Projecting the Impact of Multi-Cancer Early Detection on Late-Stage Incidence Using Multi-State Disease Modeling.

BACKGROUND: Downstaging-reduction in late-stage incidence-has been proposed as an endpoint in randomized trials of multi-cancer early detection (MCED) tests. How downstaging depends on test performance and follow-up has been studied for some cancers but is understudied for cancers without existing screening and for MCED tests that include these cancer types. METHODS: We develop a model for cancer natural history that can be fit to registry incidence patterns under minimal inputs and can be estimated for solid cancers without existing screening. Fitted models are combined to project downstaging in MCED trials given sensitivity for early- and late-stage cancers. We fit models for 12 cancers using incidence data from the Surveillance, Epidemiology, and End Results program and project downstaging in a simulated trial under variable preclinical latencies and test sensitivities. RESULTS: A proof-of-principle lung cancer model approximated downstaging in the National Lung Screening Trial. Given published stage-specific sensitivities for 12 cancers, we projected downstaging ranging from 21% to 43% across plausible preclinical latencies in a hypothetical 3-screen MCED trial. Late-stage incidence reductions manifest soon after screening begins. Downstaging increases with longer early-stage latency or higher early-stage test sensitivity. CONCLUSIONS: Even short-term MCED trials could produce substantial downstaging given adequate early-stage test sensitivity. IMPACT: Modeling the natural histories of cancers without existing screening facilitates analysis of novel MCED products and trial designs. The framework informs expectations of MCED impact on disease stage at diagnosis and could serve as a building block for designing trials with late-stage incidence as the primary endpoint.

Racial disparities in prostate cancer mortality: a model-based decomposition of contributing factors.

To investigate the relative contributions of natural history and clinical interventions to racial disparities in prostate cancer mortality in the United States, we extended a model that was previously calibrated to Surveillance, Epidemiology, and End Results (SEER) incidence rates for the general population and for Black men. The extended model integrated SEER data on curative treatment frequencies and cancer-specific survival. Starting with the model for all men, we replaced up to 9 components with corresponding components for Black men, projecting age-standardized mortality rates for ages 40-84 years at each step. Based on projections in 2019, the increased frequency of developing disease, more aggressive tumor features, and worse cancer-specific survival in Black men diagnosed at local-regional and distant stages explained 38%, 34%, 22%, and 8% of the modeled disparity in mortality. Our results point to intensified screening and improved care in Black men as priority areas to achieve greater equity.

Melanoma literacy among the general population of three western US states.

Melanoma is a significant cause of cancer death, despite being detectable without specialized or invasive technologies. Understanding barriers to preventive behaviors such as skin self-examination (SSE) could help to define interventions for increasing the frequency of early detection. To determine melanoma knowledge and beliefs across three high-incidence US states, 15,000 surveys were sent to a population-representative sample. We aimed to assess (1) melanoma literacy (i.e., knowledge about melanoma risks, attitudes, and preventive behaviors) and (2) self-reported SSE and its association with melanoma literacy, self-efficacy, and belief in the benefits of SSE. Of 2326 respondents, only 21.2% provided responses indicating high knowledge of melanoma, and 62.8% reported performing an SSE at any time in their lives. Only 38.3% and 7.3% reported being "fairly" or "very" confident about doing SSE, respectively. SSE performance among respondents was most strongly associated with higher melanoma knowledge, higher self-efficacy, and personal history of melanoma. Melanoma literacy among survey respondents was modest, with greater literacy associated with a higher likelihood of reported preventive behavior. This assessment establishes a baseline and provides guidance for public health campaigns designed to increase prevention and early detection of this lethal cancer.

Risk of cancer versus risk of cancer diagnosis? Accounting for diagnostic bias in predictions of breast cancer risk by race and ethnicity.

