ctDNA versus 18F-FDG PET-CT in predicting long-term disease control in patients with advanced melanoma undergoing immune checkpoint blockade therapy.

Abstract

9559 Background: Imaging remains the gold standard for assessing response to systemic immunotherapy in patients with advanced melanoma. Several studies have demonstrated a strong correlation between metabolic response evaluation using 18F-FDG PET-CT and long-term prognosis in patients with advanced melanoma treated with immunotherapy. Meanwhile, ctDNA kinetics has emerged as a promising alternative method to support the evaluation of patients receiving immunotherapy. Methods: We prospectively collected blood samples for liquid biopsy assessments using next-generation sequencing (NGS-Ion S5 platform; Thermo Fisher) to detect tumor somatic mutations with a 409-gene panel, and tumor mutations were tracked in plasma samples collected from advanced melanoma patients undergoing immune checkpoint blockade therapy at AC Camargo Cancer Center at baseline, Day 30 (D30), and Day 60 (D60). ctDNA was considered positive if the variant allelic fraction (VAF) exceeded 0.5% and was at least twice that in negative controls. ctDNA results were compared with Day-90 PET-CT and correlated with long-term disease control outcomes. Assessments at D30 and D60 were classified into three categories: molecular responders (MR), molecular non-responders (MNR), and negative pattern (NP), following the framework of the KEYNOTE-942 study. Results: This analysis included 15 stage IV melanoma patients treated with nivolumab (3 mg/kg) and ipilimumab (1 mg/kg). Seven patients (47%) showed an objective response on PET-CT. After a median follow-up of 26 months (range: 1–44 months), 31% of patients exhibited controlled disease. PET-CT demonstrated 78% accuracy in predicting long-term disease status (controlled vs. uncontrolled). Baseline ctDNA analysis showed that 10 patients (67%) were ctDNA-positive. The accuracy of baseline ctDNA (positive vs. negative) in predicting long-term disease control status was 71%. On D30, 13 cases were analyzed and classified as follows: 4 (MR), 6 (MNR), and 3 (NP). The accuracy of the D30 liquid biopsy analysis in predicting long-term disease status was only 31%. On D60, 11 cases were analyzed and classified as follows: 5 (MR), 4 (MNR), and 2 (NP). The accuracy of the D60 liquid biopsy analysis in predicting long-term disease status was 73%. Conclusions: ctDNA status at baseline and D60, as well as 18F-FDG PET-CT at D90, appear to have similar accuracy in predicting long-term disease control in patients with advanced melanoma treated with immune checkpoint blockade .