PRIMA: Pre-training with Risk-integrated Image-Metadata Alignment for Medical Diagnosis via LLM

TL;DR

PRIMA introduces a novel multi-modal learning framework that integrates clinical knowledge and image features for improved medical diagnosis, leveraging risk-disease correlations and advanced pre-training strategies.

Contribution

It presents a new pre-training approach combining domain-specific knowledge with multi-modal alignment, enhancing diagnostic accuracy without extensive data or computational demands.

Findings

Outperforms state-of-the-art methods in disease classification

Achieves robust performance with limited data

Effectively aligns clinical knowledge with visual features

Abstract

Medical diagnosis requires the effective synthesis of visual manifestations and clinical metadata. However, existing methods often treat metadata as isolated tags, failing to exploit the rich semantic knowledge embedded in clinical descriptions. We propose PRIMA (Pre-training with Risk-integrated Image-Metadata Alignment), a framework that integrates domain-specific knowledge into multi-modal representation learning. We first curate an expert corpus of risk-disease correlations via Retrieval-Augmented Generation (RAG) to refine Clinical ModernBERT, embedding diagnostic priors into the text encoder. To bridge the modality gap, we introduce a dual-encoder pre-training strategy utilizing DINOv3 and our refined BERT, optimized by a suite of four complementary loss functions. These losses are designed to capture multi-granular semantic alignment and handle the ambiguity of clinical correlations through soft labels. Finally, we leverage Qwen-3 to fuse these aligned features for precise disease classification. Extensive experiments demonstrate that PRIMA effectively harmonizes pixel-level features with abstract clinical expertise, significantly outperforming other state-of-the-art methods. Notably, our framework achieves superior robustness without the need for massive data collection or exhaustive computational resources. Our code will be made public upon acceptance.