Long-term cardiovascular disease risk after anthracycline and trastuzumab treatments in U.S. breast cancer survivors.

BACKGROUND: Although breast cancer survivors are at risk for cardiovascular disease (CVD) from treatment late effects, evidence to inform long-term and age-specific cardiovascular surveillance recommendations is lacking. METHODS: We conducted a retrospective cohort study of 10,211 women diagnosed with first primary unilateral breast cancer in Kaiser Permanente Washington or Colorado (aged 20+, survived ≥1 year). We estimated multivariable adjusted hazard ratios (aHR) for associations between initial chemotherapy regimen type (anthracycline and/or trastuzumab, other chemotherapies, no chemotherapy [reference]) and CVD risk, adjusted for patient characteristics, other treatments, and CVD risk factors. Cumulative incidence was calculated considering competing events. RESULTS: After 5.79 median years, 14.67% of women developed CVD (cardiomyopathy/heart failure (CM/HF), ischemic heart disease (IHD), stroke). Women treated with anthracyclines and/or trastuzumab had a higher risk of CVD compared with no chemotherapy (aHR=1.53,95%CI=1.31-1.79), persisting 5+years post-diagnosis (aHR5-<10 years=1.85,95%CI=1.44-2.39;aHR10+ years=1.83,95%CI=1.34-2.49). CM/HF risks were elevated among women treated with anthracyclines and/or trastuzumab compared with no chemotherapy, especially for ages<65 (aHR20-54years=2.97,95%CI=1.72-5.12;aHR55-64years=2.21,95%CI=1.52-3.21), differing for older women (aHR65+years=1.32,95%CI=0.97-1.78), and 5+years post-diagnosis (aHR5-<10years=1.89,95%CI=1.35-2.64;aHR10+years=2.21,95%CI=1.52-3.20). Anthracyclines and/or trastuzumab receipt was associated with increased IHD risks after 5+years (aHR5-<10years=1.51,95%CI=1.06-2.14;aHR10+years=1.86,95%CI=1.18-2.93) with no clear age effects, and stroke risk (aHR=1.33,95%CI=1.05-1.69) which did not vary by time or age. There was some evidence of long-term CM/HF and IHD risks with other chemotherapies. Among women aged<65 treated with anthracyclines and/or trastuzumab, up to 16% developed CVD by 10-years (20-54=6.91%;55-64=16.00%), driven by CM/HF (20-54=3.90%;55-64=9.78%). CONCLUSIONS: We found increased long-term risks of CM/HF and IHD among breast cancer survivors treated with anthracyclines and/or trastuzumab, and increased CM/HF risk among women aged<65.

Strategies to Increase Cervical Cancer Screening With Mailed Human Papillomavirus Self-Sampling Kits: A Randomized Clinical Trial.

Importance: Optimal strategies for increasing cervical cancer screening may differ by patient screening history and health care setting. Mailing human papillomavirus (HPV) self-sampling kits to individuals who are overdue for screening increases adherence; however, offering self-sampling kits to screening-adherent individuals has not been evaluated in the US. Objective: To evaluate the effectiveness of direct-mail and opt-in approaches for offering HPV self-sampling kits to individuals by cervical cancer screening history (screening-adherent and currently due, overdue, or unknown). Design, Setting, and Participants: Randomized clinical trial conducted in Kaiser Permanente Washington, a US integrated health care delivery system. Individuals aged 30 to 64 years with female sex, a primary care clinician, and no hysterectomy were identified through electronic health records (EHRs) and enrolled between November 20, 2020, and January 28, 2022, with follow-up through July 29, 2022. Interventions: Individuals stratified as due (eg, at the time of randomization, these individuals have been previously screened and are due for their next screening in ≤3 months) were randomized to receive usual care (patient reminders and clinician EHR alerts [n = 3671]), education (usual care plus educational materials about screening [n = 3960]), direct mail (usual care plus educational materials and a mailed self-sampling kit [n = 1482]), or to opt in (usual care plus educational materials and the option to request a kit [n = 3956]). Individuals who were overdue for screening were randomized to receive usual care (n = 5488), education (n = 1408), or direct mail (n = 1415). Individuals with unknown history for screening were randomized to receive usual care (n = 2983), education (n = 3486), or to opt in (n = 3506). Main Outcomes and Measures: The primary outcome was screening completion within 6 months. Primary analyses compared direct-mail or opt-in participants with individuals randomized to the education group. Results: The intention-to-treat analyses included 31 355 randomized individuals (mean [SD] age, 45.9 [10.4] years). Among those who were due for screening, compared with receiving education alone (1885 [47.6%]), screening completion was 14.1% (95% CI, 11.2%-16.9%) higher in the direct-mail group (914 [61.7%]) and 3.5% (95% CI, 1.2%-5.7%) higher in the opt-in group (2020 [51.1%]). Among individuals who were overdue, screening completion was 16.9% (95% CI, 13.8%-20.0%) higher in the direct-mail group (505 [35.7%]) compared with education alone (264 [18.8%]). Among those with unknown history, screening was 2.2% (95% CI, 0.5%-3.9%) higher in the opt-in group (634 [18.1%]) compared with education alone (555 [15.9%]). Conclusions and Relevance: Within a US health care system, direct-mail self-sampling increased cervical cancer screening by more than 14% in individuals who were due or overdue for cervical cancer screening. The opt-in approach minimally increased screening. To increase screening adherence, systems implementing HPV self-sampling should prioritize direct-mail outreach for individuals who are due or overdue for screening. For individuals with unknown screening history, testing alternative outreach approaches and additional efforts to document screening history are warranted. Trial Registration: ClinicalTrials.gov Identifier: NCT04679675.

