Role of artificial intelligence -based machine learning model in predicting HER2/neu gene status in breast cancer

Our study investigated the predictive efficacy of AI-based Machine Learning (ML) model for determining HER2 status in a population of 3424 breast cancer patients. Multivariate logistic regression analysis identified several independent variables that were predictive of HER2 positivity, namely age ≤ 40 years, tumor multicentricity, high tumor grade, high-grade DCIS, N3 stage disease, and negative ER status (p < 0.05). These findings suggest that patients presenting with these factors may benefit from more aggressive and targeted therapies. Furthermore, XGBoost ML model was trained using the dataset of 3324 patients, which was divided into an 80 % training set and a 20 % test set. The model achieved an impressive accuracy of 95 % on both training and test sets, as evidenced by the area under the curve (AUC) values of 0.95. The model ranked the presence of DCIS, DCIS component (major versus minor), DCIS grade, multiplicity of the tumor, and ER status as the top four variables for predicting HER2/neu status. To validate the performance of the proposed model, blind HER2 status data from an external validation cohort of 100 cases were utilized. Notably, the model demonstrated a sensitivity of 90.5 %, indicating its ability to accurately identify HER2-positive cases, and a specificity of 84.4 %, suggesting its capability to correctly classify HER2-negative cases. These results highlight the promising predictive efficacy of AI-based ML in determining HER2 status in breast cancer patients. The model's ability to accurately identify HER2-positive cases can assist in guiding treatment decisions, ensuring that patients receive appropriate and targeted therapies. However, further research with larger datasets is necessary to validate and generalize these findings.