OBJECTIVES: Cancer risk prediction may be subject to detection bias if utilization of screening is related to cancer risk factors. We examine detection bias when predicting breast cancer risk by race/ethnicity. METHODS: We used screening and diagnosis histories from the Breast Cancer Surveillance Consortium to estimate risk of breast cancer onset and calculated relative risk of onset and diagnosis for each racial/ethnic group compared with non-Hispanic White women. RESULTS: Of 104,073 women aged 40-54 receiving their first screening mammogram at a Breast Cancer Surveillance Consortium facility between 2000 and 2018, 10.2% (n = 10,634) identified as Asian, 10.9% (n = 11,292) as Hispanic, and 8.4% (n = 8719) as non-Hispanic Black. Hispanic and non-Hispanic Black women had slightly lower screening frequencies but biopsy rates following a positive mammogram were similar across groups. Risk of cancer diagnosis was similar for non-Hispanic Black and White women (relative risk vs non-Hispanic White = 0.90, 95% CI 0.65 to 1.14) but was lower for Asian (relative risk = 0.70, 95% CI 0.56 to 0.97) and Hispanic women (relative risk = 0.82, 95% CI 0.62 to 1.08). Relative risks of disease onset were 0.78 (95% CI 0.68 to 0.88), 0.70 (95% CI 0.59 to 0.83), and 0.95 (95% CI 0.84 to 1.09) for Asian, Hispanic, and non-Hispanic Black women, respectively. CONCLUSIONS: Racial/ethnic differences in mammography and biopsy utilization did not induce substantial detection bias; relative risks of disease onset were similar to or modestly different than relative risks of diagnosis. Asian and Hispanic women have lower risks of developing breast cancer than non-Hispanic Black and White women, who have similar risks.

Test sensitivity in a prospective cancer screening program: A critique of a common proxy measure.

The true sensitivity of a cancer screening test, defined as the frequency with which the test returns a positive result if the cancer is present, is a key indicator of diagnostic performance. Given the challenges of directly assessing test sensitivity in a prospective screening program, proxy measures for true sensitivity are frequently reported. We call one such proxy empirical sensitivity, as it is given by the observed ratio of screen-detected cancers to the sum of screen-detected and interval cancers. In the setting of the canonical three-state Markov model for progression from preclinical onset to clinical diagnosis, we formulate a mathematical relationship for how empirical sensitivity varies with the screening interval and the mean preclinical sojourn time and identify conditions under which empirical sensitivity exceeds or falls short of true sensitivity. In particular, when the inter-screening interval is short relative to the mean sojourn time, empirical sensitivity tends to exceed true sensitivity, unless true sensitivity is high. The Breast Cancer Surveillance Consortium (BCSC) has reported an estimate of 0.87 for the empirical sensitivity of digital mammography. We show that this corresponds to a true sensitivity of 0.82 under a mean sojourn time of 3.6 years estimated based on breast cancer screening trials. However, the BCSC estimate of empirical sensitivity corresponds to even lower true sensitivity under more contemporary, longer estimates of mean sojourn time. Consistently applied nomenclature that distinguishes empirical sensitivity from true sensitivity is needed to ensure that published estimates of sensitivity from prospective screening studies are properly interpreted.

Estimating stage-specific sensitivity for cancer screening tests.

OBJECTIVES: When evaluating potential new cancer screening modalities, estimating sensitivity, especially for early-stage cases, is critical. There are methods to approximate stage-specific sensitivity in asymptomatic populations, both in the prospective (active screening) and retrospective (stored specimens) scenarios. We explored their validity via a simulation study. METHODS: We fit natural history models to lung and ovarian cancer screening data that permitted estimation of stage-specific (early/late) true sensitivity, defined as the probability subjects screened in the given stage had positive tests. We then ran simulations, using the fitted models, of the prospective and retrospective scenarios. Prospective sensitivity by stage was estimated as screen-detected divided by screen-plus interval-detected cancers, where stage is defined as stage at detection. Retrospective sensitivity by stage was estimated based on cancers detected within specified windows before clinical diagnosis with stage defined as stage at clinical diagnosis. RESULTS: Stage-specific true sensitivities estimated by the lung cancer natural history model were 47% (early) and 63% (late). Simulation results for the prospective setting gave estimated sensitivities of 81% (early) versus 62% (late). In the retrospective scenario, early/late sensitivity estimates were 35%/57% (1-year window) and 27%/49% (2-year window). In the prospective scenario, most subjects with negative early-stage screens presented as other than early-stage interval cases. Results were similar for ovarian cancer, with estimated prospective sensitivity much greater than true sensitivity for early stage, 84% versus 25%. CONCLUSIONS: Existing methods for approximating stage-specific sensitivity in both prospective and retrospective scenarios are unsatisfactory; improvements are needed before they can be considered to be reliable.

Modeling the impact of novel systemic treatments on lung cancer screening benefits.