Development and evaluation of a method to assess breast cancer risk using a longitudinal history of mammographic density: a cohort study.

BACKGROUND: Women with dense breasts have an increased risk of breast cancer. However, breast density is measured with variability, which may reduce the reliability and accuracy of its association with breast cancer risk. This is particularly relevant when visually assessing breast density due to variation in inter- and intra-reader assessments. To address this issue, we developed a longitudinal breast density measure which uses an individual woman's entire history of mammographic density, and we evaluated its association with breast cancer risk as well as its predictive ability. METHODS: In total, 132,439 women, aged 40-73 yr, who were enrolled in Kaiser Permanente Washington and had multiple screening mammograms taken between 1996 and 2013 were followed up for invasive breast cancer through 2014. Breast Imaging Reporting and Data System (BI-RADS) density was assessed at each screen. Continuous and derived categorical longitudinal density measures were developed using a linear mixed model that allowed for longitudinal density to be updated at each screen. Predictive ability was assessed using (1) age and body mass index-adjusted hazard ratios (HR) for breast density (time-varying covariate), (2) likelihood-ratio statistics (ΔLR-χ2) and (3) concordance indices. RESULTS: In total, 2704 invasive breast cancers were diagnosed during follow-up (median = 5.2 yr; median mammograms per woman = 3). When compared with an age- and body mass index-only model, the gain in statistical information provided by the continuous longitudinal density measure was 23% greater than that provided by BI-RADS density (follow-up after baseline mammogram: ΔLR-χ2 = 379.6 (degrees of freedom (df) = 2) vs. 307.7 (df = 3)), which increased to 35% (ΔLR-χ2 = 251.2 vs. 186.7) for follow-up after three mammograms (n = 76,313, 2169 cancers). There was a sixfold difference in observed risk between densest and fattiest eight-category longitudinal density (HR = 6.3, 95% CI 4.7-8.7), versus a fourfold difference with BI-RADS density (HR = 4.3, 95% CI 3.4-5.5). Discriminatory accuracy was marginally greater for longitudinal versus BI-RADS density (c-index = 0.64 vs. 0.63, mean difference = 0.008, 95% CI 0.003-0.012). CONCLUSIONS: Estimating mammographic density using a woman's history of breast density is likely to be more reliable than using the most recent observation only, which may lead to more reliable and accurate estimates of individual breast cancer risk. Longitudinal breast density has the potential to improve personal breast cancer risk estimation in women attending mammography screening.

Impact of different financial incentive structures on a web-based health survey: do timing and amount matter?

Aim: Financial incentives improve response to electronic health surveys, yet little is known about how unconditional incentives (guaranteed regardless of survey completion), conditional incentives, and various combinations of incentives influence response rates. We compared electronic health survey completion with two different financial incentive structures. Methods: We invited women aged 30-64 years enrolled in a U.S. healthcare system and overdue for Pap screening to complete a web-based survey after receiving a mailed human papillomavirus (HPV) self-sampling kit in a pragmatic trial. HPV kit returners (n = 272) and non-returners (n = 1,083) were allocated to one of two different incentive structures: (1) Unconditional: $5 pre-incentive only (n = 653); (2) Combined: $2 pre-incentive plus $10 post-incentive conditional on completion (n = 702). Chi-square tests evaluated whether survey completion differed by incentive structure within kit return groups or was modified by kit return status. For each incentive-by-kit status group, the cost-per-survey response was calculated as: ([number invited*pre-incentive amount] + [number responses*post-incentive amount]) / number responses. Results: Overall, survey response was higher in kit returners vs. kit non-returners (42.6% vs. 11.0%, P < 0.01), and survey response was higher in the combined (20.1%) vs. unconditional (14.4%) incentive group (P = 0.01). Kit return status did not modify the association between incentive type and survey response (P = 0.52). Among respondents, time to survey completion did not differ by incentive type among either kit returners or non-returners. Among returners, the cost-per-survey response was similar between groups ($13.57 unconditional; $14.15 combined); among non-returners, the cost was greater in the unconditional ($57.78) versus the combined ($25.22) group. Conclusion: A combined incentive can be cost-effective for increasing survey response in health services research, particularly in hard-to-reach populations.

Impact of Surveillance Mammography Intervals Less Than One Year on Performance Measures in Women With a Personal History of Breast Cancer.

