ctDNA Status Is Prognostic for DFS With Celecoxib in Stage III Resected Colon Cancer

Positive circulating tumor DNA (ctDNA) status was strongly associated with worse disease-free survival (DFS) outcomes vs negative ctDNA status in patients with stage III resected colon cancer, although patients with ctDNA-positive disease achieved significantly better DFS outcomes with celecoxib (Celebrex) vs placebo, according to findings from a subgroup analysis of the phase 3 CALGB (Alliance)/SWOG 80702 trial (NCT01150045), which were presented at the 2025 Gastrointestinal Cancers Symposium.1

Among patients with ctDNA-negative colon cancer following surgery and before treatment with celecoxib or placebo (n = 767), the estimated 3-year DFS rate was 86.5% (95% CI, 84.0%-89.1%). Among those with ctDNA-positive disease (n = 173), this estimated rate was 33.7% (95% CI, 27.1%-41.8%; log-rank P < .0001).

ctDNA status was also prognostic of overall survival (OS) benefit in this patient population. In the ctDNA-negative population, the estimated 5-year OS rate was 91.5% (95% CI, 89.5%-93.6%). In the ctDNA-positive population, this estimated rate was 52.6% (95% CI, 45.3%-61.0%; log-rank P < .0001).

“In a subset of patients enrolled in CALGB/SWOG 80702, ctDNA status after surgery and prior to starting adjuvant therapy was highly prognostic of DFS and OS,” Jonathan Nowak, MD, PhD, lead study author, said in a press briefing ahead of the presentation of the data.

“These results were overall expected and are broadly consistent with [findings from] prior studies that have looked at prognosis in colon cancer,” Nowak emphasized during a subsequent presentation of the data in an oral abstract session.

Nowak is an investigator at the Hale Family Center for Pancreatic Cancer Research at Dana-Farber Cancer Institute and an assistant professor of pathology at Harvard Medical School in Boston, Massachusetts.

Trial Background and Rationale

“Approximately one-third of patients diagnosed with colon cancer have regional lymph node involvement,” Nowak said in the press briefing. “Despite optimal surgery and adjuvant chemotherapy, approximately 20% to 70% of patients with stage III disease will have a recurrence. Additional strategies beyond standard chemotherapy are needed to reduce the risk of recurrent disease and improve survival. One promising option is to use post-resection ctDNA status, which can tell us if any residual micrometastatic disease is present in order to help guide adjuvant treatment decisions.”

Prior research has shown that patients with colon cancer and colon cancer survivors who receive aspirin or COX-2 inhibitors experience favorable DFS and OS outcomes. For instance, a prospective, observational study showed that among 843 evaluable patients with stage III colon cancer who were enrolled in an adjuvant chemotherapy trial, those who used COX-2 inhibitors had improved 5-year DFS rate (HR, 0.47; 95% CI, 0.24-0.91) and OS rate (HR, 0.26; 95% CI, 0.08-0.81) rates compared with nonusers.2

Trial Design and Previously Reported Findings

The CALGB/SWOG 80702 trial formally tested whether COX-2 inhibition with celecoxib improved survival in patients with colon cancer.1 To be eligible for trial enrollment, patients needed to have resected stage III colon adenocarcinoma without metastatic disease, and at least 1 pathologically confirmed positive lymph node or N1c disease per American Joint Committee on Cancer version 7 criteria. Patients were not eligible for enrollment if they used nonsteroidal anti-inflammatory drugs at any dose more often than 2 times per week or aspirin more than 325 mg 3 times per week. However, patients using low-dose aspirin not exceeding a dose of 100 mg per day were permitted to enroll.

Patients were randomly assigned to receive daily treatment with either celecoxib at 400 mg or placebo in combination with either 6 or 12 treatments of FOLFOX (leucovorin calcium, fluorouracil, and oxaliplatin) as follows: placebo plus 12 FOLFOX treatments (arm A), celecoxib plus 12 FOLFOX treatments (arm B), placebo plus 6 FOLFOX treatments (arm C), or celecoxib plus 6 FOLFOX treatments (arm D). Celecoxib or placebo was continued for 3 years from study drug initiation.

