Ki-67 index cut-off values and survival prognostication in neuroendocrine neoplasms: Machine learning-based insights

#4403

Introduction: The current grading system for gastroenteropancreatic neuroendocrine neoplasms (GEP-NEN) is based on Ki-67 index and/or mitotic rate. By definition, high-grade NEN (neuroendocrine tumour grade 3 [NET G3] and neuroendocrine carcinoma [NEC]) have a Ki-67 >20%, while NET G1 have a Ki-67 <3% and NET G2 range between 3% and 20%.

Aim(s): This retrospective analysis aimed to identify the ideal prognostic subdivision of NEN.

Materials and methods: Ki-67 values at diagnosis were extracted from pathology reports of GEP-NEN patients treated at the Medical University of Vienna. The R package rpart v4.1.23 (Therneau and Atkinson, 2023) was used to recursively split the overall survival (OS) data into distinct subgroups based on Ki-67 and morphology in order to find the optimal cut-off values (a survival tree algorithm).

Conference:

Presenting Author:

Authors: Melhorn P, Berchtold L, Mazal P, Raderer M, Kiesewetter B,

Keywords: neuroendocrine neoplasm, prognosis, ki-67 index,