(P 114) PROGNOSTIC IMPACT OF CIRCULATING TUMOR DNA DETECTION IN FIRST-LINE TREATMENT OF ADVANCED GASTROINTESTINAL STROMAL TUMOR

Objective: Detection of circulating tumor DNA (ctDNA) is a novel and noninvasive approach in the clinical management of patients with advanced gastrointestinal stromal tumor (GIST). Most studies so far have focused on patients after progression on imatinib and the impact of ctDNA detection in first-line treatment with imatinib remains unknown. The aim of this study was to investigate whether ctDNA detection could be used as a biomarker to predict outcome and to monitor treatment response and disease relapse in first-line treatment of advanced GIST.

Methods: The study cohort consisted of 28 patients with inoperable, advanced GIST from Oslo University Hospital in Norway. All patients had tumors with a KIT or PDGFRA mutation and were included prior to the start of first-line imatinib treatment. Plasma samples were collected at baseline, after 2-4 weeks, at each evaluation and at disease progression. The median number of samples analyzed per patient was 4 (range 3-9). Sequencing libraries were made from 15 ng of cfDNA using the Invitae LiquidPlex assay, providing targeted exon sequencing of 28 cancer-related genes.

Results: ctDNA was detected in 16 of 28 (57%) baseline samples collected prior to the start of systemic treatment. The median mutated allele frequency in the positive samples was 2.5% (range 0.1-53.0%). A reduction in ctDNA levels was seen in all patients after start of imatinib treatment. Of 15 evaluable patients with detectable ctDNA at baseline, 7 (47%) had no detectable levels within 90 days of treatment. Disease progression by RECIST v1.1 was recorded in 11 patients after a median time of 27 months (range 3-72 months). Detection of ctDNA at baseline was a predictor of inferior progression-free survival (Fig. 1a; P=0.037; HR 4.50; 95% CI 0.96-21.02). Six patients died during follow-up and all had detectable ctDNA at baseline (Fig 1b; P=0.016). There was no association between ctDNA dynamics after start of imatinib and progression-free or overall survival. Plasma samples were available from five patients at the time of disease progression, of which two had detectable ctDNA. The primary KIT exon 11 mutation was detected in both samples and one sample had additional secondary mutations in exons 13, 17 and 18. None of the samples from the time point prior to radiological progression were ctDNA-positive.

Conclusion: Patients with detectable ctDNA prior to the start of first-line imatinib treatment had inferior outcome compared to patients without detectable ctDNA. ctDNA levels decreased in all patients after starting imatinib, but ctDNA dynamics was not associated with outcome. Primary and secondary mutations were detected at the time of disease progression, but not in samples collected prior to radiological progression. Our findings support the use of ctDNA detection as a prognostic biomarker in advanced GIST, while its role in monitoring early disease relapse might be questionable.