PRONTO nutritional risk identification protocol: Exploratory findings on survival prediction and comparison with the...

The aim of the study is to evaluate the accuracy of PRONTO and its association with mortality at 6 and 12 mo in outpatient and inpatient cancer patients.

A secondary exploratory analysis of a single-center cohort including patients with cancer ≥20 y, both outpatients and inpatients (N = 390). The PRONTO protocol was retrospectively reconstructed from variables available in the original dataset, including information on SARC-F and Patient-Generated Subjective Global Assessment (PG-SGA). Patients with a positive response to any of these questions for PRONTO were classified as "at nutritional risk," and the PG-SGA short-form (PG-SGA SF ≥4) was used for comparisons. The area under the receiver operating characteristic (AUROC), sensitivity, specificity, predictive values was calculated to assess the accuracy and Cox regression analyses were used to assess PRONTO's association with mortality.

PRONTO identified 84.4% at nutritional risk, while PG-SGA SF identified 76.4%. PRONTO had fair performance (AUC: 0.81, 95% CI: 0.77; 0.85), moderate sensitivity (86.3%) and high positive predictive value (95.3%), low specificity (77.0%), and low negative predictive value (51.1%). PRONTO-identified risk status was independently associated with 6- and 12-mo mortality (_adjusted HR: 2.96, 95% CI: 1.05: 8.32, adjusted HR: 2.88, 95% CI: 1.20; 6.63, respectively). The PRONTO presented high sensitivity and was independently associated with mortality.

These findings highlight its potential utility as a rapid protocol to identify nutritional risk, and its value as a prognostic tool. Given that PRONTO was retrospectively reconstructed and not prospectively administered by its intended users, we advocate for further prospective validation.