Background:Patients with breast cancer often experience health-related quality of life (HRQoL) impairments that remain difficult to predict on an individual level. Prediction models can aid in understanding individual survivorship trajectories. However, current prognostic models are based on fixed intervals, limiting their utility in clinical follow-up schedules. Objective:This study aimed to develop and externally validate time-dynamic machine learning (ML) models that predict clinically relevant HRQoL impairments in nonmetastatic patients with breast cancer. Methods:Using the pooled multicohort EORTC (European Organisation for Research and Treatment of Cancer) BALANCE (big data in patients with breast cancer) dataset (n=6316) containing repeated HRQoL measurements (EORTC QLQ [Quality of Life Core Questionnaire]-C30), we constructed over 70,000 patient assessment pairs. ML algorithms were trained using the earlier HRQoL assessment and clinical data to predict dichotomized impairments in QLQ-C30 domains at the later assessment between 2 weeks and 5 years ahead, reflecting the range of follow-up intervals available in the dataset. The best performing model was determined via the area under the receiver operating characteristic curve in the internal validation, and externally validated in an independent cohort of the BALANCE dataset, in which the calibration and predictive performance in risk groups (patients: postmenopause, with financial difficulties, with obesity, with 2 or more comorbidities, with lower educational status, and with frailty) were also evaluated. Results:ML models showed good discrimination (area under the receiver operating characteristic curve 0.64-0.84) across most domains, especially for persistent symptoms such as fatigue, financial difficulties, or functioning scales. Gradient boosting models performed best, but tended to be overconfident, with poor calibration for low-prevalence symptoms such as diarrhea or constipation. Model performance varied by risk group (eg, lower education and frailty), though no group consistently performed poorly. Performance remained stable across time windows, with prior HRQoL being the strongest predictor at the respective scale level, while clinical variables such as the type of treatment were less important for prediction. Conclusions:Time-dynamic ML models can support personalized HRQoL prediction in breast cancer care. Future improvements should focus on calibration and fairness to enable equitable, clinically meaningful implementation.

OBJECTIVES:To keep pace with rapid medical innovation, health care must be organized to enable systematic learning from every patient. The 'outpatient innovation clinic' was established to support this goal. This paper presents 12 years of experience with the outpatient innovation clinic, highlighting its approach, achievements, and challenges. METHODS:Established in 2013 at the Division of Imaging and Oncology at University Medical Center Utrecht, the Netherlands, the outpatient innovation clinic provides an infrastructure to systematically enroll patients into prospective cohort studies. Patients are invited for a consultation with a clinical researcher, and informed consent is asked for the reuse of medical data, the collection of patient-reported outcome measures (PROMs), randomization into trials, biobanking, and recontact. The infrastructure serves as a platform for evidence-based innovation, following the idea, development, exploration, assessment, and long-term evaluation (IDEAL) framework. RESULTS:Between 2013 and 2025, eight prospective (international) multicenter observational cohorts of patients with primary tumors, metastatic cancer, or neurovascular malformations were established. Out of a total of 45,099 participants, 15,691 were enrolled at the innovation clinic. Of these, 13,172 (84%) provided additional informed consent for the collection of PROMs, and 11,833 (79%) for randomization into future trials within cohorts, either to serve as controls or to receive an invitation for participation. A large real-world data repository was generated and is used for technical product development and testing, hypothesis-generating studies, training and evaluation of (deep learning) algorithms, long-term PROMs analysis, and randomized treatment comparisons. Since 2014, five cohort-embedded trials have been completed, four are ongoing, and one is in preparation. CONCLUSION:The outpatient innovation clinic offers a successful infrastructure for (deep) learning from every patient and for evaluating medical innovations within the IDEAL framework, while safeguarding patients' autonomy over the reuse of their medical data.

