Saved in:
Bibliographic Details
Main Authors: Moradbeiki, Pardis, Ghadiri, Nasser, Zahabi, Sayed Jalal, Wiil, Uffe Kock, Brockhattingen, Kristoffer Kittelmann, Ebrahimi, Ali
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2508.19319
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915465600696320
author Moradbeiki, Pardis
Ghadiri, Nasser
Zahabi, Sayed Jalal
Wiil, Uffe Kock
Brockhattingen, Kristoffer Kittelmann
Ebrahimi, Ali
author_facet Moradbeiki, Pardis
Ghadiri, Nasser
Zahabi, Sayed Jalal
Wiil, Uffe Kock
Brockhattingen, Kristoffer Kittelmann
Ebrahimi, Ali
contents Accurate sarcopenia diagnosis via ultrasound remains challenging due to subtle imaging cues, limited labeled data, and the absence of clinical context in most models. We propose MedVQA-TREE, a multimodal framework that integrates a hierarchical image interpretation module, a gated feature-level fusion mechanism, and a novel multi-hop, multi-query retrieval strategy. The vision module includes anatomical classification, region segmentation, and graph-based spatial reasoning to capture coarse, mid-level, and fine-grained structures. A gated fusion mechanism selectively integrates visual features with textual queries, while clinical knowledge is retrieved through a UMLS-guided pipeline accessing PubMed and a sarcopenia-specific external knowledge base. MedVQA-TREE was trained and evaluated on two public MedVQA datasets (VQA-RAD and PathVQA) and a custom sarcopenia ultrasound dataset. The model achieved up to 99% diagnostic accuracy and outperformed previous state-of-the-art methods by over 10%. These results underscore the benefit of combining structured visual understanding with guided knowledge retrieval for effective AI-assisted diagnosis in sarcopenia.
format Preprint
id arxiv_https___arxiv_org_abs_2508_19319
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MedVQA-TREE: A Multimodal Reasoning and Retrieval Framework for Sarcopenia Prediction
Moradbeiki, Pardis
Ghadiri, Nasser
Zahabi, Sayed Jalal
Wiil, Uffe Kock
Brockhattingen, Kristoffer Kittelmann
Ebrahimi, Ali
Image and Video Processing
Artificial Intelligence
Computer Vision and Pattern Recognition
Accurate sarcopenia diagnosis via ultrasound remains challenging due to subtle imaging cues, limited labeled data, and the absence of clinical context in most models. We propose MedVQA-TREE, a multimodal framework that integrates a hierarchical image interpretation module, a gated feature-level fusion mechanism, and a novel multi-hop, multi-query retrieval strategy. The vision module includes anatomical classification, region segmentation, and graph-based spatial reasoning to capture coarse, mid-level, and fine-grained structures. A gated fusion mechanism selectively integrates visual features with textual queries, while clinical knowledge is retrieved through a UMLS-guided pipeline accessing PubMed and a sarcopenia-specific external knowledge base. MedVQA-TREE was trained and evaluated on two public MedVQA datasets (VQA-RAD and PathVQA) and a custom sarcopenia ultrasound dataset. The model achieved up to 99% diagnostic accuracy and outperformed previous state-of-the-art methods by over 10%. These results underscore the benefit of combining structured visual understanding with guided knowledge retrieval for effective AI-assisted diagnosis in sarcopenia.
title MedVQA-TREE: A Multimodal Reasoning and Retrieval Framework for Sarcopenia Prediction
topic Image and Video Processing
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2508.19319