Adjuvant Aspirin Misses DFS End Point in Stage III CRC

Three years of adjuvant low-dose aspirin did not significantly improve disease-free survival (DFS) over placebo in patients with stage III colorectal cancer (CRC) who had undergone curative resection and received standard adjuvant chemotherapy, according to the primary analysis of the double-blind, placebo-controlled, phase 3 EPISODE-III (JCOG1503C) trial (jRCTs031180009) presented at the 2026 American Society of Clinical Oncology (ASCO) Annual Meeting.¹

The 3-year DFS rate was 78.8% (95% CI, 74.7%-82.3%) with aspirin 100 mg daily for 3 years vs 75.4% (95% CI, 71.1%-79.2%) with placebo, which was a 3.4-percentage-point numerical advantage that translated to an HR of 0.84 (95% CI, 0.65-1.09; one-sided P = .099), short of the prespecified one-sided α = 0.05 boundary. Relapse-free survival (RFS) trended in the same direction (HR, 0.87; 95% CI, 0.66-1.14), and overall survival (OS) data were immature (HR, 1.02; 95% CI, 0.65-1.60), with the final OS analysis planned for 2028. A prespecified biomarker analysis of PI3K/PIK3CA-altered tumors where 2 recently reported randomized trials have suggested benefit is ongoing.

“Low-dose aspirin did not significantly improve DFS in unselected stage III colorectal cancer,” lead investigator Atsuo Takashima, MD, PhD, of the National Cancer Center Hospital in Tokyo, said when presenting the data on behalf of the Colorectal Cancer Study Group in the Japan Clinical Oncology Group (JCOG). Biomarker analyses of the prospectively collected biobank samples “are ongoing and will inform a planned collaborative meta-analysis of randomized aspirin trials," he said.

Decades of observational data have suggested that post-diagnosis aspirin use may improve outcomes in colorectal cancer, but randomized confirmatory evidence has been mixed. The phase 3 ASCOLT trial (NCT00565708) in unselected colorectal cancer reported a DFS HR of 0.91 (95% CI, 0.73-1.13).² Two recent randomized trials in patients with PI3K-altered tumors, the phase 3 ALASCCA (NCT02647099) and phase 3 SAKK 41/13 (NCT02467582) trials, have suggested a more substantial benefit in a biomarker-defined subgroup.^3,4 EPISODE-III tested whether 100 mg of aspirin daily for 3 years, started concurrently with adjuvant chemotherapy, could improve outcomes in an unselected stage III CRC population.

EPISODE-III was conducted by JCOG and enrolled 882 patients with stage III CRC across Japanese centers between March 2018 and October 2022. Eligible patients had R0 resection with D2/D3 dissection, were 20 to 80 years old, had ECOG performance status of 0 or 1, were within 8 weeks after surgery, and had no history of regular antiplatelet or nonsteroidal anti-inflammatory drug use, gastrointestinal ulcer, asthma, inflammatory bowel disease, or bleeding.

Patients were randomly assigned 1:1 to receive 100 mg of aspirin daily or matched placebo for 3 years, started concurrently with adjuvant chemotherapy (capecitabine, CAPOX [capecitabine and oxaliplatin], or modified FOLFOX6 [leucovorin, fluorouracil, and oxaliplatin]), and stratified by institution, sex, stage (IIIA/IIIB/IIIC), and planned oxaliplatin use.

The primary end point was DFS, defined as time from randomization to relapse, secondary cancer, or death from any cause. Secondary end points included OS, RFS, safety, and relative dose intensity. The trial was designed to accrue 279 DFS events to provide 80% power to detect an assumed HR of 0.741, which corresponded to a 3-year DFS of 80% with aspirin vs 74% with placebo, at a one-sided α of 0.05.

Baseline characteristics were balanced. The median age was 65 years, and 51% of patients were male in the placebo arm, compared with 64 years and 51% in the aspirin arm, respectively. Additionally, 69% and 68%, respectively, had left-sided primary tumors and 19% and 19% had stage IIIA disease, 63% and 62% stage IIIB, and 18% and 19% stage IIIC. The planned adjuvant chemotherapy was capecitabine in 30% and 29%, CAPOX in 69% and 70%, and FOLFOX in less than 1% for both arms.

