Diagnostic Performance and Agreement of MST and NUTRISCORE Compared with GLIM Criteria in Ambulatory Cancer Patients: Results from the OncoNutridos Study

Background/Objectives: Disease-related malnutrition is highly prevalent in oncology and is associated with poor clinical outcomes. Early detection through nutritional screening is essential; however, the optimal screening tool for ambulatory cancer patients remains uncertain. This study aimed to evaluate the agreement and diagnostic performance of the Malnutrition Screening Tool (MST) and NUTRISCORE compared with the Global Leadership Initiative on Malnutrition (GLIM) criteria in a large nationwide cohort of ambulatory cancer patients. Methods: In this multicenter, observational, cross-sectional nationwide study, adult patients attending oncology day hospitals for intravenous antineoplastic treatment between April and November 2021 were included. Nutritional risk was assessed using MST (cut-off ≥ 2) and NUTRISCORE (cut-off ≥ 5). Malnutrition was diagnosed according to GLIM criteria. Agreement between tools was assessed with Cohen’s kappa, and diagnostic performance was evaluated by sensitivity, specificity, accuracy, positive predictive value, and negative predictive value. Analyses were stratified by tumor nutritional risk and cancer stage. Results: Among 4440 patients from 86 hospitals, 50.7% met the GLIM criteria for malnutrition; 72.5% had moderate and 27.5% severe malnutrition. MST identified 37.5% of patients as being at nutritional risk, compared with 17.3% identified by NUTRISCORE. Agreement between MST and NUTRISCORE was moderate overall (κ = 0.48; 95% CI, 0.45–0.51), but varied markedly according to tumor nutritional risk, ranging from high agreement in high-risk tumors (κ = 0.82) to low agreement in low-risk tumors (κ = 0.28). Relative to GLIM, MST was more sensitive than NUTRISCORE (0.51 vs. 0.27), whereas NUTRISCORE was more specific (0.92 vs. 0.76) and had a higher positive predictive value (0.77 vs. 0.68). Negative predictive value was low for both tools. Conclusions: GLIM-defined malnutrition was highly prevalent in this large cohort of ambulatory patients with cancer. MST provided greater case detection, whereas NUTRISCORE showed a more conservative profile with higher specificity but substantially lower sensitivity. These findings suggest that the choice of screening tool should consider clinical context and tumor-related nutritional risk, and that neither instrument alone reliably excludes malnutrition in outpatient oncology settings.