2.0, 95% CI 1.4-2.9) to have had a biopsy. Biopsy results were also associated with race/ethnicity.

DISCUSSION

We view these observations as hypothesis-generating rather than definitive. If confirmed, the associations we observed between negative biopsies and insurance status may reflect disparities in the timeliness and effectiveness of follow-up of suspicious lesions found via mammography. Our findings may also be relevant to the well-known association of breast cancer stage at diagnosis with low income and minority race/ethnicity.

Magnetic Resonance-Guided Large-Core Breast Biopsy Inside a 1.5-T Magnetic Resonance Scanner Using an Automatic System

RATIONALE AND OBJECTIVE

The aim of this study was to investigate the feasibility and the precision of magnetic resonance (MR)-guided large-core breast biopsies (LCBB) by using the second prototype of an automatic system (ROBITOM II), which is used to localize lesions while operating at the isocenter of a 1.5-T whole-body scanner.

METHODS AND MATERIALS

In comparison to the first prototype, ROBITOM II is equipped with a dedicated double breast coil and a high-speed trocar setting unit. In vitro experiments (n = 25) with grapefruit phantoms, which contained multiple vitamin E capsules (12 x 7 mm in size) as artificial lesions, were performed. Four patients with MR-detectable breast lesions underwent biopsy. A trocar was positioned in front of the lesion and inserted into the breast. Specimens were harvested with a coaxial technique by using a 14-G core needle biopsy gun.

RESULTS

In all 25 in vitro experiments, capsule material was detected in the specimen cylinder. In 4 patients, the coaxial needle was detected exactly at the expected position. Between 8 and 16 tissue cylinders were harvested. Histologic evaluation resulted in 1 invasive ductal carcinoma and 1 papilloma, which were confirmed after open surgery. One patient who had a proven breast cancer was biopsied for exclusion of multifocal disease. She showed fibrocystic changes, whereas open surgery revealed 3 small areas of ductal carcinoma in situ (DCIS). Another patient showed fibroadenoma after biopsy. This patient is in the follow-up period, which has lasted between 3 and 4 months up until now.

CONCLUSIONS

In this pilot patient study, the feasibility of manipulator-assisted large-core breast biopsy inside a 1.5-T whole-body scanner was demonstrated by using ROBITOM II. The precision of the device was confirmed with in vitro experiments. Although these findings are preliminary and the follow-up period is rather short, they nevertheless represent a successful proof-of-principle of LCBB with ROBITOM II.

Estimating the cumulative risk of a false-positive test in a repeated screening program

The goal of screening tests for a chronic disease such as cancer is early detection and treatment with a consequent reduction in mortality from the disease. Screening tests, however, might produce false positive and false-negative results. With an increasing number of screening tests, it is clear that the risk of a false-positive screen, a finding with potentially significant emotional, financial, and health costs, also increases. Elmore et al. (1998, New England Journal of Medicine 338, 1089-1096), Christiansen et al. (2000, Journal of the National Cancer Institute 92, 1657-1666), and Gelfand and Wang (2000, Statistics in Medicine 19, 1865-1879) investigated this problem under the somewhat unrealistic assumption that the choice of making the decision to drop out at the kth screen does not depend upon the results of the earlier k - 1 screens. In this article we obtain sufficient and necessary conditions for their assumption to hold and use one of them to provide a method for testing the validity of the assumption. A new model which does not depend on their assumption is introduced. The maximum likelihood estimator of the cumulative risk of receiving a false-positive screen under the new model is derived and its asymptotic normality is proved. The extension of the new model by incorporating covariate information is also considered. We apply our testing method and the new model to data from the breast cancer screening trial of the Health Insurance Plan of Greater New York.

The cumulative risk of a false-positive recall in the Norwegian Breast Cancer Screening Program

BACKGROUND

Biennial breast cancer screening for women ages 50-69 years is recommended by the World Health Organization. It has been claimed that the cumulative risk of a false-positive recall is a significant disadvantage in breast cancer screening programs. The primary objective of this study was to estimate the cumulative risk of a false-positive recall during a screening period of 20 years in women ages 50-51 years who are screened biennially in a population-based screening program. A secondary objective was to estimate the cumulative risk of undergoing fine-needle aspiration cytology, core needle biopsy, and open biopsy with benign morphology in the same group of women.

METHODS

The Norwegian Breast Cancer Screening Program invites all women ages 50-69 years who reside in the country to a 2-view mammography biennially. A nationwide data base that covers all of the invited women includes individual information about all screening activity. Results from three screening rounds in four counties were the basis for this study. False-positive recalls due to abnormal mammograms among 83,416 women who participated all the 3 screening rounds were the basis for the estimations.

RESULTS

It was calculated that women ages 50-51 years who participate in biennial screening run a cumulative risk of 20.8% for a false-positive recall during a screening period of 2 decades. The cumulative risk of undergoing fine-needle aspiration cytology was estimated at 3.9%, and the risk of undergoing core needle biopsy or open biopsy with benign morphology was 1.5% and 0.9%, respectively.

CONCLUSIONS

False-positive recalls are a disadvantage in a breast cancer screening programs, but the cumulative risk seemed to be acceptable in the Norwegian Breast Cancer Screening Program. It is important to communicate the existence and extent of this risk to the target group.