On the face of it Bristol-Myers Squibb’s failure yesterday to show a progression-free survival benefit for Opdivo plus Yervoy in first-line kidney cancer hands a victory to Exelixis, whose Cabometyx has a win under its belt in this setting.
Closer inspection of the data, however, reveals that all is not lost for Bristol, and casts doubts over the numerical superiority of Cabometyx in the NCI’s corresponding Cabosun trial (see table below). Remarkably, however, Bristol stock opened down 2% this morning, equivalent to the loss of nearly $2bn of valuation.
Bristol’s data, from the Checkmate-214 trial, also confirm earlier doubts about running immuno-oncology trials in all-comer populations of kidney cancer patients (Asco-GU – Roche’s renal success spells danger for all-comer studies, February 22, 2017). Still, Exelixis had no intention of losing momentum, today filing Cabometyx for first-line renal cancer on the back of Cabosun.
Perhaps Bristol has been punished for running a large, phase III study; Checkmate-214 enrolled 1,070 intermediate and poor-risk, first-line renal cancer patients. In contrast, Exelixis’s Cabosun was a phase II trial in just 157 subjects with similar disease, and both trials used Pfizer’s Sutent as comparator.
Both companies showed a benefit in terms of overall remission, without disclosing p values, but only Exelixis managed to demonstrate a PFS improvement with statistical significance. Cabometyx yielded a non-significant numerical overall survival benefit, a dataset that in Checkmate-214 has yet to mature.
|Bristol vs Exelixis in 1st-line renal cancer|
|Bristol-Myers Squibb, Checkmate-214 (NCT02231749)|
|Opdivo + Yervoy||41.6%||11.56 mth||not reached|
|Sutent||26.5%||8.38 mth||not reached|
|Relative change||15.1 points||3.18 mth||not reached|
|p value||not given (stat sig)||0.03 (not significant)||not reached|
|NCI/Exelixis, Cabosun (NCT01835158)|
|Cabometyx||46%||8.2 mth||30.3 mth|
|Sutent||18%||5.6 mth||21.8 mth|
|Relative change||28 points||2.6 mth||8.5 mth|
|p value||not given (stat sig)||0.012 (stat sig)||not significant|
|Note: in Cabosun ORR was a secondary endpoint; in Checkmate-214 all three measures were co-primaries.|
The headscratcher in the two datasets – with the caveat about across-study comparisons, of course – is that Opdivo plus Yervoy failed to hit significance despite showing a larger PFS benefit over Sutent than Cabometyx demonstrated with significance.
Moreover, the Sutent comparator arm in Cabosun seemed to have performed less well than historical norms, likely flattering the Cabometyx dataset further; meanwhile, Sutent’s eight-month PFS benefit in Checkmate-214 came in broadly in line with expectations.
The p value for PFS in Bristol’s trial was non-significant at 0.03 because of multiplicity corrections; Checkmate-214 had until February specified only PFS and OS as co-primary endpoints, before remission rate was added. Still, though this new measure was hit yesterday it is unlikely to suffice for approval.
That said, all is not lost; immuno-oncology studies often have difficulty showing a PFS benefit because of pseudoprogression, and hope still rests on the median OS readout, though this might not come until 2019.
Taking a gamble
Exelixis is taking a gamble in filing Cabometyx on PFS and remission data alone, especially given doubts raised earlier about the relevance of these measures in a small trial. Updated OS data – the numbers last year showed a hazard ratio whose upper bound worryingly drifted well over 1.0 – are expected at next month’s Esmo meeting.
Evercore ISI’s Umer Raffat said it was not clear whether Bristol had taken an interim look at OS, but said Checkmate-214 looked encouraging. With Roche’s Tecentriq plus Avastin beating Sutent in PD-L1-positive patients in the Immotion-150 trial, future analyses of this and other biomarkers in Checkmate-214 will be awaited keenly.
Stifel analysts reckon Exelixis has sufficiently strong results to support Cabometyx approval, though they admit that Checkmate-214 will reignite physician and investor debate about Casbosun’s deviation from historical norms.
Sellside consensus is for all lines of renal cancer to account for $958m of Cabometyx’s 2022 revenues, according to EvaluatePharma, while Opdivo is expected to be selling $8.2bn by that time, $444m of this coming from kidney cancer.
With most of Checkmate-214’s statistical powering apparently devoted to OS Bristol still has all to play for, though the possibility is strong that follow-on therapies will contaminate the result – demonstrating nothing more than the real-world situation.
And Exelixis might have the upper hand for now, but the FDA demanding that Cabometyx show an OS benefit in a larger trial must at this point represent the nightmare scenario.