At the primary analysis, 124 DFS events had occurred in the placebo arm vs 107 in the aspirin arm. The observed effect size was smaller than the prespecified assumption of HR 0.741, and the result did not cross the trial's one-sided α = 0.05 boundary.

Prespecified sensitivity analyses for DFS were directionally consistent: unstratified HR of 0.85 (95% CI, 0.66-1.11), censored at post-protocol therapy HR of 0.81 (95% CI, 0.62-1.05), and HR of 0.85 (95% CI, 0.66-1.11) in the eligible-patient population (n = 876).

RFS followed the same pattern, with 113 events on placebo and 100 on aspirin (HR, 0.87; 95% CI, 0.66-1.14). OS data were immature, with 38 deaths on placebo and 40 on aspirin.

Among the 605 patients with left-sided primary tumors, the DFS HR was 0.73 (95% CI, 0.53-1.00), whereas among 276 patients with right-sided primary tumors, the HR was 1.19 (95% CI, 0.75-1.89). Among the 169 patients with stage IIIA disease, the DFS HR was 0.32 (95% CI, 0.10-1.00), although event counts in that subgroup were limited.

Aspirin was well tolerated. During the adjuvant chemotherapy period, adverse event rates were largely driven by chemotherapy and balanced between arms, including any-grade neutropenia (73% placebo vs 75% aspirin), thrombocytopenia (68% vs 68%), nausea (40% vs 42%), and sensory neuropathy (63% vs 67%).

Lower GI hemorrhage was numerically more frequent with aspirin: 7 events (2%) on aspirin vs 1 (0.2%) on placebo during chemotherapy (1 grade 3 or higher event in the aspirin arm), and 6 (1%) vs 4 (1%) after chemotherapy completion (3 grade 3 or higher events in the aspirin arm vs 0 in placebo). GI ulcers occurred in 3 patients (1%) on aspirin after chemotherapy completion vs 0 on placebo. One treatment-related death occurred on aspirin during the chemotherapy phase, which was due to ischemic heart disease. The median relative dose intensity of assigned aspirin or placebo was 98.6% in both arms.

EPISODE-III joins ASCOLT in failing to show a statistically significant DFS benefit for adjuvant aspirin in an unselected colorectal cancer population. The signal for a clinically meaningful aspirin effects now appears to be concentrated in molecularly defined subgroups, particularly tumors with PI3K pathway alterations.

ALASCCA reported a recurrence-rate reduction in PIK3CA-altered tumors,³ and SAKK 41/13 reported a 43% numerical DFS improvement in a similar biomarker-selected population, although the latter trial closed prematurely after only 19 events.⁴ Takashima said the prospectively collected EPISODE-III biobank samples will enable a PI3K/PIK3CA biomarker analysis, and the investigators are planning participation in a collaborative meta-analysis of randomized aspirin trials in colorectal cancer.

References

1. Takashima A, Hirano Y, Shiozawa M, et al. Adjuvant aspirin for stage III colorectal cancer after curative resection: primary analysis of the randomized double-blind placebo-controlled phase III trial (EPISODE-III: JCOG1503C). J Clin Oncol. 2026;44(suppl 16):LBA3508. doi:10.1200/JCO.2026.44.16_suppl.LBA3508
2. Chia JWK, Segelov E, Deng Y, et al. Aspirin after completion of standard adjuvant therapy for colorectal cancer (ASCOLT): an international, multicentre, phase 3, randomised, double-blind, placebo-controlled trial. Lancet Gastroenterol Hepatol. 2025;10(3):198-209. doi:10.1016/S2468-1253(24)00387-X
3. Martling A, Hed Myrberg I, Nilbert M, et al. Low-dose aspirin for PI3K-altered localized colorectal cancer. N Engl J Med. 2025;393(11):1051-1064. doi:10.1056/NEJMoa2504650
4. Güller U, Hayoz S, Horber D, et al. Adjuvant aspirin treatment in PIK3CA-mutated colon cancer patients: the SAKK 41/13 prospective randomized placebo-controlled double-blind trial. Clin Cancer Res. 2025;31(15):3142-3149. doi:10.1158/1078-0432.CCR-24-4048