A Transfer Learning Radiomics Model for Predicting Response to Initial Transarterial Embolization in Patients with Gastroenteropancreatic Neuroendocrine Tumor Liver Metastases

Sun Yat-sen University

Status: Completed

Conditions: Neuroendocrine Tumors, NET

Study type: Observational

Funder types: Other

Identifiers

NCT06853457

IIT-2023-983

Details and patient eligibility

About

To develop and validate a CT-based transfer learning radiomics model for predicting response to initial TAE in GEP-NETLM patients and compare its performance with traditional radiomics and clinical models.

Enrollment: 257 patients

Sex: All

Ages: 18 to 78 years old

Volunteers: No Healthy Volunteers

Inclusion criteria

• (a) Clinical diagnosed as gastroenteropancreatic neuroendocrine tumor liver metastases; (b) Received initial transarterial embolization therapy; (c) Underwent multi-phase ceCT scans pre-TAE (≤1 month) and post-TAE (4-6 weeks).

Exclusion criteria

• (a) Neuroendocrine carcinoma (NEC) or other malignancies. (b) Received other liver metastasis treatments. (c) Lack of Multi-phase ceCT scans records. (d) CT images with artifacts or no visible lesions.

Trial design

257 participants in 2 patient groups

training set

Description:

The training set consisted of patients from January 2014 to December 2020.

testing set

Description:

the testing set included patients from January 2021 to September 2022.