Increased incidence and poor prognosis of breast cancer in postmenopausal women with high body mass index attending a mammography screening program in the province of Modena (Italy)

sia status, and family history suggestive of hereditary breast/ovarian cancer (HBOC). The aims of this study were to externally validate the nomogram using an independent patient cohort and to evaluate the predictive ability of BRCA status compared to family history suggestive of HBOC. Methods: We examined themedical records of all consecutive patients who underwent primary surgery for epithelial ovarian cancer at our institution between July 1992 and August 2009. Patients were included if the 7 variables from the MSKCC nomogram were available and their germline BRCA mutation status was known. Concordance index (CI) and calibration plot were used to validate the existing model on our external data. To compare predictive capabilities of BRCA and HBOC, a base competing risk regressionmodelwas formedwith theMSKCC data (excluding HBOC) and applied to our local data to generate a risk score for each patient. This risk score was then used with either HBOC or BRCA to build two new models from our validation data, whose CIs were compared through bootstrap resamples. Results: A total of 304 patients were included in the final analysis. HBOC was present in 89 patients (29%), and a BRCA mutation was confirmed in 37 patients (12%). The CI of the original nomogram on our external patients was 0.670. The newmodels built on our data showed no difference in bootstrap-corrected CIs (0.677 for both the BRCA and the HBOC models, P= 0.994). Conclusions: TheMSKCC all-stage epithelial ovarian cancer nomogram was externally validated in an independent population. Knowledge of BRCA status rather than just HBOC family history was not shown to improve the prognostic value of this tool, therebymaintaining excellent accessibility of this nomogram in clinical practice.