Low Geriatric Nutritional Risk Index as a Poor Prognostic Marker for Second-Line Pembrolizumab Treatment in Patients with Metastatic Urothelial Carcinoma: A Retrospective Multicenter Analysis

Abstract

<b><i>Background:</i></b> We evaluated the prognostic efficacy of the Geriatric Nutritional Risk Index (GNRI) in second-line pembrolizumab (PEM) therapy for patients with metastatic urothelial carcinoma (mUC). <b><i>Patients and Methods:</i></b> From January 2018 to October 2019, 52 mUC patients, treated previously with platinum-based chemotherapy, underwent second-line PEM therapy. Peripheral blood parameters were measured at the start of treatment: serum neutrophil-to-lymphocyte ratio (NLR), serum albumin, serum C-reactive protein (CRP), and body height and weight. PEM was intravenously administered (200 mg every 3 weeks). The patients were organized into two groups based on their GNRI (&#x3c;92 [low GNRI] and ≥92 [high GNRI]), and the data were retrospectively analyzed. Adverse events (AEs) were evaluated and imaging studies assessed for all patients. Analyses of survival and recurrence were performed using Kaplan-Meier curves. Potential prognostic factors affecting cancer-specific survival (CSS) were assessed by univariate and multivariate Cox regression analyses. <b><i>Results:</i></b> patients’ baseline characteristics, except for their BMI and objective response rate, did not significantly differ between the two groups. The median total number of cycles of PEM therapy was significantly higher for the high-GNRI group (<i>n</i> [range]: 6 [2–20] vs. 3 [1–6]). The median CSS with second-line PEM therapy was 3.6 months (95% confidence interval [CI]: 2.5–6.1) and 11.8 months (95% CI: 6.2–NA) in the low-GNRI and the high-GNRI group (<i>p</i> &#x3c; 0.01), respectively. Significant differences in CSS between the low- and high-CRP or -NRL groups were not found. Multivariate Cox proportional-hazards regression analysis revealed that a poor Eastern Cooperative Oncology Group performance status, visceral metastasis, and a low GNRI were significant prognostic factors for short CSS (95% CI: 1.62–6.10, HR: 3.14; 95% CI: 1.13–8.11, HR: 3.03; 95% CI: 1.32–8.02, HR: 3.25, respectively). Of the AEs, fatigue showed a significantly higher incidence in the low-GNRI group. <b><i>Conclusions:</i></b> For mUC patients receiving second-line PEM therapy, the GNRI is a useful predictive biomarker for survival outcome.