Abstract 4623: External validation of the BOADICEA model for predicting ovarian cancer risk: The Breast Cancer Family Registry

Abstract Purpose: In the U.S., ovarian cancer is the most fatal of the gynecological cancers with an overall five-year survival rate of 47.6%. While women with a BRCA1 or BRCA2 mutation are at a much higher risk of developing ovarian cancer, mutations in these genes account for less than 50 percent of the familial aggregation of ovarian cancer. Being able to identify women at greatest risk, including those with increased familial risk without a BRCA1 or BRCA2 mutation, is critical for clinical decision-making; however, there is a lack of risk prediction models for ovarian cancer and those that do exist have modest discriminatory power. Therefore, we externally validated the BOADICEA model, a breast and ovarian cancer risk prediction model based on a woman's multigenerational family history and genetic information, for predicting ovarian cancer risk in an independent, prospective cohort of women. Methods: We used data from the Breast Cancer Family Registry (BCFR), a cohort of families with breast and ovarian cancer at baseline that have been prospectively followed. We assessed the 10-year performance of the BOADICEA model (version 3) for ovarian cancer risk overall, and by known BRCA1 or BRCA2 mutation statuses. We included women who did not have an ovarian cancer diagnosis or a bilateral oophorectomy prior to baseline, and who had sufficient data to calculate the 10-year BOADICEA risk score. We assessed model calibration using the ratio of the expected (E) to the observed (O) number of ovarian cancer cases in the cohort (E/O), and model discrimination by the concordance statistics (C-statistic) derived from the receiver operating characteristic curves. Results: There were 125 prospective ovarian cancer cases over a median of 12.3 years of follow-up among 18,534 women eligible for this analysis. For the overall cohort, the BOADICEA model was well calibrated with an E/O of 0.87, 95% confidence interval (CI) (0.70, 1.08). There was a suggestion of model underprediction in the top quartile of assigned risk (1.33% observed risk vs. 1.09% predicted risk), however it was not statistically significant (p=0.15). The C-statistic was 0.77, 95% CI (0.73, 0.82). For known BRCA1 or BRCA2 mutation carriers combined, the E/O was 0.91, 95% CI (0.63, 1.32) and the C-statistic was 0.75, 95% CI (0.67, 0.83). For non-carriers, the E/O was 0.85, 95% CI (0.65, 1.11) and the C-statistic was 0.68, 95% CI (0.59, 0.74). Conclusion: The BOADICEA model is well-calibrated in predicting ovarian cancer risk over 10 years and has good discriminatory power for women at increased familial risk of breast and ovarian cancer, with or without a known mutation in BRCA1 or BRCA2. Therefore, BOADICEA has clinical utility for evaluating ovarian cancer risk based on a woman's family cancer history and genetic information. Citation Format: Jennifer S. Ferris, Jeanine M. Genkinger, Mary Beth Terry, Yuyan Liao, Robert J. MacInnis, Irene L. Andrulis, Saundra S. Buys, Mary B. Daly, Esther M. John, John L. Hopper. External validation of the BOADICEA model for predicting ovarian cancer risk: The Breast Cancer Family Registry [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4623.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.