Checkpoint Inhibitor Monotherapy in Potentially Trial-Eligible or Trial-Ineligible Patients With Metastatic NSCLC in the German Prospective CRISP Registry Real-World Cohort (AIO-TRK-0315).

INTRODUCTION: Patients with metastatic NSCLC (mNSCLC) treated with immune checkpoint inhibitors in clinical practice may often not meet the strict inclusion criteria of clinical trials. Our aim was to assess the trial eligibility of patients with mNSCLC treated with pembrolizumab monotherapy in real-world and to compare the outcome of "trial-ineligible" and "potentially trial-eligible" patients.

METHODS: Data from the prospective, clinical research platform CRISP were used to compare patient characteristics, treatment, and outcome of patients with programmed cell death-ligand 1 tumor proportion score greater than or equal to 50% tumors treated with pembrolizumab monotherapy who are deemed either "potentially trial-eligible" or "trial-ineligible" according to inclusion and exclusion criteria of the registrational studies (KEYNOTE-024 and -042).

RESULTS: Of 746 patients included, 343 patients (46.0%) were classified as "trial-ineligible" and had significantly worse outcomes compared with "potentially trial-eligible" patients (n = 403, 54.0%): median progression-free survival: 6.2 (95% confidence interval [CI]: 5.2-8.4) versus 10.3 (95% CI: 8.4-13.8) months, hazard ratio (trial-ineligible versus potentially trial-eligible) of 1.43 (95% CI: 1.19-1.72), p less than 0.001; median overall survival: 15.9 (95% CI: 11.4-20.3) versus 25.3 (95% CI: 19.8-30.4) months, hazard ratio of 1.36 (95% CI: 1.10-1.67), p equals 0.004.

CONCLUSIONS: Our data reveal that a considerable proportion of patients with mNSCLC are not eligible to participate in a clinical trial and were found to have worse outcomes than potentially trial-eligible patients, whose outcomes were comparable with those obtained from pivotal clinical trials. This is of substantial clinical relevance for physicians discussing outcomes to be expected with their patients and stresses the need for real-world effectiveness analyses.