Validity of models for predicting BRCA1 and BRCA2 mutations.

Area of Science:

• Genetics
• Oncology
• Medical Informatics

Background:

• Deleterious mutations in BRCA1 and BRCA2 genes significantly increase susceptibility to breast and ovarian cancers.
• Multiple models exist for estimating mutation probabilities, but their comparative accuracy and limitations are not fully understood.

Purpose of the Study:

• To systematically evaluate and quantify the predictive accuracy of seven widely used models for BRCA1/BRCA2 mutation carrier status.
• To compare the performance of models including BRCAPRO, family history assessment tool, Finnish, Myriad, National Cancer Institute, University of Pennsylvania, and Yale University.

Main Methods:

• A cross-sectional validation study was conducted using independent patient data not involved in model development.
• The study included multicenter data from the Cancer Genetics Network, encompassing population-based samples and genetic counseling clinic attendees.
• Model performance was assessed using the c-statistic, sensitivity, and specificity to discriminate between mutation carriers and non-carriers.

Main Results:

• Significant variation in predictive accuracy was observed among the seven models, with top performers achieving a c-statistic around 80%.
• BRCAPRO demonstrated the highest overall c-statistic, though performance margins over other models were often narrow.
• All models exhibited notable false-negative and false-positive rates, particularly in non-high-risk populations, across various probability thresholds.

Conclusions:

• All evaluated models provide adequate discrimination for identifying potential BRCA1/BRCA2 mutation carriers, supporting individualized genetic counseling.
• Discrimination capabilities differ across models and patient populations, highlighting the need for careful model selection.
• Three recently published models were not included in this validation study.