{"total":23,"items":[{"citing_arxiv_id":"2607.05880","ref_index":1,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Harrison.Rad 1.5 Technical Report: A radiology foundation model that can draft reports from images, priors and clinical context","primary_cat":"cs.CV","submitted_at":"2026-07-07T06:23:08+00:00","verdict":"CONDITIONAL","verdict_confidence":"MODERATE","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Harrison.Rad 1.5 is a radiology-specific multimodal LLM that passes simulated FRCR 2B Short Case examinations and outperforms general-purpose frontier models on plain-film radiography reporting tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.21290","ref_index":5,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"NoduLoCC2026: Lung Nodule Localization and Classification Contest from Chest X-Ray Images","primary_cat":"cs.CV","submitted_at":"2026-06-19T10:13:38+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"A new contest provides a chest X-ray dataset for lung nodule classification and localization; the best entry reaches 0.72 balanced accuracy and 0.79 AUC on classification but only predicts the correct nodule count on 53% of external test images with 12.83 mm median error.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.21020","ref_index":7,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"CheXpercept: A Benchmark for Evaluating Expert-Level Lesion Perception in Chest X-rays","primary_cat":"cs.CV","submitted_at":"2026-06-19T01:10:24+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"CheXpercept is a sequential multi-level perception benchmark showing VLMs perform adequately only on coarse lesion detection in chest X-rays while degrading sharply on finer tasks, with medical VLMs offering no advantage over general models.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.12590","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Analyzing and Improving Fine-grained Preference Optimization in Medical LVLMs","primary_cat":"cs.CV","submitted_at":"2026-06-10T18:35:36+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"Proposes bidirectional token-wise KL regularizer and visual-contrastive grounding objective to create fine-grained on-policy preference pairs for medical LVLMs by minimally editing model outputs.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.06407","ref_index":8,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"A Vision-language Framework for Comparative Reasoning in Radiology","primary_cat":"cs.CV","submitted_at":"2026-06-04T17:12:47+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Introduces MedReCo-DB dataset of 690k+ images and entity-aware models MedReCo/MedReCo-VLM that improve reference retrieval and comparative change interpretation in radiology across multiple centers and modalities.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.26691","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Mind the Tool Failures: Achieving Synergistic Tool Gains for Medical Agents","primary_cat":"cs.AI","submitted_at":"2026-05-26T08:29:38+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"A GRPO-based RL framework with probabilistic risk minimization, disagreement-aware synergy rewards, and entropy-guided sampling enables instance-level tool selection that closes the single-oracle risk gap on medical benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.23629","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"DDX-TRACE: A Benchmark for Medical Diagnostic Trajectories in VLMs","primary_cat":"cs.CV","submitted_at":"2026-05-22T13:41:10+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"DDX-TRACE is a physician-adjudicated benchmark for evaluating VLMs on evidence-supported diagnostic trajectories rather than final answers alone in multimodal neuroradiology.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.20469","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"HalluCXR: Benchmarking and Mitigating Hallucinations in Medical Vision-Language Models for Chest Radiograph Interpretation","primary_cat":"cs.CV","submitted_at":"2026-05-19T20:30:32+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"HalluCXR benchmark shows 61.9-82.3% hallucination rates across VLMs on MIMIC-CXR images, identifies patterns such as length-based risk and over-fabrication of common findings, and demonstrates ensemble mitigation that cuts fabrication by up to 84.8%.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.05810","ref_index":7,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"CXR-ContraBench: Benchmarking Negated-Option Attraction in Medical VLMs","primary_cat":"cs.CV","submitted_at":"2026-05-07T07:46:17+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Medical VLMs frequently select negated options that contradict visible chest X-ray findings, achieving only ~30% accuracy on direct presence probes, but a post-hoc consistency verifier raises accuracy above 95%.