Search papers, labs, and topics across Lattice.
This paper introduces a hierarchical mixture-of-experts (MoE) model for the classification of interstitial lung disease (ILD) that effectively integrates imaging data with structured electronic health records (EHR). By employing a two-stage gating mechanism, the model assigns patient-specific weights to predictions from both modalities and decomposes the EHR branch into clinically relevant feature groups, thereby enhancing interpretability and specialization. The proposed approach outperformed existing methods, achieving a mean AUC of 0.8750 under strict patient-level cross-validation, demonstrating its potential for improved diagnostic accuracy in complex medical scenarios.
Achieving a mean AUC of 0.8750, this hierarchical MoE model outperforms traditional methods by effectively integrating imaging and EHR data for ILD classification.
Mixture-of-experts (MoE) models combine specialized predictors under learned routing, offering a principled mechanism for leveraging heterogeneity in medical data. We present a hierarchical multimodal MoE for interstitial lung disease (ILD) classification that integrates a frozen, pre-trained imaging expert with structured electronic health records (EHR) via two-stage gating. A modality-level gate assigns patient-specific weights to imaging and EHR predictions, while a sub-gating module decomposes the EHR branch into clinically defined feature groups with learned, group-specific contributions. This design preserves stable imaging representations while enabling input-dependent clinical weighting and explicit EHR specialization. Under strict patient-level cross-validation, the model achieved the highest mean AUC among the evaluated methods (0.8750 +- 0.0443), compared with 0.8646 for imaging-only REN and 0.7685 for SwinUNETR. The framework extends interpretability across anatomical regions, imaging--EHR utilization, and clinically defined EHR feature groups.
