Transcript:Nicolas Girard, MD: Clearly PD-L1 [programmed death-ligand 1] status is now necessary in the management of stage III non—small cell lung cancer, because we can prescribe durvalumab only in PD-L1–positive tumors. So now we developed this kind of reflex testing at the time of initial diagnosis. We have discussed collaboration with our pathologists, and once the tumor is locally advanced or metastatic, we have a systematic testing for PD-L1. This is the only way to define the role of treatment strategy before we actually stop the treatment in PD-L1–positive tumors. We will plan everything—chemotherapy, radiotherapy, and immunotherapy at the initiation of the treatment. In PD-L1–negative tumors, we will look for clinical trials opened for patients with stage III disease to try to give the patient access to innovative strategies, because currently we do not have access to durvalumab as a standard therapy.

In the PACIFIC trial we do not have the PD-L1 status for all the patients. It was not required for entering the trial, and at the end we have the information about PD-L1 status, only in 66% of the tumors. At the end, all the data that we have about PD-L1 status is then exploratory because we do not have any stratification on PD-L1 in this trial. At the end we can explore. And we know from the returns in terms of PFS [progression-free survival] and the OS [overall survival] that the benefit of durvalumab is observed both in cases of PD-L1 below 25% and above 25%. This 25% cutoff for PD-L1 expression was actually prespecified in the protocol of the trial.

The European agency required us to perform an additional analysis based on PD-L1, looking at PD-L1—negative versus PD-L1–positive tumors, similar to the 25% cutoff analysis. We cannot exclude any bias regarding the balance between prognosis factor in those 2 groups of tumors. Again, at the end we miss the information for 30% of the tumors, and we do not know whether there is still a balance between the prognosis factors in those analyses based on PD-L1. At the end we’re looking at PD-L1–negative tumors. We still have a benefit in terms of PFS for durvalumab versus placebo, but we do not observe this benefit in terms of OS. And actually looking at the survival curves of patients, the overall survival curves, we can see that the survival of patients in the placebo arm is actually better in PD-L1–negative tumors, which is a finding that was not previously reported. We never identified PD-L1 negativity as a favorable and promising factor.

So it’s very difficult to integrate scientific things off those data. There are probably imbalances in prognostic factors. So we probably need to do additional studies focusing on PD-L1—negative tumors to understand what is the actual benefit of durvalumab. In terms of PFS we know that there is a benefit, but also terms of overall survival.

Well, PD-L1 is obviously not the best marker in this situation. We know that PD-L1 is very important for the management of patients with metastatic non—small cell lung cancer. But in nonmetastatic disease we now have evidence that accumulates showing that PD-L1 status may not be the best biomarker for immunotherapy. We also have data in early-stage disease in the perioperative setting, especially the neoadjuvant settings. Early disease, eligible for surgery, and treating patients with neoadjuvant immunotherapy.

We know that we have a very high efficacy in terms of pathological response, and this does not seem to be related to PD-L1 status. So maybe PD-L1 status in nonmetastatic non—small lung cancer is not the best biomarker to predict the efficacy or the inefficacy of immunotherapy. What are the potential biomarkers? I don’t know. But we know that tumor mutation burden [TMB] is a predictor of the efficacy of immunotherapy. Tumor mutation burden is more related to cancer cells, and there’s enough neoantigen and the immunogenicity of cancer cells. So maybe in early-stage disease, when the immune system is not exhausted by disease progression, TMB is a better biomarker. Again, TMB was not collected as part of the PACIFIC trial. We probably need to have additional studies to explore additional biomarkers, including TMB, or also potential signature, or immunomonitoring. The biomarker, or EDL biomarker for immunotherapy, currently is not known. Probably we will need multiple biomarkers to have a better prediction of the efficacy of immunotherapy in this setting

Transcript Edited for Clarity