Purpose:Positive margins after breast-conserving surgery (BCS) are an important risk factor for local tumor recurrence and the need for re-excision in women with breast cancer. It remains unclear which factors collectively influence positive margins after BCS and whether the outcomes vary among hospitals. This study investigated the occurrence and risk factors of positive margins after BCS in women with breast cancer in two Dutch hospitals. Methods:Data were collected retrospectively from medical records of women who were diagnosed with newly invasive breast carcinoma or carcinoma in situ and underwent primary BCS between January 1, 2018 and December 31, 2020 in two Dutch hospitals. Results:A total of 423 cases (410 patients) were included, with a median age of 58 years (interquartile range 51-66). On average, the positive margin rate after BCS was 8.0%, which was significantly higher in the low-volume hospital (14.9%) than in the high-volume hospital (5.9%). Invasive lobular carcinoma (odds ratio [OR] = 4.97, confidence interval [CI] = 1.91-12.94), postoperative tumor size (OR = 1.07, CI = 1.03-1.11), low hospital volume (OR = 3.90, CI = 1.58-9.66), and lumpectomy size (OR = 0.97, CI = 0.94-1.00) were significantly associated with positive margins after BCS. Conclusion:Positive margin rate after BCS was low, with 8.0% among all cases and varied significantly between the hospitals. Patient-, tumor-, radiological-, and surgery-related factors may contribute to these variations. An optimized collaborative multidisciplinary approach to breast cancer care, particularly between radiologists and surgeons, should be strived to achieve improved oncological outcomes in women after BCS.

Background: Long-term patient-reported outcomes (PROs) from unselected breast cancer populations are scarce, but highly important to identify knowledge gaps and unmet needs among survivors. We investigated PROs of women during five years following treatment for invasive breast cancer or ductal carcinoma in situ (DCIS).

Objectives The objectives of the study are to investigate infection risk in offspring born to women with systemic lupus erythematosus (SLE) compared with offspring born to women without SLE and examine the mediating role of preterm birth.Design This is a register-based cohort study.Setting Liveborn singletons born in Sweden, 2006–2021, were included in the study.Participants 1248 infants born to mothers with SLE (≥2 International Classification of Diseases-coded visits in the National Patient Register (NPR) and Medical Birth Register, with ≥1 visit before pregnancy) and 34 886 infants born to women without SLE from the general population were included.Primary and secondary outcome measures The primary outcome was any visit for infection in the NPR or anti-infectives in the Prescribed Drug Register. The secondary outcome was hospitalised infection. Infection risks within 72 hours, within 1 month and within 1 year were estimated.Results SLE offspring had a higher risk of infection in the first 72 hours compared with non-SLE (2.1% vs 1.2%; risk ratios (RR) (95% CI) 1.62 (1.09 to 2.42)), the first month (5.2% vs 4.5%; RR 1.12 (0.88 to 1.43)) and first year of life (38.2% vs 37.2%; RR 1.09 (1.01 to 1.17)). The hospitalised infection risk for SLE offspring was similar to that of non-SLE (5.8% vs 5.5%, first year of life). The percentage of the total effect of maternal SLE on infant infection mediated through preterm birth was 86% for infection in the first 72 hours and 27% in the first year of life.Conclusions The risk of infection in SLE offspring is most increased in the first 3 days after birth, and a proportion of this association can be explained by preterm birth. To prevent early neonatal infections, maternal SLE could be considered as a risk factor before allowing early discharge from postnatal care.

Objective: To investigate sickness benefits following delivery in mothers with systemic lupus erythematosus (SLE) and mothers without SLE. Method: SLE and non-SLE mothers, matched by age and month of delivery, with a singleton liveborn (2004-2008), were identified from the Swedish Lupus Linkage cohort. Work loss (sum of sick leave and disability pension) was studied from 1 year prenatally to 3 years postpartum. Adjusted logistic regression models of covariates associated with > 30 days of work loss in the first and second years postpartum were estimated in SLE mothers. Results: Among 130 SLE mothers and 440 non-SLE mothers, SLE mothers were more likely to have work loss from the prenatal year (42% vs 16%) to 3 years postpartum (49% vs 15%). In SLE mothers, work loss was on average 61 +/- 112 days (mean +/- sd) in the prenatal year and 38 +/- 83 days in the first year postpartum, which increased to 71 +/- 114 days in the third year postpartum. Having > 30 days of sick leave in the year of delivery [odds ratio (OR) 4.4, 95% confidence interval (CI) 1.5-12.9] and <= 12 years of education (OR 2.6, 95% CI 1.1-6.0) were associated with work loss in the first year postpartum. No covariates were associated with work loss in the second year postpartum. Conclusion: SLE mothers more often had work loss in the prenatal year to 3 years postpartum compared to non-SLE mothers. Lower education and sick leave in the year of delivery were associated with a higher odds of work loss in the first year postpartum in SLE.

