External Validation of the Breast Cancer Surgery Risk Calculator (BCSRc): A Predictive Model for Postoperative...

Area of Science:

• Oncology
• Surgical Outcomes Research
• Medical Informatics

Background:

• The breast cancer surgical risk calculator (BCSRc) is a tool for predicting acute complications after breast cancer surgery.
• It aids in preoperative risk stratification and informed patient-physician decision-making.
• This study aimed to externally validate the BCSRc models on a new patient cohort.

Purpose of the Study:

• To externally validate the four predictive models within the breast cancer surgical risk calculator (BCSRc).
• To assess the performance, accuracy, and calibration of the BCSRc on a large, independent cohort of breast cancer patients.

Main Methods:

• Utilized a large cohort from the National Surgical Quality Improvement Program (2019-2020) for external validation.
• Employed logistic regression models to predict overall, infectious, hematologic, and internal organ complications.
• Evaluated model performance using area under the curve, Brier score, and Hosmer-Lemeshow tests.

Main Results:

• External validation demonstrated strong performance with area under the curve values around 0.70 for all models.
• Accuracy (Brier scores) ranged from 0.003 to 0.04, indicating good predictive accuracy.
• Hosmer-Lemeshow tests showed good model calibration, with all p-values > 0.05.

Conclusions:

• The breast cancer surgical risk calculator (BCSRc) demonstrates excellent external validation measures.
• The tool effectively stratifies patients by complication risk, enhancing preoperative decision-making.
• Updated coefficients will be implemented on the BCSRc platform (www.breastcalc.org).