OBJECTIVE: When multiple surveillance mammograms are performed within an annual interval, the current guidance for one-year follow-up to determine breast cancer status results in shared follow-up periods in which a single breast cancer diagnosis can be attributed to multiple preceding examinations, posing a challenge for standardized performance assessment. We assessed the impact of using follow-up periods that eliminate the artifactual inflation of second breast cancer diagnoses. MATERIALS AND METHODS: We evaluated surveillance mammograms from 2007-2016 in women with treated breast cancer linked with tumor registry and pathology outcomes. Second breast cancers included ductal carcinoma in situ or invasive breast cancer diagnosed during one-year follow-up. The cancer detection rate, interval cancer rate, sensitivity, and specificity were compared using different follow-up periods: standard one-year follow-up per the American College of Radiology versus follow-up that was shortened at the next surveillance mammogram if less than one year (truncated follow-up). Performance measures were calculated overall and by indication (screening, evaluation for breast problem, and short interval follow-up). RESULTS: Of 117971 surveillance mammograms, 20% (n = 23533) were followed by another surveillance mammogram within one year. Standard follow-up identified 1597 mammograms that were associated with second breast cancers. With truncated follow-up, the breast cancer status of 179 mammograms (11.2%) was revised, resulting in 1418 mammograms associated with unique second breast cancers. The interval cancer rate decreased with truncated versus standard follow-up (3.6 versus 4.9 per 1000 mammograms, respectively), with a difference (95% confidence interval [CI]) of -1.3 (-1.6, -1.1). The overall sensitivity increased to 70.4% from 63.7%, for the truncated versus standard follow-up, with a difference (95% CI) of 6.6% (5.6%, 7.7%). The specificity remained stable at 98.1%. CONCLUSION: Truncated follow-up, if less than one year to the next surveillance mammogram, enabled second breast cancers to be associated with a single preceding mammogram and resulted in more accurate estimates of diagnostic performance for national benchmarks.

Comparison of Mammography AI Algorithms with a Clinical Risk Model for 5-year Breast Cancer Risk Prediction: An Observational Study.

Background Although several clinical breast cancer risk models are used to guide screening and prevention, they have only moderate discrimination. Purpose To compare selected existing mammography artificial intelligence (AI) algorithms and the Breast Cancer Surveillance Consortium (BCSC) risk model for prediction of 5-year risk. Materials and Methods This retrospective case-cohort study included data in women with a negative screening mammographic examination (no visible evidence of cancer) in 2016, who were followed until 2021 at Kaiser Permanente Northern California. Women with prior breast cancer or a highly penetrant gene mutation were excluded. Of the 324 009 eligible women, a random subcohort was selected, regardless of cancer status, to which all additional patients with breast cancer were added. The index screening mammographic examination was used as input for five AI algorithms to generate continuous scores that were compared with the BCSC clinical risk score. Risk estimates for incident breast cancer 0 to 5 years after the initial mammographic examination were calculated using a time-dependent area under the receiver operating characteristic curve (AUC). Results The subcohort included 13 628 patients, of whom 193 had incident cancer. Incident cancers in eligible patients (additional 4391 of 324 009) were also included. For incident cancers at 0 to 5 years, the time-dependent AUC for BCSC was 0.61 (95% CI: 0.60, 0.62). AI algorithms had higher time-dependent AUCs than did BCSC, ranging from 0.63 to 0.67 (Bonferroni-adjusted P < .0016). Time-dependent AUCs for combined BCSC and AI models were slightly higher than AI alone (AI with BCSC time-dependent AUC range, 0.66-0.68; Bonferroni-adjusted P < .0016). Conclusion When using a negative screening examination, AI algorithms performed better than the BCSC risk model for predicting breast cancer risk at 0 to 5 years. Combined AI and BCSC models further improved prediction. © RSNA, 2023 Supplemental material is available for this article.

Risk of second primary cancer among women in the Kaiser Permanente Breast Cancer Survivors Cohort.

BACKGROUND: Breast cancer survivors are living longer due to early detection and advances in treatment and are at increased risk for second primary cancers. Comprehensive evaluation of second cancer risk among patients treated in recent decades is lacking. METHODS: We identified 16,004 females diagnosed with a first primary stage I-III breast cancer between 1990 and 2016 (followed through 2017) and survived ≥ 1 year at Kaiser Permanente (KP) Colorado, Northwest, and Washington. Second cancer was defined as an invasive primary cancer diagnosed ≥ 12 months after the first primary breast cancer. Second cancer risk was evaluated for all cancers (excluding ipsilateral breast cancer) using standardized incidence ratios (SIRs), and a competing risk approach for cumulative incidence and hazard ratios (HRs) adjusted for KP center, treatment, age, and year of first cancer diagnosis. RESULTS: Over a median follow-up of 6.2 years, 1,562 women developed second cancer. Breast cancer survivors had a 70% higher risk of any cancer (95%CI = 1.62-1.79) and 45% higher risk of non-breast cancer (95%CI = 1.37-1.54) compared with the general population. SIRs were highest for malignancies of the peritoneum (SIR = 3.44, 95%CI = 1.65-6.33), soft tissue (SIR = 3.32, 95%CI = 2.51-4.30), contralateral breast (SIR = 3.10, 95%CI = 2.82-3.40), and acute myeloid leukemia (SIR = 2.11, 95%CI = 1.18-3.48)/myelodysplastic syndrome (SIR = 3.25, 95%CI = 1.89-5.20). Women also had elevated risks for oral, colon, pancreas, lung, and uterine corpus cancer, melanoma, and non-Hodgkin lymphoma (SIR range = 1.31-1.97). Radiotherapy was associated with increased risk for all second cancers (HR = 1.13, 95%CI = 1.01-1.25) and soft tissue sarcoma (HR = 2.36, 95%CI = 1.17-4.78), chemotherapy with decreased risk for all second cancers (HR = 0.87, 95%CI = 0.78-0.98) and increased myelodysplastic syndrome risk (HR = 3.01, 95%CI = 1.01-8.94), and endocrine therapy with lower contralateral breast cancer risk (HR = 0.48, 95%CI = 0.38-0.60). Approximately 1 in 9 women who survived ≥ 1 year developed second cancer, 1 in 13 developed second non-breast cancer, and 1 in 30 developed contralateral breast cancer by 10 years. Trends in cumulative incidence declined for contralateral breast cancer but not for second non-breast cancers. CONCLUSIONS: Elevated risks of second cancer among breast cancer survivors treated in recent decades suggests that heightened surveillance is warranted and continued efforts to reduce second cancers are needed.