BACKGROUND: Since low-dose computed tomography (LDCT) screening was shown to be effective in the National Lung Screening Trial (NLST), novel targeted therapies and immunotherapies for advanced lung cancer have become available. This study investigated the impact of these treatment advances on the expected benefits of LDCT screening. METHODS: A microsimulation model of LDCT screening for high-risk individuals under standard systemic treatments (chemotherapy and radiation therapy) and novel treatments (immunotherapy and targeted therapy) was used. The model assumed a reduction in advanced-stage disease consistent with the NLST, and given the stage at diagnosis, it projected survival. The disease-specific relative mortality reduction (MR) due to LDCT screening was projected in the trial setting and in a population eligible for LDCT screening under the current US Preventive Services Task Force (USPSTF) recommendations. RESULTS: The availability of novel treatments reduced the MR in the LDCT arm of the NLST from 15% to 13.5% and the number of lung cancer deaths prevented from 310 to 224 per 100,000 persons screened. Over 10 years, population LDCT screening based on USPSTF recommendations prevented 374 lung cancer deaths per 100,000 under standard treatments (13.3% MR) and 236 per 100,000 under fully adopted novel treatments (10.6% MR). The number needed to screen to avert one death over 10 years was 270 under standard treatments and 440 under novel treatments. CONCLUSIONS: The transition from standard systemic treatments to novel treatments is expected to reduce the relative and absolute mortality benefits of LDCT screening. Benefit-harm tradeoffs of LDCT screening are likely to change as novel treatments become widespread.

Evaluation of Prostate Cancer Screening Strategies in a Low-Resource, High-risk Population in the Bahamas.

Importance: The benefit of prostate-specific antigen screening may be greatest in high-risk populations, including men of African descent in the Caribbean. However, organized screening may not be sustainable in low- and middle-income countries. Objective: To evaluate the expected population outcomes and resource use of conservative prostate-specific antigen screening programs in the Bahamas. Design Setting and Participants: Prostate cancer incidence from GLOBOCAN and prostate-specific antigen screening data for 4300 men from the Bahamas were used to recalibrate 2 decision analytical models previously used to study prostate-specific antigen screening for Black men in the United States. Data on age and results obtained from prostate-specific antigen screening tests performed in Nassau from 2004 to 2018 and in Freeport from 2013 to 2018 were used. Data were analyzed from January 15, 2021, to March 23, 2022. Interventions: One or 2 screenings for men aged 45 to 60 years and conservative criteria for biopsy (prostate-specific antigen level >10 ng/mL) and curative treatment (Gleason score ≥8) were modeled. Categories of Gleason scores were 6 or lower, 7, and 8 or higher, with higher scores indicating higher risk of cancer progression and death. Main Outcomes and Measures: Projected numbers of tests and biopsies, prostate cancer (over)diagnoses, lives saved, and life-years gained owing to screening from 2022 to 2040. Results: In this decision analytical modeling study, screening histories from 4300 men (median age, 54 years; range, 13-101 years) tested between 2004 and 2018 at 2 sites in the Bahamas were used to inform the models. Screening once at 60 years of age was projected to involve 40 000 to 42 000 tests (range between models) and prevent 500 to 600 of 10 000 to 14 000 prostate cancer deaths. Screening at 50 and 60 years doubled the number of tests but increased lives saved by only 15% to 16%. Among onetime strategies, screening once at 60 years of age involved the fewest tests per life saved (74-84 tests) and curative treatments per life saved (1.2-2.8 treatments). Conclusions and Relevance: The findings of this decision analytical modeling study of prostate cancer screening in the Bahamas suggest that limited screening offered modest benefits that varied with screening ages and number of tests. The results can be combined with data on capacity constraints and evaluated relative to competing national public health priorities.

Cancer Modeling as Learning Experience.

Cancer modeling has become an accepted method for generating evidence about comparative effectiveness and cost-effectiveness of candidate cancer control policies across the continuum of care. Models of early detection policies require inputs concerning disease natural history and screening test performance, which are often subject to considerable uncertainty. Model validation against an external data source can increase confidence in the reliability of assumed or calibrated inputs. When a model fails to validate, this presents an opportunity to revise these inputs, thereby learning new information about disease natural history or diagnostic performance that could both enhance the model results and inform real-world practices. We discuss the conditions necessary for validly drawing conclusions about specific inputs such as diagnostic performance from model validation studies. Doing so requires being able to faithfully replicate the validation study in terms of its design and implementation and being alert to the problem of non-identifiability, which could lead to explanations for failure to validate other than those identified. See related article by Rutter et al., p. 775.