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"is different, targeting answer-space polarity confusion in medical VLMs through benchmark construction and deterministic verification rather than distributional decoding adaptation. Recent medical VLMs and report-oriented resources motivate the relevance of this failure mode but do 3/42 CXR-ContraBench not isolate it. LLaVA-Med [16], MedGemma [28], CheXagent [7], Med-Flamingo [23], BiomedCLIP [43], MedKLIP [37], RadFM [38], XrayGPT [30], RaDialog [25], and R2GenGPT [35] improve biomedical vision- language capability, while ReXrank [45], CoCa-CXR [6], CorBenchX [48], M3D [2], and DeepTumorVQA [5] broaden evaluation across report generation, error correction, or 3D diagnosis. Our contribution is complementary: we make both presence-side semantic reversal and absence-side contradiction measurable"},{"citing_arxiv_id":"2604.22989","ref_index":8,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"CheXmix: Unified Generative Pretraining for Vision Language Models in Medical Imaging","primary_cat":"cs.CV","submitted_at":"2026-04-24T20:03:04+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"CheXmix combines masked autoencoder pretraining with early-fusion generative modeling to outperform prior models on chest X-ray classification by up to 8.6% AUROC, inpainting by 51%, and report generation by 45% on GREEN.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.18967","ref_index":65,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"CXRMate-2: Structured Multimodal Temporal Embeddings and Tractable Reinforcement Learning for Clinically Acceptable Chest X-ray Radiology Report Generation","primary_cat":"cs.CV","submitted_at":"2026-04-21T01:30:51+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"CXRMate-2 improves chest X-ray report generation via temporal embeddings and tractable RL, delivering metric gains and 45% acceptability in radiologist review with no significant preference difference on most findings.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.18250","ref_index":7,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Medical Image Understanding Improves Survival Prediction via Visual Instruction Tuning","primary_cat":"cs.CV","submitted_at":"2026-04-20T13:27:39+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"A vision-language model pre-trained via instruction tuning on CT-report pairs improves survival prediction accuracy over baselines, especially when clinical data alone is weak, while also producing text answers to clinical questions.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.13598","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Enhancing Reinforcement Learning for Radiology Report Generation with Evidence-aware Rewards and Self-correcting Preference Learning","primary_cat":"cs.LG","submitted_at":"2026-04-15T08:08:06+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"ESC-RL improves RL for radiology reports via group-wise evidence-aware rewards (GEAR) and LLM-driven self-correcting preference learning (SPL), reaching state-of-the-art on two chest X-ray datasets.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"trained on large-scale image-report pairs (e.g., MA VL (Phan et al., 2024)) to extract disease- grounded response maps (DRMs). Specifically, we obtain predicted DRMs Mpred conditioned on I and the generated refined report R, and ground- truth DRMs Mgt conditioned on I and the ground- truth reportR ∗: Mpred =ψ vl(I, R)∈R H×W×K , Mgt =ψ vl(I, R∗)∈R H×W×K , (6) where K is the number of diseases, and (H, W) is the spatial resolution of the response map. Evidence-aware Alignment Reward.For group GTP, we enforce spatial consistency between Mpred and Mgt using an IoU-based loss, which promotes consistent spatial coverage between pre- dicted and ground-truth evidence maps, LTP =1− 1 |GTP| X k∈GTP 2P h,w Mpred"},{"citing_arxiv_id":"2604.09450","ref_index":7,"ref_count":2,"confidence":0.9,"is_internal_anchor":false,"paper_title":"ECHO: Efficient Chest X-ray Report Generation with One-step Block Diffusion","primary_cat":"cs.LG","submitted_at":"2026-04-10T16:07:14+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"ECHO introduces one-step block diffusion via Direct Conditional Distillation and Response-Asymmetric Diffusion to generate chest X-ray reports faster than autoregressive models while improving clinical metrics.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"In recent years, Vision-Language Models (VLMs) [3, 19, 24, 26, 31, 42, 46], where visual features are aligned with language instructions to enable complex cross-modal understanding, have demonstrated significant progress in many fields, notably medical image analysis [6, 21, 25, 34, 37, 39, 44, 52, 59]. Within this field, automated chest X-ray report generation (CXR-RG) [7, 23, 35, 43, 48, 54] has emerged as a critical application. As one of the most common clinical imaging exams, CXR's high volume places a heavy diag- nostic burden on radiologists, creating strong demand for high-throughput automated reporting systems to ease workloads. Despite the promising performance achieved, canonical autoregressive (AR) VLMs often"},{"citing_arxiv_id":"2604.13060","ref_index":1,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Dental-TriageBench: Benchmarking Multimodal Reasoning for Hierarchical Dental Triage","primary_cat":"cs.CL","submitted_at":"2026-03-18T07:43:49+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":8.0,"formal_verification":"none","one_line_summary":"Dental-TriageBench is the first expert-annotated multimodal benchmark for hierarchical dental triage and shows a substantial performance gap between 19 MLLMs and junior dentists, especially on multi-domain referral cases.