Machine learning prediction of pathological complete response and overall survival of breast cancer patients in the Montefiore Health System in the Bronx

Abstract Background. Predicting response to treatment and survival in breast cancer patients remains a crucial challenge. This study employed four machine-learning models to predict pathological complete response (pCR) and overall survival (OS) up to 7.5 years. Methods. Demographics, staging, tumor subtypes, income, insurance status, and data from radiology reports were obtained from 475 breast cancer patients on neoadjuvant chemotherapy from 01/01/2017 to 12/31/2021 in the Montefiore Health System in the Bronx. Logistic regression, Neural Network, Random Forest, and Gradient Boosted Regression models were used to predict outcomes (pCR and OS) with five-fold cross validation. Results. pCR was not associated with age, race, ethnicity, differentiation, income, and insurance status (p > 0.05). ER-/HER2 + showed the highest pCR rate, followed by triple negative, ER+/HER2+, and ER+/HER2- (p < 0.05), tumor staging (p = 0.011), tumor size (p < 0.003) and background parenchymal enhancement (BPE) (p < 0.03) were associated with pCR. Machine-learning models ranked ER+/HER2-, ER-/HER2+, tumor size, and BPE as top predictors of pCR (AUC = 0.74–0.76). OS was associated with race, pCR status, tumor subtype, and insurance status (p < 0.05), but not ethnicity and incomes (p > 0.05). Machine-learning models ranked tumor stage, pCR, nodal stage, and triple negative subtype as top predictors of OS (AUC = 0.83–0.85). When grouping race and ethnicity by tumor subtypes, neither OS nor pCR were different due to race and ethnicity for each tumor subtype (p > 0.05). Conclusion. Tumor subtypes and imaging characteristics were top predictors of pCR. Insurance status, race, tumor subtypes and pCR were associated with OS. Machine-learning models accurately predicted pCR and OS using clinicopathological and radiological data.