[68Ga]Ga-DOTANOC PET/CT & radiomics: Potential role in discriminating tumor grade in well-differentiated primary pancreatic neuroendocrine tumors

#3131

Introduction: The possibility to non-invasively assess the primary lesion tumor grade by PET/CT derived parameters is extremely appealing, potentially overcoming the limits of bioptical sampling.

Aim(s): To assess whether 68Ga-DOTANOC PET/CT radiomic features (RFs) can distinguish G2 from G1 pancreatic neuroendocrine tumor (panNET).

Materials and methods: Two sets of patients (pts), a training (TC) and a validation cohort (VC), with G1-G2 primary panNET and a pre-surgical 68Ga-DOTANOC PET/CT were included. Tumor grade was assessed by surgery on the whole tumor mass in the TC, on the bioptic sample in the VC. In the TC, 1st-order and 2nd-order RFs were computed from SUV maps for the whole tumor volume. 60 single and 1770 coupled RFs were considered and correlated with primary tumor grade. The discrimination capability was assessed both for single and coupled RFs. The Linear Discriminant Analysis was used to build a discriminative radiomic model and ROC curve and the AUC were computed. The single and coupled RFs showing the highest p-values and AUC were finally selected and secondary tested on the VC.

Conference: 18th Annual ENETS Concerence (2021)

Presenting Author: Calabrò D

Authors: Calabrò D, Bevilacqua A, Malavasi S, Baiocco S, Ricci C,

Keywords: panNET, DOTANOC, PET/CT, radiomics,