Objective: The estimand framework offers a structured approach to define the treatment effect to be estimated in a clinical study. Defining the estimand upfront helps formulating the research question and informs study design, data collection and statistical analysis methods. Since the Trials within Cohorts (TwiCs) design has unique characteristics, the objective of this study is to describe considerations and provide guidance for formulating estimands for TwiCs studies.Methods: The key attributes of an estimand are the target population, treatments that are compared, the endpoint, intercurrent events and their handling, and the population-level summary measure. The estimand framework was applied retrospectively to two TwiCs studies: the SPONGE and UMBRELLA Fit trial. The aim is to demonstrate how the estimand framework can be implemented in TwiCs studies, thereby focusing on considerations relevant for defining the estimand. Three estimands were defined for both studies. For the SPONGE trial, estimators were derived.Results: Intercurrent events considered to occur exclusively or more frequently in TwiCs studies compared to conventional randomized trials included intervention refusal after randomization, misalignment of timing of routine cohort measurements and the intervention period, and participants in the control arm initiating treatments similar to the studied intervention. Considerations for handling refusal after randomization related to decisions on whether the target population should include all eligible participants or the subpopulation that would accept (or undergo) the intervention when offered. Considerations for handling treatment initiation in the control arm and misalignments of timing related to decisions on whether such events should be considered part of treatment policy or whether interest is in a hypothetical scenario where such events do not occur.Conclusion: The TwiCs study design has unique features that pose specific considerations when formulating an estimand. The examples in this study can provide guidance in the definition of estimands in future TwiCs studies.

Abstract Background Planning for return to work (RTW) is relevant among sub‐groups of metastatic breast cancer (mBC) survivors. RTW and protective factors for RTW in patients with mBC were determined. Methods Patients with mBC, ages 18–63 years, were identified in Swedish registers, and data were collected starting 1 year before their mBC diagnosis. The prevalence of working net days (WNDs) (>90 and >180) during the year after mBC diagnosis (y1) was determined. Factors associated with RTW were assessed using regression analysis. The impact of contemporary oncological treatment of mBC on RTW and 5‐year mBC‐specific survival was compared between those diagnosed in 1997–2002 and 2003–2011. Results Of 490 patients, 239 (48.8%) and 189 (36.8%) had >90 and >180 WNDs, respectively, during y1. Adjusted odds ratios (AORs) of WNDs >90 or >180 during y1 were significantly higher for patients with age ≤50 years (AOR180 = 1.54), synchronous metastasis (AOR90 = 1.68, AOR180 = 1.67), metastasis within 24 months (AOR180 = 1.51), soft tissue, visceral, brain as first metastatic site (AOR90 = 1.47) and sickness absence <90 net days in the year before mBC diagnosis, suggesting limited comorbidities (AOR90 = 1.28, AOR180 = 2.00), respectively. Mean (standard deviation) WNDs were 134.9 (140.1) and 161.3 (152.4) for patients diagnosed with mBC in 1997–2002 and 2003–2011, respectively (p = 0.046). Median (standard error) mBC‐specific survivals were 41.0 (2.5) and 62.0 (9.6) months for patients diagnosed with mBC in 1997–2002 and 2003–2011, respectively (p < 0.001). Conclusions RTW of more than 180 WNDs was associated with younger age, early development of metastases and limited comorbidities during the year before the diagnosis of mBC. Patients diagnosed with mBC in 2003 or later had more WNDs and better survival than those diagnosed earlier.

