Contralateral breast cancer risks for BRCA1, BRCA2, PALB2, CHEK2, and ATM pathogenic variant carriers in two large cohort studies Authors: Isaac Allen, Nancy Whish, Siddhartha Yadav, Nicholas J. Boddicker, Caroline Drukker, Irma van de Beek, Diana M. Eccles, Clare Turnbull, Paul Pharoah, Antonis C. Antoniou, Douglas F. Easton, Fergus J. Couch, Marjanka K. Schmidt, Marc Tischkowitz Published: 2026 ABSTRACT BackgroundAll breast cancer (BC) patients have a risk of developing contralateral BC (CBC). This will be higher in women carrying a BC susceptibility gene pathogenic variant (PV), but specific risk estimates are lacking. We addressed this in a study-level meta-analysis of two large, recent cohort studies that used Breast Cancer Risk after Diagnostic Gene Sequencing (BRIDGES) project and Cancer Risk Estimates Related to Susceptibility (CARRIERS) consortium data.MethodsWe used population risks from the Netherlands to estimate 10-year cumulative CBC risks for BRCA1, BRCA2, PALB2, CHEK2, and ATM PV carriers. We estimated pooled hazard ratios (HRs) in meta-analyses using the generic inverse-variance method and assuming random effects. We stratified by first BC estrogen receptor (ER) status.ResultsAmong 462 BRCA1, 590 BRCA2, 796 CHEK2, 297 PALB2, and 345 ATM PV carriers, we estimated 10-year cumulative CBC risks of 10.1% (95% confidence interval (CI):7.5–13.6%) for BRCA1, 10.2% (95%CI:7.4–13.9%) for BRCA2, 7.1% (95%CI:3.4–14.3%) for PALB2, 8.0% (95%CI:5.7–11.1%) for CHEK2 (any PV), 8.5% (95%CI:5.9–12.3%) for CHEK2 del.1100 C, and 4.5% (95%CI:2.5–8.0%) for ATM PV carriers. We found elevated HRs in BRCA1 (HR:2.8, 95%CI:2.1–3.6), BRCA2 (HR:2.8, 95%CI:2.1–3.7), and CHEK2 PV carriers (all PVs: HR:2.1, 95%CI:1.6–2.9. del.1100 C PVs only: HR:2.3, 95%CI:1.6–3.3). We found no significantly elevated risks in PALB2 (HR:1.9, 95%CI:0.9–3.8) or ATM (HR:1.2, 95%CI:0.7–2.1) PV carriers. We saw increased risks in BRCA1/2 and CHEK2 PV carriers following ER-positive BC, and in BRCA1/2 and PALB2 PV carriers following ER- negative BC.ConclusionBC survivors carrying BRCA1, BRCA2, and CHEK2 PVs and ER-negative BC survivors carrying PALB2 PVs are at two-fold or higher CBC risks. These estimates may assist clinical decision making for management of the contralateral breast. Introduction BRCA1, BRCA2, PALB2, CHEK2, and ATM PVs are associated with elevated breast cancer (BC) risks [1, 2]. Contralateral BC (CBC) risks have also been shown to be elevated in BRCA1 and BRCA2 PV carriers, although CBC counts were small even in the largest studies to date [3–6]. Research into CBC risks associated with PALB2, CHEK2, or ATM PVs is more limited [4, 5]. Precise CBC risk estimates for BRCA1, BRCA2, PALB2, CHEK2, and ATM PV carriers would partly address this gap, which could have clinical management implications. For example, contralateral risk-reducing mastectomy uptake following BC is higher in BRCA1, BRCA2, PALB2, CHEK2, and ATM PV carriers than those without PVs in BC susceptibility genes [7]. CBC risk estimates based on large cohorts with long-term follow-up could help assess the benefits of these procedures and aid the counseling and personalized risk management process for BC survivors carrying PVs in BC susceptibility genes. Two large, recent cohort studies by Morra et al. [5] and Yadav et al. [4] estimated CBC risks in BRCA1, BRCA2, PALB2, CHEK2, and ATM PV carriers. Morra et al. followed 34,401 BC survivors of European ancestry, 5.3% of whom carried such PVs (330 BRCA1, 420 BRCA2, 656 CHEK2, 200 PALB2, and 229 ATM PV carriers), for CBC over a median of 10.9 years of follow-up. Yadav et al. followed 15,104 BC survivors treated with ipsilateral surgery, of predominantly European ancestry, 4.4% of whom carried such PVs (132 BRCA1, 170 BRCA2, 140 CHEK2, 97 PALB2, and 116 ATM PV carriers), for CBC over a median of 11.0 years of follow-up. Morra et al. drew their data from the Breast Cancer Risk after Diagnostic Gene Sequencing (BRIDGES) project, and Yadav et