Aspirin halved colorectal cancer recurrence in patients with PI3K pathway mutations

In 2012, an analysis of two large US cohorts reported something that looked too convenient to be true: among people whose colorectal tumours carried a PIK3CA mutation, regular aspirin use after diagnosis was associated with a multivariate hazard ratio for cancer-related death of 0.18 (95% CI, 0.06 to 0.61; P<0.001) [s2]. Among those with wild-type PIK3CA, aspirin was associated with nothing at all — a hazard ratio of 0.96 (95% CI, 0.69 to 1.32; P=0.76), with a P value of 0.009 for the interaction [s2].

An eighty-cent generic drug, a single genetic marker, and an effect size that large is precisely the pattern that observational studies produce and randomised trials usually erase. The randomised test was published in the New England Journal of Medicine in September, and it did not erase it [s1].

What ALASCCA did

The trial was double-blind, randomised and placebo-controlled, enrolling patients with stage I, II or III rectal cancer, or stage II or III colon cancer, whose tumours carried somatic alterations in PI3K pathway genes [s1]. Patients were assigned 1:1 to 160 mg of aspirin or matched placebo once daily for three years [s1].

Eligibility split into two prespecified groups. Group A were patients with PIK3CA hotspot mutations in exon 9 or 20 — the population the 2012 cohort analysis had flagged. Group B were those with other moderate- or high-impact somatic variants in PIK3CA, PIK3R1 or PTEN, a wider slice of the same signalling pathway [s1].

The primary endpoint was colorectal cancer recurrence in group A, in a time-to-event analysis [s1]. Recurrence in group B, disease-free survival and safety were secondary [s1].

The screening burden is worth noting. Alterations in PI3K pathway genes were detected in 1,103 of 2,980 patients with complete genomic data — 37.0% [s1]. Of 515 patients with group A alterations and 588 with group B alterations, 314 and 312 respectively were randomised [s1]. Roughly five patients had to be sequenced to randomise one.

The result

Among group A patients, the estimated three-year cumulative incidence of recurrence was 7.7% with aspirin and 14.1% with placebo (hazard ratio, 0.49; 95% CI, 0.24 to 0.98; P = 0.04) [s1].

Among group B, it was 7.7% versus 16.8% (hazard ratio, 0.42; 95% CI, 0.21 to 0.83) [s1].

Three-year disease-free survival was 88.5% with aspirin versus 81.4% with placebo in group A (hazard ratio, 0.61; 95% CI, 0.34 to 1.08), and 89.1% versus 78.7% in group B (hazard ratio, 0.51; 95% CI, 0.29 to 0.88) [s1].

The confidence intervals deserve attention. The primary endpoint's upper bound is 0.98 and its P value 0.04 — a positive result, but one that sits close to the threshold and would not survive much loss of data. The disease-free survival interval in group A crosses 1.0 (upper bound 1.08) [s1], meaning the trial did not demonstrate a disease-free survival benefit in its primary population even while demonstrating a recurrence benefit. That is not a contradiction — recurrence and disease-free survival are different endpoints with different event counts — but it does mean the strongest claim the data support is about recurrence.

The group B result, on a secondary endpoint in a broader biomarker population, is if anything the cleaner one, with both recurrence and disease-free survival intervals excluding 1.0 [s1].

The harms

Severe adverse events occurred in 16.8% of aspirin recipients and 11.6% of placebo recipients [s1]. That is a difference of 5.2 percentage points against a recurrence reduction of 6.4 percentage points in group A — a ratio that will matter a great deal to any individual patient and that the trial's headline numbers obscure. Aspirin's principal harm is bleeding, and three years of daily dosing in a post-surgical oncology population is not a low-risk exposure.

Why the biomarker framing is the story

Aspirin is already known to reduce the incidence of colorectal adenoma and colorectal cancer among high-risk people [s1]. What ALASCCA tests is something narrower and more interesting: whether a cheap, old, non-targeted drug behaves like a targeted therapy when the target is specified in advance.

The 2012 cohort study proposed exactly that — that PIK3CA mutation might serve as a predictive molecular biomarker for adjuvant aspirin therapy, on the mechanistic reasoning that aspirin's inhibition of PTGS2 (cyclooxygenase-2) downregulates PI3K signalling [s2]. Thirteen years later, a randomised trial designed around that hypothesis found an effect in the direction and roughly the magnitude predicted.

What is not established

The trial has no wild-type comparison arm. Every randomised patient had a PI3K pathway alteration [s1], so ALASCCA cannot itself demonstrate that the benefit is specific to those patients — the interaction that made the 2012 finding compelling remains observational.

Three-year follow-up on a three-year treatment tells us about early recurrence, not about long-term survival, and the trial reports neither overall survival nor what happens after aspirin stops.

The 160 mg dose is not a dose that has been optimised; it is the dose that was chosen.

The trial was funded by the Swedish Research Council and others, and registered as NCT02647099 [s1].

What to watch: whether guideline bodies treat PI3K pathway sequencing as actionable in localised colorectal cancer, and whether the effect holds with longer follow-up.

This article is informational and is not medical advice. Aspirin carries bleeding risks and decisions about it are made with a clinician.

Sources

• [s1] "Low-Dose Aspirin for PI3K-Altered Localized Colorectal Cancer," New England Journal of Medicine, published online September 2025. https://doi.org/10.1056/NEJMoa2504650
• [s2] "Aspirin use, tumor PIK3CA mutation, and colorectal-cancer survival," New England Journal of Medicine, October 2012. https://doi.org/10.1056/NEJMoa1207756