Multimodal Graph-based Classification of Esophageal Motility Disorders

2026-05-13

AI Summary

This study addresses the diagnostic challenges in esophageal motility disorders arising from the complexity of high-resolution impedance manometry (HRIM) data and inter-clinician variability in interpretation. To this end, the authors propose a novel multimodal fusion approach that, for the first time, applies graph neural networks (GNNs) to model HRIM signals by constructing a spatiotemporal graph representation of esophageal physiology. This GNN-derived representation is integrated with patient-specific textual embeddings extracted via a large language model to enable multiclass classification of swallowing events. The proposed method significantly outperforms baseline models relying solely on HRIM features across all classification tasks, and its graph-based modeling demonstrates markedly superior performance compared to conventional visual analysis techniques.

Abstract

Diagnosing esophageal motility disorders pose significant challenges due to the complexity of high-resolution impedance manometry (HRIM) data and variability in clinical interpretation. This work explores the feasibility of a multimodal Machine Learning (ML)-based classification approach that combines HRIM recordings with patient-specific information and incorporates a graph-based modeling of esophageal physiology. We analyze HRIM recordings with corresponding patient information from 104 patients with esophageal motility disorders. Patient data includes demographic, clinical, and symptom information extracted from structured questionnaires and free-text notes using keyword detection and large language model-based processing. HRIM data is represented as spatio-temporal graphs, where nodes correspond to pressure values along the esophagus and edges encode spatial adjacency and impedance dynamics. A graph neural network (GNN) is applied to learn physiologically meaningful representations, which are fused with patient embeddings for multi-category, multi-class classification of swallow events. The impact of patient features and graph-based modeling is evaluated by ablation studies and comparison to vision-based classifier baselines. The proposed multimodal approach indicates improvements over models that rely solely on HRIM-derived features across all classification categories. Additionally, the graph-based modeling provides gains compared to vision-based baselines. Our experiments systematically assess the complementary contribution of multiple modalities, as well as demonstrate the feasibility of our proposed graph-based approach. Our initial findings demonstrate that integrating patient-level data with graph-based representations of HRIM signals appears to be a promising direction for more accurate classification of esophageal motility disorders.

Problem

• esophageal motility disorders
• high-resolution impedance manometry
• clinical interpretation variability
• diagnostic challenges

Innovation

• graph neural network
• multimodal learning
• esophageal motility disorders
• high-resolution impedance manometry
• patient embedding