Estimation of Breast Cancer Overdiagnosis in a U.S. Breast Screening Cohort.

BACKGROUND: Mammography screening can lead to overdiagnosis-that is, screen-detected breast cancer that would not have caused symptoms or signs in the remaining lifetime. There is no consensus about the frequency of breast cancer overdiagnosis. OBJECTIVE: To estimate the rate of breast cancer overdiagnosis in contemporary mammography practice accounting for the detection of nonprogressive cancer. DESIGN: Bayesian inference of the natural history of breast cancer using individual screening and diagnosis records, allowing for nonprogressive preclinical cancer. Combination of fitted natural history model with life-table data to predict the rate of overdiagnosis among screen-detected cancer under biennial screening. SETTING: Breast Cancer Surveillance Consortium (BCSC) facilities. PARTICIPANTS: Women aged 50 to 74 years at first mammography screen between 2000 and 2018. MEASUREMENTS: Screening mammograms and screen-detected or interval breast cancer. RESULTS: The cohort included 35 986 women, 82 677 mammograms, and 718 breast cancer diagnoses. Among all preclinical cancer cases, 4.5% (95% uncertainty interval [UI], 0.1% to 14.8%) were estimated to be nonprogressive. In a program of biennial screening from age 50 to 74 years, 15.4% (UI, 9.4% to 26.5%) of screen-detected cancer cases were estimated to be overdiagnosed, with 6.1% (UI, 0.2% to 20.1%) due to detecting indolent preclinical cancer and 9.3% (UI, 5.5% to 13.5%) due to detecting progressive preclinical cancer in women who would have died of an unrelated cause before clinical diagnosis. LIMITATIONS: Exclusion of women with first mammography screen outside BCSC. CONCLUSION: On the basis of an authoritative U.S. population data set, the analysis projected that among biennially screened women aged 50 to 74 years, about 1 in 7 cases of screen-detected cancer is overdiagnosed. This information clarifies the risk for breast cancer overdiagnosis in contemporary screening practice and should facilitate shared and informed decision making about mammography screening. PRIMARY FUNDING SOURCE: National Cancer Institute.

Impact of cancer screening on metastasis: A prostate cancer case study.

BACKGROUND: Trials of cancer screening present results in terms of deaths prevented, but metastasis is also a key endpoint that screening seeks to prevent. We developed a framework for projecting overall (de novo and progressive) metastases prevented in a screening trial using prostate cancer screening as a case study. METHODS: Mechanistic simulation model in which screening shifts a fraction of cases that would be metastatic at diagnosis to being non-metastatic. This shift increases the incidence of non-overdiagnosed, organ-confined cases. We use estimates of the risk of metastatic progression for these cases to project how many progress to metastasis after diagnosis and tally the projected de novo and progressive metastatic cases with and without screening. We use data on stage shift from the European Randomized Study of Screening for Prostate Cancer (ERSPC) and data on the risk of metastatic progression from the Scandinavian Prostate Cancer Group-4 trial. We estimate the relative risk and absolute risk reductions in metastatic disease at diagnosis and compare these with reductions in overall metastases. RESULTS: Assuming no effect of screening beyond initial stage shift at diagnosis, the model projects a 43% reduction in metastasis at diagnosis but a 22% reduction in the cumulative probability of metastasis over 12 years in favor of screening. These results are consistent with the empirical findings from the ERSPC. CONCLUSION: Any reduction in metastatic disease at diagnosis under screening is likely to be an overly optimistic predictor of the impact of screening on overall metastasis and disease-specific mortality.

A Latent Disease Model to Reduce Detection Bias in Cancer Risk Prediction Studies.

In studies of cancer risk, detection bias arises when risk factors are associated with screening patterns, affecting the likelihood and timing of diagnosis. To eliminate detection bias in a screened cohort, we propose modeling the latent onset of cancer and estimating the association between risk factors and onset rather than diagnosis. We apply this framework to estimate the increase in prostate cancer risk associated with black race and family history using data from the SELECT prostate cancer prevention trial, in which men were screened and biopsied according to community practices. A positive family history was associated with a hazard ratio (HR) of prostate cancer onset of 1.8, lower than the corresponding HR of prostate cancer diagnosis (HR = 2.2). This result comports with a finding that men in SELECT with a family history were more likely to be biopsied following a positive PSA test than men with no family history. For black race, the HRs for onset and diagnosis were similar, consistent with similar patterns of screening and biopsy by race. If individual screening and diagnosis histories are available, latent disease modeling can be used to decouple risk of disease from risk of disease diagnosis and reduce detection bias.