al. from the Cancer Risk Estimates Related to Susceptibility (CARRIERS) consortium. Further details of these studies, such as risk estimates provided and further inclusion criteria, are in the supplementary material. Morra et al. observed 103 CBCs in PV carriers, while Yadav et al. observed 90 CBCs in PV carriers. Our objective was to examine the consistency of the relative risk estimates for CBC in the two studies and pool the relative risk estimates in a study-level meta-analysis, to derive more precise estimates than were available from the individual studies. The combined relative risk estimates have then been combined with external population-based CBC risk estimates to provide gene-specific cumulative CBC risks. Methods We included the studies by Morra et al. [5] and Yadav et al. [4] in our meta-analysis, as they are the largest recent population-based studies that report CBC risk estimates in BRCA1, BRCA2, PALB2, CHEK2, and ATM PV carriers. Exposure, outcome, and measure of association We defined carrying a BRCA1, BRCA2, PALB2, CHEK2, or ATM PV as separate exposures, and a CBC diagnosis as the outcome. We selected the hazard ratio (HR) as the measure of association. Statistical analysis We performed all statistical analyses in R version 4.3.1 [8]. For both studies, we extracted the CBC HR as the principal summary measure. Both studies presented CBC HRs both adjusted and unadjusted for covariates such as first BC pathology and receipt of treatments. The adjusted HRs were the ones extracted. We used the random-effects generic inverse variance method with DerSimonian-Laird estimators [9, 10] to perform the meta-analyses. We calculated standard errors for the HRs by dividing the differences between the natural logarithms of the 95% confidence interval boundaries by 3.92, as recommended by the Cochrane Handbook for Systematic Reviews of Interventions [9]. We performed meta-analyses of CBC HRs separately for PV carriers of each gene above, both without stratifying by initial BC subtype and restricted to survivors of estrogen receptor (ER)-positive first BC. We performed a meta-analysis of c.1100del (p.Thr367fs) CHEK2 PV carriers in addition to a meta-analysis of all CHEK2 PV carriers. We performed meta-analyses for CBC risk in BRCA1, BRCA2, and PALB2 PV carriers with ER-negative first BC. We did not perform a meta-analysis of CBC risks in CHEK2 or ATM PV carriers with ER-negative first BC as HRs were not provided by both studies (PVs in these genes are not typically associated with ER-negative BC). We assessed between-study heterogeneity using Cochran’s Q test [11]. We estimated 10-year cumulative CBC risks in BRCA1, BRCA2, PALB2, CHEK2 (all PVs and the c.1100del variant alone), and ATM PV carriers following BC of any subtype and ER-positive BC. We also estimated 10-year cumulative CBC risks in BRCA1, BRCA2, and PALB2 PV carriers following ER-negative BC. We did this by taking estimates of cumulative CBC risks among BC survivors, ER-positive BC survivors, and ER-negative BC survivors from the Netherlands Cancer Institute (NKI), based on published national registry data as the baseline estimates [12]. We assumed these baseline risks approximated the risks in BRCA1, BRCA2, PALB2, CHEK2, and ATM PV non-carriers. We subtracted these baseline estimates from 1 and raised the results to the power of the corresponding HRs. We then subtracted the results from 1 and multiplied by 100 to give percentage cumulative risks. For example, a HR of 2 applied to a baseline cumulative risk of 10.00% would correspond to an absolute cumulative risk of 19.0%. As a sensitivity analysis, we recalculated each of the cumulative risk estimates for ER-positive and ER-negative BC survivors with the baseline NKI risks separately restricted to those diagnosed with BC of the relevant ER status at age under 40, 40 to 54, 55 to 69, and 70 to 84 years. Results Cumulative risk estimates We estimated 10-year cumulative CBC risks of 10.1% (95%CI:7.5–13.6%) for BRCA1, 