Differential neurotoxicity patterns across neuroendocrine tumor grades: A real-world analysis of 12,847 patients.

Abstract

640 Background: Neuroendocrine tumors (NETs) exhibit substantial heterogeneity across WHO grades, yet treatment-related neurotoxicity is poorly characterized. We hypothesized that higher-grade NETs have increased neurotoxicity risk, which may inform monitoring and early intervention. Methods: Retrospective analysis of 12,847 NET patients (Grade 1: 5,234; Grade 2: 6,018; Grade 3: 1,595) treated systemically in the TriNetX Global Health Research Network (2015–2024). NET cases were identified using ICD-10 codes for malignant carcinoid/neuroendocrine tumors (C7A.xx by primary site, C7B.xx for metastatic, and C25.x for pancreatic NETs). Grade (G1: Ki-67 ≤2%; G2: 3–20%; G3: >20%) was obtained from structured pathology and tumor registry fields mapped within TriNetX; free-text clinical notes were not accessed. Neurotoxicity was classified as mild–moderate (Grade 1–2) or severe (Grade 3–4). Primary endpoints were incidence and severity by grade; secondary endpoints included hospitalization and long-term sequelae. Propensity score matching adjusted for age, sex, comorbidities, prior treatments, and tumor burden. Cox models assessed time-to-event outcomes. Results: Neurotoxicity incidence increased with tumor grade: 18.2% (Grade 1), 31.7% (Grade 2), 47.9% (Grade 3). Adjusted OR for Grade 3 vs Grade 1: 3.84 (95% CI 3.42–4.31, p<0.001). Severe events were predominant in Grade 3 (23.1% vs 3.4% in Grade 1; p<0.001), including encephalopathy (8.9% vs 1.2%) and seizures (6.7% vs 0.8%). Hospitalizations were highest in Grade 3 (18.3% vs 2.1% Grade 1). PRRT and immunotherapy contributed to grade-dependent neurotoxicity; combination therapies further amplified risk (OR 2.67, 95% CI 1.89–3.77). Five-year follow-up revealed persistent neurologic disability in 19.4% of Grade 3 patients versus 2.3% of Grade 1, with significantly increased healthcare utilization. Conclusions: In this largest real-world NET neurotoxicity analysis, grade-dependent patterns were pronounced, with Grade 3 tumors at highest risk for severe and long-term neurologic complications. These findings support grade-stratified monitoring, early intervention strategies, and risk-based clinical decision-making, offering a framework for personalized neurotoxicity management in NET patients.