OBJECTIVE:The objective of this study is to determine the prevalence and predictors of sickness absence (SA) and disability pension (DP) in women with metastatic breast cancer (mBC).METHODS:Data were obtained from Swedish registers concerning 1,240 adult women diagnosed 1997-2011 with mBC, from 1 year before (y-1) to 2 (y1) and 2 (y2) years after diagnosis. SA and DP prevalence was calculated. Odds ratios (AOR) were determined for factors associated with using long-term (SA > 180 days or DP > 0 days) sickness benefits.RESULTS:Prevalence of SA and DP was 56.0% and 24.8% during y-1, 69.9% and 28.9% during y1, and 64.0% and 34.7% during y2, respectively. Odds of using long-term sickness benefits were higher y1 and y2 in patients using long-term sickness benefits the year before diagnosis (AOR = 3.82, 95% CI 2.91-5.02; AOR = 4.31, 95% CI 2.96-6.29, respectively) and y2 in patients with mBC diagnosis 1997-2000 (AOR = 1.84, 95% CI 1.10-3.08) and using long-term sickness benefits the year after diagnosis (AOR = 22.10, 95% CI 14.33-34.22).CONCLUSIONS:The prevalence of sickness benefit utilisation was high and increased after mBC diagnosis, particularly for patients using long-term sickness benefits prior to diagnosis. Additional study is needed to determine factors that might reduce the need for sickness benefits and enhance work ability in these patients.

Cyclophosphamide (CPA) dosing by body surface area (BSA, m 2 ) has been questioned as a predictor for individual drug exposure. This study investigated phosphoramide mustard-hemoglobin (PAM-Hb, pmol g −1 Hb) as a biomarker of CPA exposure in 135 female breast cancer patients receiving CPA during three courses based on BSA: 500 mg/m 2 (C500 group, n = 67) or 600 mg/m 2 (C600 group, n = 68). The inter-individual difference was calculated for both groups by dividing the highest through the lowest PAM-Hb value of each course. The inter-occasion difference was calculated in percentage for each individual by dividing their PAM-Hb value through the group mean per course, and subsequently dividing this ratio of the latter through the previous course. A multivariable linear regression (MLR) was performed to identify factors that explained the variation of PAM-Hb. During the three courses, the inter-individual difference changed from 3.5 to 2.1 and the inter-occasion difference ranged between 13.3% and 11.9% in the C500 group. In the C600 group, the inter-individual difference changed from 2.7 to 2.9 and the inter-occasion difference ranged between 14.1% and 11.7%. The MLR including BSA, age, GFR, and albumin explained 17.1% of the variation of PAM-Hb and was significantly better then the model including only BSA. These factors should be considered when calculating the first dose of CPA for breast cancer patients.

Objective To investigate the risk of gestational diabetes mellitus (GDM) associated with systemic lupus erythematosus (SLE) by comparing pregnancies in women with SLE to general population controls. Methods We identified singleton pregnancies among women with SLE and general population controls in the Swedish Medical Birth Register (MBR; 2006–2016), sampled from the population-based Swedish Lupus Linkage (SLINK) cohort (1987–2012). SLE was defined by ≥ 2 International Classification of Diseases (ICD)-coded visits in the National Patient Register (NPR) and MBR, with ≥ 1 visit before pregnancy. GDM was defined by ≥ 1 ICD-coded visit in the NPR or MBR. Glucocorticoid (GC) and hydroxychloroquine (HCQ) dispensations within 6 months before and during pregnancy were identified in the Prescribed Drug Register. Risk ratios (RRs) and 95% CIs of GDM associated with SLE were estimated using modified Poisson regression models, stratified by parity and adjusted for maternal age at delivery, year of birth, and obesity. Results We identified 695 SLE pregnancies including 18 (2.6%) with GDM and 4644 non-SLE pregnancies including 65 (1.4%) with GDM. Adjusted RRs of GDM associated with SLE were 1.11 (95% CI 0.38–3.27) for first deliveries and 2.03 (95% CI 1.21–3.40) for all deliveries. Among SLE pregnancies, GDM occurred in 7/306 (2.3%) with ≥ 1 GC before and/or during pregnancy, 11/389 (2.8%) without GC, 7/287 (2.4%) with ≥ 1 HCQ before and/or during pregnancy, and in 11/408 (2.7%) without HCQ. Conclusion When looking at all deliveries, SLE was associated with a 2-fold higher risk of GDM. GDM occurrence did not differ by GC or HCQ.

