Feasibility of circulating tumor DNA (ctDNA) to guide organ preservation in patients with node-negative rectal cancer undergoing neoadjuvant chemotherapy, excision, and observation in the phase II CCTG CO.28 trial

180 Background: CCTG CO.28 evaluated 3 months of neoadjuvant CAPOX or FOLFOX followed by transanal excision surgery (TES) in patients (pts) with T1-T3, N0 rectal cancers with the goal of organ preservation. Total mesorectal excision (TME) was recommended for inadequate downstaging or high-risk features on TES pathology. Detection of ctDNA is associated with a high rate of recurrence in high risk resected colorectal cancer. However, the sensitivity of ctDNA in early-stage rectal cancer is unclear. Methods: Tissue CGP was performed retrospectively on pretreatment resected tumor using Foundation OneCDx (F1CDx), followed by MRD detection using FoundationOneTracker. Briefly, variants were selected from F1CDx using an algorithm that filters out non-tumor derived variants (e.g. germline). A tumor-informed personalized multiplex PCR-next generation sequencing assay, consisting of up to 16 variants, was used to detect ctDNA and quantify plasma mean variant allelic frequency (VAF) and mean tumor molecules per mL of plasma (MTM/mL). FoundationOneTracker was performed on blood samples collected pre/post-neoadjuvant chemotherapy. Buffy coat sequencing was performed at FMI to validate the monitoring variant selection algorithm. Results: 52 pts underwent F1CDx and FoundationOneTracker analysis. KRAS mutations were observed in 60% (n =31) of cases; MSI-H and BRAF V600E mutations were not detected. From 1,580 distinct variants detected in tissue and paired buffy coat sequencing, 985 had VAF >30% in the buffy coat specimen and were considered germline. The variant selection algorithm filtered 99.8% (n = 983) of germline variants; the remaining 2 were not selected for primer design. A median of 7 variants (range 2-16) were tracked per pt. Detectable ctDNA was found in 0% (0/6), 19% (6/32) and 55% (6/11) of treatment naïve T1, T2 and T3ab cases, respectively, and 83% of those which were initially positive (10/12) cleared ctDNA post chemotherapy. Three pts had detectable ctDNA pre-excision, including 1 who was negative pretreatment. From 42 pts with pre/post-treatment plasma, 19 pts were recommended for TME; 2 had detectable ctDNA, 5 had ctDNA clearance, and 12 had no ctDNA detection. Conclusions: Node-negative rectal cancer exhibited ctDNA shed in a minority of pts, with most clearing ctDNA on chemotherapy before their TES. The dynamic ctDNA signal we demonstrate could supplement clinical factors in informing organ preservation in very early-stage rectal cancer. Larger studies are needed to determine if this information can be used to personalize therapy. Clinical trial information: NCT03259035 . [Table: see text]