Nutrition and inflammation are closely related to prognosis in breast cancer patients. However, current nutritional and inflammatory measures predict disease free survival (DFS) of breast cancer are still different, and the most predictive measures remain unknown. This study aimed to compare the predictive effects of commonly used nutritional and inflammatory measures on DFS and to improve existing nutritional or inflammatory measures in order to develop a new model that is more effective for predicting postoperative recurrence and metastasis in breast cancer patients. The clinical data of 536 female breast cancer patients with invasive ductal carcinoma who underwent surgery at Shaoxing People’s Hospital from January 2012 to December 2018 were retrospectively evaluated. The predictive effects of nutritional and inflammatory indicators on DFS were evaluated. Machine learning was used to evaluate and rank laboratory indicators, select relatively important variables to modify nutritional or inflammatory indicators with the best predictive power, and evaluate their predictive role in patients’ postoperative recurrence and metastasis. Among various metrics predicting DFS, the CONUT score emerged paramount with an area under the curve (AUC) of 0.667. Interestingly, the combination of the erythrocyte levels with the CONUT score (ECONUT) achieved the highest AUC (0.722). The Kaplan-Meier survival analysis showed that the group exhibiting high ECONUT scores experiencing a notably poorer DFS. ECONUT was identified as an independent risk factor for postoperative DFS (P < 0.001). The ECONUT model could provide an effective assessment tool for predicting DFS in breast cancer patients.
Read full abstract