Fig. 3

Breast cancer polygenic risk score in DCIS patients with and without a breast cancer subsequent event. a Schematic overview of the study design. Treatment consisted of surgery and adjuvant radiation or endocrine therapy. b Comparison of breast cancer PRS score distribution between patients with (red) or without (black) a breast cancer subsequent event (BCSE). Dashed vertical lines represent mean normalized PRS values for each respective group. A logistic regression was constructed with each PRS and DCIS size, grade, age and ancestry for outcome of 5-year recurrence, the resulting Pseudo-R-squared, p-value and Bonferroni corrected q-values are listed for each PRS. Distributions were generated using kernel density estimates of histograms. c Forest plot representation of hazard ratios (square) and 95% confidence intervals (error-bars), for each tested breast cancer PRS, obtained from a Cox Proportional-Hazard model accounting for DCIS size, grade, and age, the ancestry of the patient (Additional file 1: Figure S4). The dotted line represents a log hazard ratio of 1, or having no effect on the outcome. The q-values represent Bonferroni corrected p-values for the effective number of tests. Significant hazard ratios (q < 0.05) are indicated in bold text. d Evaluation of discrimination of Cox proportional hazard model for BCSE vs no BCSE outcome using Harrel’s C-index (y-axis) for models only using available risk factors versus available risk factors and breast cancer PRS, colored by the significance of hazard ratios for breast PRS (q < 0.05, light green). e Same as c but for non-cancer PRS. f Cox proportional hazard estimate of breast cancer subsequent event (BCSE)—free survival for two PRS over time in years. Curves are obtained by varying PRS (solid colored lines from blue as lowest and red as highest PRS percentile), as compared to each model baseline (dashed line) while keeping all other covariates the same. Each case and control was weighted by the epidemiological incidence of BCSE treated with surgery and endocrine therapy (15% at 10 years) [24]