75IN - Tools to Evaluate Tumour Response to Therapy: Imaging

Abstract

ABSTRACT With neuroendocrine tumor, liver metastases are by far the most common site of recurrence of the disease. Liver mets pattern are now better known (hypervascular 60-83%, hypovascular 5%, and micronodular 22%). In the detection of liver metastases, MRI is more sensitive that CT scan, and adding diffuse weighted sequences to MRI can improve this sensitivity to 78–80%. Once detected, current ENETS guidelines recommend a staged approach to the management of liver metastases, depending on liver extend. If only one lobe is affected, surgery should be considered. If one lobe is predominantly affected, surgery and local ablative treatment are recommended. When a diffuse pattern is present, medical therapy, systemic chemotherapy and endovascular therapy are recommended, along with molecular targeted therapy. Then it is a challenge to evaluate tumour response due to these various therapies. Imaging can be effective in measuring reducing of tumour size, with RECIST. RECIST criteria, are accurate in assessing tumor progression but sometimes inaccurate in assessing tumor response after locoregional therapy or under molecular targeted therapy. There is poor correlation between tumor necrosis and conventional methods of response assessment, which poses questions of how best to quantify efficacy of targeted therapy. Choi criteria might also be considered as an alternative to RECIST to evaluate the effects of necrosis in patients with advanced well-differentiated neuroendocrine tumors because it takes in account necrosis. They are based on internal changes, tumor size (>10% decrease) OR tumor density (>15 HU decrease) Should we abandon RECIST? No, RECIST is still used in many clinical trials and is useful for assessing tumor progression. Then ENETS Modified Criteria are on study, associating RECIST and CHOI criteria. We probably should go back from RECIST 1.1 to RECIST, because RECIST 1.1 studies 2 targets per organ and Net liver mets have a large discrepancy of response. To choose 5 targets probably will lower discrepancy. Functional tools, mainly obtained with MR imaging, show interesting results but lack standardization. In the future, they will be combined to morphologic criteria Disclosure: All authors have declared no conflicts of interest.