One of the prevailing themes of immuno-oncology sessions at Asco was the growing applicability of PD-1 blockade across multiple cancers, but as the markets opened on Monday it was clear that in some cases this was failing to meet inflated expectations.
The already wounded Bristol-Myers Squibb slipped another 1.3% on doubts over toxicity and the identification of relevant biomarkers. But safety fears must be set against response durability, which in some patients has been stunning. The right patients need to be identified, and oncologists testified that the understanding of the way biomarkers evolve has come on in leaps and bounds.
Dr Padmanee Sharma, of the University of Texas MD Anderson Center, set the stage on Saturday in a discussion at which she argued robustly against using PD-L1 expression on tumours and antigen-presenting cells as a biomarker for the activity of PD-1/PD-L1-blocking agents like Bristol’s nivolumab.
The key is that immune response changes over time, so tumours might initially be PD-L1-negative but go on to express this antigen later, especially in response to blockade with other checkpoint inhibitors. Thus, she stated, PD-L1 expression at a single time point might not reflect an evolving immune response in the tumour microenvironment.
The matter is vitally important because confusion had arisen over why some PD-L1-negative patients do respond to PD-1 inhibition. Meanwhile, other studies had shown a strong preference towards PD-L1-positives, prompting fears that the market might be smaller if significant amounts of patients were thus excluded from it.
But Dr Sharma was adamant: “I would not like to deprive a PD-L1-negative patient of an anti-PD-L1 treatment,” she said. She added that T-cell infiltration into tumours – another hot theme of Asco – might be a far more reliable prognosis marker.
Disappointing the market
Even before Monday’s share price dip Bristol had already disappointed, losing $5.5bn of market cap when the Asco abstracts went live (Asco preview – The selloff begins early, May 19, 2014).
Dr Sharma had an upbeat message for Bristol bulls, saying combination strategies like that of Bristol’s Yervoy and nivolumab might provide a path for “considering all patients for immunotherapy, regardless of the expression of ... PD-L1”.
In a late-breaking abstract on Monday Bristol revealed updated results from a complex phase I melanoma study investigating various ways of combining the two treatment approaches. New data on survival suggested that sequential therapy with Yervoy and nivo was just as good as four combinations of the two agents – offering the hope of using sequential treatment to spare patients of some of the worst toxicities.
Though patient numbers were small, this supports the theory of induction of PD-L1 expression – something backed up by the fact that substantial activity was observed in the trial in PD-L1-low/negative tumours. The headline numbers – one and two-year survival rates of 85% and 79% – were described by the investigators as “unprecedented”, and median overall survival was 39.7 months.
Still, there is a toxicity cost: there was one drug-related death from the consequences of colitis in a newly disclosed 41-patient cohort. Safety will remain a key focus, especially since the hazards of stimulating the immune system are becoming increasingly clear.
Bristol’s problem at Asco was that it has not lived up to earlier expectations, but also that it has failed to beat the progress shown by its competitors Merck & Co and Roche.
Merck, for instance, unveiled a 411-patient melanoma study with its anti-PD-1, pembrolizumab, showing a 69% survival rate at 12 months. Prior Yervoy exposure did not cause a significant difference in survival – another important fact as regards sequencing of these burgeoning immune therapies.
The most important message here was safety: the lead investigator, Dr Antoni Ribas, from UCLA, said pembrolizumab looked to be “one of the most benign therapies I’ve ever seen” in this setting.
Roche, meanwhile, presented phase I data on its anti-PD-L1 MAb MPDL3280A (RG7446) in previously treated urothelial bladder cancer, showing a 43% tumour shrinkage rate. On a call on Monday ISI Group’s Mark Schoenebaum called these data “striking” and said they made Roche Asco’s big pharma winner in immuno-oncology.
He also stressed how Asco had shown that there were more and more tumours in which PD-1/PD-L1 blockade worked. Dr Steven O’Day, an Asco expert discussing the developments on Monday, said this was now “truly a brave new world of immunotherapy. We are bursting out of melanoma and into solid tumours, some of which are very hard to treat.”
This, and possible data in haematological cancers, too, at the Ash meeting, will be something to which PD-1 bulls will cling as pharma stocks enter the dreaded post-Asco summer slump.