Both Roche and Novartis have endeavoured to show that the durability of impressive response rates generated by Braf and Mek inhibitors in melanoma can be improved by adding PD-1 blockade. But while the former looks to have succeeded, its rival countryman has failed.
The Combi-I trial, which used Novartis’s unapproved PD-1 agent spartalizumab on top of Tafinlar and Mekinist, did not meet its primary endpoint, the company quietly announced late last Friday. This was supposed to be spartalizumab’s route to market so this is clearly disappointing; however, a closer look at the Roche data in this setting suggests that the Novartis trial might not have missed by much.
Roche's Imspire150 and the Combi-I study were similar in design and size, both recruiting untreated, metastatic Braf positive melanoma patients, and pitting a triple combination of PD-1 blockade and Braf and Mek inhibition against doublet kinase inhibitor therapy.
Roche won approval for Tecentriq on the back of the Imspire150 trial in June, which tested its anti-PD-L1 asset Tecentriq on top of Cotellic and Zelboraf. Full data were detailed at AACR this year. Both pivotal studies used investigator-assessed PFS for a primary endpoint; however, Imspire150 also underwent an independent review. Revealingly, this reading of the data failed to find a statistically significant difference between the study arms.
|Scraping a win? Imspire150 results|
|Investigator assessed||Independent review|
|Tecentriq + Cotellic + Zelboraf||Placebo + Cotellic + Zelboraf||Tecentriq + doublet||Placebo + doublet|
|HR=0.78 (p=0.0249)||Not significant|
|mOS (months, data not mature)||28.8||25.1|
|Source: US drug label, AACR.|
These results do not confirm a clear win for adding Tecentriq in this setting. Overall survival remains an important readout, and while a longer duration of response for the triplet therapy is certainly encouraging, a survival benefit would give this finding a boost.
So while Novartis’s setback leaves Roche a clear run at this space, the implications of the failure could yet prove uncomfortable.
Novartis has yet to release any data from Combi-I and it will be interesting to see how close this trial came to succeeding. Differences in the studies will also need to be scrutinised, as subtle shifts in design or recruitment could have played a role in the different outcomes (Roche and Novartis hope to push forward again in melanoma, August 23, 2019).
A cross-trial comparison of the kinase inhibitor backbone used is also notable. This hints at better underlying efficacy from the Roche doublet on certain measures, although on the important overall survival reading Tafinlar and Mekinist seem to triumph.
|Braf and Mek kinase inhibitor combos in front-line melanoma|
|Novartis: Tafinlar + Mekinist (n=211)||Roche: Cotellic + Zelboraf (n=247)|
|Median PFS (months)||9.3||12.3|
|Overall response rate||66%||70%|
|Median duration of response (months)||9.2||13|
|Overall survival (months)||25.1||22.3|
|Source: US drug labels.|
Whatever the reason for Combi-I’s failure, Novartis is now left even further behind in the competitive anti-PD-(L)1 space. Huge commercial success was not anticipated for such a latecomer, of course, but having an internal anti-PD-1 agent has long been considered strategically important, for combination reasons.
The agent is in a number of other early studies with other experimental mechanisms; the next important readout could be from another later line melanoma study, possibly in 2022.
Of course the worst-case scenario also has to be considered here: spartalizumab is itself a dud. When and if the company finally approaches regulators for approval, having a failed phase III in the bag might make for uncomfortable questions.
This story has been corrected to state that Tecentriq received full approval in advanced melanoma, not accelerated approval.