Med-StepBench: A Hierarchical Reasoning Framework for Evaluating Hallucinations in Medical Vision-Language Models

• URL: http://arxiv.org/abs/2605.10002v1
• Date: Mon, 11 May 2026 05:26:59 GMT
• Title: Med-StepBench: A Hierarchical Reasoning Framework for Evaluating Hallucinations in Medical Vision-Language Models
• Abstract summary: Existing medical hallucination benchmarks primarily focus on 2D imaging with one-shot diagnostic questions.<n>We introduce Med-StepBench, the first large-scale benchmark for step-wise hallucination detection in 3D oncological PET/CT.
• Score: 8.20487574753681
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large vision-language models (VLMs) demonstrate strong performance in medical image understanding, but frequently generate clinically plausible yet incorrect statements, raising significant safety concerns. Existing medical hallucination benchmarks primarily focus on 2D imaging with one-shot diagnostic questions, offering limited insight into whether predictions are grounded in correct localization and abnormality identification, allowing critical reasoning errors to remain hidden behind seemingly correct diagnoses. We introduce Med-StepBench, the first large-scale benchmark for step-wise hallucination detection in 3D oncological PET/CT, comprising over 12,000 images and more than 1,000,000 image-statement pairs across volumetric and multi-view 2D data, which decomposes clinical reasoning into four expert-designed diagnostic stages. Using clinician-verified annotations, we perform the first step-level evaluation of general-purpose and medical VLMs, revealing systematic failure modes obscured by aggregate accuracy metrics. Furthermore, we show that current VLMs are highly susceptible to adversarial yet clinically plausible intermediate explanations, which significantly amplify hallucinations despite contradictory visual evidence. Together, our findings highlight fundamental limitations in grounding multi-step clinical reasoning and establish Med-StepBench as a rigorous benchmark for developing safer and more reliable medical VLMs.