Economic Evaluation of Mailed Home-Based Human Papillomavirus Self-sampling Kits for Cervical Cancer Screening.

Importance: Human papillomavirus (HPV) self-sampling addresses barriers to cervical cancer screening, and mailed self-sampling kits have been reported to increase screening uptake. International research suggests mailed kits are cost-effective in certain settings. However, the cost-effectiveness of mailing HPV self-sampling kits for increasing screening uptake has not been evaluated in the US. Objective: To conduct an economic evaluation of a mailed HPV self-sampling intervention among underscreened women enrolled in an integrated US health care system. Design, Setting, and Participants: This economic evaluation involved a cost-effectiveness analysis of results from a randomized clinical trial of 19 851 women aged 30 to 64 years enrolled in a health plan from Kaiser Permanente Washington (KPWA), a US-based integrated health care system. Women were identified through electronic medical records, and eligible participants were enrolled in a health plan for at least 3 years and 5 months, had a primary care clinician, had not received a Papanicolaou test for at least 3 years and 5 months, and had not received a hysterectomy. Enrollment occurred from February 25, 2014, to August 29, 2016, with follow-up through February 25, 2018. The current economic evaluation was conducted between August 2, 2021, and July 30, 2022. Intervention delivery costs were calculated from both the KPWA and Medicare perspectives and were based on either wellness visit or Papanicolaou test-only visit costs. Intervention: Participants in the control group received usual care, which comprised patient reminders and ad hoc outreach for screening. Participants in the intervention group received usual care plus a mailed HPV self-sampling kit. Main Outcome and Measures: The primary economic outcome was the incremental cost-effectiveness ratio (ICER) for increased screening uptake, defined as the incremental difference in cost (intervention group minus control group) divided by the difference in the number of participants completing screening (intervention group minus control group) within 6 months of randomization. Results: Among 19 851 women (mean [SD] age, 50.1 [9.5] years; 76.7% White), 9960 were randomized to the intervention group, and 9891 were randomized to the control group. Baseline ICERs ranged from $85.84 (95% CI, $85.68-$85.99) using KPWA wellness visits as the cost basis to $146.29 (95% CI, $146.20-$146.38) using Medicare Papanicolaou test-only visits as the cost source. Subgroups of participants aged 50 to 64 years and participants most recently overdue for screening achieved cost-effectiveness at lower levels of willingness to pay for an additional completed screening than other subgroups. Conclusions and Relevance: In this economic evaluation, mailing HPV self-sampling kits to women overdue for cervical cancer screening was cost-effective for increased screening uptake relative to usual care. These results support mailing HPV kits as an efficient outreach strategy for increasing screening rates among eligible women in US health care systems.

Variation in second breast cancer risk after primary invasive cancer by time since primary cancer diagnosis and estrogen receptor status.

BACKGROUND: In women with previously treated breast cancer, occurrence and timing of second breast cancers have implications for surveillance. The authors examined the timing of second breast cancers by primary cancer estrogen receptor (ER) status in the Breast Cancer Surveillance Consortium. METHODS: Women who were diagnosed with American Joint Commission on Cancer stage I-III breast cancer were identified within six Breast Cancer Surveillance Consortium registries from 2000 to 2017. Characteristics collected at primary breast cancer diagnosis included demographics, ER status, and treatment. Second breast cancer events included subsequent ipsilateral or contralateral breast cancers diagnosed >6 months after primary diagnosis. The authors examined cumulative incidence and second breast cancer rates by primary cancer ER status during 1-5 versus 6-10 years after diagnosis. RESULTS: At 10 years, the cumulative second breast cancer incidence was 11.8% (95% confidence interval [CI], 10.7%-13.1%) for women with ER-negative disease and 7.5% (95% CI, 7.0%-8.0%) for those with ER-positive disease. Women with ER-negative cancer had higher second breast cancer rates than those with ER-positive cancer during the first 5 years of follow-up (16.0 per 1000 person-years [PY]; 95% CI, 14.2-17.9 per 1000 PY; vs. 7.8 per 1000 PY; 95% CI, 7.3-8.4 per 1000 PY, respectively). After 5 years, second breast cancer rates were similar for women with ER-negative versus ER-positive breast cancer (12.1 per 1000 PY; 95% CI, 9.9-14.7; vs. 9.3 per 1000 PY; 95% CI, 8.4-10.3 per 1000 PY, respectively). CONCLUSIONS: ER-negative primary breast cancers are associated with a higher risk of second breast cancers than ER-positive cancers during the first 5 years after diagnosis. Further study is needed to examine the potential benefit of more intensive surveillance targeting these women in the early postdiagnosis period.

Performance of Statistical and Machine Learning Risk Prediction Models for Surveillance Benefits and Failures in Breast Cancer Survivors.

