# Evo-MedAgent: Beyond One-Shot Diagnosis with Agents That Remember, Reflect, and Improve Source: [Evo-MedAgent: Beyond One-Shot Diagnosis with Agents That Remember, Reflect, and Improve]() Collection index: [llms.txt]() Data updated: 2026-04-18T05:00:04.772Z Authors: Weixiang Shen, Bailiang Jian, Jun Li, Che Liu, Johannes Moll, Xiaobin Hu, Daniel Rueckert, Hongwei Bran Li, Jiazhen Pan Published: 2026-04-15T23:12:02.000Z ## Abstract Tool-augmented large language model (LLM) agents can orchestrate specialist classifiers, segmentation models, and visual question-answering modules to interpret chest X-rays. However, these agents still solve each case in isolation: they fail to accumulate experience across cases, correct recurrent reasoning mistakes, or adapt their tool-use behavior without expensive reinforcement learning. While a radiologist naturally improves with every case, current agents remain static. In this work, we propose Evo-MedAgent, a self-evolving memory module that equips a medical agent with the capacity for inter-case learning at test time. Our memory comprises three complementary stores: (1)~\\emph{Retrospective Clinical Episodes} that retrieve problem-solving experiences from similar past cases, (2)~an \\emph{Adaptive Procedural Heuristics} bank curating priority-tagged diagnostic rules that evolves via reflection, much like a physician refining their internal criteria, and (3)~a \\emph{Tool Reliability Controller} that tracks per-tool trustworthiness. On ChestAgentBench, Evo-MedAgent raises multiple-choice question (MCQ) accuracy from 0.68 to 0.79 on GPT-5-mini, and from 0.76 to 0.87 on Gemini-3 Flash. With a strong base model, evolving memory improves performance more effectively than orchestrating external tools on qualitative diagnostic tasks. Because Evo-MedAgent requires no training, its per-case overhead is bounded by one additional retrieval pass and a single reflection call, making it deployable on top of any frozen model. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Clinical Task Execution on MedAgentBench OOD v2](): Clinical Task Execution; MedAgentBench OOD v2; Accuracy: 86.7. - [Clinical Task Execution on MedAgentBench (val)](): Clinical Task Execution; MedAgentBench (val); Accuracy: 62.5. - [Clinical Task Execution on MedAgentBench (test)](): Clinical Task Execution; MedAgentBench (test); Accuracy: 67.8. - [Clinical Task Execution on MedAgentBench OOD](): Clinical Task Execution; MedAgentBench OOD; Accuracy: 31.3. - [Clinical Task Execution on MedAgentBench v2 (test)](): Clinical Task Execution; MedAgentBench v2 (test); Accuracy: 69.3. - [Clinical Task Execution on MedAgentBench v2 (val)](): Clinical Task Execution; MedAgentBench v2 (val); Accuracy: 68.8. ## Original sources and implementation [Original paper]() [PDF]()