Importance Cardiovascular disease (CVD) is common in patients treated for breast cancer, especially in patients treated with systemic treatment and radiotherapy and in those with preexisting CVD risk factors. Coronary artery calcium (CAC), a strong independent CVD risk factor, can be automatically quantified on radiotherapy planning computed tomography (CT) scans and may help identify patients at increased CVD risk. Objective To evaluate the association of CAC with CVD and coronary artery disease (CAD) in patients with breast cancer. Design, Setting, and Participants In this multicenter cohort study of 15 915 patients with breast cancer receiving radiotherapy between 2005 and 2016 who were followed until December 31, 2018, age, calendar year, and treatment-adjusted Cox proportional hazard models were used to evaluate the association of CAC with CVD and CAD. Exposures Overall CAC scores were automatically extracted from planning CT scans using a deep learning algorithm. Patients were classified into Agatston risk categories (0, 1-10, 11-100, 101-399, >400 units). Main Outcomes and Measures Occurrence of fatal and nonfatal CVD and CAD were obtained from national registries. Results Of the 15 915 participants included in this study, the mean (SD) age at CT scan was 59.0 (11.2; range, 22-95) years, and 15 879 (99.8%) were women. Seventy percent (n = 11 179) had no CAC. Coronary artery calcium scores of 1 to 10, 11 to 100, 101 to 400, and greater than 400 were present in 10.0% (n = 1584), 11.5% (n = 1825), 5.2% (n = 830), and 3.1% (n = 497) respectively. After a median follow-up of 51.2 months, CVD risks increased from 5.2% in patients with no CAC to 28.2% in patients with CAC scores higher than 400. After adjustment, CVD risk increased with higher CAC score (hazard ratio [HR]CAC = 1-10 = 1.1; 95% CI, 0.9-1.4; HRCAC = 11-100 = 1.8; 95% CI, 1.5-2.1; HRCAC = 101-400 = 2.1; 95% CI, 1.7-2.6; and HRCAC>400 = 3.4; 95% CI, 2.8-4.2). Coronary artery calcium was particularly strongly associated with CAD (HRCAC>400 = 7.8; 95% CI, 5.5-11.2). The association between CAC and CVD was strongest in patients treated with anthracyclines (HRCAC>400 = 5.8; 95% CI, 3.0-11.4) and patients who received a radiation boost (HRCAC>400 = 6.1; 95% CI, 3.8-9.7). Conclusions and Relevance This cohort study found that coronary artery calcium on breast cancer radiotherapy planning CT scan results was associated with CVD, especially CAD. Automated CAC scoring on radiotherapy planning CT scans may be used as a fast and low-cost tool to identify patients with breast cancer at increased risk of CVD, allowing implementing CVD risk-mitigating strategies with the aim to reduce the risk of CVD burden after breast cancer. Trial Registration ClinicalTrials.gov Identifier: NCT03206333.

Abstract Background Advances in the treatment of metastatic breast cancer (mBC) have led to improved life expectancy. Many cancer survivors desire to return to paid work to enhance their sense of well-being. For patients with mBC, little is known about how the diagnosis impacts ability to work or the factors that increase the need for sickness benefits. Patients and methods Data were collected from two Swedish national registers, for females ages 18 to 63 years in the Stockholm-Gotland healthcare region with a new diagnosis of mBC from 1997 through 2011. Type of first-line palliative treatment was identified in medical records of a subset of the study population. Use of sickness absence (SA) and disability pension (DP) by these patients during the year before and one and two years after mBC diagnosis was determined from a third register. Regression analysis was performed to ascertain which covariate factors were associated with long-term (> 30 days) SA. Results A total of 1,240 patients were evaluated the year before and the first year after mBC diagnosis; only 805 patients were still alive and evaluated the second year after diagnosis. The proportions of patients having SA and DP were 56.0% and 24.8% the year before, 69.9% and 28.9% the first year after, and 64.0% and 34.7% the second year after diagnosis, respectively. Adjusted odds of having long-term SA were significantly higher at 1 and 2 years after diagnosis for patients with age < 45 years (AOR = 3.43 and AOR = 1.70, respectively), early calendar year of diagnosis (AOR = 1.72 and AOR = 1.79, respectively), metachronous mBC (AOR = 4.85 and AOR = 4.52, respectively), and SA ≥ 90 days the year before diagnosis (AOR = 3.44 and AOR = 1.98, respectively). Odds were also significantly higher the second after diagnosis for patients treated with chemotherapy (AOR = 1.81) or radiotherapy (AOR = 2.23), compared to those treated with hormonal therapy, Conclusions Rates of SA and DP increase after a diagnosis of mBC. Women who are younger, develop metachronous mBC, use SA heavily before mBC, and receive chemotherapy or radiotherapy have a greater need for sickness benefits after an mBC diagnosis.

