# Comparative performance of the patient-generated subjective global assessment, European Society for Clinical Nutrition and Metabolism criteria, and Global Leadership Initiative on Malnutrition criteria in patients with colorectal cancer: a multicenter study utilizing Bayesian inference

## TL;DR

This study compares three methods for diagnosing malnutrition in colorectal cancer patients and finds that one method is better at identifying malnutrition while another better predicts survival.

## Contribution

The study provides a comparative analysis of three malnutrition criteria using Bayesian inference in colorectal cancer patients.

## Key findings

- PG-SGA has high sensitivity and specificity for diagnosing malnutrition in colorectal cancer patients.
- GLIM criteria better predict survival outcomes compared to the other methods.
- All three nutritional assessment methods are independent risk factors for overall survival.

## Abstract

Malnutrition is a common complication among patients with colorectal cancer and has a significant impact on prognosis. However, the lack of a diagnostic gold standard for malnutrition complicates clinical nutritional intervention. This study aimed to compare the diagnostic value of the Patient-Generated Subjective Global Assessment (PG-SGA), European Society for Clinical Nutrition and Metabolism (ESPEN), and Global Leadership Initiative on Malnutrition (GLIM) criteria in identifying malnutrition in patients with colorectal cancer and to assess their ability to predict survival outcomes.

This study retrospectively reviewed data from 3,182 patients diagnosed with colorectal cancer in the Investigation on Nutrition Status and Clinical Outcome of Patients with Common Cancers in China database collected between July 2013 and March 2022. Using Bayesian principles, we calculated the sensitivity and specificity of the PG-SGA, ESPEN, and GLIM criteria for diagnosing malnutrition among patients with colorectal cancer. We also analyzed the relation between nutritional diagnosis and patient survival.

Our findings revealed that the PG-SGA has high sensitivity [0.80, 95% credible interval (CrI): 0.61–0.94] and specificity (0.99, 95% CrI, 0.99–1.00) for diagnosing malnutrition in patients with colorectal cancer. The ESPEN criteria showed high sensitivity (0.84, 95% CrI, 0.80–0.86), whereas the GLIM criteria exhibited high specificity (0.81, 95% CrI, 0.79–0.82). All three nutritional assessment methods were identified as independent risk factors for overall survival. Statistically significant differences in survival periods existed among the GLIM-defined nutritional status subgroups.

The PG-SGA demonstrated superior sensitivity and specificity in diagnosing malnutrition among patients with colorectal cancer. By contrast, the GLIM criteria performed better in predicting survival outcomes. Malnutrition is a significant risk factor that influences the survival of patients with colorectal cancer.