WCLC 2022 | KEYNOTE-091 and IMPOWER010: adjuvant chemotherapy for NSCLC

At WCLC, here we have seen the update of the IMPower010 data. IMPower010 is randomized Phase III study that compare one year of adjuvant atezolizumab versus observation in patient with stage IB, II and III, non-small cell lung cancer after resection and all the patients receive at least one cycle of chemotherapy. We know that the trial is positive, mainly in the patients with 1+ or more PD-L1 NSCLC. What we have seen here is the updated overall-survival data, now we have a follow-up of 46 months. It’s a bit more than one year compared to the previous data that we have seen, and the good thing is that the survival curves divide more and more with time, suggesting that the benefit of the adjuvant immunotherapy is really something that stay, in time, compared to chemotherapy, where the benefit is really within the five first years.

After five years, there might be some detrimental effect of the chemo because of the late side effect, although concern for chemo, very few patients. There are a lot of discussion between the IMPower010 data and the PEARLS study. PEARLS is the adjuvant trial with pembrolizumab. It’s also known as KEYNOTE-091. It’s roughly a similar study, regarding patient selection, so it’s stage IB/II, III, non-small cell lung cancer after resection and there patients receive either one year of pembrolizumab or placebo.

Between these two major trial in the field, there are some discrepancies in the subgroups. For the PD-L1 negative, for example, there is no benefit for atezolizumab, where there is a benefit for pembrolizumab and a lot of discussion around that. One of my main comment here is that if we remember the PACIFIC trail with maintenance durvalumab after definitive chemo in stage III, known small cell lung cancer, you might remind that for the PD-L1 negative, there were no effects of durvalumab. Although, in fact, the performance of the control arm, in PACIFIC, for the PD-L1 negative was really much higher than expected and much higher than for the PD-L1 positive, so at the end, we most of all think that the PD-L1 negative patients of PACIFIC with durvalumab had no benefit, just because, by chance, the control alarm over performed.

It would be very interesting, in IMpower to see the data for the PD-L1 negative and the other way around for the PD-L1 positive, the 50+, IMPower would identify as significant. Although KEYNOTE-029 is not significant in this population, we know that the benefit of immunotherapy for this population is really driven by the very, very high expressor, the 90+ percent PD-L1. This is something that we know from the data of single agent pembrolizumab or cemiplimab in the metastatic patients.

There again, will be very interesting to have the breakdown in the high PD-L1, between the PD-L1 from 50% to 90% and the 90 plus, just to see if there is no imbalance. Finally, we treat a micro-metastatic disease that we cannot really assess because, by definition, this is something we cannot see. What we can say is that there are some discrepancies between the PD-L1 in the resected tumors and the PD-L1 of the relapse or the PD-L1 of circulating tumor cells, and it’s also something that we can have to keep in mind with adjuvant immunotherapy.

We treat a disease we cannot see, and the biology of this and seeing disease might not be exactly similar to the primary tumor resected so that the benefit of immunotherapy could be a bit different. I see, in that field, the circulating tumor DNA would be a key biomarker to better understand who are the patients that really benefit from an adjuvant treatment.