BACKGROUND: Machine learning (ML) approaches facilitate risk prediction model development using high-dimensional predictors and higher-order interactions at the cost of model interpretability and transparency. We compared the relative predictive performance of statistical and ML models to guide modeling strategy selection for surveillance mammography outcomes in women with a personal history of breast cancer (PHBC). METHODS: We cross-validated seven risk prediction models for two surveillance outcomes, failure (breast cancer within 12 months of a negative surveillance mammogram) and benefit (surveillance-detected breast cancer). We included 9,447 mammograms (495 failures, 1,414 benefits, and 7,538 nonevents) from years 1996 to 2017 using a 1:4 matched case-control samples of women with PHBC in the Breast Cancer Surveillance Consortium. We assessed model performance of conventional regression, regularized regressions (LASSO and elastic-net), and ML methods (random forests and gradient boosting machines) by evaluating their calibration and, among well-calibrated models, comparing the area under the receiver operating characteristic curve (AUC) and 95% confidence intervals (CI). RESULTS: LASSO and elastic-net consistently provided well-calibrated predicted risks for surveillance failure and benefit. The AUCs of LASSO and elastic-net were both 0.63 (95% CI, 0.60-0.66) for surveillance failure and 0.66 (95% CI, 0.64-0.68) for surveillance benefit, the highest among well-calibrated models. CONCLUSIONS: For predicting breast cancer surveillance mammography outcomes, regularized regression outperformed other modeling approaches and balanced the trade-off between model flexibility and interpretability. IMPACT: Regularized regression may be preferred for developing risk prediction models in other contexts with rare outcomes, similar training sample sizes, and low-dimensional features.

Patient and Nodule Characteristics Associated With a Lung Cancer Diagnosis Among Individuals With Incidentally Detected Lung Nodules.

BACKGROUND: Pulmonary nodules are a common incidental finding on CT imaging. Few studies have described patient and nodule characteristics associated with a lung cancer diagnosis using a population-based cohort. RESEARCH QUESTION: Does a relationship exist between patient and nodule characteristics and lung cancer among individuals with incidentally detected pulmonary nodules, and can this information be used to create exploratory lung cancer prediction models with reasonable performance characteristics? STUDY DESIGN AND METHODS: We conducted a retrospective cohort study of adults older than 18 years with lung nodules of any size incidentally detected by chest CT imaging between 2005 and 2015. All patients had at least 2 years of complete follow-up. To evaluate the relationship between patient and nodule characteristics and lung cancer, we used binomial regression. We used logistic regression to create prediction models, and we internally validated model performance using bootstrap optimism correction. RESULTS: Among 7,240 patients with a median age of 67 years, 56% of whom were women, with a median BMI of 28 kg/m2, 56% of whom were ever smokers, 31% of whom had prior nonlung malignancy, with a median nodule size 5.6 mm, 57% of whom had multiple nodules, and 40% of whom had an upper lobe nodule, 265 patients (3.7%; 95% CI, 3.2%-4.1%) had a diagnosis of lung cancer. In a multivariate analysis, age, sex, BMI, smoking history, and nodule size and location were associated with a lung cancer diagnosis, whereas prior malignancy and nodule number and laterality were not. We were able to construct two prediction models with an area under the curve value of 0.75 (95% CI, 0.72-0.80) and reasonable calibration. INTERPRETATION: Lung cancer is uncommon among individuals with incidentally detected lung nodules. Some, but not all, previously identified factors associated with lung cancer also were associated with this outcome in this sample. These findings may have implications for clinical practice, future practice guidelines, and the development of novel lung cancer prediction models for individuals with incidentally detected lung nodules.

Effect of Patient Characteristics on Uptake of Screening Using a Mailed Human Papillomavirus Self-sampling Kit: A Secondary Analysis of a Randomized Clinical Trial.

Importance: Mailing human papillomavirus (HPV) self-sampling kits increases cervical cancer screening participation, but effects may differ across subpopulations. Subpopulation data can inform US health care system implementation. Objective: To identify patient characteristics that modify effectiveness of a mailed kit intervention at increasing screening. Design, Setting, and Participants: This was a secondary analysis of data from the Home-Based Options to Make Cervical Cancer Screening Easy (HOME) randomized clinical trial conducted from 2014 to 2018 at Kaiser Permanente Washington. Data analysis was performed from March 2018 to May 2022. Individuals aged 30 to 64 years with female sex, health plan enrollment longer than 3 years and 5 months, a current primary care clinician, and no Papanicolaou test within the prior 3 years and 5 months were identified through electronic medical records and randomized (1:1) to the control or intervention group. Interventions: The control group received usual care Papanicolaou screening reminders and outreach. The intervention group received usual care plus an unsolicited mailed HPV self-sampling kit. Main Outcomes and Measures: Screening uptake was captured within 6 months after randomization. Baseline patient characteristics (age, race, ethnicity, travel time to clinic, income, body mass index, tobacco use, health plan enrollment duration, time since last Papanicolaou test, mammography, comorbidities, and colorectal cancer screening adherence) were extracted from the electronic medical record. Results: Of 19 734 individuals (mean [SD] age, 50.1 [9.5] years; 14 129 [71.6%] White), 9843 were randomized to the intervention group, and 9891 were randomized to the control group. Screening uptake was 26.3% (2592 of 9843 individuals) in the intervention group vs 17.4% (1719 of 9891 individuals) in the control group (relative risk [RR], 1.51; 95% CI, 1.43-1.60). Although absolute differences in uptake by group varied little by screening history, relative effects were greater with longer vs shorter time since last Papanicolaou test (no prior Papanicolaou test: RRs, 1.85-3.25; ≥10 years: RR, 2.78; 5-10 years: RRs, 1.69-1.86; <5 years: RRs 1.29-1.37). Relative effects were greater in participants overdue (RR, 2.03; 95% CI, 1.73-2.38) vs up-to-date with mammography (RR, 1.53; 95% CI, 1.41-1.67), although absolute difference was greater in the up-to-date group. Differences by age were not significant, with RRs of 1.33 to 1.48 across 5-year age groups in participants 30 to 54, vs 1.60 (95% CI, 1.40-1.82) in participants 55 to 59 and 1.77 (95% CI, 1.56-2.01) in participants 60 to 64 years. Among those mailed kits, there were differences in kit use vs in-clinic screening by age, race, plan enrollment duration, underscreening duration, and colorectal cancer screening adherence. Conclusions and Relevance: In this secondary analysis of a randomized clinical trial, clinically important improvements in screening uptake were observed for all subgroups. Differences in magnitude of intervention effect and kit use highlighted opportunities to optimize HPV self-sampling for priority groups. Trial Registration: ClinicalTrials.gov Identifier: NCT02005510.

