Aspirin's CRC Benefit Narrows to PIK3CA Exon 9/20 Mutations

Few cancer therapies cost less than a daily aspirin, which is why the question of whether it keeps colorectal cancer from returning has stayed open for 15 years. The translational arm of the ASCOLT trial has answered part of it, and the answer is narrower than the guideline that arrived ahead of it.¹

A Null Parent Trial With Subgroups Still Open

ASCOLT itself reported no benefit. Across 66 centers in 11 countries and territories, 1587 patients with Dukes' C or high-risk Dukes' B colon cancer, or Dukes' B or C rectal cancer, were randomly assigned after standard adjuvant chemotherapy to aspirin 200 mg daily or placebo for 3 years. Five-year disease-free survival (DFS) reached 77.0% (95% CI, 73.6%-80.0%) with aspirin against 74.8% (95% CI, 71.3%-77.9%) with placebo (hazard ratio [HR], 0.91; 95% CI, 0.73-1.13; P = .38).²

That confidence interval left room for a reduction in events up to 27%, and epidemiologic work had long suggested any benefit might concentrate in tumors carrying PIK3CA mutations or overexpressing cyclooxygenase-2 (COX-2). Two later randomized trials, SAKK 41/13 and ALASCCA, reported reduced recurrence with adjuvant aspirin in patients with PI3K-pathway mutations. The ASCOLT investigators note that the National Comprehensive Cancer Network (NCCN) guidance was subsequently updated to recommend adding low-dose aspirin to adjuvant therapy for patients with any somatic PI3K-pathway mutation, not just PIK3CA. ASCOLT had prespecified both of its subgroups before any of that landed.¹

Tumor Tissue From 465 of 1587 Patients

The translational substudy (NCT00565708) drew on patients in Australia and New Zealand plus sites in Singapore, Malaysia, and Taiwan able to supply tumor tissue. Of 778 participants who began study medication, tissue was adequate for molecular analysis in 465, split 218 to placebo and 247 to aspirin. PIK3CA and PTEN were profiled by next-generation sequencing in 289 tumors; where that failed, Sanger sequencing covered exons 9 and 20 in a further 108.

The translational cohort was not a miniature of the parent trial: because most participants came from Australia and New Zealand, 71% were of Caucasian background against 28% in the main trial. Among the 397 patients assessable for PIK3CA, 69 (17%) carried a mutation, more often in right-sided tumors (25%) than left-sided (15%) or rectal (10%) disease (P = .008).

Neither subgroup delivered. Among the 69 patients with a PIK3CA mutation in any exon, there were 8 DFS events on placebo and 8 on aspirin (HR, 0.93; 95% CI, 0.35-2.47). The 45 patients with exon 9/20 mutations gave 7 vs 4 events (HR, 0.72; 95% CI, 0.21-2.46), and the expanded group of 84 with PIK3CA or PTEN mutations ran the other way (HR, 1.23; 95% CI, 0.47-3.19). COX-2 overexpression, in 307 tumors (69%), produced 28 vs 34 events and an HR of 0.99 (95% CI, 0.60-1.63).

Those confidence intervals are wide enough to hide a real effect, and the authors say so rather than claiming a negative result. With 86 DFS events spread across subgroups, the substudy was underpowered to estimate efficacy on its own and was never designed to.

The Pooled Answer Is Specific to Exon 9/20

Combining ASCOLT's translational data with SAKK 41/13 and ALASCCA changes the picture. Across 471 patients with PIK3CA exon 9/20 mutations and 79 events, aspirin cut DFS events by an estimated 39% (HR, 0.61; 95% CI, 0.39-0.96). Pooling any PIK3CA-pathway mutation gave an HR of 0.62 (95% CI, 0.44-0.87).

The broader PI3K-pathway group did not hold up as cleanly. For other PIK3CA exons, PIK3R1, or PTEN alterations, the pooled HR was 0.59 (95% CI, 0.34-1.00), but the 2 contributing trials disagreed (interaction P = .049). Applying ALASCCA's own criteria reproduced the split: significant for hot-spot exon 9/20 mutations, heterogeneous, and nonsignificant for everything else.

“While the combined analyses reported in this paper provides some support for this recommendation relating to [PIK3CA] exon 9/20 mutations, we believe it would be premature to include the broader group of other PI3K related mutations based on a single RCT,” the authors wrote, referring to the updated NCCN guidance.

The limitations were disclosed. ASCOLT started a decade before the other 2 trials, when staging imaging was weaker and adjuvant oxaliplatin newer; its patients began aspirin only after chemotherapy rather than soon after surgery; doses differed; and SAKK 41/13 excluded rectal cancer. Bayer supplied the tablets, but the authors state they had no role in design, analysis, interpretation, or writing.

For an oncology pathway committee, the distinction is between a mutation class with pooled randomized support and one resting on a single trial. PIK3CA testing is already routine in much of colorectal practice, so the incremental cost of acting on exon 9/20 is near zero, and the drug is generic. A prospective meta-analysis of 10 eligible adjuvant aspirin trials totaling 7043 participants, most still in follow-up, is where the wider question gets settled.

References

1. Segelov E, Li S, Li I, et al. Adjuvant aspirin for colorectal cancer with PIK3CA-mutated and COX-2 overexpressed tumours: the ASCOLT translational research study and meta-analysis. EBioMedicine. 2026;130:106389. doi:10.1016/j.ebiom.2026.106389

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