CT-RATE

License

CC BY-NC-SA 4.0
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

Citation

Paper BibTeX:

@article{hamamci2024developing,
  title={Developing generalist foundation models from a multimodal dataset for 3d computed tomography},
  author={Hamamci, Ibrahim Ethem and Er, Sezgin and Wang, Chenyu and Almas, Furkan and Simsek, Ayse Gulnihan and Esirgun, Sevval Nil and Doga, Irem and Durugol, Omer Faruk and Dai, Weicheng and Xu, Murong and others},
  journal={arXiv preprint arXiv:2403.17834},
  year={2024}
}

Dataset description

CT-RATE is a large-scale multimodal dataset pairing non-contrast chest CT volumes with radiology text reports, multi-abnormality labels, and metadata. It contains 25,692 chest CT volumes (expanded to 50,188 through reconstructions) from 21,304 unique patients, enabling research on vision–language learning and abnormality detection in 3D medical imaging.

Number of CT volumes: 3134

Contrast: Non-contrast

CT body coverage: Chest

Does the dataset include any ground truth annotations?: No

Original GT annotation targets: -

Number of annotated CT volumes: -

Annotator: -

Acquisition centers: -

Pathology/Disease: 18 distinct abnormalities

Original dataset download link: https://huggingface.co/datasets/ibrahimhamamci/CT-RATE

Original dataset format: nifti