Development and validation of a preoperative CT-based risk scoring system for predicting recurrence-free survival in patients undergoing curative surgery for gastric cancer

Zhengqi Zhu,Haipeng Gong,Jianan Gu,Yongfeng Dai,Chunyan Yang,Mimi Mao,Anyi Song,Feng Feng
DOI: https://doi.org/10.1016/j.ejrad.2024.111303
IF: 4.531
2024-01-11
European Journal of Radiology
Abstract:Purpose The objective of this study was to establish and validate a preoperative risk scoring system that incorporated both clinical and computed tomography(CT) variables to predict recurrence-free survival (RFS) in gastric cancer(GC) patients who underwent curative resection. Method We retrospectively included consecutive patients with surgically confirmed GC who underwent preoperative CT scans between October 2017 and January 2022. Multivariate Cox regression analysis was employed in the derivation set to identify clinical and CT variables associated with RFS and to construct a risk score. This risk score was subsequently validated in an independent test set. Results A total of 346 patients were included in the study, with 213 in the derivation set and 133 in the test set. Five variables, namely ctEMVI, ctBorrmann, visceral obesity, sarcopenia, and NLR, were independently associated with RFS. In the test set, the preoperative risk score exhibited a c-index of 0.741, which outperformed the predictive accuracy of pathological tumor staging (c-index of 0.673, p = 0.021) at various time points. The preoperative risk score effectively stratified patients into low and high-risk groups. Conclusion The developed preoperative risk scoring system demonstrated the ability to predict RFS following curative resection in GC patients.
radiology, nuclear medicine & medical imaging
What problem does this paper attempt to address?