e14121 Background: Sickness absence (SA) and disability pension (DP) from diagnosis of primary breast cancer (BC) among patients who in a later stage will develop a recurrence or metastatic disease is unknown. This study explores the prevalence and risk factors of SA and DP in this study population before and after primary breast cancer diagnosis. Methods: Longitudinal register data on 1,310 female patients living in Sweden (age 20 to 63 years) diagnosed with primary BC between 1996 to 2011, were analyzed for annual prevalence of SA days and DP starting 2 years pre- to 5 years postdiagnosis. Logistic regressions were used to explore associations between primary BC characteristics and future SA. Results: 579 (44.2%) had loco-regional recurrence after a median of 2.5 years (interquartile range (IQR) = 1.3-4.3) and 731 (55.8%) had a metastatic disease after a median of 2.3 years (IQR = 1.3-4.1). 320 (24.4%) of 1,310 were still relapse-free 5 years postdiagnosis. SA was high during year 1 postdiagnosis but decreased steadily through year 5 (67.9%, 40.3%, 29.0%, 23.9%, 19.4%, respectively), while DP steadily increased during this time period (16.4%, 17.5%, 19.7%, 24.9%, 28.7%, respectively). The annual prevalence of SA over 180 days among patients who later were diagnosed with metastatic disease was constantly higher compared to patients who later were diagnosis of loco-regional recurrence. Pre-diagnosis SA, age < 50 years at the time of diagnosis, higher tumor stage, chemotherapy and future metastatic disease were associated with higher odds ratios for SA (odds ratio range, 1.56 to 4.42) Conclusions: In conclusion, longitudinal rates of SA and DP in a cohort of women with early BC, of whom a part developed disease relapse during follow-up were higher compared to previous studies. These are unique findings as previously published studies excluded patients with disease relapse. Patients at increased risk for SA, should be assessed and triaged to optimize chances for a smooth transition to return to work after oncological treatment.

Aims Heart failure (HF) patients diagnosed with breast cancer (BC) may have a higher risk of death, and different HF presentation and treatment than patients without BC. Methods and results A total of 14 998 women with incident HF (iHF) or prevalent HF (pHF) enrolled in the Swedish HF Registry within and after 1 month since HF diagnosis, respectively, between 2008 and 2013. Patients were linked with the National Patient-, Cancer-, and Cause-of-Death Registry. Two hundred and ninety-four iHF and 338 pHF patients with BC were age-matched to 1470 iHF and 1690 pHF patients without BC. Comorbidity and treatment characteristics were compared using the chi(2) tests for categories. Cox proportional hazard models assessed the hazard ratio (HR) and 95% confidence intervals (95% Cls) of all-cause and cardiovascular mortality among HF patients with and without BC. In the pHF group, BC patients had less often myocardial infarction (21.6% vs. 28.6%, P < 0.01) and received less often aspirin (47.6% vs. 55.1%, P= 0.01), coronary revascularization (11.8% vs. 16.2%, P< 0.01), or device therapy (0.9% vs. 3.0%, P= 0.03). After median follow-up of 2 years, risk of all-cause mortality (iHF: HR= 1.04, 95% CI = 0.83-1.29 and pHF: HR =0.94, 95% CI= 0.79-1.12), cardiovascular mortality (iHF: HR= 0.94, 95% CI=0.71-1.24 and pHF: HR= 0.89, 95% CI =0.71-1.10), and HF mortality (iHF: HR = 0.80, 95% CI = 0.34-1.90 and pHF: HR = 0.75, 95% CI= 0.43-1.29) were similar for patients with and without BC in the iHF and pHF groups. Conclusion Risk of all-cause and cardiovascular mortality in HF patients did not differ by BC status. Differences in pre-existing myocardial infarction and HF treatment among pHF patients with and without BC may suggest differences in pathogenesis of HF.