Correction: Incorporating Breast Cancer Recurrence Events Into Population-Based Cancer Registries Using Medical Claims: Cohort Study.

[This corrects the article DOI: 10.2196/18143.].

Incorporating Breast Cancer Recurrence Events Into Population-Based Cancer Registries Using Medical Claims: Cohort Study.

BACKGROUND: There is a need for automated approaches to incorporate information on cancer recurrence events into population-based cancer registries. OBJECTIVE: The aim of this study is to determine the accuracy of a novel data mining algorithm to extract information from linked registry and medical claims data on the occurrence and timing of second breast cancer events (SBCE). METHODS: We used supervised data from 3092 stage I and II breast cancer cases (with 394 recurrences), diagnosed between 1993 and 2006 inclusive, of patients at Kaiser Permanente Washington and cases in the Puget Sound Cancer Surveillance System. Our goal was to classify each month after primary treatment as pre- versus post-SBCE. The prediction feature set for a given month consisted of registry variables on disease and patient characteristics related to the primary breast cancer event, as well as features based on monthly counts of diagnosis and procedure codes for the current, prior, and future months. A month was classified as post-SBCE if the predicted probability exceeded a probability threshold (PT); the predicted time of the SBCE was taken to be the month of maximum increase in the predicted probability between adjacent months. RESULTS: The Kaplan-Meier net probability of SBCE was 0.25 at 14 years. The month-level receiver operating characteristic curve on test data (20% of the data set) had an area under the curve of 0.986. The person-level predictions (at a monthly PT of 0.5) had a sensitivity of 0.89, a specificity of 0.98, a positive predictive value of 0.85, and a negative predictive value of 0.98. The corresponding median difference between the observed and predicted months of recurrence was 0 and the mean difference was 0.04 months. CONCLUSIONS: Data mining of medical claims holds promise for the streamlining of cancer registry operations to feasibly collect information about second breast cancer events.

Measuring problem prescription opioid use among patients receiving long-term opioid analgesic treatment: development and evaluation of an algorithm for use in EHR and claims data.

Objective: Opioid surveillance in response to the opioid epidemic will benefit from scalable, automated algorithms for identifying patients with clinically documented signs of problem prescription opioid use. Existing algorithms lack accuracy. We sought to develop a high-sensitivity, high-specificity classification algorithm based on widely available structured health data to identify patients receiving chronic extended-release/long-acting (ER/LA) therapy with evidence of problem use to support subsequent epidemiologic investigations. Methods: Outpatient medical records of a probability sample of 2,000 Kaiser Permanente Washington patients receiving ≥60 days' supply of ER/LA opioids in a 90-day period from 1 January 2006 to 30 June 2015 were manually reviewed to determine the presence of clinically documented signs of problem use and used as a reference standard for algorithm development. Using 1,400 patients as training data, we constructed candidate predictors from demographic, enrollment, encounter, diagnosis, procedure, and medication data extracted from medical claims records or the equivalent from electronic health record (EHR) systems, and we used adaptive least absolute shrinkage and selection operator (LASSO) regression to develop a model. We evaluated this model in a comparable 600-patient validation set. We compared this model to ICD-9 diagnostic codes for opioid abuse, dependence, and poisoning. This study was registered with ClinicalTrials.gov as study NCT02667262 on 28 January 2016. Results: We operationalized 1,126 potential predictors characterizing patient demographics, procedures, diagnoses, timing, dose, and location of medication dispensing. The final model incorporating 53 predictors had a sensitivity of 0.582 at positive predictive value (PPV) of 0.572. ICD-9 codes for opioid abuse, dependence, and poisoning had a sensitivity of 0.390 at PPV of 0.599 in the same cohort. Conclusions: Scalable methods using widely available structured EHR/claims data to accurately identify problem opioid use among patients receiving long-term ER/LA therapy were unsuccessful. This approach may be useful for identifying patients needing clinical evaluation.

Lifetime Benefits and Harms of Prostate-Specific Antigen-Based Risk-Stratified Screening for Prostate Cancer.

