Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
CXRMate-2, a chest X-ray report generation model, was developed using structured multimodal conditioning and reinforcement learning with a composite reward to improve semantic alignment with radiologist reports. The model achieved state-of-the-art performance on MIMIC-CXR, CheXpert Plus, and ReXgradient datasets, showing significant gains over strong benchmarks like MedGemma 1.5 (4B). A blinded retrospective evaluation by radiologists found that CXRMate-2's reports were deemed acceptable in 45% of cases, with no significant preference difference for seven out of eight analyzed findings compared to radiologist reports, suggesting a pathway to clinical utility.
AI can now generate chest X-ray reports that radiologists find as good as their own nearly half the time, hinting at a near-future where AI assists in diagnosis.
Chest X-ray (CXR) radiology report generation (RRG) models have shown rapid progress, yet their clinical utility remains uncertain due to limited evaluation by radiologists. We present CXRMate-2, a state-of-the-art CXR RRG model that integrates structured multimodal conditioning and reinforcement learning with a composite reward for semantic alignment with radiologist reports. Across the MIMIC-CXR, CheXpert Plus, and ReXgradient datasets, CXRMate-2 achieves statistically significant improvements over strong benchmarks, including gains of 11.2% and 24.4% in GREEN and RadGraph-XL, respectively, on MIMIC-CXR relative to MedGemma 1.5 (4B). To directly compare CXRMate-2 against radiologist reporting, we conduct a blinded, randomised qualitative retrospective evaluation. Three consultant radiologists compare generated and radiologist reports across 120 studies from the MIMIC-CXR test set. Generated reports were deemed acceptable (defined as preferred or rated equally to radiologist reports) in 45% of ratings, with no statistically significant difference in preference rates between radiologist reports and acceptable generated reports for seven of the eight analysed findings. Preference for radiologist reports was driven primarily by higher recall, while generated reports were often preferred for readability. Together, these results suggest a credible pathway to clinically acceptable CXR RRG. Improvements in recall, alongside better detection of subtle findings (e.g., pulmonary congestion), are likely sufficient to achieve non-inferiority to radiologist reporting. With these targeted advances, CXR RRG systems may be ready for prospective evaluation in assistive roles within radiologist-led workflows.