A PET/CT image-based deep learning approach for survival prognosis and clinical management of postoperative adjuvant chemotherapy in esophageal squamous cell carcinoma patients
Description
This dataset contains CT, PET, and PET/CT images of patients diagnosed with esophageal squamous cell carcinoma (ESCC). It is released in conjunction with the study titled “A PET/CT image-based deep learning approach for survival prognosis and clinical management of postoperative adjuvant chemotherapy in esophageal squamous cell carcinoma patients”. Images of ESCC patients who received two different treatment therapies: namely surgery alone (SA) and surgery followed by postoperative adjuvant chemotherapy (SPOCT), are included. Moreover, the dataset also provides the ESCCPro-SA and ESCCPro-SPOCT models and code, proposed in the aforementioned study, to enable future researchers to replicate the models. Specifically, the “ESCC_OA_Data.rar” file consists of two folders, namely “Patients_SA” and “Patients_SPOCT”, which correspond to the data of ESCC patients who underwent each respective therapy. Each image within the dataset follows a specific naming convention. The initial number in the file name represents the anonymized patient identification (ranging from 1 to 646), followed by the type of image (CT, PET, or PET/CT), and the corresponding image slice number. The final number in the file name is specifically used to differentiate between patients treated with SA (numbers 1-2) and patients treated with SPOCT (numbers 3-4). Please cite the reference if using this dataset: A PET/CT image-based deep learning approach for survival prognosis and clinical management of postoperative adjuvant chemotherapy in esophageal squamous cell carcinoma patients.
Files
Steps to reproduce
The code to reproduce the models has been provided in this dataset.