Related papers

• MedReaMM: Evaluating Large Multimodal Models on Expert-Level Clinical Diagnostic Synthesis [61.266743965495884]
  We introduce MedReaMM, a benchmark designed to evaluate models' ability to synthesize heterogeneous clinical evidence.<n>MedReaMM comprises 625 expert-validated cases with an average of 2.79 medical images per case and a total of 1,042 standardized diagnoses annotated with ICD-11 codes.<n>We evaluate 23 Large Multimodal Models (LMMs) and find that most achieve diagnostic accuracy scores below 50%, underscoring a substantial gap in multimodal diagnostic synthesis capability.
  arXiv  Detail & Related papers  (2026-08-23T09:37:53Z)
• Evaluating Multi-Turn Multimodal Diagnostic Reasoning on Challenging Real-World Clinical Cases [59.5328455571863]
  ClinMM-Bench is the largest multi-turn multimodal clinical diagnostic evaluation benchmark to date.<n>It contains 1,089 challenging real-world clinical cases and 3,760 medical images across eight specialties.<n> proprietary models achieved the highest overall diagnostic accuracy, but the proportion of completely correct diagnoses remained limited.
  arXiv  Detail & Related papers  (2026-07-28T16:19:03Z)
• RadSight: Towards Perceptually Reliable Multimodal Radiology Image Understanding [29.104092621171606]
  Perception-Bench is a large-scale benchmark comprising 1.13 million samples.<n>Our analysis on Perception-Bench reveals that existing MLLMs lack the ability to capture even the most basic lesion attributes.<n>Motivated by this, we propose RadSight, a perception-driven MLLM built upon a dual 2D/3D encoder architecture.
  arXiv  Detail & Related papers  (2026-07-24T13:35:12Z)
• MultiDx: A Multi-Source Knowledge Integration Framework towards Diagnostic Reasoning [66.94527468532843]
  We propose MultiDx, a two-stage diagnostic reasoning framework.<n>It first generates suspected diagnoses and reasoning paths by leveraging knowledge from web search, SOAP-formatted case, and clinical case database.<n>Then it integrates multi-perspective evidence through matching, voting, and differential diagnosis to generate the final prediction.
  arXiv  Detail & Related papers  (2026-04-27T08:46:29Z)
• Benchmarking Multi-turn Medical Diagnosis: Hold, Lure, and Self-Correction [72.89352076103889]
  Large language models (LLMs) achieve high accuracy in medical diagnosis when all clinical information is provided in a single turn.<n>We introduce MINT, a high-fidelity, multi-turn medical diagnosis benchmark comprising 1,035 cases.<n>We uncover three persistent behavioral patterns that significantly impact diagnostic decisions.
  arXiv  Detail & Related papers  (2026-04-06T00:23:10Z)
• MedObvious: Exposing the Medical Moravec's Paradox in VLMs via Clinical Triage [20.835664121303534]
  Vision Language Models (VLMs) are increasingly used for tasks like medical report generation and visual question answering.<n>In clinical practice, interpretation begins with pre-diagnostic sanity checks.<n>Existing benchmarks largely assume this step is solved, and therefore miss a critical failure mode.<n>We introduce MedObvious, a 1,880-task benchmark that isolates input validation as a set-level consistency capability.
  arXiv  Detail & Related papers  (2026-03-24T17:59:54Z)
• M3CoTBench: Benchmark Chain-of-Thought of MLLMs in Medical Image Understanding [66.78251988482222]
  Chain-of-Thought (CoT) reasoning has proven effective in enhancing large language models by encouraging step-by-step intermediate reasoning.<n>Current benchmarks for medical image understanding generally focus on the final answer while ignoring the reasoning path.<n>M3CoTBench aims to foster the development of transparent, trustworthy, and diagnostically accurate AI systems for healthcare.
  arXiv  Detail & Related papers  (2026-01-13T17:42:27Z)
• Benchmarking Egocentric Clinical Intent Understanding Capability for Medical Multimodal Large Language Models [48.95516224614331]
  We introduce MedGaze-Bench, the first benchmark leveraging clinician gaze as a Cognitive Cursor to assess intent understanding across surgery, emergency simulation, and diagnostic interpretation.<n>Our benchmark addresses three fundamental challenges: visual homogeneity of anatomical structures, strict temporal-causal dependencies in clinical, and implicit adherence to safety protocols.
  arXiv  Detail & Related papers  (2026-01-11T02:20:40Z)
• Simulating Viva Voce Examinations to Evaluate Clinical Reasoning in Large Language Models [51.91760712805404]
  We introduce VivaBench, a benchmark for evaluating sequential clinical reasoning in large language models (LLMs)<n>Our dataset consists of 1762 physician-curated clinical vignettes structured as interactive scenarios that simulate a (oral) examination in medical training.<n>Our analysis identified several failure modes that mirror common cognitive errors in clinical practice.
  arXiv  Detail & Related papers  (2025-10-11T16:24:35Z)
• TemMed-Bench: Evaluating Temporal Medical Image Reasoning in Vision-Language Models [54.48710348910535]
  Existing medical reasoning benchmarks primarily focus on analyzing a patient's condition based on an image from a single visit.<n>We introduce TemMed-Bench, the first benchmark designed for analyzing changes in patients' conditions between different clinical visits.
  arXiv  Detail & Related papers  (2025-09-29T17:51:26Z)
• Beyond Classification Accuracy: Neural-MedBench and the Need for Deeper Reasoning Benchmarks [21.203358914772465]
  Recent advances in vision-language models (VLMs) have achieved remarkable performance on standard medical benchmarks, yet their true clinical reasoning ability remains unclear.<n>We introduce Neural-MedBench, a compact yet reasoning-intensive benchmark designed to probe the limits of multimodal clinical reasoning in neurology.
  arXiv  Detail & Related papers  (2025-09-26T12:20:01Z)
• EH-Benchmark Ophthalmic Hallucination Benchmark and Agent-Driven Top-Down Traceable Reasoning Workflow [43.82288530883818]
  EH-Benchmark is a novel ophthalmology benchmark designed to evaluate hallucinations in Medical Large Language Models.<n>We categorize hallucinations based on specific tasks and error types into two primary classes: Visual Understanding and Logical Composition.<n>Our framework significantly mitigates both types of hallucinations, enhancing accuracy, interpretability, and reliability.
  arXiv  Detail & Related papers  (2025-07-24T12:07:36Z)
• DrVD-Bench: Do Vision-Language Models Reason Like Human Doctors in Medical Image Diagnosis? [1.1094764204428438]
  We propose DrVD-Bench, the first benchmark for clinical visual reasoning.<n>DrVD-Bench consists of three modules: Visual Evidence, Reasoning Trajectory Assessment, and Report Generation Evaluation.<n>Our benchmark covers 20 task types, 17 diagnostic categories, and five imaging modalities-CT, MRI, ultrasound, radiography, and pathology.
  arXiv  Detail & Related papers  (2025-05-30T03:33:25Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.