ENETS Abstract Search
Prediction of Symptomatic Mesenteric Mass in Patients with Small Intestinal Neuroendocrine Tumors Using a CT Radiomics Approach
Introduction: A mesenteric mass surrounded by fibrosis is a hallmark feature of small intestinal neuroendocrine tumors (SI-NETs) that can induce severe abdominal complications. To improve clinical outcome, there is a need for personalized treatment strategies based on accurate prediction of development of abdominal complications.
Aim(s): To identify patients prone to develop abdominal complications due to mesenteric metastases using a radiomics approach based on CT.
Materials and methods: SI-NET patients with a metastatic mesenteric mass, and a CT scan between 2008-2018 were retrospectively included from the Rotterdam NET-database. The mesenteric mass (MM) and surrounding mesentery (SM) were manually segmented on the CT scan Subsequently, 826 radiomics features were extracted. Additionally, the MM location and several patient characteristics were evaluated as features. Classification was performed using a combination of machine learning approaches. Evaluation was performed through a 100x random-split cross-validation. The performance of the models was compared to an expert multidisciplinary tumor board (MTB).
Conference: 17th Annual ENETSConcerence (2020)
Presenting Author: Blazevic A
To read the full abstract, please log into your ENETS Member account.