BACKGROUND: Studies conducted in Swedish populations have shown that men with lowest prostate-specific antigen (PSA) levels at ages 44-50 years and 60 years have very low risk of future distant metastasis or death from prostate cancer. This study investigates benefits and harms of screening strategies stratified by PSA levels. METHODS: PSA levels and diagnosis patterns from two microsimulation models of prostate cancer progression, detection, and mortality were compared against results of the Malmö Preventive Project, which stored serum and tracked subsequent prostate cancer diagnoses for 25 years. The models predicted the harms (tests and overdiagnoses) and benefits (lives saved and life-years gained) of PSA-stratified screening strategies compared with biennial screening from age 45 years to age 69 years. RESULTS: Compared with biennial screening for ages 45-69 years, lengthening screening intervals for men with PSA less than 1.0 ng/mL at age 45 years led to 46.8-47.0% fewer tests (range between models), 0.9-2.1% fewer overdiagnoses, and 3.1-3.8% fewer lives saved. Stopping screening when PSA was less than 1.0 ng/mL at age 60 years and older led to 12.8-16.0% fewer tests, 5.0-24.0% fewer overdiagnoses, and 5.0-13.1% fewer lives saved. Differences in model results can be partially explained by differences in assumptions about the link between PSA growth and the risk of disease progression. CONCLUSION: Relative to a biennial screening strategy, PSA-stratified screening strategies investigated in this study substantially reduced the testing burden and modestly reduced overdiagnosis while preserving most lives saved. Further research is needed to clarify the link between PSA growth and disease progression.

Prostate cancer mortality and metastasis under different biopsy frequencies in North American active surveillance cohorts.

BACKGROUND: Active surveillance (AS) is an accepted means of managing low-risk prostate cancer. Because of the rarity of downstream events, data from existing AS cohorts cannot yet address how differences in surveillance intensity affect metastasis and mortality. This study projected the comparative benefits of different AS schedules in men diagnosed with prostate cancer who had Gleason score (GS) ≤6 disease and risk profiles similar to those in North American AS cohorts. METHODS: Times of GS upgrading were simulated based on AS data from the University of Toronto, Johns Hopkins University, the University of California at San Francisco, and the Canary Pass Active Surveillance Cohort. Times to metastasis and prostate cancer death, informed by models from the Scandinavian Prostate Cancer Group 4 trial, were projected under biopsy surveillance schedules ranging from watchful waiting to annual biopsies. Outcomes included the risk of metastasis, the risk of death, remaining life-years (LYs), and quality-adjusted LYs. RESULTS: Compared with watchful waiting, AS biopsies reduced the risk of prostate cancer metastasis and prostate cancer death at 20 years by 1.4% to 3.3% and 1.0% to 2.4%, respectively; and 5-year biopsies reduced the risk of metastasis and prostate cancer death by 1.0% to 2.4% and 0.6% to 1.6%, respectively. There was little difference between annual and 5-year biopsy schedules in terms of LYs (range of differences, 0.04-0.16 LYs) and quality-adjusted LYs (range of differences, -0.02 to 0.09 quality-adjusted LYs). CONCLUSIONS: Among men diagnosed with GS ≤6 prostate cancer, obtaining a biopsy every 3 or 4 years appears to be an acceptable alternative to more frequent biopsies. Reducing surveillance intensity for those who have a low risk of progression reduces the number of biopsies while preserving the benefit of more frequent schedules.

ESTIMATING AND COMPARING CANCER PROGRESSION RISKS UNDER VARYING SURVEILLANCE PROTOCOLS.

Outcomes after cancer diagnosis and treatment are often observed at discrete times via doctor-patient encounters or specialized diagnostic examinations. Despite their ubiquity as endpoints in cancer studies, such outcomes pose challenges for analysis. In particular, comparisons between studies or patient populations with different surveillance schema may be confounded by differences in visit frequencies. We present a statistical framework based on multistate and hidden Markov models that represents events on a continuous time scale given data with discrete observation times. To demonstrate this framework, we consider the problem of comparing risks of prostate cancer progression across multiple active surveillance cohorts with different surveillance frequencies. We show that the different surveillance schedules partially explain observed differences in the progression risks between cohorts. Our application permits the conclusion that differences in underlying cancer progression risks across cohorts persist after accounting for different surveillance frequencies.

A Framework for Treatment Decision Making at Prostate Cancer Recurrence.

