Texture analysis in the prediction of progression-free survival (PFS) in patients with progressive neuroendocrine tumors (NETs) treated with lanreotide autogel (LAN) in the prospective, open-label, non-comparative, phase 2 CLARINET FORTE trial

#3636

Introduction: Tumor texture analysis may provide new imaging biomarkers for PFS prediction in patients with NETs.

Aim(s): Assess whether tumor texture parameters predict PFS in patients with NETs.

Materials and methods: CLARINET FORTE (NCT02651987) enrolled patients with progressive NETs (n=51 midgut; n=48 pancreatic) receiving standard LAN dose (120mg/28 days); on study entry, patients received LAN 120mg/14 days. CT analysis assessed mean parameters (kurtosis, entropy, uniformity) for all target lesions by patient from screening to Week 24. Higher kurtosis and entropy and lower uniformity indicate increased tumor heterogeneity. Receiver operating characteristic curves of baseline parameters determined area under the curve (AUC) and Youden index (optimal cut-off value to discriminate patients on PFS). Kaplan-Meier and univariate Cox regression analysis assessed associations between PFS and parameters dichotomized by cut-off value.

Conference:

Presenting Author:

Authors: Dromain C, Truong Thanh X, Grandoulier A, Pavel M,

Keywords: texture analysis, progression-free survival, neuroendocrine tumor,