The immune checkpoint inhibitor buzz began almost 3 years ago to the day, with the first presentation of Opdivo data from a Phase I trial, at the annual ASCO meeting in 2012. Since then, two PD-1 inhibitors have been approved for the treatment of melanoma (Keytruda and Opdivo in the second half of 2014) and NSCLC (Opdivo in early 2015). Furthermore, development of Opdivo, Keytruda and other PD-1/PD-L1 pathway inhibitors is ongoing, not only in melanoma and NSCLC, but in other niche indications bladder cancer, glioblastoma, and Hodgkin's lymphoma.
PD-1 blockade has been heralded as practice changing in melanoma, and building on the success of Yervoy, PD-1 pathway inhibitors have demonstrated strong efficacy in multiple Phase III clinical trials, both as single agents and in combination with other therapies.
During the ASCO 2015 meeting, the first data from randomized trials of PD-1 inhibitors in NSCLC were presented. Based on the currently available data, can we argue that PD-1 pathway inhibitors are practice changing in NSCLC.
All three randomized trials presented, Phase III trials of Opdivo in squamous-cell NSCLC (CheckMate-017) and in non-squamous-cell NSCLC (CheckMate-057) and Phase II trial (POPLAR) of atezolizumab (formerly known as MPDL3280A) suggest that PD-1/PD-L1 pathway inhibition confers an approximate 2-3 month MOS benefit in previously-treated NSCLC patients compared with docetaxel. Additionally, a positive correlation between PD-L1 expression levels and MOS benefit was observed in the CheckMate-057 and POPLAR trials, but not in CheckMate-017.
Considering solely the MOS data, one could indeed argue that PD-1/PD-L1 inhibitors can and should be considered practice changing, and be regarded as the new standards of care for doublet-chemotherapy-treated NSCLC. The MOS data however are marred by contradictory (and in the case of CheckMate-057 and POPLAR disappointing) PFS data.
Surprisingly, in the intent-to-treat population of the CheckMate-057 and POPLAR trials, PFS in the PD-1/PD-L1 inhibitor cohort was inferior to that of docetaxel, with a hazard ratio of just under 1. In contrast, in CheckMate-017, Opdivo demonstrated a statistically (but not clinically) significant PFS benefit versus docetaxel. A positive correlation between PD-L1 expression and PFS was observed with atezolizumab and Opdivo in CheckMate-057.
Opdivo and atezolizumab performed as expected in terms of OS in the intent-to-treat population, but the PFS benefit expected based on the very promising early-Phase data did not materialize in these trials. Opdivo performed as expected in CheckMate-017, although the less than one month PFS improvement was marginal. Owing to the mechanism of action and the delayed response expected of immunotherapies, this finding was not unexpected.
The OS benefit and PD-L1 expression correlations of anti-PD-1/PD-L1 inhibitors in NSCLC have overshadowed the PFS data at ASCO 2015. The PFS data were, the elephant in the room, and these data now raise important questions:
- Which patients should receive a PD-1/PD-L1 inhibitor? Based on the PFS data, it is clear that not all patients are deriving a benefit from PD-1/PD-L1 inhibition. Squamous-cell and non-squamous-cell, PD-L1-positive patients derive the most benefit from Opdivo; PD-L1-positive patients derive most benefit from atezolizumab. PD-L1-negative patients derive some benefit from PD-1/PD-L1 inhibitor, but this is dependent on the methodology of testing and thresholds used to define PD-L1 positivity, meaning that it is not clear whether or not these patients should be treated with PD-1/PD-L1 inhibitors.
- Will the data affect regulatory approval(s)? The primary end point for the CheckMate-017 and CheckMate-057 trials is OS, and indeed, the OS benefit seen in CheckMate-017 was sufficient for the FDA to approve the drug for squamous-cell NSCLC (in March 2015). Atezolizumab was granted Breakthrough Therapy Designation by the FDA for the treatment of PD-L1-positive NSCLC, and as such, the Phase II POPLAR data could facilitate its accelerated approval. In the first-line setting however, all Phase III trials, with one exception, have PFS as the primary end point. Will PD-1/PD-L1 inhibitors who fail to demonstrate a PFS benefit, especially atezolizumab and MEDI4736 ± tremelimumab evaluated in unselected patient populations, gain first-line label expansions in this setting on the basis of OS secondary end point data, If approved, how important will histology and PD-L1 expression be in drug selection? Perhaps such drugs will be approved in the patient subpopulations suggested to have the greatest PFS and OS benefit, but how will payers respond? Given the financial burden of these therapies, it is likely that payers will push to limit use of PD-1/PD-L1 inhibitors in patients unlikely to gain PFS and OS benefits. With these questions in mind, a review of the ongoing Phase III trials suggests that Keytruda has a significant advantage over other PD-1/PD-L1 inhibitors, as it is the only therapy with a Phase III trial powered to demonstrate an OS benefit as a primary end point (KEYNOTE-042).
- Can combinations of PD-1/PD-L1 inhibitors with other therapies demonstrate superiority in delaying disease progression compared with single agent PD-1/PD-L1 inhibitors? A number of Phase I clinical trials are ongoing and will provide early signs of activity of such regimens. The sample size of such trials is typically small, but early data suggest that atezolizumab/chemotherapy combinations appear to be most active, with approximately two thirds of patients responding to treatment. PD-1/PD-L1 and CTLA-4 dual inhibition results in a reduction in tumor burden in NSCLC patients, but not to the same extent seen in melanoma. PFS data from the atezolizumab/chemotherapy trial were not reported, but it is hoped that the addition of chemotherapy will result in a rapid and prolonged delay in tumor progression, by combining the best attributes of both treatment modalities. Should this hypothesis be proven in the atezolizumab first-line Phase III trials, its regulatory approval will likely be more certain, and would be a key factor in driving prescriptions and uptake of atezolizumab drug combinations over single-agent PD-1/PD-L1 inhibitors.
- Are PD-1/PD-L1 inhibitors practice changing in NSCLC? Yes, but only for some patients. Data presented at ASCO 2015 strongly support Opdivo use in previously-treated, squamous-cell NSCLC and PD-L1-positive, non-squamous-cell NSCLC. Similarly, currently available data also support the biomarker-driven prescribing of atezolizumab and Keytruda in previously-treated NSCLC. PD-1/PD-L1 inhibitor use in the first-line setting will also likely be biomarker-driven, unless combination approaches offer superior delays in disease progression than those expected from currently available data.
Orestis Mavroudis-Chocholis, Ph.D., is a principal business insights analyst on the Oncology team at Decision Resources Group.