Personalized Treatment Outcome Prediction from Scarce Data via Dual-Channel Knowledge Distillation and Adaptive Fusion

TL;DR

This paper introduces CFKD-AFN, a novel method that improves personalized treatment outcome predictions for scarce data by leveraging low-fidelity simulation data through dual-channel knowledge distillation and adaptive fusion, with demonstrated accuracy gains and interpretability.

Contribution

The paper presents a new cross-fidelity knowledge distillation and adaptive fusion network that effectively combines low- and high-fidelity data for improved predictions in precision medicine.

Findings

Significant prediction accuracy improvements (6.67% to 74.55%) over state-of-the-art methods.

Robustness to varying sizes of high-fidelity datasets.

Extension to an interpretable model for clinical decision support.

Abstract

Personalized treatment outcome prediction based on trial data for small-sample and rare patient groups is critical in precision medicine. However, the costly trial data limit the prediction performance. To address this issue, we propose a cross-fidelity knowledge distillation and adaptive fusion network (CFKD-AFN), which leverages abundant but low-fidelity simulation data to enhance predictions on scarce but high-fidelity trial data. CFKD-AFN incorporates a dual-channel knowledge distillation module to extract complementary knowledge from the low-fidelity model, along with an attention-guided fusion module to dynamically integrate multi-source information. Experiments on treatment outcome prediction for the chronic obstructive pulmonary disease demonstrates significant improvements of CFKD-AFN over state-of-the-art methods in prediction accuracy, ranging from 6.67\% to 74.55\%, and strong robustness to varying high-fidelity dataset sizes. Furthermore, we extend CFKD-AFN to an interpretable variant, enabling the exploration of latent medical semantics to support clinical decision-making.