10.2% (95%CI:7.4–13.9%) for BRCA2, 7.1% (95%CI:3.4–14.3%) for PALB2, 8.0% (95%CI:5.7–11.1%) for CHEK2 (any PV), 8.5% (95%CI:5.9–12.3%) for CHEK2 del.1100 C, and 4.5% (95%CI:2.5–8.0%) for ATM PV carriers following a first BC diagnosis irrespective of the ER status of the BC, extrapolating from the baseline NKI risk of 3.8% (95%CI:3.7–4.0%). When restricted to survivors of ER-positive BC, the corresponding estimates were 10.4% (95%CI:5.9–18.2%), 11.1% (7.3–16.8%), 4.4% (95%CI:0.7–24.8%), 7.6% (95%CI:5.2–11.1%), 8.8% (95%CI:5.8–13.3%), and 5.1% (95%CI:2.5–10.0%), with a baseline risk of 3.8% (95%CI:3.6–4.0%). Among ER-negative BC survivors, we estimated cumulative 10-year CBC risks of 12.0% (95%CI:7.8–18.1%), 8.6% (95%CI:3.9–18.6%), and 11.4% (95%CI:5.6–22.7%) in BRCA1, BRCA2, and PALB2 PV carriers, respectively. The baseline risk was 4.2% (95%CI:3.8–4.6%). Cumulative CBC risk plots over 10 years of follow-up, stratified by PV carrier status and by ER status of the first BC, are shown in Fig. 1.Fig. 1Pooled contralateral breast cancer hazard ratios in following breast cancer in BRCA1, BRCA2, PALB2, CHEK2, and ATM pathogenic variant carriers. Pooled contralateral breast cancer hazard ratios in following breast cancer in BRCA1, BRCA2, PALB2, CHEK2, and ATM pathogenic variant carriers. The 10-year cumulative CBC risk estimate for those diagnosed with ER-positive BC in the NKI dataset increases from 2.5% when diagnosed at age under 40 years to 3.5% when diagnosed at age 40 to 54 and 4.5% when diagnosed at age 55 to 69, before decreasing to 3.7% when diagnosed at age 70 to 84 years. In contrast, the corresponding estimates for ER-negative BC survivors are 5.0%, 4.8%, 3.8%, and 3.2%. Since the 10-year cumulative risk estimates in carrier groups were generated by directly extrapolating from baseline NKI risks (Methods section) under the assumption that the HR estimates do not vary by the age at diagnosis, the 10-year cumulative risk estimates for carriers of PVs following ER-positive BC when the baseline NKI risks were stratified by age at first BC diagnosis analogously respectively peaked at 12.2% (95%CI:6.7–21.6%), 13.0% (95%CI:8.4–19.9%), 5.2% (95%CI:0.8–29.2%), 8.9% (95%CI:5.9–13.3%), 10.3% (6.6–15.8%) and 6.0% (2.9–11.9%) in carriers of BRCA1, BRCA2, PALB2, CHEK2 (all PVs and del.1100 C), and ATM PV carriers when the baseline risks contained data only on those aged 55 to 69 at first BC diagnosis. Similarly, under the assumption of constant HR estimates by the age at first BC diagnosis, 10-year CBC risk estimates following ER-negative BC peaked at 14.2% (8.1–24.0%) in BRCA1, 10.2% (95%CI:4.0–24.7%) in BRCA2, and 13.6% (95%CI:5.7–29.9%) in PALB2 PV carriers when the baseline risks contained data only on those diagnosed with BC at under age 40. Ten-year cumulative CBC risk estimates for BRCA1, BRCA2, PALB2, CHEK2 (any PV and del1100C PV only) and ATM PV carriers following ER-positive BC, and for BRCA1, BRCA2, and PALB2 PV carriers following ER-negative BC, stratified by age group used to generate the baseline estimates, may be seen the Supplementary Material (Table S1 and all supplementary figures). Summary hazard ratios CBC risk was significantly increased in BRCA1, BRCA2, and CHEK2 PV carriers among survivors of an initial BC of any subtype (BRCA1: HR:2.8, 95%CI:2.1–3.6. BRCA2: HR:2.8, 95%CI:2.1–3.7. CHEK2 (all PVs): HR:2.1, 95%CI:1.6–2.9. CHEK2 (1100 C.del only): HR:2.3, 95%CI:1.6–3.3) and of positive ER status (BRCA1: HR:2.9, 95%CI:1.6–4.9. BRCA2: HR:3.0, 95%CI:2.0–4.5. CHEK2 (all PVs): HR:2.0, 95%CI:1.4–2.9. CHEK2 (1100 C.del only): HR:2.4, 95%CI:1.6–3.5). We found significant evidence for an elevation in CBC risks among BRCA1, BRCA2, and PALB2 carriers with an ER-negative first BC (BRCA1: HR:3.0, 95%CI:2.1–4.3. BRCA2: HR:2.1, 95%CI:1.0–4.4. PALB2: HR:2.9, 95%CI:1.5–5.5). Notably, we did not see a significantly increased CBC risk in PALB2 or ATM PV carriers with either a first BC of any subtype or an ER-positive first BC (PALB2 