External Validation of an Ensemble Model for Automated Mammography Interpretation by Artificial Intelligence.

Importance: With a shortfall in fellowship-trained breast radiologists, mammography screening programs are looking toward artificial intelligence (AI) to increase efficiency and diagnostic accuracy. External validation studies provide an initial assessment of how promising AI algorithms perform in different practice settings. Objective: To externally validate an ensemble deep-learning model using data from a high-volume, distributed screening program of an academic health system with a diverse patient population. Design, Setting, and Participants: In this diagnostic study, an ensemble learning method, which reweights outputs of the 11 highest-performing individual AI models from the Digital Mammography Dialogue on Reverse Engineering Assessment and Methods (DREAM) Mammography Challenge, was used to predict the cancer status of an individual using a standard set of screening mammography images. This study was conducted using retrospective patient data collected between 2010 and 2020 from women aged 40 years and older who underwent a routine breast screening examination and participated in the Athena Breast Health Network at the University of California, Los Angeles (UCLA). Main Outcomes and Measures: Performance of the challenge ensemble method (CEM) and the CEM combined with radiologist assessment (CEM+R) were compared with diagnosed ductal carcinoma in situ and invasive cancers within a year of the screening examination using performance metrics, such as sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC). Results: Evaluated on 37 317 examinations from 26 817 women (mean [SD] age, 58.4 [11.5] years), individual model AUROC estimates ranged from 0.77 (95% CI, 0.75-0.79) to 0.83 (95% CI, 0.81-0.85). The CEM model achieved an AUROC of 0.85 (95% CI, 0.84-0.87) in the UCLA cohort, lower than the performance achieved in the Kaiser Permanente Washington (AUROC, 0.90) and Karolinska Institute (AUROC, 0.92) cohorts. The CEM+R model achieved a sensitivity (0.813 [95% CI, 0.781-0.843] vs 0.826 [95% CI, 0.795-0.856]; P = .20) and specificity (0.925 [95% CI, 0.916-0.934] vs 0.930 [95% CI, 0.929-0.932]; P = .18) similar to the radiologist performance. The CEM+R model had significantly lower sensitivity (0.596 [95% CI, 0.466-0.717] vs 0.850 [95% CI, 0.766-0.923]; P < .001) and specificity (0.803 [95% CI, 0.734-0.861] vs 0.945 [95% CI, 0.936-0.954]; P < .001) than the radiologist in women with a prior history of breast cancer and Hispanic women (0.894 [95% CI, 0.873-0.910] vs 0.926 [95% CI, 0.919-0.933]; P = .004). Conclusions and Relevance: This study found that the high performance of an ensemble deep-learning model for automated screening mammography interpretation did not generalize to a more diverse screening cohort, suggesting that the model experienced underspecification. This study suggests the need for model transparency and fine-tuning of AI models for specific target populations prior to their clinical adoption.

Women's Breast Cancer Screening Confidence by Screening Modality and Breast Density: A Breast Cancer Surveillance Consortium Survey Study.

Objective: Little is known about women's confidence in their breast cancer screening. We sought to characterize breast cancer screening confidence by imaging modality and clinically assessed breast density. Materials and Methods: We undertook a cross-sectional survey of women ages 40-74 years who received digital mammography (DM), digital breast tomosynthesis (DBT), and/or breast magnetic resonance imaging (MRI) with a normal screening exam in the prior year. The main outcome was women's confidence (Very, Somewhat, A little, Not at all) in their breast cancer screening detecting any cancer. Multivariable logistic regression identified correlates of being very confident in breast cancer screening by screening modality group: Group 1) DM vs. DBT and Group 2) DM or DBT alone vs. with supplemental MRI. Results: Overall, 2329 of 7439 (31.3%) invitees participated, with 30%-61% being very confident in their screening across modality and density subgroups. Having dense versus nondense breasts was associated with lower odds of being very confident (Group 1: odds ratio [OR]: 0.58; 95% confidence interval [CI]: 0.46-0.79; Group 2: OR: 0.56; 95% CI: 0.40-0.79). There were no differences by modality within Group 1, but for Group 2, women undergoing MRI had higher odds of being very confident (OR: 1.69; 95% CI: 1.21-2.37). Other correlates of greater screening confidence were as follows: Group 1-being offered a screening test choice and cost not influencing modality received, and Group 2-decision satisfaction and worry. Conclusions: Women with dense breasts had lower screening confidence regardless of screening modality and those undergoing MRI had higher confidence regardless of density. The importance of informing women about screening options is underscored by observed associations between screening choice, decision satisfaction, and screening confidence. ClinicalTrials.gov: NCT02980848.