BACKGROUND: Of the 50,000 men in the US who elect for radical prostatectomy for prostate cancer, 24% to 40% will have a prostate-specific antigen (PSA) recurrence (PSA-R) within 10 years. Deciding whether to administer salvage therapy (ST) at PSA-R presents challenges, as treatment reduces the risk of progression to clinical metastasis but incurs unnecessary side effects should the man die before metastasis. We have developed a new harm-benefit framework using a clinical cohort to inform shared decision making between patients and physicians at PSA-R. METHODS: Records of 1,045 Johns Hopkins University Hospital patients diagnosed between 1984 and 2013 who had PSA-R following radical prostatectomy were analyzed using marginal structural models to estimate the baseline risk of metastasis and the effect of ST (radiation therapy with or without hormone therapy) while accounting for selection into ST on the basis of PSA growth. The estimated model predicts the harm-benefit tradeoffs of ST within patient subgroups. The benefit of ST is the absolute reduction in the risk of metastasis within 10 years; the harm is the frequency of cancers that would not have metastasized in the patient's lifetime in the absence of ST (overtreatment). RESULTS: The adjusted hazard ratio associated with ST was 0.41 (95% CI, 0.31 to 0.55). Providing ST to all men at PSA-R reduced the risk of metastasis from 43% to 23% but led to 31% of men being overtreated (harm/benefit = 31/(43-23) = 1.6). Providing ST to men with Gleason score >7 reduced the risk of metastasis from 67% to 39%, with 13% of men being overtreated (harm/benefit = 13/(67-39) = 0.5). CONCLUSIONS: A quantitative framework that evaluates primary harms and benefits of ST after PSA-R will facilitate informed decision making. Immediate ST may be more appropriate in patient subgroups at elevated risk of metastasis.

Subsequent Breast Cancer Risk Following Diagnosis of Atypical Ductal Hyperplasia on Needle Biopsy.

IMPORTANCE: Atypical ductal hyperplasia (ADH) is a known risk factor for breast cancer. Published risk estimates are based on cohorts that included women whose ADH was diagnosed before widespread use of screening mammograms and did not differentiate between the methods used to diagnose ADH, which may be related to the size of the ADH focus. These risks may overestimate the risk in women with presently diagnosed ADH. OBJECTIVE: To examine the risk of invasive cancer associated with ADH diagnosed using core needle biopsy vs excisional biopsy. DESIGN: A cohort study was conducted comparing the 10-year cumulative risk of invasive breast cancer in 955 331 women undergoing mammography with and without a diagnosis of ADH. Data were obtained from 5 breast imaging registries that participate in the National Cancer Institute-funded Breast Cancer Surveillance Consortium. EXPOSURES: Diagnosis of ADH on core needle biopsy or excisional biopsy in women undergoing mammography. MAIN OUTCOMES AND MEASURES: Ten-year cumulative risk of invasive breast cancer. RESULTS: The sample included 955 331 women with 1727 diagnoses of ADH, 1058 (61.3%) of which were diagnosed by core biopsy and 635 (36.8%) by excisional biopsy. The mean (interquartile range) age of the women at diagnosis was 52.6 (46.9-60.4) years. From 1996 to 2012, the proportion of ADH diagnosed by core needle biopsy increased from 21% to 77%. Ten years following a diagnosis of ADH, the cumulative risk of invasive breast cancer was 2.6 (95% CI, 2.0-3.4) times higher than the risk in women with no ADH. Atypical ductal hyperplasia diagnosed via excisional biopsy was associated with an adjusted hazard ratio (HR) of 3.0 (95% CI, 2-4.5) and, via core needle biopsy, with an adjusted HR of 2.2 (95% CI, 1.5-3.4). Ten years after an ADH diagnosis, an estimated 5.7% (95% CI, 4.3%-10.1%) of the women had a diagnosis of invasive cancer. Women with ADH diagnosed on excisional biopsy had a slightly higher risk (6.7%; 95% CI, 3.0%-12.8%) compared with those with ADH diagnosed via core needle biopsy (5%; 95% CI, 2.2%-8.9%). CONCLUSIONS AND RELEVANCE: Current 10-year risks of invasive breast cancer after a diagnosis of ADH may be lower than those previously reported. The risk associated with ADH is slightly lower for women whose ADH was diagnosed by needle core biopsy compared with excisional biopsy.