@article{zhu2026ct,
  title = {{CT-Bench}: A Comprehensive Benchmark for Multimodal {AI} in Computed Tomography Analysis},
  author = {Zhu, Qingqing and Jin, Qiao and Mathai, Tejas S. and Fang, Yin and Wang, Zhizheng and Yang, Yifan and Sarfo-Gyamfi, Maame and Hou, Benjamin and Gu, Ran and Balamuralikrishna, Praveen T. S. and Wang, Kenneth C. and Summers, Ronald M. and Lu, Zhiyong},
  journal = {Machine Learning for Biomedical Imaging},
  volume = {2026},
  number = {Special Issue on MICCAI Open Data 2026},
  pages = {739--757},
  year = {2026},
  issn = {2766-905X},
  doi = {10.59275/j.melba.2026-b141},
  url = {https://melba-journal.org/2026:036},
  code = {https://github.com/CelestialShine/CT-Bench},
  abstract = {Publicly available computed tomography (CT) resources with lesion-level image, spatial, and textual annotations remain limited, restricting reproducible development and evaluation of multimodal medical AI. We present CT-Bench, a CT lesion image-text resource and visual question answering benchmark constructed by extending DeepLesion with curated lesion-level text, standardized lesion-size representations, structured attributes, and QA tasks. CT-Bench contains 20,335 lesion instances from 7,795 CT studies and 3,793 patients, including CT key-slice images, bounding-box annotations, PACS-derived lesion descriptions, lesion size measurements, and associated metadata. The resource also includes a multiple-choice QA benchmark with 2,850 question-answer pairs across seven lesion analysis tasks: image-to-description matching, multi-slice CT-to-description matching, description-to-image retrieval, description-to-bounding-box localization, lesion size estimation, image-based attribute recognition, and multi-slice CT attribute recognition. Hard negative answer choices are included to evaluate fine-grained image-text alignment and lesion-level reasoning. The resource is available with documentation, metadata files, and reproducibility code. Validation includes annotation review by trained annotators and medical experts, hard-negative verification, human-reader comparison, and baseline experiments with general and medical vision-language models.},
}