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

2025-10-30

AI Summary

To address the performance limitations of personalized treatment effect prediction in few-shot and rare-disease cohorts—stemming from scarcity of high-quality clinical data—this paper proposes a Cross-Fidelity Knowledge Distillation and Adaptive Fusion Network. Methodologically, it introduces a dual-channel knowledge distillation module that jointly leverages abundant low-fidelity synthetic data and scarce high-fidelity clinical trial data; incorporates an attention-guided dynamic fusion mechanism to enable complementary modeling of heterogeneous, multi-source features; and constructs an interpretable variant to uncover latent medical semantics. Evaluated on chronic obstructive pulmonary disease (COPD) treatment response prediction, the method achieves 74.55% accuracy—surpassing the state-of-the-art by 6.67%—while demonstrating strong robustness across varying scales of high-fidelity data. This work advances clinically actionable, interpretable, and generalizable decision support for precision medicine.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Data Mining & Knowledge Management: Knowledge Acquisition from the Web
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

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.

Problem

• Predicting treatment outcomes for rare patient groups with scarce data
• Enhancing prediction accuracy using low-fidelity simulation data
• Dynamically integrating multi-source information via knowledge distillation

Innovation

• Leverages low-fidelity simulation data to enhance trial predictions
• Uses dual-channel knowledge distillation for complementary knowledge extraction
• Employs attention-guided fusion to dynamically integrate multi-source information