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2511.15825","ref_index":14,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"IMACT-CXR: An Interactive Multi-Agent Conversational Tutoring System for Chest X-Ray Interpretation","primary_cat":"cs.AI","submitted_at":"2025-11-19T19:32:29+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"IMACT-CXR presents an integrated multi-agent system using AutoGen, Bayesian Knowledge Tracing, gaze feedback, and vision-language models to provide interactive tutoring for chest X-ray interpretation with preliminary evidence of improved learner performance.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2510.07191","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Resolution scaling governs DINOv3 transfer performance in chest radiograph classification","primary_cat":"cs.CV","submitted_at":"2025-10-08T16:25:04+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"DINOv3 at 512x512 resolution with ConvNeXt-B outperforms prior initializations for adult chest X-ray classification but shows no benefit in pediatric cohorts or at 1024 resolution.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2509.20490","ref_index":4,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"RadAgents: Multimodal Agentic Reasoning for Chest X-ray Interpretation with Radiologist-like Workflows","primary_cat":"cs.MA","submitted_at":"2025-09-24T19:08:01+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"RadAgents is a multi-agent framework coupling clinical priors with task-aware multimodal reasoning and radiologist-like workflows, plus grounding and retrieval-augmentation for conflict resolution in chest X-ray interpretation.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2506.09067","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Enhancing the Safety of Medical Vision-Language Models by Synthetic Demonstrations","primary_cat":"cs.CV","submitted_at":"2025-06-08T16:26:51+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Synthetic clinical demonstrations at inference time improve safety of Med-VLMs against visual and textual jailbreaks while preserving general performance on medical tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2504.07415","ref_index":7,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"RA-RRG: Multimodal Retrieval-Augmented Radiology Report Generation with Key Phrase Extraction","primary_cat":"cs.CV","submitted_at":"2025-04-10T03:14:01+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"RA-RRG extracts key phrases with LLMs, retrieves them via multimodal similarity, and conditions report generation on them to achieve SOTA CheXbert scores and competitive RadGraph F1 on MIMIC-CXR and IU X-ray while supporting multi-view inputs.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2408.16213","ref_index":12,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"M4CXR: Exploring Multi-task Potentials of Multi-modal Large Language Models for Chest X-ray Interpretation","primary_cat":"cs.CV","submitted_at":"2024-08-29T02:12:58+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"M4CXR is a multi-modal large language model that performs multiple tasks in chest X-ray analysis including report generation with claimed SOTA clinical accuracy using chain-of-thought prompting.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2401.02458","ref_index":51,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Data-Centric Foundation Models in Computational Healthcare: A Survey","primary_cat":"cs.LG","submitted_at":"2024-01-04T08:00:32+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":3.0,"formal_verification":"none","one_line_summary":"The paper surveys data-centric strategies for foundation models in computational healthcare and supplies a curated list of related models and datasets.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"performing and well-aligned FM can be employed as a benchmark or an evaluation tool to au- tomatically assess other FMs, eliminating the need of static benchmarks or human experts. For instance, Chen et al. [45] explore reference-free evaluation methods that prompt ChatGPT to score the quality of model-generated texts without a pre-defined ground truth. Chiang et al. [51] show that LLM evaluation can be stable over prompt formatting and consistent with human experts. Jain et al. [120] propose self-supervised evaluation strategies to assess LLM properties without benchmarks or human annotations. Ye et al. propose FLASK [329], a fine-grained LLM evaluation Data-Centric Foundation Models in Computational Healthcare: A Survey 21"},{"citing_arxiv_id":"2305.10415","ref_index":12,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"PMC-VQA: Visual Instruction Tuning for Medical Visual Question Answering","primary_cat":"cs.CV","submitted_at":"2023-05-17T17:50:16+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"PMC-VQA dataset and MedVInT model achieve better generative performance on medical VQA benchmarks by visual instruction tuning on a newly constructed large-scale dataset.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null}],"limit":50,"offset":0}