Roche must be wondering how its cross-town rival Novartis has managed to show so much progress in the otherwise disappointing PI3K inhibitor class of cancer projects. Its earliest candidate in this class, pictilisib, has stumbled badly in a breast cancer trial at a time when Novartis is preparing to file buparlisib in the same disease.
PI3Ks have largely fallen out of the spotlight as the first agent, Gilead Sciences’ Zydelig, has underperformed and as immuno-oncology has dominated the development landscape. Pictilisib’s setback will not help this state of affairs, although good news from buparlisib and Roche’s follow-up PI3K project taselisib could help improve the image.
In the phase II Peggy trial pictilisib in combination with paclitaxel was no better than placebo in slowing disease progression in the intent-to-treat population of metastatic, hormone receptor-positive, HER-negative patients; a subgroup with sensitive PIK3CA mutations was no more successful. Details of this trial were released Saturday at the European Cancer Congress in Vienna.
The project, also known as GDC-0941, remains active in combination with fulvestrant in breast cancer patients resistant to aromatase inhibitor therapy and in a non-small cell lung cancer test that uses it it as an add-on to a chemotherapy cocktail and Avastin. However, it is telling that the Swiss group has leapfrogged it with taselisib, which has entered phase III in breast cancer.
Can’t miss misses
It is telling that the usually can’t-miss Roche missed in this circumstance, although the group is obviously far from being out of shots on goal in this space. The issue so far with this class is two-fold: safety and specificity.
Zydelig was launched in leukaemia and lymphoma with black box warnings for liver, gastrointestinal and lung toxicity. Similarly to Zydelig, diarrhoea was one of the gastrointestinal side effects reported in pictilisib patients. In addition, hyperglycaemia – a problem that bedevils the Clovis Oncology lung cancer project rociletinib – and rash were adverse events reported by patients.
Zydelig, of course, had the advantage of being an agent selective to the delta isoform of PI3K, expressed in B cells and therefore a useful target in diseases like follicular B-cell non-Hodgkin's lymphoma. Pictilisib is not a selective agent, which means it cannot be as active on the alpha isoform of PI3K that is predominant in breast cancer, said Dr Fabrice Andre, an oncologist with Institut Gustave Roussy who commented on the pictilisib trial. Roche's follow-on project taselisib has a higher affinity for the alpha isoform.
That brings up the question of the Novartis agent buparlisib, formerly known as BKM120, which as Dr Andre notes is also non-selective. The world’s biggest drugmaker by sales has released very little detail on this project, other than to note that the pivotal Belle-2 trial met its primary objective by showing an improvement in progression-free survival over fulvestrant alone. No safety data were revealed.
In fact, Novartis has released little on this trial or indeed earlier studies. Design papers on Belle-3 and 4 have been published, and the second-quarter 2015 delivery of Belle-2 would have left ample time to submit a paper for the ECC.
In the second-quarter earnings release, in which the Belle-2 results were announced, Novartis said it would be discussing the details of the findings before proceeding with a filing. Those anxious for progress for PI3Ks need to hope that there is not a surprise hiding in these data.