CVPR Findings2026
DELRER: Disease Evolution-Informed Longitudinal Radiology Report Generation
Kaiyu Wang, Bing Wang, Changchun Li, You Lu, Yaning Wang, Huimao Zhang, Ximing Li
Abstract
Auto Radiology Report Generation (RRG) can effectively help relieve the workload of radiologists. Based on this, Longitudinal Radiology Report Generation (LRRG) utilizes patients' longitudinal follow-up radiology examinations to provide a comprehensive description of patients' current status. In clinical practice, radiologists often estimate a patient's current expected state based on previous radiology images and disease evolution patterns. Inspired by this clinical thought process, we propose a novel LRRG method, namely Disease Evolution-informed Longitudinal radiology Report genERation (DELRER). Given longitudinal data, our basic idea is to model the disease-specific evolution represented by radiology images, then estimate the current status of diseases as direct auxiliary signals. Experimental results show that by imitating the thought process of estimating the status of diseases, DELRER can exceed other state-of-the-art methods in diagnostic accuracy, while generating high-quality diagnostic reports.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
- PromptMRG: Diagnosis-Driven Prompts for Medical Report Generation cites Combining Automatic Labelers and Expert Annotations for Accurate Radiology Report Labeling Using BERT
- PromptMRG: Diagnosis-Driven Prompts for Medical Report Generation cites Generating Radiology Reports via Memory-driven Transformer
Builds on14
- An Image is Worth 16x16 Words: Transformers for Image Recognition at ScaleAlexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn et al.ICLR 2021 · 21,477 citations
- BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language ModelsJunnan Li, Dongxu Li, Silvio Savarese, Steven C. H. HoiICML 2023 · 7,873 citations
- Generating Radiology Reports via Memory-driven TransformerZhihong Chen, Yan Song, Tsung-Hui Chang, Xiang WanEMNLP 2020 · 552 citations
- Combining Automatic Labelers and Expert Annotations for Accurate Radiology Report Labeling Using BERTAkshay Smit, Saahil Jain, Pranav Rajpurkar, Anuj Pareek et al.EMNLP 2020 · 212 citations
- PromptMRG: Diagnosis-Driven Prompts for Medical Report GenerationHaibo Jin, Haoxuan Che, Yi Lin, Hao ChenAAAI 2024 · 168 citations
Related papers
- TIM: Temporal Decoupling with Iterative Mutual-Refinement Model for Longitudinal Radiology Report GenerationYiheng Dong, Yi Lin, Shilong Huang, Xiyan Yang et al.CVPR 2026
- BiOTPrompt: Bidirectional Optimal Transport Guided Prompting for Disease Evolution-aware Radiology Report GenerationTengfei Liu, Yijian Fan, Boyue Wang, Yongli Hu et al.CVPR 2026
- MARE: Multimodal Analogical Reasoning for Disease Evolution-Aware Radiology Report GenerationQingqing Gao, Tengfei Liu, Xiaoyan Li, Xiaodan Zhang et al.AAAI 2026
- HC-LLM: Historical-Constrained Large Language Models for Radiology Report GenerationTengfei Liu, Jiapu Wang, Yongli Hu, Mingjie Li et al.AAAI 2025 · 6 citations
- Enhanced Contrastive Learning with Multi-view Longitudinal Data for Chest X-ray Report GenerationKang Liu, Zhuoqi Ma, Xiaolu Kang, Yunan Li et al.CVPR 2025
