QureRadEmbed improves radiology report comparison and search accuracy
QureRadEmbed: Structuring Radiological Similarity through Attribute and Reasoning Supervision
Artificial Intelligence
Summary
Radiology reports have many details that affect how similar diseases look, like which lung lobe is affected or how severe the disease is. The authors created QureRadEmbed, a tool that better understands these fine details by learning from both structured similarity scores and reasoning-based descriptions. This lets it compare radiology reports more accurately and quickly than previous methods. It also helps to find important information in reports even with very few labeled examples.
What this means in practice
- •For radiology imaging teams: Enhance radiology report search engines with more accurate and attribute-aware similarity comparisons.
- •For medical data indexing teams: Support structured report indexing and retrieval using embeddings that capture fine-grained radiological details.
Authors
Janhavi Prabhu, Sahil, Shivam Ashok Shukla, Manoj Tadepalli
Abstract
Radiological similarity depends on disease relationships and on fine details such as laterality, lobe, severity, size, and certainty. Broad biomedical similarity can overlook these qualifiers, particularly when several attributes vary together. We introduce QureRadEmbed, a 4B radiology-aware encoder trained with two complementary signals: RadSim supplies deterministic, attribute-decomposed ranking targets, while RadThought aligns reports with hierarchical evidence and reasoning descriptions. A three-stage curriculum combines these signals with report triplets, finding perturbations, and single- and cross-attribute contrasts. The final model achieves 0.996 mean ordering accuracy across ten controlled synthetic attributes and raises Spearman correlation with the designed joint-attribute targets from 0.501 to 0.976. On external findings-to-impression retrieval, Recall@1 reaches 10.5% on Open-I, 12.4% on testing XR, and 42.9% on testing CT, compared with 6.6%, 5.7%, and 31.4% for its backbone. Frozen embeddings support finding extraction with only 100 labeled testing-XR reports (macro-F1 0.481 versus 0.412 for the backbone). Whole-report comparison costs 8.8 seconds per 1,000 pairs in our benchmark, versus 2,755.1 seconds for the generative evaluator GREEN. Sentence-level comparison improves sensitivity to local discrepancies, although generative evaluation remains stronger on several expert-rated and subtle-error tasks. The results support reusable radiology-aware representations for search, structured report indexing, and efficient report comparison.