Development and Validation of a Nomogram for Predicting Breast Malignancy in Male Patients Based on Clinical and Ultrasound Features.

Abstract

Objective: This study aimed to construct a nomogram based on clinical and ultrasound (US) features to predict breast malignancy in males.

Methods: The medical records between August, 2021 and February, 2023 were retrospectively collected from the database. Patients included in this study were randomly divided into training and validation sets in a 7:3 ratio. The models for predicting the risk of malignancy in male patients with breast lesions were virtualized by the nomograms.

Results: Among the 71 enrolled patients, 50 were grouped into the training set, while 21 were grouped into the validation set. After the multivariate analysis was done, pain, BI-RADS category, and elastography score were identified as the predictors for malignancy risk and were selected to generate the nomogram. The C-index was 0.931 for the model. Concordance between predictions and observations was detected by calibration curves and was found to be good in this study. The model achieved a net benefit across all threshold probabilities, which was shown by the decision curve analysis (DCA) curve.

Conclusion: We successfully constructed a nomogram to evaluate the risk of breast malignancy in males using clinical and US features, including pain, BI-RADS category, and elastography score, which yielded good predictive performance.