Clinical Outcome of NSCLC Patients Treated With Immune Checkpoint Inhibitor Therapy: A Retrospective Cohort Study at the Pius-Hospital Oldenburg

Introduction Based on genetic alterations and newly identified pathways, cancer treatment evolved in the last years towards novel therapies such as immune-checkpoint-inhibitor therapies. Nivolumab was the first immune checkpoint inhibitor approved for NSCLC, mostly in 2nd line, followed by pembrolizumab, also in 1st line. As licensing studies always enrol a very selected patient population, it is important to look at outcome parameters and toxicity in real world patients. As these antibodies have been administered at the lung cancer center Oldenburg, Pius-Hospital since 2015, it is possible to retrospectively describe this cohort of patients.

Methods The retrospective cohort includes all pts (n = 132) with histologically confirmed diagnosis of NSCLC stage IV treated with I/O in 1st, 2nd or further lines at the Pius-Hospital Oldenburg. Clinical data and tumor characteristics were systematically captured out of the data bank of the lung cancer center database, including mutation status and PD-L1-Score. The clinical parameters comprise age, gender, smoking history, ECOG-status, type of first line treatment with their best response rate (BRR) and progression free survival (PFS), according to RECIST. Immune-related adverse events (IRAE) were retrospectively taken out of the patientsʼ records and classified.

Results 104 pts were treated with Nivolumab in 2nd or further line of treatment. Median survival of this patient cohort was 19.6 months, ORR was 30.9% and median PFS was 6.1 months. There was a numerically improved outcome in high PD-L1 expressers. Most pts (16 out of 28) receiving pembrolizumab were treated in 1st line and had a PD-L1 expression of > 50%. Median PFS and OS were 5.8 months and 17.5 months respectively. Interestingly the rate of IRAE grade 3 – 5 were 7.1% (2 out of 28 pts) in the pembrolizumab group, classified both as grade 5. The IRAE rate was 17.9%. Details of the adverse event patterns will be presented at the meeting.

Conclusion Data in a group of patients treated with I/O suggest that the expected ORR, PFS and OS are reproduced in the real world. Although not statistically significant, PD-L1 expression might be associated with improved outcome in NSCLC patients treated with I/O across lines. The differences in the side effects are unexpected and most likely due to patient selection.