Plasma-informed minimal residual disease (MRD) assay: A multicenter prospective study in Japanese patients with stage II colorectal cancer.

Abstract

161 Background: ̃80% of stage II colorectal cancer (CRC) can be cured by surgery alone. However, adjuvant chemotherapy is recommended for patients with high risk features such as bowel obstruction, < 12 lymph nodes examined and T4 tumors. Traditional pathological grading and biomarkers such as carcinoembryonic antigen has limited sensitivity. Several reports indicated circulating tumor DNA (ctDNA) may represent a promising prognostic factor to assess MRD as a factor for prediction of recurrence after surgery. Here, we present a proof-of-concept study for the development of a novel plasma-based highly sensitive Next Generation Sequencing (NGS) panel using SafeSEQ technology in operable CRC Japanese patients. Methods: This multicenter prospective study recruited patients diagnosed as operable clinical stage II CRC (n = 46) with pre- and post- (4̃6 weeks) operative plasma samples collected between Nov, 2019 and Jan, 2021. ctDNA were extracted and a 14-gene NGS panel was used to analyze single nucleotide variants (SNVs) and Indels covered by gene-specific amplicons. MRD variant was defined as same variant detected in both pre- and post- operative plasma samples. Tissue NGS by a 500-gene panel was also performed in a small number of tissue samples (n = 5) to compare the concordance of plasma and tissue variants. Results: Pre- and post-operative ctDNA status of 46 patients were analyzed. ctDNA positive was observed in 69.6% (32/46, 95%CI 55.2, 81.0) pre- and 34.8% (16/46, 95%CI 22.7, 49.3) post- samples. AKT1, CTNNB1, NRAS, POLE and PPP2R1A mutation were not detected in this study. TP53 mutation was most frequently detected in both pre- (22/46) and post- (11/46) samples, whereas APC mutation was ranked 2nd in pre- (15/46) but none in post- samples. A combined 96 variants were detected in all samples, in which 76 of them were < 0.5% mutant allele frequency (MAF). MRD variants were detected in 17.4% (8/46, 95%CI 8.82, 30.99) post- samples. Evaluation of positive percentage agreement between tissue and pre- plasma samples in three patients show that a total 7 variants detected in plasma, and 3 of them were detected in tissue samples. Conclusions: This study assesses the feasibility of a plasma-informed NGS panel by evaluating pre- and post- operative plasma samples. The presence of variants with < 0.5% MAF detected in this study indicate a highly sensitive method is required for accurate MRD detection. Further observation is required to explore the relationship between MRD variant and clinical outcome such as 2-year progression-free survival.