Prognosis-Related Nutritional Score for Cancer Patients (PRNS): a clinical nutritional score derived from a retrospective cohort study

BACKGROUND: Nutritional assessment and quality of life (QOL) have become important indices for therapeutic efficacy in patients with malignancies. We aim to develop and validate an easy-to-use questionnaire with prognostic value to assess nutritional status in hospitalized cancer patients. METHODS: A comprehensive survey focused on patient-generated subjective global assessment (PG-SGA) and 30-item European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire (EORTC QLQ-C30 Chinese version) was performed in a cohort of 22,776 patients derived from the INSCOC study. Among them, 1948 patients were followed for 3 years after admission. An observational, retrospective, cross-sectional cohort study was conducted in accordance with TRIPOD statement. Breiman's random forest model was applied to calculate variable importance (VIMP) for items in PG-SGA and EORTC QLQ-C30 (Chinese version) for nutritional recommendation. Cox regression model was employed to construct Prognosis-Related Nutritional Score for Cancer Patients (PRNS). Kaplan–Meier Survival curve, ROC and DCA were calculated to evaluate prognostic value of nutritional status categorized by PRNS, and compared with PG-SGA. RESULTS: Nutritional status was classified into 4 levels by PRNS scores: well nourished (≤ 4.5 points), mild malnourished (5–7.5 points), moderate malnourished (8–14.5 points), and severe malnourished (≥ 15 points). Significant median overall survival differences were found among nutritional status groups stratified by the PRNS (all Ps < 0.05). Compared with PG-SGA, PRNS had better prognostic value for survival stratified by nutritional status. The external, internal validity, test–retest reliability and rater reliability were satisfactory. CONCLUSIONS: We systematically developed and validated PRNS as a nutrition screening tool for cancer patients. Compared with PG-SGA, PRNS has better prognostic value and simpler operation. TRIAL REGISTRATION: Investigation on Nutrition Status and its Clinical Outcome of Common Cancers, ChiCTR1800020329. Registered 24 December 2018—Retrospectively registered, http://www.chictr.org.cn/showproj.aspx?proj=31813 SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12967-022-03696-x.