With positive trials of two immune checkpoint inhibitors, Opdivo (nivolumab) and Keytruda (pembrolizumab), in second line NSCLC and compared with Taxotere (docetaxel), it should come as a surprise to nobody that another checkpoint inhibitor, Tecentriq (atezolizumab) has also proven superior to Taxotere in the OAK trial of previously treated NSCLC patients, as reported in a press release today.  Perhaps the biggest surprise is that this result actually has the potential to shake up the field even with Tecentriq as a late third entrant into the race.

The trial in question is called OAK, which is a very straightforward head to head phase III trial of Tecentriq, a PD-L1 inhibitor, vs. standard Taxotere in 1225 patients who had received one or two lines of prior chemotherapy and were not restricted by level of PD-L1 expression.

This trial is very similar to trials with the PD-1 inhibitor Opdivo in patients with previously treated squamous NSCLC (Checkmate 017) and another with Opdivo in previously treated non-squamous NSCLC (Checkmate 057), without restriction by PD-L1 status. Both of those trials demonstrated a significant improvement in overall survival compared with Taxotere, leading to the approval of Opdivo in previously treated patients with advanced NSCLC, regardless of PD-L1 status.  In addition, the Keynote-010 trial of the PD-L1 inhibitor Keytruda vs. Taxotere also demonstrated a very similar survival benefit but was restricted to patients with PD-L1 expression.  At this time, the approval of Keytruda is only specifically for patients who test positive for PD-L1 with a threshold level of >50% expression, based on an earlier trial with Keytruda that demonstrated clearly greatest benefit in the much smaller minority of patients with high level PD-L1 expression using a 50% cutoff (about 28% of patients).

Enter Tacentriq.  As noted above, it might theoretically demonstrate differences in efficacy and/or tolerability compared with PD-1 inhibitors like Opdivo and Keytruda because it inhibits the PD-L1 receptor on the tumor side instead of the PD-1 receptor on the T-cell (immune cell) side.

From Ribas, NEJM, 2012

 

We had already learned that the smaller phase II POPLAR trial, with the same overall design as OAK (Tecentriq vs. Taxotere in previously treated advanced NSCLC), demonstrated the same pattern of survival benefit as the Opdivo and Keytruda trials -- an approximately 3 month improvement in median overall survival vs. Taxotere, a "hazard ratio" (measure of outcome across all time points) for survival of an approximately 30% improvement.  Though it used a different test of PD-L1 expression, looking at PD-L1 expression in the host immune cells around the tumor in addition to the tumor itself, it showed the same pattern of most other work with checkpoint inhibitors in NSCLC. Specifically, the trial showed an overall benefit in the broad population regardless of PD-L1 expression, but there was clear association of PD-L1 expression on tumor cells and/or immune cells with benefit from Tecentriq: specifically, the benefit was greatest in the patients with strong PD-L1 expression on either tumor cells, host immune cells, or both, and there was no benefit of Tecentriq over Taxotere in the 32% of patients with no PD-L1 expression on either tumor cells or immune cells:

To summarize, the efficacy results with Tecentriq have looked very similar to what we've seen with PD-1 inhibitors in the same setting -- the only difference being that Opdivo has actually shown a benefit in squamous NSCLC patients that wasn't associated with PD-L1 expression,  but no other checkpoint inhibitor studies have studied squamous NSCLC patients separately.

Though there has been some speculation that Tecentriq, as a PD-L1 inhibitor and not a PD-1 inhibitor, may have different and a more favorable side effect profile than the PD-1 inhibitors, when I look at the results of Opdivo, Keytruda, and Tecentriq side by side with results from different studies compared with Taxotere (not as good a comparison of these agents in the same trial, but this is the best we can do and should service us pretty well), the Tecentriq certainly doesn't look any better tolerated:

The press release reported that there was a survival benefit for Tecentriq on the OAK trial in the broad population, regardless of PD-L1 expression status. This puts it on equal footing with Opdivo, which currently has an approval for previously treated advanced NSCLC regardless of PD-L1 expression and is the overwhelming market leader as second line treatment of advanced NSCLC today.  In contrast, Keytruda has an approval only for patients who have high level PD-L1 expression, requiring patient tumor tissue to be tested beforehand, while Opdivo has no such requirement. Most oncologists have concluded that there is therefore no clear value in doing testing for PD-L1 testing and give Opdivo as the path of least resistance.  Since the FDA approved Opdivo for everyone despite the Checkmate 057 trial of Opdivo vs. Taxotere in non-squamous NSCLC being positive in the broad population but showing no improvement in survival in the patients with no PD-L1 expression, I would be surprised to see the FDA approve Tecentriq for a "PD-L1 positive only" population even in the very likely event that the benefit is confined to the approximately 2/3 of patients with some degree of PD-L1 expression by the Roche test used in the Tecentriq studies.

Why should a third entrant into the same market, likely to show the same general results as Opdivo and Keytruda, lead to a shake up? Because if Tecentriq gets an "all comers" approval for previously treated advanced NSCLC, it will be the first/only PD-1 or PD-L1 checkpoint inhibitor that will be an every three week treatment for a broad population.  Opdivo, as an every 2 week therapy, has what is likely to be a remarkably similar/functionally equivalent efficacy and side effect profile, but patients will prefer an every three week treatment (my patients LOVE seeing me as often as possible and may be an exception) , as will most oncologists, for whom access to "chair time" in their infusion centers is a very limited resource and often a bottleneck for their practice. With costs also pretty comparable, unless insurers mandate one over another in this setting, Tecentriq could quickly overtake leadership of the second line market from Opdivo.

But more changes are afoot. We know that the Keynote-024 trial of Keytruda vs. standard first line chemo showed a survival benefit for Keytruda in the most selected patients -- the 28% with high level PD-L1 expression on the companion PD-L1 test for Keytruda -- so immunotherapy will likely move to first line for selected patients once we get the full data and an approval for Keytruda in that setting. I strongly suspect that this will mean it rapidly becomes standard practice to perform PD-L1 testing as soon as possible after diagnosis with advanced NSCLC, and a minority of less than 1/3 of patients will get first line Keytruda. What about the remaining 70-75% of patients? Most will want to be treated with immunotherapy in the second line setting, and nearly all will prefer a treatment that requires fewer clinic visits and IV infusions, so Tecentriq will remain a compelling choice if it gets a broad indication.  I do expect that Keytruda's approval in second line will broaden to include patients with lower level (1-49%) PD-L1 expression based on the positive results in this group as well as the high PD-L1 expression patients in the Keynote-010 trial, so it's likely that we'll have a couple of competing options as every 3 week treatment and that Opdivo may well be squeezed out as just a shade less appealing because of the schedule.  And we may well see it become clearer that PD-L1 inhibitors are just not a leading second line choice for the ~1/3 of patients whose tumors have no PD-L1 expression.  

With immunotherapy combinations coming, all we can say is that this field will continue to evolve, as new press releases and soon real data will be emerging every few weeks that has the potential to further change the landscape. Stay tuned for more.