A Mathematical Model for Complete Morphological Regression in Primary Operable HER2-Positive Breast Cancer

Background. Breast cancer (BC) is distinguished with its biological tumour subtypes as luminal A, B, HER2-positive and triple-negative. The current clinical guidelines of the Russian Ministry of Health prescribe neoadjuvant targeted chemotherapy as combined treatment in the HER2-positive cancer subtype. An adequate model for treatment efficacy prediction in such patients had been missing to date. Aim. Development of a mathematical model and its computer realisation for complete morphological regression estimation in patients with primary operable HER2-positive breast cancer. Materials and methods. Statistically significant predictors were estimated with the treatment outcome data on 103 HER2- positive breast cancer cases with neoadjuvant targeted chemotherapy. A binary logistic regression model was developed to account for a dichotomous variable dependency on certain predictors. Results and discussion. Multivariate analysis laid out a mathematical model and software “Complete morphological regression estimation in primary operable EGFR-expressing breast cancer under neoadjuvant chemotherapy”. Our results attest that the program correctly automates a systematic estimation of complete morphological regression achieved prior to neoadjuvant targeted chemotherapy and is clinically justified for optimising treatment regimens in primary operable HER2-positive BC. Conclusion. The mathematical model and computer program developed estimate the rate of complete morphological regression achieved prior to neoadjuvant targeted chemotherapy with a high 92 % sensitivity, 97.33 % specificity and 93.21% accuracy.