Abstract
<title>Abstract</title> <p> Background This study aimed to develop and validate the C-reactive protein-albumin score (CAZ score) using a multicenter cohort, evaluating its prognostic value for cancer patients and its association with severe malnutrition and cachexia. Methods The analysis included a multicenter training cohort of 6,492 cancer patients, with external validation in Cohort A ( <italic>n</italic> = 4,548) and Cohort B ( <italic>n</italic> = 1,944). The CAZ score was calculated by normalizing baseline C-reactive protein and albumin levels into z-scores and combining them, then stratifying into four categories (normal, mild, moderate, severe). Kaplan-Meier analysis, Cox proportional hazards models, and restricted cubic spline analysis were used to assess the CAZ score’s prognostic value for overall survival (OS). Logistic regression examined its associations with severe malnutrition and cancer cachexia, with sensitivity analysis excluding short-term mortality to test robustness. Results Kaplan-Meier analysis revealed significant stepwise differences in OS across the four CAZ score categories (71.9% vs 65.9% vs 54.9% vs 42.3%, Log-rank <italic>P</italic> < 0.001). Restricted cubic spline analysis showed a significant linear dose-response relationship between the continuous CAZ score and cancer-related mortality risk ( <italic>P</italic> < 0.001). The CAZ score-based inflammatory grading system provided significant prognostic stratification, with substantial differences in OS among patients in different inflammatory burden grade groups (85.5% vs 62.9% vs 46.9%, <italic>P</italic> < 0.001). Cox models identified the CAZ score as an independent prognostic factor for OS, with severe CAZ scores correlating to a 2.102-fold higher mortality risk compared to the normal group (95% CI: 1.872–2.360, <italic>P</italic> < 0.001). Logistic regression demonstrated a dose-dependent increase in the risk of severe malnutrition and cachexia ( <italic>P</italic> < 0.001) with higher CAZ scores. All findings were consistently replicated in the two independent validation cohorts. Conclusion The CAZ score, a novel standardized composite biomarker, effectively predicts OS in cancer patients and identifies those at high risk for severe malnutrition and cachexia. </p>