Reproducibility, prognostic homology, and biological relevance of CT features in NSCLC

Published: 1 June 2023| Version 2 | DOI: 10.17632/h4c3f8wvj2.2
Jiangdian Song


This dataset aims to evaluate the reproducibility, prognostic homology, and biological relevance of CT features in NSCLC. Four CT scans of the Credence Cartridge Radiomics (CCR) phantom from different CT manufactures were included, and the manual Segmentation of the NSCLC tumor (i.e., the cartridge of shredded rubber particles which mimic the texture of NSCLC) in the phantoms was conducted. In addition, the The Cancer Genome Atlas Lung Adenocarcinoma (TCGA-LUAD) datasets were used in this study. The CT images were acquired from: The RNA-seq data was acquired from: Due to copyright restrictions, interested readers should download the source data via the links. In this dataset, GE, P, S, and T represent the CT scans from GE, Philips, Siemens, and Toshiba, respectively. The "image" folder stores CT images, and the "mask" folder stores the manual segmentation of corresponding CT images.


Steps to reproduce

1. Read the CT images in the "image" folder. 2. Add the mask images in the corresponding "mask" folder.


China Medical University


Artificial Intelligence, Non-Small Cell Lung Cancer, Radiomics