MedTVT-R1: A Multimodal LLM Empowering Medical Reasoning and Diagnosis
Yuting Zhang, Kaishen Yuan, Hao Lu, Yutao Yue, Jintai Chen, Kaishun Wu
Abstract
Hematocrit Platelet Count White Blood Cells pCO2 35.4 206 7.2 130 Electrocardiogram (ECG) Chest X-Ray Lab Test Red Blood Cells Phosphate Basophils pH 4.11 3.6 0.4 7.42 <think>From the integration of ECG, CXR, and blood test findings, several confirmed conditions can be inferred: Acute Renal Failure: Blood test results show severe hyperkalemia and elevated creatine kinase, which suggest rhabdomyolysis contributing to metabolic derangements. The metabolic acidosis, high anion gap, and lactate elevation indicate hypoperfusion and critical illness, further stressing renal function. Coronary Artery Disease: The ECG demonstrates acute ST elevation myocardial infarction (MI) with anterolateral ST elevations, reflecting significant coronary artery blockage. The presence of an acute infarct confirms ischemic damage due to blood flow impairment. Elevated liver enzymes, related to ischemia, reinforce cardiac involvement. Elevated lactate corroborates tissue hypoxia from compromised cardiac output. Hypertension: While no specific blood pressure readings are provided, existing evidence supports systemic effects of hypertension. Hyponatremia and organ dysfunction imply chronic vascular strain, and mediastinal widening on CXR, possibly linked with an acute vascular event like dissection, suggests longstanding hypertensive damage. Sepsis: Mild leukocytosis from the blood test indicates systemic inflammation, potentially due to infection. The CXR findings of right lung field opacity suggest pleural effusion or infection likely contributing to sepsis. Severe acidemia and elevated lactate further suggest tissue hypoxia and metabolic derangement typical of sepsis. Coagulopathy with prolonged PTT is consistent with disseminated intravascular coagulation, a complication of sepsis. In conclusion, acute renal failure, coronary artery disease, hypertension, and sepsis are supported by the diagnostic integration of ECG evidence of myocardial infarction, CXR findings indicating pleural effusion or vascular complications, and blood test abnormalities suggestive of multiorgan dysfunction. Immediate intervention is warranted for these confirmed conditions.</think> <answer>Acute Renal Failure; Coronary Artery Disease; Hypertension; Sepsis</answer>
Figure 1: Overview of MedTVT-R1: MedTVT-R1 seamlessly integrates Electrocardiogram (Time Series), Chest X-ray (Visual Image), and Blood Test (Tabular Data) to deliver comprehensive long-text Medical reasoning and diagnosis across various diseases.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext 1c441a5b-8440-40bc-800e-d8ad84e2a0e4Cited by top-tier papers2
- Beyond N-grams: A Hierarchical Reward Learning Framework for Clinically-Aware Medical Report GenerationYuan Wang, Shujian Gao, Jiaxiang Liu, Songtao Jiang et al.AAAI 2026 · 2 citations
- CMR-RD: Long-Tailed Adaptive VLM for Explainable CMR DiagnosisYansong Li, Zhongxi Qiu, Yun Tian, Zheng jinyu et al.CVPR 2026
Builds on15
- 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
- LoRA: Low-Rank Adaptation of Large Language ModelsEdward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu et al.ICLR 2022 · 18,833 citations
- Visual Instruction TuningHaotian Liu, Chunyuan Li, Qingyang Wu, Yong Jae LeeNeurIPS 2023 · 11,349 citations
- NExT-GPT: Any-to-Any Multimodal LLMShengqiong Wu, Hao Fei, Leigang Qu, Wei Ji et al.ICML 2024 · 786 citations
- Visual-RFT: Visual Reinforcement Fine-TuningZiyu Liu, Zeyi Sun, Yuhang Zang, Xiaoyi Dong et al.ICCV 2025 · 563 citations
Related papers
- The Impact of Auxiliary Patient Data on Automated Chest X-Ray Report Generation and How to Incorporate ItAaron Nicolson, Shengyao Zhuang, Jason Dowling, Bevan KoopmanACL 2025 · 6 citations
- ECG-R1: Protocol-Guided and Modality-Agnostic MLLM for Reliable ECG InterpretationJiarui Jin, Haoyu Wang, Xingliang Wu, Xiaocheng Fang et al.ICML 2026 · 11 citations
- Tracing the Heart's Pathways: ECG Representation Learning from a Cardiac Conduction PerspectiveTan Pan, Yixuan Sun, Chen Jiang, Qiong Gao et al.AAAI 2026
- READ: Reasoning Like a Cardiologist for Contactless Arrhythmia DiagnosisRui Lyu, Hao Feng, Anfu Zhou, Huadong Ma et al.UbiComp 2026
- MedM2G: Unifying Medical Multi-Modal Generation via Cross-Guided Diffusion with Visual InvariantChenlu Zhan, Yu Lin, Gaoang Wang, Hongwei Wang et al.CVPR 2024 · 20 citations
