Stephen Liu, MD: We saw recently, at the 2019 World Conference on Lung Cancer, the presentation of the CASPIAN data. CASPIAN is a 3-arm randomized trial for patients with untreated extensive stage small cell lung cancer randomized to chemotherapy, platinum—either carboplatin or cisplatin—with etoposide for up to 6 cycles, versus durvalumab plus chemotherapy, versus durvalumab plus tremelimumab, an anti—CTLA-4 [anti–cytotoxic T-lymphocyte–associated protein 4] antibody, plus chemotherapy. And what we saw was a comparison between chemotherapy alone and chemotherapy with durvalumab.

We’ve not yet seen the results from the tremelimumab arm. What we saw was the addition of durvalumab, the addition of an anti—PD-L1 [anti–programmed death-ligand 1] antibody to chemotherapy improved survival. We saw an improvement in median survival from 10.3 months to 13 months, with a hazard ratio 0.73. We saw an improvement in 1-year survival rate, and we saw a very favorable toxicity profile. This strategy improved survival without significantly increasing the toxicity.

What really stands out from this study is how remarkably similar it was to IMpower133. If you line up those survival curves, the 1-year survival rates with PD-L1 and chemotherapy are between 52% and 54%, versus chemotherapy, 39% in both arms. The curves split at about the same area. The hazard ratios: 0.70 with atezolizumab, 0.73 with durvalumab. These results really overlap, really validating this strategy. This was not an aberration. This was not a one-off. When you add an anti—PD-L1 antibody to chemotherapy in an all-comers population, you see a survival benefit. And, the strategy works. It’s led to an improvement in survival and really has established new standards of care.

We saw some updates from both IMpower133 and CASPIAN at ESMO 2019 [European Society of Medical Oncology annual meeting]. First, we have longer follow-up with IMpower133. What it showed was that with more follow-up, a median of 22.9 months follow-up, we see that survival benefit persists. The 18-month survival rate was 34% with atezolizumab versus 21% with chemotherapy, a difference of 13 percentage points at 18 months. We’re hoping that difference is maintained as more time passes, though we do have to wait to see, with more follow-up, how the 2-year, 3-year, 5-year survival differences will be.

What we also saw was some work in biomarkers. We know from the original presentation of IMpower133 that the use of blood-based tumor mutational burden [TMB], blood TMB, did not discriminate between patients who would respond or who would benefit from atezolizumab versus not. We saw a benefit with atezolizumab whether you were above or below the TMB threshold, whether you use 10 or 16 mutations per megabase. And so, TMB wasn’t useful in identifying the patients who are getting that benefit. But we know it’s a subset. We know a subset of patients is likely driving that benefit. And so, the next marker to look at was PD-L1 expression.

What we saw with IMpower133 is, first, most samples were not evaluable for PD-L1. Most samples were PD-L1—negative. In fact, if you look at PD-L1 expression on the tumor, 95% were PD-L1–negative. If you included PD-L1 expression on the immune cells in the microenvironment, that increased to about 50%. But it did not play a predictive role. PD-L1 expression was not useful in identifying who would derive that benefit from atezolizumab.

In a similar session, we saw from CASPIAN that PD-L1 expression results were almost identical. Again, 95% of samples were PD-L1—negative when you looked at just the tumor. When you looked at the immune cells, the rate of PD-L1 positivity went up. But PD-L1 expression did not identify who would benefit from the addition of durvalumab to chemotherapy. And so, we are left in a state where we were before—an all-comers population where we see a benefit and PD-L1 does not help identify who’s getting that benefit. I believe there is a biomarker that’s going to be much more elusive than PD-L1 or TMB. Right now, those should not play a role in selecting patients to derive benefit from this therapy.

We know that the addition of durvalumab improved survival. What we were still waiting to hear was the impact on quality of life. And so, at ESMO 2019, we saw some patient-reported outcomes [PROs] data. We saw that the arm that received durvalumab did have a slower time to deterioration, which is an important PRO. We saw that specific outcomes such as dyspnea and several other symptoms did seem to take a longer time to deteriorate with the addition of durvalumab. And so, this positive impact on patient-reported outcomes, which had previously been shown with IMpower133 as well, is very supportive of this regimen. We are improving survival. We assume that we are making patients’ lives better, and this shows us that we are.

Transcript Edited for Clarity