ABSTRACT Objective We aimed to determine the longitudinal prevalence and the predictors of sickness absence (SA) and disability pension (DP) in breast cancer (BC) women who eventually developed relapse. Methods A total of 1293 BC women, who were ages 20–63 years, diagnosed between 1996 and 2011 and by 2016 had all developed relapse, were identified in Swedish registers and were followed from two years before to five years after their primary diagnosis, while they were relapse‐free. Annual prevalence of SA and DP was calculated. Logistic regression was used to estimate adjusted odds ratios (AOR) for long‐term SA (>30 days) at one (y1) and three (y3) years post‐diagnosis. Results Prevalence of long‐term SA was 68.1% in y1 and 16.3% in y5. Prevalence of DP progressively increased from 16.3% in y1 to 29.0% in y5. Predictors of long‐term SA included age <50 years (y1:AOR = 1.79 [1.39–2.29]), TNM stage III (y1:AOR = 1.54 [1.03‐2.31]; y3:AOR = 2.21 [1.32–3.72]), metastasis (y1:AOR = 1.64 [1.26–2.12]; y3:AOR = 1.51 [1.05–2.18]), comorbidity (y1:AOR = 2.41 [1.55–3.76]; y3 AOR = 4.62 [2.49–8.57]) and any combination of radiotherapy, chemotherapy and hormonal therapy (y1:AOR = 2.05–5.71). Conclusion Among BC women who later developed relapse, those who had higher stages of BC, had comorbidity and received neoadjuvant and/or adjuvant therapy were at significantly higher risk of needing long‐term SA after their diagnosis.

IntroductionCardiovascular disease (CVD) is an important cause of death in breast cancer survivors. Some breast cancer treatments including anthracyclines, trastuzumab and radiotherapy can increase the risk of CVD, especially for patients with pre-existing CVD risk factors. Early identification of patients at increased CVD risk may allow switching to less cardiotoxic treatments, active surveillance or treatment of CVD risk factors. One of the strongest independent CVD risk factors is the presence and extent of coronary artery calcifications (CAC). In clinical practice, CAC are generally quantified on ECG-triggered cardiac CT scans. Patients with breast cancer treated with radiotherapy routinely undergo radiotherapy planning CT scans of the chest, and those scans could provide the opportunity to routinely assess CAC before a potentially cardiotoxic treatment. The Bragatston study aims to investigate the association between calcifications in the coronary arteries, aorta and heart valves (hereinafter called ‘cardiovascular calcifications’) measured automatically on planning CT scans of patients with breast cancer and CVD risk.Methods and analysisIn a first step, we will optimise and validate a deep learning algorithm for automated quantification of cardiovascular calcifications on planning CT scans of patients with breast cancer. Then, in a multicentre cohort study (University Medical Center Utrecht, Utrecht, Erasmus MC Cancer Institute, Rotterdam and Radboudumc, Nijmegen, The Netherlands), the association between cardiovascular calcifications measured on planning CT scans of patients with breast cancer (n≈16 000) and incident (non-)fatal CVD events will be evaluated. To assess the added predictive value of these calcifications over traditional CVD risk factors and treatment characteristics, a case-cohort analysis will be performed among all cohort members diagnosed with a CVD event during follow-up (n≈200) and a random sample of the baseline cohort (n≈600).Ethics and disseminationThe Institutional Review Boards of the participating hospitals decided that the Medical Research Involving Human Subjects Act does not apply. Findings will be published in peer-reviewed journals and presented at conferences.Trial registration numberNCT03206333.