Design and Validation of the GI-NEC Score to Prognosticate Overall Survival in Patients With High-Grade Gastrointestinal Neuroendocrine Carcinomas

Abstract

Background: Prognostic markers for risk stratification of patients with gastrointestinal high-grade neuroendocrine carcinomas (GI-NECs) are lacking; we designed and validated a prognostic score for overall survival (OS).

Methods: Consecutive patients diagnosed in five neuroendocrine specialist European centers were included. Patients were divided into three cohorts: a training cohort (TC), an external validation cohort (EVC), and a prospective validation cohort (PVC). Prognostic factors were identified by log-rank test, Cox-regression, and logistic regression analyses. The derived score was internally and externally validated. All statistical tests were two-sided.

Results: Of 395 patients screened, 313 were eligible (TC = 109 patients, EVC = 184 patients, and PVC = 20 patients). The derived prognostic score included five variables: presence of liver metastases, alkaline phosphatase (ALK), lactate dehydrogenase (LDH), Eastern Cooperative Oncology Group performance status (ECOG PS), and Ki67. In multivariable analysis, the score was prognostic for OS (hazard ratio [HR] = 1.86, 95% confidence interval [CI] = 1.47 to 2.35, P < .001) and had good discrimination (C-index = 0.76) and calibration (mean error = 0.021, 90th percentile = 0.037) in the TC. These results were validated in the EVC and PVC, in which our score was able to prognosticate for OS when adjusted for other prognostic variables in the multivariable analysis (HR = 1.85, 95% CI = 1.27 to 2.71, P = .001; and HR = 4.51, 95% CI = 1.87 to 10.87, P = .001, respectively). The score classified patients into two groups with incremental risk of death: group A (0-2 points, 181 patients [63.9%], median OS = 19.4 months, 95% CI = 16.1 to 25.1) and group B (3-6 points, 102 patients [36.1%], median OS = 5.2 months, 95% CI = 3.6 to 6.9).

Conclusions: The GI-NEC score identifies two distinct patient cohorts; it provides a tool for clinicians when making treatment decisions and may be used as a stratification factor in future clinical trials.