Lung tumor CT segmentation dataset

Both ESLS26 and ASLS26 are publicly available datasets for CT segmentation of lung tumors. The data comes from Siemens SOMATOM Definition AS+ spiral CT scans in Germany and has been anonymized with standardized parameters (tube voltage 120 kVp, pitch 1.2, etc.). The scan range extends from the apex of the lung to the floor of the diaphragm.The ESLS26 dataset covers the period from September 26, 2020 to April 13, 2022, totaling 730 cases, of which 439 were male. CT images of patients with early-stage lung cancer were segmented using ITK-SNAP 3.8.0 (http://itksnap.org) to obtain masks. All tumor lesions were segmented by a radiologist with over 5 years of experience, carefully avoiding bronchi and blood vessels. The segmentation results were then reviewed and revised by a radiologist with over 10 years of experience, establishing a reliable ground truth for the segmentation process. After tumor segmentation on the CT images was completed, the segmentation data was exported as a separate file and stored along with the corresponding original CT image. This process allows for rapid matching of the CT image to its segmentation in subsequent examinations and repairs.The ASLS26 dataset covers the period from February 14, 2014 to July 24, 2019, and includes 449 cases. tumor volume volumes (GTVs) delineation for each patient was performed by at least two experienced radiation oncologists using the Varian radiation delineation system. Any disagreements were resolved through consensus based on anatomical and dosimetric criteria. Data for each patient is stored in DICOM file format. The RTSTRUCT file for each dataset contains the GTVs structure manually delineated by the radiation oncologists using the Varian system. Both datasets were anonymized using the RSNA DICOM anonymizer tool (Version 18.0, https://github.com/RSNA/Anonymizer). Meta-information was recorded in a CSV table, including anonymization information and original patient and examination details.