Treatment-related thoracic soft tissue sarcomas in US breast cancer survivors: a retrospective cohort study.

BACKGROUND: Soft tissue sarcoma is a rare but serious side-effect of radiotherapy to treat breast cancer, and rates are increasing in the USA. We evaluated potential co-factors in two complimentary cohorts of US breast cancer survivors. METHODS: In this retrospective cohort study, we sourced data from the Kaiser Permanente (KP) cohort and the Surveillance, Epidemiology, and End Results (SEER) 13 registries cohort, both in the USA. The KP cohort included 15 940 women diagnosed with breast cancer from Jan 1, 1990, to Dec 31, 2016, in KP Colorado, KP Northwest (which serves Oregon and Southwest Washington state), or KP Washington, with detailed treatment data and comorbidities (including hypertension and diabetes at or before breast cancer diagnosis) from electronic medical records. The SEER cohort included 457 300 women diagnosed with breast cancer from Jan 1, 1992, to Dec 31, 2016, within the 13 SEER registries across the USA, with initial treatment data (yes vs no or unknown). Eligibility criteria in both cohorts were female breast cancer survivors (stage I-III) aged 20-84 years at diagnosis who had breast cancer surgery, and had survived at least 1 year after breast cancer diagnosis. The outcome of interest was any second thoracic soft tissue sarcoma (angiosarcomas and other subtypes) that developed at least 1 year after breast cancer diagnosis. Risk factors for thoracic soft tissue sarcoma were assessed using multivariable Poisson regression models. FINDINGS: In the KP cohort, median follow-up was 9·3 years (IQR 5·7-13·9) and 19 (0·1%) of 15 940 eligible, evaluable women developed a thoracic soft tissue sarcoma (11 angiosarcomas, eight other subtypes). Most (94·7%; 18 of 19) thoracic soft tissue sarcomas occurred in women treated with radiotherapy; thus, radiotherapy was associated with a significantly increased risk of developing a thoracic soft tissue sarcoma (relative risk [RR] 8·1 [95% CI 1·1-60·4]; p=0·0052), but there was no association with prescribed dose, fractionation, or boost. The RR of angiosarcoma after anthracyclines was 3·6 (95% CI 1·0-13·3; p=0·058). Alkylating agents were associated with an increased risk of developing other sarcomas (RR 7·7 [95% CI 1·2-150·8]; p=0·026). History of hypertension (RR 4·8 [95% CI 1·3-17·6]; p=0·017) and diabetes (5·3 [1·4-20·8]; p=0·036) were each associated with around a five-times increased risk of angiosarcoma. In the SEER cohort, 430 (0·1%) of 457 300 patients had subsequent thoracic soft tissue sarcomas (268 angiosarcomas and 162 other subtypes) after a median follow-up of 8·3 years (IQR 4·3-13·9). Most (77·9%; 335 of 430) cases occurred after radiotherapy; thus, radiotherapy was associated with a significantly increased risk of developing a thoracic soft tissue sarcoma (RR 3·0 [95% CI 2·4-3·8]; p<0·0001) and, for angiosarcomas, the RR for breast-conserving surgery plus radiotherapy versus mastectomy plus radiotherapy was 1·9 (1·1-3·3; p=0·012). By 10 years after radiotherapy, the cumulative incidence of thoracic soft tissue sarcoma was 0·21% (95% CI 0·12-0·34) in the KP cohort and 0·15% (95% CI 0·13-0·17) in SEER. INTERPRETATION: Radiotherapy was the strongest risk factor for thoracic soft tissue sarcoma in both cohorts. This finding, along with the novel findings for diabetes and hypertension as potential risk factors for angiosarcomas, warrant further investigation as potential targets for prevention strategies and increased surveillance. FUNDING: US National Cancer Institute and National Institutes of Health.

Design of a pragmatic randomized controlled trial of home-based human papillomavirus (HPV) self-sampling for increasing cervical cancer screening uptake in a U.S. healthcare system: The STEP trial.

BACKGROUND: Mailing HPV self-sampling kits to overdue individuals increases cervical cancer screening adherence; offering self-sampling to previously adherent individuals has not been evaluated in the U.S. Given heterogeneity of the U.S. health system and population, data are needed to optimize how HPV self-sampling is offered to individuals who are overdue, due after successful past screening, or have an unknown screening history. METHODS: STEP is a pragmatic randomized controlled trial set within a U.S. integrated healthcare delivery system, designed to compare different outreach approaches for offering HPV self-sampling in populations defined by prior screening behavior (previously-adherent, overdue, or unknown screening history). Over 14 months, eligible individuals were identified through electronic medical record (EMR) data and randomized to Usual Care (UC), Education (UC + educational materials about cervical cancer screening), Direct-Mail (UC + Education + a mailed self-sampling kit) or Opt-In (UC + Education + option to request a kit), depending on screening history. The primary objective is to compare screening completion by outreach approach and screening history. Secondary objectives include evaluating incremental cost-effectiveness of outreach approaches, and identifying patient preference for, and satisfaction with, HPV self-screening, and barriers to abnormal results follow-up (measured through interviews and focus groups). CONCLUSIONS: The trial was designed to generate data that U.S. health systems can use to inform primary HPV screening implementation strategies that incorporate HPV self-sampling options to improve screening access, adherence, and patient satisfaction. The objective of this report is to describe the rationale and design of this pragmatic trial.

