Uncertainty-Aware Logistic Regression with Gray-Zone Refinement for Predicting Response to Neoadjuvant Chemotherapy in Breast Cancer

Aixa Ximena Torres Fuertes, Rodrigo Romero Tello, Fatima Rosmery Jara Cuya, Jesus Aldair Sullon Silva, Ariana Mirella Villegas Suarez

Keywords: Breast cancer, neoadjuvant chemotherapy, machine learning, logistic regression, uncertainty quantification, transcriptomics

TL;DR: Predicting Chemotherapy Response in Breast Cancer

Track: Proceedings

Abstract: Predicting response to neoadjuvant chemotherapy (NAC) in breast cancer remains a clinical challenge. We developed a machine learning framework combining bibliographicallyweighted Elastic Net for dimensionality reduction with regularized Logistic Regression (LR) as the primary model, and a selective escalation strategy using a multilayer perceptron (MLP) for ambiguous predictions. From GSE205568 (n=2551), 730 robust genes were selected. LR achieved strong performance (nested-CV AUCPR = 0.82, ROC-AUC = 0.93), but uncertainty analysis identified a “gray zone” near the decision threshold, concentrating misclassif ications. Routing these cases to an MLP and aggregating outputs via stacking with isotonic recalibration improved gray-zone AUCPR by +0.24 and yielded perfect calibration (ECE ≈ 0). External validation on GSE25065 (n=198) showed that while discrimination transferred (ROC-AUC = 0.94, AUCPR = 0.76), recalibration and local threshold adjustment were required to recover clinically useful performance (F1 = 0.74, Recall = 0.95) (de Hond et al., 2023). These findings support the use of LR as a reliable baseline, augmented by explicit uncertainty detection and selective complexity to improve robustness in clinical prediction.

General Area: Applications and Practice

Specific Subject Areas: Computational Biology, Explainability & Interpretability

Data And Code Availability: Yes

Ethics Board Approval: No

Entered Conflicts: I confirm the above

Anonymity: I confirm the above

Submission Number: 225