Adding a FAK inhibitor to a KRAS G12C blocker in colorectal cancer

Design

multicentre phase 1b/2 trial with single-arm and open-label randomised components

Population

51 patients in phase 2 with previously treated KRAS G12C-mutated metastatic colorectal cancer, all Asian, China

Primary outcome

investigator-assessed confirmed objective response rate (RECIST 1.1)

Effect

randomised part 38.9% (95% CI 17.3–64.3) vs 16.7% (3.6–41.4); difference 22.2% (−7.7 to 49.1), one-sided p = 0.068

KRAS G12C inhibitors work less well in colorectal cancer than in lung cancer, and focal adhesion kinase signalling is one proposed reason. This phase 1b/2 study across Chinese tertiary hospitals established a dose of ifebemtinib 100 mg daily with garsorasib 600 mg twice daily, then tested the combination in previously treated KRAS G12C-mutated metastatic colorectal cancer.

Fifty-one patients entered phase 2. In the 15-patient single-arm part, the confirmed objective response rate was 46.7% (95% CI 21.3 to 73.4), with seven partial responses, which met the threshold to proceed. In the 36-patient randomised part, confirmed response was 38.9% (17.3 to 64.3) with the combination against 16.7% (3.6 to 41.4) with garsorasib alone — a difference of 22.2 percentage points (95% CI −7.7 to 49.1, one-sided p = 0.068).

Grade 3 treatment-related events occurred in 33% of the combination group and 28% of the monotherapy group; diarrhoea was the commonest at 18% across combination-treated patients. Serious adverse events were 30% against 22%. No grade 4 treatment-related events or treatment-related deaths occurred.

A doubling of response rate that does not reach significance in 36 patients is exactly what an underpowered randomised phase 2 produces, and it means neither that the combination works nor that it does not. The honest reading is that this justifies a properly powered trial, and the median ages differed by twelve years between arms in a study with no stratification, which is worth remembering before treating the point estimates as comparable.

• Test for KRAS G12C in metastatic colorectal cancer — the treatment landscape is moving
• Do not use this combination outside a trial; the randomised comparison was not significant
• Expect diarrhoea as the dominant grade 3 toxicity if either agent is used
• Note the 12-year median age difference between randomised arms when reading the response rates

Why it matters

It tests a specific explanation for why KRAS G12C inhibitors underperform in the colon.

Don't overread it

Thirty-six randomised patients with a non-significant difference and unbalanced arms — this is hypothesis-generating only.

The statistics, in plain English

A 22.2 percentage point difference with a confidence interval running from −7.7 to 49.1 includes the possibility of harm — it cannot distinguish a large benefit from none. A one-sided p of 0.068 fails even the more permissive one-sided threshold. With 18 patients per arm and no stratification, the arms differed substantially at baseline, which alone could account for the gap. Response rate is also a surrogate; no survival comparison is reported.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.