{
  "abstract": "We present RadMultiBench, a multimodal benchmark of 6,148 medical images paired with expert radiological impressions across five categories--Breast Cancer, Lung Benign, Lung Malignant, Pneumonia, and Tuberculosis--targeting two core clinical tasks: automated report generation and disease classification. To enable reliable, repeatable experimentation, we provide a reproducible pipeline for data ingestion, de-identification, split management, and standardized reporting of text and classification metrics, facilitating dataset versioning and continuous benchmark testing. We compare three architectures: (1) a classical CNN-RNN with Bahdanau attention (ResNet-LSTM), (2) a general-domain vision-language (V-L) prefix-projection model (CLIP-GPT), and (3) a domain-specific V-L prefix-projection model (BioViL-T). Results reveal a sharp performance divide: the ResNet-LSTM attains near-perfect classification (99.5% accuracy) and is the only model to generate coherent reports (BLEU-4 = 0.021), whereas both prefix-projection models catastrophically fail generation (BLEU-4 [≤] 0.0001); moreover, the domain-specific BioViL-T also fails classification (11.5% accuracy). From a data-engineering angle, the pipeline is productionizable with lineage, feature registries, and continuous integration to guard against silent regressions when updating models or data; from a data-science angle, the benchmark provides baselines and analyses that foreground feature alignment and architectural suitability as first-order concerns, inviting ablations on multimodal objectives, retrieval-augmented encoders, and uncertainty-aware decoding. Overall, RadMultiBench highlights that state-of-the-art status in general V-L domains does not translate automatically to high-stakes clinical tasks, and it furnishes the tooling and evidence base needed to develop robust, clinically viable multimodal systems.",
  "abstract_shown": true,
  "affiliations": [
    "North South University"
  ],
  "article_type": "New results",
  "arxiv": null,
  "author_total": 6,
  "authors": [
    {
      "name": "M. N. Alam"
    },
    {
      "name": "M. T. Rahman"
    },
    {
      "name": "D. Sarkar"
    },
    {
      "affs": [
        0
      ],
      "name": "A. F. H. Dhrubo"
    },
    {
      "name": "M. A. Qayum"
    },
    {
      "name": "M. A. Khan"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/cmhbe5x4jwmrt.bib",
  "cited_by": 0,
  "doi": "10.64898/2026.11.04.764428",
  "fulltext_url": "https://www.biorxiv.org/content/10.64898/2026.11.04.764428v1.full",
  "id": "cmhbe5x4jwmrt",
  "issns": [],
  "landing_url": "https://www.biorxiv.org/content/10.64898/2026.11.04.764428v1",
  "language": null,
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "biorxiv",
      "bioinformatics"
    ]
  ],
  "openalex": null,
  "pdf_url": "https://www.biorxiv.org/content/10.64898/2026.11.04.764428v1.full.pdf",
  "pmcid": null,
  "pmid": null,
  "published": "2026-11-04",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/cmhbe5x4jwmrt.ris",
  "sources": [
    {
      "first_seen": "2026-11-07T00:21:18Z",
      "native_id": "10.64898/2026.11.04.764428",
      "source": "biorxiv",
      "url": "https://www.biorxiv.org/content/10.64898/2026.11.04.764428v1"
    }
  ],
  "subjects": [
    {
      "field": "biology",
      "origin": "biorxiv:bioinformatics",
      "subfield": "bioinformatics"
    }
  ],
  "title": "RadMultiBench: A Reproducible Dual-Task Benchmark Exposing the Limits of Pre-Trained Vision-Language Models Across Thoracic and Breast Imaging",
  "title_text": "RadMultiBench: A Reproducible Dual-Task Benchmark Exposing the Limits of Pre-Trained Vision-Language Models Across Thoracic and Breast Imaging",
  "type": "preprint",
  "updated_at": "2026-11-08T00:21:07Z",
  "url": "https://research.draconitic.com/paper/cmhbe5x4jwmrt",
  "venue": "bioRxiv",
  "versions": []
}