Artificial intelligence applications in 68Ga-DOTA PET/CT images: Prediction of response assessment in GEP-NETs undergoing PRRT with 177Lu-DOTATOC

#3119

Introduction: Radiomics aims at identifying features contained in biomedical images, analysing them in various scenarios such as the patients' outcome prediction.

Aim(s): We developed a new radiomics predictive model of disease progression in GEP-NETs using 68Ga-DOTA-PET/CT images.

Materials and methods: We examined 324 SSTR-positive lesions, after a retrospective analysis of 38 GEP-NETs (9G1-27G2-2G3) who underwent restaging 68Ga-DOTA-PET/CT before complete PRRT with 177Lu-DOTATOC. Clinical, laboratory and radiological FU data were collected for a period of at least 6 months after the last cycle. We used LifeX software, through which 65 features were extracted from each lesion divided by the site (parenchyma, LN, BL) and compared to the singular lesions’ outcome. Pre-PRRT CgA values and histological grading were considered as additional features. A new statistical system was implemented based on the point-biserial correlation coefficient and the logistic regression analysis for the reduction and selection of the features; the DA was used, instead, to obtain the predictive model.

Conference: 18th Annual ENETS Concerence (2021)

Presenting Author: Vento A

Authors: Vento A, Laudicella R, Spataro A, Comelli A, Stefano A,

Keywords: PRRT, radiomics, artificial intelligence, 177Lu, 68Ga-DOTA, pet/ct, gep-net,