Imatinib is still better first than asciminib. Don't be fooled.

Limitations in the ASC4FIRST trial : Asciminib in Newly Diagnosed Chronic Myeloid Leukemia.

In this ASCO 2024 series, we provide insights about ASCO 2024 presentations, which is currently taking place between 31th May and 4th June. The annual congress of the American Society of Clinical Oncology is a key event in oncology. Each year, it attracts more than 30,000 attendees in Chicago.

The data from the ASC4FIRST trial were presented at ASCO, and published the same day in the NEJM, with some commenting on social media that we are somehow entering in a post-imatinib era.

Imatinib was the first TKI (tyrosine-kinase inhibitor) to be approved in oncology, and remained, so far, the best oral anti-cancer drug. Indeed, after the introduction of imatinib, the life expectancy of people diagnosed with CML progressively returned to the same level as that of people without CML (figure from this publication below).

No other drugs have shown such an effect after imatinib.

There are 3 concerns with the ASC4FIRST trial

The trial compared the new drug, asciminib, to either imatinib or a second-generation TKI, this is the trick! In other words, there were 2 buckets in the control group: half of patients got imatinib, the other half got second-generation TKIs.

1 - Nested and adjacent groups.

The primary objectives were the differences in 48 week major molecular response (48wMMR), and 2 groups were compared:
1) asciminib versus all TKIs (imatinib and 2nd generation TKIs lumped together)
2) asciminib versus imatinib

What is missing here? “Asciminib was not compared with second-generation TKIs as a primary objective […] “was a secondary objective”.

What is problematic is that we already knew second-generation TKIs were better than imatinib for this endpoint. By lumping the groups together, the trial was almost guaranteed to be positive! This is a common trick in modern trials: combining groups that make no clinical sense (see our paper on nested and adjacent subgroups here).

The relevant comparison, for this endpoint, would have been asciminib versus 2nd-gen TKIS.

And actually, this difference (66.0% 48wMMR with asciminib vs 57.8% with 2nd gen TKIs)… was not significant !

2 - Molecular milestones are not sacrosanct

The endpoint itself, 48-week MMR, is not correlated with long-term outcomes. Molecular milestones are not sacrosanct. See our paper here, led by Anushka Walia.

3 - the study was open-label, so the reporting of side effects is prone to biases (conscious or unconscious)… and should be taken with a huge grain of salt

If you want side effects, the trial should be double blind.