Asco 2020 – small-cell lung cancer and the role of CTLA-4

Much of the lung cancer focus at this year's virtual Asco meeting is on non-small cell histology, but how Roche’s Tecentriq might be challenged in front-line small-cell disease should not be ignored either. For one, it will raise fresh concerns about the deleterious effect of inhibiting CTLA-4, courtesy of further data from Astrazeneca’s Caspian trial.

Despite this several biotechs continue to investigate CTLA-4, albeit not necessarily in lung cancer, an early look at the Asco abstracts reveals. And, still in SCLC, Merck & Co seems now to be arguing that its Keynote-604 study of Keytruda was positive, though its commercial potential is likely zero.

Keynote-604, a trial combining Keytruda not with CTLA-4 inhibition but with chemo, was toplined in January as a success on progression-free but a failure on the more important overall survival endpoint. Merck said a seemingly numerically positive 20% reduction in risk of death did not meet its statistical plan’s significance threshold, but did not disclose the absolute time figures or a p value.

Its Asco abstract does, however: median OS is 10.8 versus 9.7 months, with p=0.0164 for the hazard ratio. Why is this statistically non-significant? Because, apparently, Merck had assigned very little alpha to the OS analysis, and the threshold for significance had been set at p=0.0128.

The abstract’s authors add that “a post hoc analysis” yielded a p value that did meet this threshold, but do not elaborate further. They conclude that Keytruda “significantly ... prolonged OS”, supporting its benefit in SCLC.

Ignoring the intricacies of statistical analysis, a 1.1-month survival improvement seems to offer little hope, however. It underperforms even the underwhelming two-month benefit on which Tecentriq was approved in its Impower-133 study.

For now the first-line SCLC battle is between Roche and Astra’s Imfinzi plus chemo combo, approved in March on the basis of a 2.7-month OS advantage in one cohort of the Caspian study. But it is the second active cohort, Imfinzi and chemo plus the anti-CTLA-4 MAb tremelimumab, that doctors will be scrutinising at Asco.

This is because these subjects actually did worse than chemo alone in numerical terms: median OS for the treme-containing triplet was 10.4 versus 10.5 months, while median PFS came in at 4.9 months, half a month worse than control. The fact that rates of adverse events leading to discontinuation were twice as high with the triplet than the doublet or control provides a clue.

Nevertheless, research into blocking CTLA-4 is far from dead. Innovent, for instance, is presenting data on its own CTLA-4 MAb, IBI310, with or without the PD-1 MAb sintilimab, in melanoma, while Alligator, Xencor and Alphamab have abstracts on CTLA-4-targeting bispecific antibodies.

There are some hints of activity, the most important of which will be in patients who have already failed PD-(L)1 blockade, and the hope is that many datasets will be updated at Asco; some, including Cytomx’s and Alligator’s, have no efficacy results in the abstracts.

Also undisclosed are the virtual meeting's late-breaking abstracts, whose titles are known but which Asco will not make public until May 28. There is still much to play for.

The Asco virtual conference takes place on May 29-31.

Further reading:

Asco 2020 – abstract dump triggers first moves

Asco 2020 – first-line lung cancer focus

Asco 2020 – early allogeneic Car-T promise

Asco 2020 starts to take shape