The trial had a target sample size of 2500 patients; 2526 total patients were enrolled, and 2524 patients were included.3 DFS served as the trial’s primary end point. In the primary analysis, investigators observed no statistically significant DFS difference with celecoxib vs placebo (HR, 0.89; 95% CI, 0.76-1.03; stratified log-rank P = .12). Furthermore, the effects of celecoxib treatment were not significantly different based on assigned adjuvant FOLFOX duration.

ctDNA Analysis Rationale and Baseline Characteristics

In the Gastrointestinal Cancers Symposium analysis, Nowak noted that the HR and Kaplan-Meier curves for DFS in the primary analysis indicated that a subgroup of patients may derive benefit from adjuvant celecoxib.1 

To identify whether any patient subgroup derived greater benefit from celecoxib treatment compared with the overall population, investigators retrospectively performed a ctDNA analysis using the Signatera minimal residual disease assay on banked plasma specimens from trial patients that were collected after surgery and before trial enrollment.

Of the patients included in the trial’s primary analysis, 1752 consented to biospecimen collection and had tumor tissue successfully banked, 1197 had whole exome data successfully generated, and 1011 had sufficient plasma for ctDNA testing. Among the 940 patients with evaluable ctDNA, 18.4% were ctDNA positive and 81.6% were ctDNA negative.

Clinicopathologic patient characteristics associated with ctDNA positivity were few but included higher T stage, higher N stage, and male sex. Notably, a borderline association with ctDNA positivity was observed with the assigned oral agent (celecoxib vs placebo), despite similar numbers of patients were assigned to each of the treatment arms. In the total population, 50.4% of patients received celecoxib and 49.6% of patients received placebo. In the ctDNA-negative population, the rates of celecoxib and placebo user were 48.9% and 51.1%, respectively. However, in the ctDNA-positive population, the rate of celecoxib users was slightly higher than the rate of placebo users, at 57.2% vs 42.8%. Importantly, in the abstract session, Nowak explained that the plasma that was used for ctDNA testing was collected before patients initiated the study treatment.

Additional Efficacy Findings

When patients were stratified by treatment with celecoxib vs placebo, investigators showed that patients with ctDNA-negative disease at the time of study enrollment had similar DFS outcomes regardless of whether they received celecoxib or placebo, with estimated 3-year DFS rates of 87.4% (95% CI, 84.0%-91.0%) and 85.6% (95% CI, 82.0%-89.4%), respectively (HR, 0.76; 95% CI, 0.54-1.08; P = .1293). However, among patients with ctDNA-positive disease, a significant DFS benefit was seen in the celecoxib arm vs the placebo arm, with estimated 3-year DFS rates of 41.0% (95% CI, 32.2%-52.2%) vs 22.6% (95% CI, 14.3%-35.5%), respectively (HR, 0.55; 95% CI, 0.39-0.80; P = .0013).

Similarly, OS outcomes were not significantly different between patients with ctDNA-negative disease who received celecoxib or placebo, with estimated 5-year OS rates of 91.8% (95% CI, 88.9%-94.7%) and 91.3% (95% CI, 88.4%-94.3%), respectively (HR, 0.86; 95% CI, 0.55-1.35; P = .5098). However, ctDNA positivity appeared to predict an OS benefit of adjuvant celecoxib over placebo. Among patients with ctDNA-positive disease, the estimated 5-year OS rates were 61.6% (95% CI, 52.4%-72.4%) vs 39.9% (95% CI, 29.6%-53.8%), respectively (HR, 0.58; 95% CI, 0.38-0.90; P = .0135).

In the press briefing, Nowak highlighted that findings from this subgroup analysis “also held true in multivariable adjusted models for both DFS and OS when the factors that are commonly known to predict colon cancer survival [were] included in the model.”

