LUMEN: Longitudinal Multi-Modal Radiology Model for Prognosis and Diagnosis

2026-02-24

AI Summary

This work addresses the challenges radiologists face in manually comparing longitudinal medical images—such as chest X-rays—for time-consuming analysis and difficult prognosis assessment. To this end, the authors propose LUMEN, a novel framework that introduces the first instruction-following dataset specifically designed for longitudinal imaging. By leveraging multi-image joint modeling and multi-task instruction tuning, LUMEN enhances large vision-language models to perform both diagnostic reasoning and prognosis prediction. The approach integrates multimodal alignment, longitudinal image modeling, and visual question answering techniques, achieving significant performance gains over existing baselines on the MIMIC-CXR and Medical-Diff-VQA benchmarks. Notably, it demonstrates strong clinical potential in prognosis-focused visual question answering tasks.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Large Vision Models
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Large language models for search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Large vision-language models (VLMs) have evolved from general-purpose applications to specialized use cases such as in the clinical domain, demonstrating potential for decision support in radiology. One promising application is assisting radiologists in decision-making by the analysis of radiology imaging data such as chest X-rays (CXR) via a visual and natural language question-answering (VQA) interface. When longitudinal imaging is available, radiologists analyze temporal changes, which are essential for accurate diagnosis and prognosis. The manual longitudinal analysis is a time-consuming process, motivating the development of a training framework that can provide prognostic capabilities. We introduce a novel training framework LUMEN, that is optimized for longitudinal CXR interpretation, leveraging multi-image and multi-task instruction fine-tuning to enhance prognostic and diagnostic performance. We conduct experiments on the publicly available MIMIC-CXR and its associated Medical-Diff-VQA datasets. We further formulate and construct a novel instruction-following dataset incorporating longitudinal studies, enabling the development of a prognostic VQA task. Our method demonstrates significant improvements over baseline models in diagnostic VQA tasks, and more importantly, shows promising potential for prognostic capabilities. These results underscore the value of well-designed, instruction-tuned VLMs in enabling more accurate and clinically meaningful radiological interpretation of longitudinal radiological imaging data.

Problem

• longitudinal imaging
• prognosis
• diagnosis
• radiology
• vision-language models

Innovation

• longitudinal imaging
• vision-language model
• instruction fine-tuning
• prognostic VQA
• multi-modal radiology