PV carriers: Any subtype: HR:1.9, 95%CI:0.9–3.8. ER positive: HR: 1.2, 95%CI:0.2–7.0. ATM PV carriers: Any subtype: HR:1.2, 95%CI:0.7–2.1. ER positive: HR: 1.3, 95%CI:0.7–2.6). We found no significant evidence of heterogeneity in any of the meta-analyses. All HR estimates are shown in Fig. 2.Fig. 2Ten-year contralateral breast cancer cumulative risk curves for BRCA1, BRCA2, PALB2, CHEK2, and ATM pathogenic variant carriers following breast cancer.a First breast cancer of any subtype. b Estrogen receptor positive breast cancer. c Estrogen receptor negative breast cancer. Ten-year contralateral breast cancer cumulative risk curves for BRCA1, BRCA2, PALB2, CHEK2, and ATM pathogenic variant carriers following breast cancer. a First breast cancer of any subtype. b Estrogen receptor positive breast cancer. c Estrogen receptor negative breast cancer. Discussion In this study, we presented a meta-analysis of CBC risks in two of the largest population-based cohort studies of BRCA1, BRCA2, PALB2, CHEK2, and ATM PV carriers published to date. We found significantly elevated CBC risks in BRCA1, BRCA2, and CHEK2 PV carriers, consistent previous with observations in females of predominantly European ancestry, although the estimates were lower than previously reported [3, 6, 13]. In addition, our 10-year cumulative CBC risk estimates in BRCA1 and BRCA2 PV carriers were more precise than previous estimates [3, 6]. Interestingly, the magnitudes of our CBC relative risk estimates in BRCA1 and BRCA2 PV carriers, relative to noncarriers, were lower than the corresponding first BC risks [2]. This may be partly because a greater proportion of BRCA1 and BRCA2 PV carriers undergo risk-reducing interventions following, rather than prior to, their first BC diagnosis [3, 14], which Morra et al. did not adjust for. However, further research into risk-reducing intervention rates in large cohorts of BRCA1 and BRCA2 PV carriers would be needed to confirm this, and risk-reducing measures do not fully eliminate the risk of a CBC in any case [15]. Another partial explanation may be index-event bias, whereby the magnitude of an association between CBC risk and a risk factor for first BC would be attenuated due to the population at risk being survivors of the initial BC [16]. Index-event bias could also partly account for why we did not observe significant evidence for an elevation in CBC risks among ATM carriers or in PALB2 PV carriers not stratified by ER status of the first BC, despite these carrier groups being at increased risks of first BC, and be the reason for the smaller estimate of PRS313 for CBC compared to the first BC [2, 17]. However, the absence of significant associations for ATM and PALB2 might be due to limited power: the confidence intervals for these estimates are wide. Systemic therapies for the initial BC may also have reduced CBC risk relative to the risk of the first BC, with Kramer et al. reporting CBC HRs of 0.46 (95%CI:0.41–0.52) and 0.70 (95%CI:0.62–0.80) in BC survivors respectively treated with endocrine therapy and adjuvant chemotherapy [12]. However, it should be noted that both Yadav et al. and Morra et al. adjusted their HR estimates for receipt of endocrine therapy, with Morra et al. further adjusting for receipt of trastuzamab and neo-adjuvant or adjuvant chemotherapy. We observed significant increases in CBC risk among ER-positive BC survivors carrying BRCA1, BRCA2, and CHEK2 PVs, and significantly elevated CBC risks among ER-negative BC survivors carrying BRCA1, BRCA2, and PALB2 PVs. If CBC risks are indeed elevated in PALB2 PV carriers only following ER-negative BC rather than ER-positive BC, this could have implications for the clinical management of BC survivors carrying PALB2 PVs. However, these results were based on a total of only 7 CBC events in PALB2 PV carriers following ER-positive BC and 9 CBC events in PALB2 PV carriers following ER-negative BC across both studies, so additional studies in large international