Keywords:
Artificial Intelligence, Interventional vascular, Oncology, CT, Radioembolisation, Cancer, Metastases
Authors:
P. Schindler, M. Masthoff, M. Köhler, K. Rahbar, L. Stegger, W. Heindel, M. Wildgruber, W. Roll
DOI:
10.26044/ecr2023/C-16195
Purpose
Colorectal cancer (CRC) is one of the leading causes of global cancer burden and a highly lethal disease in the metastatic state. Novel multimodal treatment strategies, including systemic therapy in combination with various local-ablative treatment options have shown promising results to improve patient outcomes. While radiofrequency ablation, microwave ablation or transarterial chemoembolization (TACE) are frequently used in earlier stages of the disease, transarterial radioembolization (TARE) with Yttrium-90-loaded (Y-90) microspheres may be also considered as a treatment option with non-curative intent. As TARE is a costly and highly demanding therapy with possible side effects, careful patient selection is crucial.
Previous studies have investigated the predictive value of pre-treatment imaging for outcomes of TARE in metastatic colorectal cancer (mCRC). Based on the assumption of higher Y-90 microspheres accumulation in hypervascular metastases, some studies revealed a favorable outcome of patients with increased arterial perfusion in liver metastases measured by computed tomography (CT) perfusion(1). In contrast, in another study, radiographic vascular appearance of liver tumors, regardless of pre-treatment imaging modality, did not affect survival after TARE(2). Moreover, the distribution pattern of Technetium-99m-labelled macroaggregated albumin (Tc-99m-MAA) uptake by colorectal liver metastases in pre-TARE evaluation failed as a predictor of response(3).
Despite progress in conventional liver imaging techniques, imaging features beyond what is humanly visible are required to make further diagnostic improvements. As tumor heterogeneity is one of the major problems limiting the efficacy of targeted therapies and compromising treatment outcomes, several recent studies found promising results on radiomic feature (RF)-based analysis in oncologic imaging for outcome prediction in several entities. The application of radiomic analysis for outcome prediction in interventional oncology have been focused on patients who underwent TACE for hepatocellular carcinoma. However, there is only limited evidence for the potential of radiomic analysis to predict treatment response to TARE(4).
Therefore, this study aimed to evaluate the benefit of a radiomics-based model for predicting response and survival of patients with colorectal liver metastases treated with TARE.