Added value of tumor–stroma ratio to postsurgery circulating tumor DNA and pTN stage in risk stratification of patients with stage III colon cancer treated with adjuvant chemotherapy

TL;DR

This study shows that combining tumor-stroma ratio with circulating tumor DNA and tumor stage improves risk prediction for colon cancer patients after surgery and chemotherapy.

Contribution

The study demonstrates that tumor-stroma ratio adds value to circulating tumor DNA and pTN stage in risk stratification of stage III colon cancer patients.

Findings

Postsurgery ctDNA-positive patients have a 65% 3-year recurrence risk.

Within ctDNA-negative patients, those with pT4/N2 and stroma-high tumors have a 40% recurrence risk.

One-third of patients with ctDNA-negative, pT1-3N1, and stroma-low tumors have <3% recurrence risk.

Abstract

Patients with stage III colon cancer (CC) are routinely treated with resection followed by adjuvant chemotherapy (ACT). Half of patients are cured by surgery alone and overtreated with ACT, yet another ∼30% experience disease recurrence. Upfront risk stratification requires biomarkers beyond the conventional pathological stage (pTN). Detection of postsurgery circulating tumor DNA (ctDNA) is indicative of minimal residual disease and prognostic of recurrence. However, its negative predictive value is insufficient to guide treatment de-escalation. This study aimed to investigate whether adding tumor–stroma ratio (TSR) can improve patient risk stratification.

This study included 206 patients from PLCRC-PROVENC3 based on radical resection of stage III CC followed by ACT. Postsurgery ctDNA status was determined using Labcorp® Plasma Detect™. On a hematoxylin–eosin resection section, the TSR was scored by trained observers and dichotomized as stroma-low (≤50% stroma) versus stroma-high (>50%). The primary outcome was time to recurrence in univariable and multivariable Cox analyses to inform risk-group stratification, reporting hazard ratios (HR) and recurrence risks (RR).

Postsurgery ctDNA was the strongest predictor [n = 26, 3-year RR 65.4% (95% CI 41.3% to 79.6%) versus 16.8% (95% CI 11.0% to 22.3%); HR 6.1 (95% CI 3.4-10.8)], followed by pT4/pN2 [n = 80, HR 3.0 (95% CI 1.7-5.2)] and a stroma-high tumor [n = 88, HR 3.0 (95% CI 1.7-5.2)]. Within the ctDNA-negative subgroup (n = 180), we identified a low-risk group [pT1-3N1 and stroma-low; n = 72, 3-year RR 2.9% (95% CI 0% to 6.7%)], intermediate-risk group [either pT4/N2 or stroma-high; n = 68, 3-year RR 17.2% (95% CI 7.4% to 26.0%), HR 8.2 (95% CI 1.9-36.2)], and high-risk group [pT4/N2 and stroma-high; n = 40, 3-year RR 40.3% (95% CI 22.9% to 53.9%), HR 21.5 (95% CI 5.0-92.3)].

Adding TSR to ctDNA and pTN stage improved risk stratification of stage III CC patients receiving ACT. One-third of the patients had none of the biomarkers and could be considered for de-escalation based on their very low RR. In addition to ctDNA-positive patients, ctDNA-negative patients with a pT4/N2 stroma-high tumor may require treatment escalation to reduce their high RR.

•Tumor–stroma ratio and ctDNA have complementary value in stratifying RR after ACT in stage III CC patients.•Postsurgery ctDNA-positive patients have the highest 3-year RR of 65%.•Within the ctDNA-negative group, pathological T4/N2 stage and a stroma-high tumor identifies a high-risk group with 40% RR.•Based on ctDNA-negativity, pT1-3N1 and stroma-low, one-third of the patients are identified as very low risk with <3% RR.

Tumor–stroma ratio and ctDNA have complementary value in stratifying RR after ACT in stage III CC patients.

Postsurgery ctDNA-positive patients have the highest 3-year RR of 65%.

Within the ctDNA-negative group, pathological T4/N2 stage and a stroma-high tumor identifies a high-risk group with 40% RR.

Based on ctDNA-negativity, pT1-3N1 and stroma-low, one-third of the patients are identified as very low risk with <3% RR.

Figures