番茄社区

Adjuvant atezolizumab (Tecentriq) significantly improved disease-free survival (DFS) compared with best supportive care after chemotherapy in patients with resected stage II-IIIA non-small cell lung cancer (NSCLC), according to the presented at the 2021 virtual .

In this exclusive 番茄社区 video, , a thoracic oncology specialist at City of Hope in Duarte, California, wonders about the true practical implications.

Following is a transcript of his remarks:

Hi, I'm Dr. Jack West, and I'm an associate clinical professor in medical oncology at the City of Hope Comprehensive Cancer Center in the Los Angeles area. In the lung cancer sessions at ASCO 2021, arguably the most anticipated presentation was by Dr. Heather Wakelee on the , which looked at adjuvant atezolizumab immunotherapy compared to supportive care after adjuvant chemotherapy for patients with resected stage IB-IIIA non-small cell lung cancer.

And for years, we've had adjuvant chemotherapy as a standard of care for patients with stage IB-IIIA disease established by a modest survival benefit. And this study looked at the potential value of immunotherapy with atezolizumab given every 3 weeks for up to a year in this randomized trial, randomizing just over a thousand patients who had received some chemotherapy -- ideally up to four cycles, but anywhere from one to four cycles was accepted -- and then randomizing patients to atezolizumab or supportive care.

Now in this study the primary endpoint was not overall survival, but was disease-free survival. And there was a hierarchical statistical design, so that the first testing done was disease-free survival in patients with stage II-IIIA disease. And with PD-L1 positive non-small cell lung cancer, patients could be enrolled on this study with any degree of PD-L1 expression on their tumor from undetectable or less than 1% to high.

But the first assessment was for PD-L1 positive and stage II to III -- if that was positive, which it was, with a hazard ratio of 0.66; favoring the atezolizumab arm -- then there would be a testing done for all randomized patients regardless of PD-L1 status, with stage II-IIIA disease. And this analysis was also positive with a hazard ratio of 0.79 -- so statistically significant, but with a less dramatic improvement in disease-free survival.

And there were subset analyses done looking at forest plots that showed that the subset of patients with a tumor PD-L1 over 50% -- about a third of patients -- had the clearest benefit in terms of disease-free survival, with a hazard ratio of 0.43 in that subset of patients with stage II-IIIA disease -- in contrast, in the patients with a PD-L1 of 1% or higher, so positive, but not as positive, the hazard ratio was 0.66.

Now it's important to note that this analysis includes the patients with high PD-L1 as also being higher than 1%. So it's kind of like Russian nesting dolls. And I would say that the bigger question is what the patients with low PD-L1 in the 1-49% group did -- we don't have that, but you can calculate that that hazard ratio is not as impressive; it's going to be over 0.8 if they essentially dilute the more favorable results with the patients with high PD-L1. And then when looked at separately, the patients with PD-L1 less than 1%, that hazard ratio is 0.97, and so really not as impressive.

The third hierarchical comparison is disease-free survival in the entire intent-to-treat population for stage IB-IIIA, and that did not achieve statistical significance with a hazard ratio of 0.81. Because of that, the subsequent potential analysis for overall survival is not done and may never be done. I don't think it will be.

There was a presentation of the interim overall survival data, and that showed some trend of more favorable survival with a hazard ratio of 0.77 for patients with 1% or higher PD-L1 and stage II-IIIA disease, but the broader population really did not see any good hint of improvement in overall survival, albeit with early data.

And I would say the real question is what to do with this information. Overall there does seem to be a subpopulation here who benefited meaningfully, at least in terms of disease-free survival by receiving adjuvant atezolizumab for up to a year. The questions are, does the improvement in disease-free survival translate to overall survival -- that hasn't been shown yet.

And while we do have some precedent we're giving for using disease-free survival as an endpoint in the adjuvant setting for resected lung cancer, based on the trial of adjuvant osimertinib, the hazard ratio in that trial, , was 0.21. So remarkably good, and really of a magnitude that is good enough that many of us were more inclined to waive the surrogate endpoint question of whether it would translate to overall survival difference.

Here we're not seeing that magnitude of benefit, and so it's not clear if that will translate to overall survival improvement. At the same time, we also have a statistical analysis that's focused on all PD-L1 positive patients, but that really seems to concentrate the benefit at the 50% or higher mark, and where the low PD-L1 group, while they are included in what was technically a positive analysis, dilute the favorable results in the high PD-L1 group.

We don't have an FDA approval at this time, and there is some ambiguity. I personally would be far more inclined to favor offering or recommending this for patients who have high PD-L1, even if it's FDA approved for a PD-L1 threshold of 1% or higher. I would also really prefer to have my results presented in a way that doesn't nest them so that the 1% and higher is contaminated by including the threshold of 50% as well.

I would like to see 1-49% as another group. And I would also really like to see that all trials of adjuvant whatever intervention include overall survival analysis as a component of it, rather than sidestep that with a statistical design that obviates overall survival analysis, if it doesn't meet multiple different thresholds for disease free survival.

So a controversial trial, but I would say that immunotherapy for early-stage non-small cell lung cancer is likely to be incorporated for at least a subset. I would favor it in a somewhat enriched way.

This is just one of many trials entering this space that I would continue to encourage people to watch and learn from in the coming years.

Comment