Related Experiment Video
Updated: Jun 9, 2026
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Prognostic evaluation using nutrition-inflammation biomarkers from routine blood tests in metastatic breast cancer: a
Wen-Xiong Nong1, Sheng-Kai Huang2, Wen-Hai Zhang1
1Department of Breast Surgery, Eighth Affiliated Hospital of Guangxi Medical University, Guigang City People's Hospital, Guigang, Guangxi, China.
Objective:
To evaluate and compare the prognostic value of multiple nutrition- and inflammation-related indices derived from routine blood tests in patients with metastatic breast cancer, and to identify the most clinically applicable indicators.
Methods:
This retrospective study analyzed data from 163 newly diagnosed breast cancer patients with a single distant metastasis. Seventeen nutrition- and inflammation-related indices were assessed, including ALI, NLR, dNLR, SII, SIRI, PLR, MLR, NMLR, PNI, GLR, AGR, GNRI, mGNRI, TP, PA, TRF, and ALB. Kaplan-Meier survival analysis and Cox regression models were used to evaluate the associations between these indices and overall survival (OS). The predictive performance of each index was assessed using the concordance index (C-index), and the Boruta algorithm was applied to identify key variables. Restricted cubic spline (RCS) regression was performed to explore potential nonlinear relationships between indices and OS, and subgroup analyses were conducted to examine index performance across different clinical characteristics.
Results:
Multivariate Cox regression analysis identified MLR, SIRI, ALI, AGR, and PA as independent predictors of OS in patients with metastatic breast cancer. Patients with MLR ≥ 0.33 had a 3.94-fold increased risk of death, and those with SIRI ≥ 1.70 had a 3.32-fold increased risk. In contrast, patients with ALI ≥ 53.99, AGR ≥ 1.11, and PA ≥ 181 had 73%, 76%, and 77% reductions in mortality risk, respectively. Inflammation-related indices demonstrated stronger predictive value for short-term (1-year) outcomes, whereas nutrition-related indices were more effective in predicting medium- to long-term (3-5 years) survival. ALI and MLR showed consistent prognostic performance across all clinical subgroups, while SIRI, AGR, and PA exhibited significant interactions in specific subgroups.
Conclusion:
Nutrition- and inflammation-related indices derived from routine blood tests-particularly MLR, SIRI, ALI, AGR, and PA-serve as effective tools for prognostic assessment in metastatic breast cancer. Among these, ALI and MLR have universal applicability across patient subgroups, offering clinicians valuable guidance for optimizing patient management strategies and improving survival and quality of life.