Combined prognostic value of post-surgery circulating tumor DNA and tumor-stroma ratio in patients with stage III colon cancer treated with adjuvant chemotherapy.

Abstract

3134 Background: Patients with stage III colon cancer (CC) are routinely treated with resection followed by adjuvant chemotherapy (ACT). About half of patients are cured by surgery and hence overtreated with ACT, yet another ~30% experience recurrence and are currently undertreated. Only ~20% of patients are cured by ACT and we are unable to identify these patients. Prognostic value of circulating tumor DNA (ctDNA) and the tumor-stroma ratio (TSR) has been shown in separate studies. This study aimed to integrate these biomarkers with pTNM substage to better predict outcome in stage III CC patients treated with ACT. Methods: Patients with stage III CC who received radical resection followed by ACT were selected from the Prospective Dutch ColoRectal Cancer cohort (PLCRC) substudy PROVENC3 (Rubio-Alarcon AACR 2024). Blood was collected between surgery and ACT, to determine ctDNA status using Labcorp Plasma Detect. Based on a diagnostic H&E slide from the CC resection, the TSR was determined by a trained observer according to the United study (Polack ESMO open 2024). A stroma content of ≤50% was considered low and >50% high. The primary outcome was recurrence risk (RR), calculated from date of resection. Results: In the overall cohort (N = 207), the 3-year RR was 23.4% [17.3-29.1]. In total, 88 patients (43%) were stroma-high and had a higher recurrence risk (3-year RR 33.1% [22.5-42.3]) than the 119 stroma-low patients (3-year RR 16.0% [8.9-22.5]; HR 2.7 [1.6-4.6]). CtDNA was detectable after surgery in 28 patients (13.5%; HR 5.8 [3.3-10]), of whom 11 (39%) were stroma-high. T4/N2 stage was observed in 82 patients (HR 2.9 [1.7-5.0]), of whom 46 (56%) were stroma-high. TSR (HR 2.6 [1.5-4.6]) had added prognostic value to ctDNA (HR 7.6 [4.3-13]) and pTNM substage (HR 2.9 [1.7-5.0]) in a multivariable cox model (LRT p<0.001). Patients with no detectable ctDNA and stroma-low T1-3N1 CC were at low recurrence risk (N = 71; 3-year RR 2.9% [0-6.8]). In comparison, patients with no detectable ctDNA and a tumor that was either stroma-high or T4/N2 were considered intermediate risk (N = 68; 3-year RR 17.2% [7.4-26.0]; HR 5.4 [1.5-19]). Patients with detectable ctDNA and/or stroma-high T4/N2 CC had a high risk (N = 68; 3-year RR 50.2% [36.7-60.8]; HR 19 [5.9-62]). Conclusions: The tumor-stroma ratio has added value to post-surgery ctDNA and pTNM substage in predicting outcome in stage III CC patients treated with ACT. The recurrence risk in the third of patients with no detectable ctDNA and stroma-low T1-3N1 CC was only 3%. It is of interest to investigate whether this low risk would persist in a cohort treated with surgery only, to suggest whether these patients could be spared ACT in the future. The third of patients with detectable ctDNA, and/or stroma-high T4/N2 CC, had a 50% recurrence risk despite ACT, highlighting the need for alternative adjuvant treatment options for these patients.