{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T08:20:44Z","timestamp":1788337244399,"version":"build-2803163510"},"reference-count":40,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T00:00:00Z","timestamp":1788307200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Digit. Health"],"abstract":"<jats:p>Recent advances in large language models (LLMs) and vision transformers have enabled multimodal systems that integrate clinical text with medical imaging for diagnostic decision-making. While these systems show promising results on benchmark datasets in well-resourced research settings, their applicability in low-resource healthcare environments where diagnostic disparities are most severe remains limited and poorly understood. This mini review synthesizes key developments in LLM\u2013vision fusion architectures from 2018 to 2026, with a focus on radiology-oriented visual question answering (VQA) and report generation systems viewed from a deployment perspective. Rather than comprehensively cataloguing multimodal medical AI, we synthesize the evolution of LLM\u2013vision fusion architectures and discuss complementary deployment-enabling strategies, including parameter-efficient adaptation, post-training quantization, federated learning, and multilingual support, where they directly improve the feasibility of radiology AI in resource-constrained healthcare settings. Rather than focusing solely on performance benchmarks, we examine these approaches through a deployment-oriented lens, highlighting trade-offs between representational capacity, computational efficiency, interpretability, and memory footprint. We argue that current progress remains substantially shaped by model scaling and benchmark optimization, which often do not address the memory, connectivity, and annotation constraints of low-resource healthcare systems. While cross-modal transformer architectures provide strong representational alignment, their computational demands and reliance on large curated datasets limit real-world deployment. In contrast, emerging directions including parameter-efficient fine-tuning, post-training quantization, federated learning, and modular agent-based systems offer more tractable pathways toward clinical integration under hardware and data constraints. To bridge the gap between benchmark performance and clinical utility, we identify concrete challenges in data scarcity, multilingual coverage, and calibration, and propose a shift toward lightweight, interpretable, and hardware-aware multimodal AI. This perspective highlights the need to move beyond scaling-centric design toward models that can run on 4\u20138\u2009GB VRAM, operate offline, and generalize across languages and imaging equipment.<\/jats:p>","DOI":"10.3389\/fdgth.2026.1862236","type":"journal-article","created":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T07:57:21Z","timestamp":1788335841000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Multimodal medical diagnosis: a mini review of LLM\u2013vision fusion models in low-resource healthcare settings"],"prefix":"10.3389","volume":"8","author":[{"given":"Kahakashan","family":"Ashraf","sequence":"first","affiliation":[{"name":"Department of Computer Science and Engineering, Chittagong University of Engineering and Technology","place":["Chattogram, Bangladesh"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Md. Hamid","family":"Hosen","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, East Delta University","place":["Chattogram, Bangladesh"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nuzhat Tabassum","family":"Farah","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, East Delta University","place":["Chattogram, Bangladesh"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Md Kishor","family":"Morol","sequence":"additional","affiliation":[{"name":"ELITE Research Lab, LLC","place":["Queens, NY, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tze Hui","family":"Liew","sequence":"additional","affiliation":[{"name":"Faculty of Information Science and Technology, Multimedia University","place":["Bukit Beruang, Melaka, Malaysia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dip","family":"Nandi","sequence":"additional","affiliation":[{"name":"ELITE Research Lab, LLC","place":["Queens, NY, United States"]},{"name":"Department of Computer Science and Engineering, American International University-Bangladesh","place":["Dhaka, Bangladesh"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mashiour","family":"Rahman","sequence":"additional","affiliation":[{"name":"ELITE Research Lab, LLC","place":["Queens, NY, United States"]},{"name":"Department of Computer Science and Engineering, American International University-Bangladesh","place":["Dhaka, Bangladesh"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Abdullah","family":"Al Jubair","sequence":"additional","affiliation":[{"name":"ELITE Research Lab, LLC","place":["Queens, NY, United States"]},{"name":"Department of Computer Science and Engineering, American International University-Bangladesh","place":["Dhaka, Bangladesh"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,2]]},"reference":[{"key":"B1","year":""},{"key":"B2","doi-asserted-by":"publisher","first-page":"20552076241281626","DOI":"10.1177\/20552076241281626","article-title":"The impact of limited access to digital health records on doctors and their willingness to adopt electronic health record systems","volume":"10","author":"Bouh","year":"2024","journal-title":"Digit Health"},{"key":"B3","doi-asserted-by":"publisher","first-page":"853","DOI":"10.1093\/inthealth\/ihae073","article-title":"Large multimodal models: boon or burden for low- and middle-income countries","volume":"17","author":"Malpani","year":"2025","journal-title":"Int Health"},{"key":"B4","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1016\/j.media.2017.07.005","article-title":"A survey on deep learning in medical image analysis","volume":"42","author":"Litjens","year":"2017","journal-title":"Med Image Anal"},{"key":"B5","author":"Rajpurkar","year":""},{"key":"B6","doi-asserted-by":"publisher","first-page":"1234","DOI":"10.1093\/bioinformatics\/btz682","article-title":"BioBERT: