Abstract
PURPOSE: Cachexia has a very high prevalence in patients with cancer, and lacks effective screening tools yet. Global Leadership Initiative on Malnutrition (GLIM) is a novel malnutrition assessment tool, with increased important roles in malnutrition diagnosis for patients with cancer. However, whether GLIM can be used as an effective screening tool remains unknown. METHODS: We performed a multicenter cohort study including 8,478 solid tumor patients from 40 clinical centers throughout China. Cachexia was diagnosed based on the 2011 international cancer cachexia consensus. The receiver operating characteristic curves (ROC) and decision curve analysis (DCA) were developed to determine the efficacy and clinical net benefit of GLIM and Patient-Generated Subjective Global Assessment (PG-SGA) in the detection of cancer cachexia, respectively. RESULTS: According to the consensus guidelines, 1,441 (17.0%) cancer patients were diagnosed with cachexia among 8,478 patients in the present study. The sensitivity of one-step GLIM and two-step GLIM for detecting cachexia were 100 and 88.8%, respectively, while that of PG-SGA was 86.2%. The accuracies of one-step GLIM and two-step GLIM reached 67.4 and 91.3%, which were higher than that of PG-SGA (63.1%). The area under the curves (AUCs) of one-step GLIM (0.835) and two-step GLIM (0.910) were higher than PG-SGA (0.778) in patients with cancer. The DCA also revealed that two-step GLIM had better clinical effect than PG-SGA between 20-50% threshold probabilities. CONCLUSION: GLIM could be used as an effective tool in screening cancer cachexia, two-step GLIM criteria show more accurate while one-step GLIM criteria is more sensitive. TRIAL REGISTRATION: ChiCTR1800020329.