MM‑CT‑ILD: Multi‑modal Lung Cancer Therapy‑related Interstitial Lung Disease Dataset for Deep Learning‑Assisted Diagnosiss

This dataset includes 1086 chest CT images and corresponding clinical data from 267 lung cancer patients, with complete longitudinal follow-up records covering disease progression at different time points from pre-treatment to post-treatment.All original DICOM images were standardized using the SimpleITK tool and uniformly resampled to 1×1×1 mm³ voxel spacing.

Only CT pixel data and image geometric information were retained, while comprehensive and irreversible desensitization was performed simultaneously. We adopted a patient-level unified random offset scheme, assigning a unique random seed to each patient. This scheme strictly preserves the relative time intervals of all events for the same patient to ensure the scientific validity of the research data, while completely cutting off the association between the original dates and real patient identities.This dataset contains the following data files:Chest CT image folder: Stored in subdirectories by patient ID, containing a total of 1086 NIfTI (.nii.gz) files, all of which have completed metadata desensitizationClinical annotation file: Annotated_result.xlsx, containing 1086 records with fields including patient ID, CT filename, file path, presence or absence of inflammation, and three-class annotation results (0: normal, 1: CT-ILD, 2: other pneumonias)DICOM parameter file: Dicom_param.xlsx, containing technical parameters of all CT scans such as tube voltage, slice thickness, pixel spacing, window width and window levelClinical laboratory result file: Lab_result.xlsx, containing multiple laboratory indicators of patients including blood routine, biochemistry, coagulation function and corresponding test datesTreatment plan file: Treatment_plan.xlsx, containing patients' tumor diagnosis dates, treatment regimens and medication informationCode folder: Contains core scripts for DICOM-to-NIfTI format conversion and dataset statistical analysisSupplementary document: Data_Usage_Agreement_DUA.pdf, the official data usage agreement specifying privacy and data redistribution constraintsThere are no missing values in all tabular data.