In the multivariable analysis, among ctDNA-negative patients, the HR for DFS with celecoxib vs placebo was 0.76 (95% CI, 0.53-1.08; adjusted P = .1262), and the HR for OS was 0.84 (95% CI, 0.53-1.34; adjusted P = .4593). Among ctDNA-positive patients, the HR for DFS with celecoxib vs placebo was 0.63 (95% CI, 0.43-0.92; adjusted P = .0167), and the HR for OS was 0.63 (95% CI, 0.40-0.98; adjusted P = .0419).

Given the anti-inflammatory properties of celecoxib and the differing tumor microenvironments between microsatellite-stable (MSS) and microsatellite instability–high disease, investigators evaluated the effect of microsatellite status on outcomes with celecoxib vs placebo.

Specifically, in patients with MSS tumors that were ctDNA negative, the estimated 3-year survival rates were 87.7% (95% CI, 84.2%-91.5%) with celecoxib and 85.4% (95% CI, 81.6%-89.5%) with placebo (HR, 0.75; 95% CI, 0.52-1.09; P = .1307). In patients with MSS tumors that were ctDNA positive, a benefit with celecoxib was observed, with an estimated 3-year survival rate of 39.7% (95% CI, 30.6%-51.5%) vs 22.3% (95% CI, 13.9%-35.8%) with placebo (HR, 0.59; 95% CI, 0.40-0.86; P = .0055).

Furthermore, Nowak highlighted in the abstract session that research such as the phase 3 ALASCCA trial (NCT02647099) has shown that the presence of activating PIK3CA mutations in tumors may be predictive of benefits of adjuvant NSAID use. In this CALGB/SWOG 80702 subgroup analysis, a benefit with celecoxib use was observed among ctDNA-positive patients with PIK3CA wild-type tumors. In the ctDNA-positive population, the estimated 3-year survival rates were 40.3% (95% CI, 30.7%-52.9%) with celecoxib and 24.2% (95% CI, 14.9%-39.3%) with placebo (HR, 0.61; 95% CI, 0.41-0.92; P = .0166). Of patients who were ctDNA negative, the estimated 3-year survival rates were 86.7% (95% CI, 82.7%-91.0%) with celecoxib and 84.8% (95% CI, 80.7%-89.0%) with placebo (HR, 0.81; 95% CI, 0.56-1.18; P = .2777).

Limitations and Next Steps

Nowak noted that limitations of this research include the fact that it is a post hoc analysis. Moreover, ctDNA status was not used to select patients for treatment with adjuvant celecoxib, and only a subgroup of patients in the CALGB/SWOG 80702 trial were included.

“These results suggested a potential role for ctDNA in determining which patients should consider celecoxib in addition to standard FOLFOX chemotherapy as an adjuvant therapy option,” Nowak emphasized during the abstract session.

Nowak concluded by explaining that sensitivity and subgroup analyses of this trial are ongoing, as well as studies evaluating the predictive value of ctDNA for 3 vs 6 months of adjuvant FOLFOX in this patient population.

Disclosures: Dr Nowak reports performing consulting or advisory roles for Leica B and receiving research funding from Natera.

References

1. Nowak JA, Shi Q, Twombly T, et al. Prognostic and predictive role of circulating tumor DNA (ctDNA) in stage III colon cancer treated with celecoxib: findings from CALGB (Alliance)/SWOG 80702. J Clin Oncol. 2025;43(4):LBA14. Presented at: 2025 Gastrointestinal Cancers Symposium; January 23-25, 2025; San Francisco, CA. LBA14.
2. Ng K, Meyerhardt JA, Chan AT, et al. Aspirin and COX-2 inhibitor use in patients with stage III colon cancer. J Natl Cancer Inst. 2014;107(1):345. doi:10.1093/jnci/dju345
3. Meyerhardt JA, Shi Q, Fuchs CS, et al. Effect of celecoxib vs placebo added to standard adjuvant therapy on disease-free survival among patients with stage II colon cancer: the CALGB/SWOG 80702 (Alliance) randomized clinical trial. JAMA. 2021;325(13):1277-1286. doi:10.1001/jama.2021.2454