A framework that strengthens clinical alignment by separating and enhancing training-time alignment and inference-time alignment is proposed and experiments show that complementary visual representations with a multi-encoder design and concept-level auxiliary learning help preserve clinically meaningful information.
Abstract
Medical image captioning is a technique that accelerates early-stage diagnostic workflows and enhances the interpretability of medical diagnostic AI systems. However, unlike general image captioning, clinically reliable captioning remains challenging due to grayscale-based modalities, subtle anatomical cues, specialized medical phrasing, and variations in data quality. Despite recent advances in large vision-language models, fluent outputs do not necessarily guarantee sufficient alignment with clinical concept spaces or evaluation criteria. To address this issue, we propose a framework that strengthens clinical alignment by separating and enhancing training-time alignment and inference-time alignment. We build a medical image captioning pipeline that integrates single/dual vision encoders based on BioMedCLIP and SigLIP2, a Q-Former, and a LLaMA-based decoder, and examine the contribution of auxiliary learning for UMLS concept/type prediction. At inference, we apply single-embedding-based reranking to select the best caption among candidates, while at training we introduce MedPAIR-SCST, which combines clinically relevant rewards to shift the generative distribution toward improved clinical alignment. Our experiments show that complementary visual representations with a multi-encoder design and concept-level auxiliary learning help preserve clinically meaningful information. Furthermore, inference-time reranking provides a practical way to improve semantic and clinical alignment without additional training, whereas MedPAIR-SCST goes beyond selection by directly improving the model's distribution to generate more consistent and clinically grounded captions. These findings suggest that jointly leveraging selection-based alignment and reinforcement-learning-based alignment can promote more trustworthy medical image captioning even in data-constrained settings.
In order to produce meaningful textual interpretations of intricate clinical images, medical image captioning has become a significant field of study at the nexus of computer vision and natural language processing. Due to the lack of explicit modeling of medical entities, current methods frequently fail to produce descriptions that are both semantically valid and clinically useful, despite notable advances in deep learning and vision language modeling. Typical captioning methods in particular, fall short of being able to retrieve fine grained diagnostic information and maintain semantic consistency with clinical findings as they focus on global features. This paper addresses these limitations by presenting an entity-aware medical image captioning approach which aims to identify and incorporate clinically relevant entities into the caption generation, including but not limited to, diagnostic finding, anatomical structures, or diagnostic characteristics. The proposed method utilizes entity level representations as a means of guiding the captioning process thereby ensuring a tighter semantic consistency between visual modalities and the resultant textual output. Consequently, this leads to more comprehensible, informative and clinically relevant generated reports. Additionally, inclusion of entity awareness can aid the model in effectively understanding the relationships between medical concepts leading to captions more consistent with medical expertise. The results demonstrate that explicit modeling with structured semantic information within vision-language frameworks are crucial and that entity-aware methods have the potential to greatly improve captioning. This work has the ability to help advance health intelligence applications which will serve to better assist clinical decision making, scale the processing of medical images, and facilitate accurate medical documentation.
Shaik Rafi, Syed Rizwana, P. Drutika et al.· IEEE Access· 0 citations
Multimodal large language models (MLLMs) hold immense potential to revolutionize clinical practice, yet deploying them in the medical domain is fundamentally a vision-centric challenge: models must absorb knowledge from heterogeneous 2D and 3D medical images, and evaluation protocols must align with radiologists'clinical practice and provide an accurate, fine-grained and factualness-driven assessment. In this paper, we introduce ClinFusion, a vision-centric MLLM designed for holistic medical understanding that systematically addresses these limitations. We propose a compositional and cascaded vision encoder architecture featuring a Cascade Spatial-Aware Locality Fusion operator that unifies diverse 2D and native 3D medical image understanding within a fused encoder. We further introduce a vision-grounded evaluation framework, including MedIF-Bench for instruction-following assessment and a region-of-interest-grounded method for clinically aligned and factualness-driven report generation evaluation. We show that ClinFusion sets a new state-of-the-art across a comprehensive suite of 2D and 3D multimodal medical benchmarks---spanning visual question answering, report generation, and instruction following---as well as textual medical tasks, outperforming leading open-source medical MLLMs (\textit{e.g.}, Hulu-Med, Lingshu) on 20 out of 24 benchmarks and demonstrating multimodal capabilities better than powerful proprietary models such as GPT-5.2 and Gemini-3-Flash on 13 out of 16 benchmarks, and can be further augmented with agentic tool use for retrieval-augmented and tool-assisted clinical workflows. A blinded evaluation by board-certified radiologists confirms that ClinFusion produces the highest-ranked reports, and validates our RoI-grounded metric as achieving the strongest correlation with expert judgment among all automatic evaluation metrics examined.
Hangjie Yuan, Yichen Qian, Zhiwei Tang et al.· 0 citations
LocAnyMed-CoT-20K is derived, a rationale-augmented subset that connects anatomical context, visual observations, and spatial conclusions through structured reasoning and further improves cross-source generalization through fine-tuning.
Zi-hao Wang, Tong Liu, Zhiwei Wang et al.· 0 citations
A novel multimodal RAG framework tailored for MedVQA is proposed, which leverages multimodal data, including medical images, reports, and generated captions, to provide more accurate clinical answers, and introduces a training paradigm that uses captions as auxiliary supervision, enhancing cross-modal alignment via contrastive learning.
Mai A. Shaaban, M. Zarei, Adnan Khan et al.· 0 citations
Model editing promises a fast, targeted way to correct post-deployment mistakes in medical vision-language models (VLMs) without costly retraining. However, existing multimodal model editing benchmarks focus on general-purpose tasks and do not reflect realistic clinical domain requirements and variability. To address this, we introduce M3Bench, a clinically grounded benchmark for multimodal model editing that evaluates whether an edit remains reliable, precise, and generalizable under the challenges of image and text variation, modality and protocol shifts, clinical knowledge composition, and temporal progression. M3Bench contains 16,276 questions spanning diverse anatomy, modalities, and specialties, and supports both single and sequential edits. By evaluating 4 representative editors across 6 medical and general VLMs, we find that no method excels across all criteria. Gradient-based editors achieve strong transfer but suffer from catastrophic locality violations, whereas memory-based methods preserve locality but lack compositional generality and exhibit high backbone-dependent hyperparameter sensitivity. We further attribute these failures to the latent space geometry of VLMs and how different editing methods shift its landscape. Overall, M3Bench establishes a rigorous clinical stress test for multimodal model editing and offers actionable guidance for safer post-deployment adaptation. The benchmark is publicly available at https://github.com/BioMed-AI-Lab-U-Michgan/M3Bench .