Public datasets
The following is the growing list of publicly available datasets that use DICOM for storing results of image analysis or interpretation:
- TCIA QIN-HEDNECK collection: whole body PET/CT images of patients before and after therapy. Includes SUV-correction, segmentations and segmentation-based measurements stored as DICOM SEG and SR for a subset of population. See details in this manuscript: https://peerj.com/articles/2057/.
- TCIA NSCLC Radiogenomics collection: chest CT images and segmentations of nodules stored as DICOM SEG.
- TCIA-LGG and TCIA-GBM collection segmentations: segmentation of sub-structures of brain tumors stored as DICOM SEG objects.
- TCIA QIN-Prostate-Repeatability collections : segmentations and segmentation-based measurements stored as DICOM objects for a dataset evaluating repeatability of multiparametric MRI in the prostate. See details in this manuscript: TBD (in press).
- TCIA LIDC-IDRI segmentations and characterizations of lung nodules: standardized representation of the annotations and nodule characterization for the LIDC project. See details in this manuscript: https://peerj.com/preprints/27378v1/.
- Crowds Cure Cancer: Data collected at the RSNA 2017 annual meeting: linear annotations of the maximum dimension of the tumor in a range of cancer types. Annotations are available as DICOM SR-TID1500 objects.