a pre-trained biomedical language representation model for biomedical text mining","volume":"36","author":"Lee","year":"2020","journal-title":"Bioinformatics"},{"key":"B7","author":"Alsentzer","year":""},{"key":"B8","doi-asserted-by":"publisher","first-page":"136","DOI":"10.1038\/s41746-020-00341-z","article-title":"Fusion of medical imaging and electronic health records using deep learning: a systematic review and implementation guidelines","volume":"3","author":"Huang","year":"2020","journal-title":"NPJ Digit Med"},{"key":"B9","author":"Ben Abacha","year":""},{"key":"B10","author":"Zhang","year":""},{"key":"B11","doi-asserted-by":"publisher","first-page":"172","DOI":"10.1038\/s41586-023-06291-2","article-title":"Large language models encode clinical knowledge","volume":"620","author":"Singhal","year":"2023","journal-title":"Nature"},{"key":"B12","doi-asserted-by":"publisher","first-page":"180251","DOI":"10.1038\/sdata.2018.251","article-title":"A dataset of clinically generated visual questions and answers about radiology images","volume":"5","author":"Lau","year":"2018","journal-title":"Sci Data"},{"key":"B13","author":"Kovaleva","year":""},{"key":"B14","doi-asserted-by":"publisher","first-page":"19826","DOI":"10.1038\/s41598-021-98390-1","article-title":"MedFuseNet: an attention-based multimodal deep learning model for visual question answering in the medical domain","volume":"11","author":"Sharma","year":"2021","journal-title":"Sci Rep"},{"key":"B15","doi-asserted-by":"publisher","first-page":"50626","DOI":"10.1109\/ACCESS.2020.2980024","article-title":"CGMVQA: a new classification and generative model for medical visual question answering","volume":"8","author":"Ren","year":"2020","journal-title":"IEEE Access"},{"key":"B16","author":"Wang","year":""},{"key":"B17","first-page":"28541","article-title":"LLaVA-Med: training a large language-and-vision assistant for biomedicine in one day","volume-title":"Advances in Neural Information Processing Systems","author":"Li","year":"2023"},{"key":"B18","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1186\/s40537-024-01054-w","article-title":"Transformer-enabled multi-modal medical diagnosis for tuberculosis classification","volume":"12","author":"Kumar","year":"2025","journal-title":"J Big Data"},{"key":"B19","doi-asserted-by":"publisher","first-page":"10704","DOI":"10.1038\/s41598-025-95666-8","article-title":"A vision attention driven language framework for medical report generation","volume":"15","author":"Varol Ar\u0131soy","year":"2025","journal-title":"Sci Rep"},{"key":"B20","author":"Chen","year":""},{"key":"B21","doi-asserted-by":"publisher","first-page":"e70220","DOI":"10.1002\/eng2.70220","article-title":"Vision-language models for automated chest X-ray interpretation: leveraging ViT and GPT-2","volume":"7","author":"Islam","year":"2025","journal-title":"Eng Rep"},{"key":"B22","author":"Moor","year":""},{"key":"B23","author":"Zhou","year":""},{"key":"B24","author":"Liu","year":""},{"key":"B25","doi-asserted-by":"publisher","first-page":"9","DOI":"10.1038\/s41746-025-01362-4","article-title":"BiomedCLIP: a multimodal biomedical foundation model pretrained from fifteen million scientific image\u2013text pairs","volume":"8","author":"Zhang","year":"2025","journal-title":"NPJ Digit Med"},{"key":"B26","author":"Bannur","year":""},{"key":"B27","author":"Tu","year":""},{"key":"B28","author":"Wu","year":""},{"key":"B29","author":"Chen","year":""},{"key":"B30","first-page":"1","article-title":"Privacy-preserving chest X-ray report generation via multimodal federated learning with ViT and GPT-2","author":"Hossain","year":"2025","journal-title":"Biomed Mater Devices"},{"key":"B31","doi-asserted-by":"publisher","first-page":"119","DOI":"10.1038\/s41746-020-00323-1","article-title":"The future of digital health with federated learning","volume":"3","author":"Rieke","year":"2020","journal-title":"NPJ Digit Med"},{"key":"B32","author":"Houlsby","year":""},{"key":"B33","author":"Hu","year":""},{"key":"B34","author":"El Mir","year":""},{"key":"B35","doi-asserted-by":"publisher","first-page":"27570","DOI":"10.1109\/ACCESS.2025.3540388","article-title":"BVQA: connecting language and vision through multimodal attention for open-ended question answering","volume":"13","author":"Bhuyan","year":"2025","journal-title":"IEEE Access"},{"key":"B36","doi-asserted-by":"publisher","first-page":"317","DOI":"10.1038\/s41597-019-0322-0","article-title":"MIMIC-CXR, a de-identified publicly available database of chest radiographs with free-text reports","volume":"6","author":"Johnson","year":"2019","journal-title":"Sci Data"},{"key":"B37","doi-asserted-by":"publisher","first-page":"15929","DOI":"10.1109\/ACCESS.2025.3528335","article-title":"A multimodal framework for radiology report summarization to improve clinical performance","volume":"13","author":"Sultan","year":"2025","journal-title":"IEEE Access"},{"key":"B38","author":"Irvin","year":""},{"key":"B39","first-page":"34892","article-title":"Visual instruction tuning","volume-title":"Advances in Neural Information Processing Systems","author":"Liu","year":"2023"},{"key":"B40","doi-asserted-by":"publisher","first-page":"2176","DOI":"10.1038\/s41591-021-01595-0","article-title":"Underdiagnosis bias of artificial intelligence algorithms applied to chest radiographs in under-served patient populations","volume":"27","author":"Seyyed-Kalantari","year":"2021","journal-title":"Nat Med"}],"container-title":["Frontiers in Digital Health"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1862236\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T07:57:22Z","timestamp":1788335842000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1862236\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,2]]},"references-count":40,"alternative-id":["10.3389\/fdgth.2026.1862236"],"URL":"https:\/\/doi.org\/10.3389\/fdgth.2026.1862236","relation":{},"ISSN":["2673-253X"],"issn-type":[{"value":"2673-253X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,2]]},"article-number":"1862236"}}