consortia will be needed to confirm these findings. Furthermore, additional associations between BRCA1, BRCA2, PALB2, CHEK2, or ATM PV carrier status and CBC risk cannot be ruled out, as some of our analyses may have been underpowered due to the low CBC counts in Morra et al. and Yadav et al. The estimated 10-year CBC cumulative risk estimates in BRCA1 and BRCA2 PV carriers are somewhat lower than those reported by other very large cohort studies [3, 6]. This may be due to ascertainment criteria differences, as previous studies drew their data from BC family cohorts [6] and genetic testing facilities that select for those at higher BC risk due to criteria such as family history [3]. However, the estimates reported within each study varied widely, with Morra et al. estimating 10-year cumulative risks of 7.2% in BRCA1 PV carriers and 5.4% in BRCA2 PV carriers and Yadav et al. observing corresponding risks of 33% and 27%. This could be partly because only Yadav et al. censored at contralateral prophylactic mastectomy, which they observed in 57 patients. It is also possible that differences in the age distribution at first BC diagnosis between the studies could explain some of the difference in cumulative risk estimates. A second weakness of our study is that CBC risk estimates stratified by age at first BC diagnosis were not reported by both Morra et al. and Yadav et al., and therefore, we could not assess the variation in either pooled HR estimates or in CBC cumulative risk estimates by age at first BC diagnosis. Similarly, we could not assess the effects of other first BC risk factors, such as family history, ethnicity, tumor size, morphology, or HER2 status of the first BC, or receipt of hormonal therapy or radiotherapy to treat the first breast tumor [18], due to CBC risk estimates stratified by these factors not being reported by both studies. Finally, we extrapolated our cumulative CBC risk estimates from baseline rates from the NKI, which contained data on BC patients diagnosed between 2003 and 2010. This does not exactly match the diagnosis periods in Yadav et al. or Morra et al. For example, Morra et al. used data on women diagnosed between 1942 and 2018 (median: 2003; IQR: 1999–2006). The primary strengths of this study are the large sample sizes of the studies included in the meta-analysis, the long follow-up of these studies, and the lack of heterogeneity observed in any of the meta-analyses. This combined analysis provides more precise estimates of CBC risk in germline BC susceptibility gene PV carriers, which could aid in personalized decision making on risk management strategies. Our pooled estimates, in conjunction with previous studies [3, 6, 19, 20], indicate that clinicians should discuss CBC risk-reducing options following a first BC diagnosis in females carrying BRCA1 or BRCA2 PVs. However, further studies in larger cohorts would be needed to provide optimal clinical recommendations regarding CBC risk management in CHEK2 PV carriers, or in PALB2 PV carriers following ER-negative first BC. Finally, the lack of significant evidence for an association between ATM PVs and CBC risk in both this combined analysis and in the separate studies by Yadav et al. and Morra et al. could be of clinical relevance, although larger studies would be required to further confirm this finding. Conclusion In this study, we presented pooled relative risk and 10-year cumulative risk estimates of CBC in BRCA1, BRCA2, PALB2, CHEK2, and ATM PV carriers following BC, in BRCA1, BRCA2, PALB2, CHEK2, and ATM PV carriers following ER-positive BC, and in BRCA1, BRCA2, and PALB2 PV carriers following ER-negative BC. Further studies into the variation in CBC risk among BRCA1, BRCA2, PALB2, CHEK2, and ATM PV carriers by age, first BC pathology, and first BC treatment is required to improve optimize the CBC prevention management of women with hereditary breast cancer. Supplementary information Supplementary information Supplementary information