Fleischner Society Guideline Recommendations for Incidentally Detected Pulmonary Nodules and the Probability of Lung Cancer.

PURPOSE: The Fleischner Society aims to limit further evaluations of incidentally detected pulmonary nodules when the probability of lung cancer is <1% and to pursue further evaluations when the probability of lung cancer is ≥1%. To evaluate the internal consistency of guideline goals and recommendations, the authors evaluated stratum-specific recommendations and 2-year probabilities of lung cancer. METHODS: A retrospective cohort study (2005-2015) was conducted of individuals enrolled in one of two integrated health systems with solid nodules incidentally detected on CT. The 2017 Fleischner Society guidelines were used to define strata on the basis of smoking status and nodule size and number. Lung cancer diagnoses within 2 years of nodule detection were ascertained using cancer registry data. Confidence interval (CI) inspection was used to determine if stratum-specific probabilities of lung cancer were different than 1%. RESULTS: Among 5,444 individuals with incidentally detected lung nodules (median age, 66 years; 54% women; 57% smoked; median nodule size, 5.5 mm; 55% with multiple nodules), 214 (3.9%; 95% CI, 3.4%-4.5%) were diagnosed with lung cancer within 2 years. For 7 of 12 strata (58%), 2,765 patients (51%), and 194 lung cancer cases (91%), there was alignment between Fleischner Society goals and recommendations. Alignment was indeterminate for 5 strata (42%), 2,679 patients (49%), and 20 lung cancer cases (9%) because CIs for the probability of lung cancer spanned 1%. CONCLUSIONS: Fleischner Society guideline goals and recommendations align at least half the time. It is uncertain whether alignment of guideline goals and recommendations occurs more often.

Accuracy of tumor registry versus pharmacy dispensings for breast cancer adjuvant endocrine therapy.

PURPOSE: Accounting for endocrine therapy use for breast cancer treatment is important for studies of survivorship. We evaluated the accuracy of Surveillance, Epidemiology, and End Results (SEER) breast cancer endocrine therapy data compared with pharmacy dispensings from an integrated health system. METHODS: We included women with non-metastatic hormone receptor positive primary breast cancer diagnosed between 1995 and 2017 enrolled in Kaiser Permanente Washington, linking their data with SEER. We used pharmacy dispensings for endocrine therapy within one year following diagnosis as our reference standard. We calculated kappa (concordance), positive predictive value (PPV), and negative predictive values (NPV) overall and stratified by woman and tumor characteristics of interest. RESULTS: Of 5,055 women, mean age at diagnosis was 62 years (interquartile range = 53-71); 53% had localized stage, 56% received lumpectomy with radiation, and 31% received chemotherapy. SEER data alone identified 67% of women as having received endocrine therapy; this increased to 75% with pharmacy dispensings. SEER's concordance with pharmacy dispensings was 0.68 (PPV = 91%; NPV = 76%). PPV did not vary by tumor or women characteristics; however, NPV declined with younger age at diagnosis (64% in < 45 years vs. 86% in 75+ years), increasing tumor stage (49% in regional stage vs. 91% in DCIS), and chemotherapy treatment (41% in those with chemotherapy vs. 83% in those without chemotherapy). CONCLUSION: Pharmacy dispensings enable more complete endocrine therapy capture, particularly in women with more advanced tumors or who receive chemotherapy. We determined woman, tumor, and treatment characteristics that contribute to underascertainment of endocrine therapy use in tumor registries.

Decision quality and regret with treatment decisions in women with breast cancer: Pre-operative breast MRI and breast density.

PURPOSE: We evaluated self-report of decision quality and regret with breast cancer surgical treatment by pre-operative breast MRI use in women recently diagnosed with breast cancer. METHODS: We conducted a survey with 957 women aged 18 + with stage 0-III breast cancer identified in the Breast Cancer Surveillance Consortium. Participants self-reported receipt of pre-operative breast MRI. Primary outcomes were process measures in the Breast Cancer Surgery Decision Quality Instrument (BCS-DQI) (continuous outcome) and Decision Regret Scale (dichotomized outcome as any/none). Generalized estimating equations with linear and logit link were used to estimate adjusted associations between breast MRI and primary outcomes. All analyses were also stratified by breast density. RESULTS: Survey participation rate was 27.9% (957/3430). Study population was primarily > 60 years, White, college educated, and diagnosed with early-stage breast cancer. Pre-operative breast MRI was reported in 46% of women. A higher proportion of women who were younger age (< 50 years), commercially insured, and self-detected their breast cancer reported pre-operative breast MRI use. In adjusted analysis, pre-operative breast MRI use compared with no use was associated with a small but statistically significantly higher decision quality scores (69.5 vs 64.7, p-value = 0.043). Decision regret did not significantly differ in women who reported pre-operative breast MRI use compared with no use (54.2% v. 48.7%, respectively, p-value = 0.11). Study results did not vary when stratified by breast density for either primary outcome. CONCLUSIONS AND RELEVANCE: Breast MRI use in the diagnostic work-up of breast cancer does not negatively alter women's perceptions of surgical treatment decisions in early survivorship. CLINICAL TRIALS REGISTRATION NUMBER: NCT03029286.