Elizabeth Hedgeman1, Mette Nørgaard2, Tapashi Dalvi3, Lars Pedersen4, Hanh Pham Hansen5, Jill Walker6, Anita Midha7, Norah Shire8, Anne-Marie Boothman9, Jon P Fryzek10, James Rigas11, Anders Mellemgaard12, Torben R Rasmussen13, Stephen Hamilton-Dutoit14, Deirdre Cronin-Fenton15. 1. EpidStrategies, Rockville, MD, USA. Electronic address: ehedgeman@gmail.com. 2. Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark. Electronic address: mn@clin.au.dk. 3. AstraZeneca, Gaithersburg, MD, USA. Electronic address: tapashi.dalvi@astrazeneca.com. 4. Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark. Electronic address: lap@clin.au.dk. 5. Institute of Pathology, Aarhus University Hospital, Aarhus, Denmark. Electronic address: HANHHA@rm.dk. 6. AstraZeneca, Cambridge, United Kingdom. Electronic address: Jill.Walker@astrazeneca.com. 7. AstraZeneca, Cambridge, United Kingdom. Electronic address: anita.midha1@astrazeneca.com. 8. AstraZeneca, Gaithersburg, MD, USA. Electronic address: Norah.Shire@astrazeneca.com. 9. AstraZeneca, Cambridge, United Kingdom. Electronic address: anne-marie.boothman@astrazeneca.com. 10. EpidStrategies, Rockville, MD, USA; Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark. Electronic address: jfryzek@epidstrategies.com. 11. AstraZeneca, Gaithersburg, MD, USA. Electronic address: James.Rigas@astrazeneca.com. 12. Bornholms Hospital, Copenhagen, Denmark. Electronic address: a.mellemgaard@dadlnet.dk. 13. Danish Lung Cancer Group, Odense, Denmark; Department of Respiratory Medicine, Aarhus University Hospital, Aarhus, Denmark. Electronic address: torbrasm@rm.dk. 14. Institute of Pathology, Aarhus University Hospital, Aarhus, Denmark. Electronic address: stephami@rm.dk. 15. Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark. Electronic address: dc@clin.au.dk.
Abstract
BACKGROUND: PD-L1 expression on tumor cells (TCs) or immune cells (ICs) may be used as a prognostic marker for survival in patients with NSCLC. We characterized PD-L1 expression on TCs or ICs in a patient cohort with NSCLC to determine associations between PD-L1 expression and overall survival (OS), according to EGFR and KRAS mutation status. METHODS: Danish patients aged >18 years diagnosed with NSCLC before 2014 on first- (N = 491), second- (N = 368), or third-line (N = 498) therapy were included. Data were extracted from population-based medical registries. Tumor samples from pathology archives were tested for biomarkers. High PD-L1 expression was defined as expression on ≥25 % of TCs or ICs based on first diagnostic biopsy or surgical resection. KRAS and EGFR mutation status were tested using PCR-based assays. Cox regression analysis was used to compute adjusted HRs and associated 95 % CIs. RESULTS: PD-L1 TC and IC ≥ 25 % were observed in 24.3 %-31.0 % and 11.7-14.7 % of patients, respectively. EGFR and KRAS mutations were detected in 4.7 %-8.8 % and 26.5 %-30.7 % of patients, respectively. PD-L1 TC ≥ 25 % was not associated with survival advantage in first- (HR = 0.96, 95 % CI: 0.75-1.22), second- (1.08, 0.81-1.42), or third-line (0.94, 0.74-1.20) therapy. PD-L1 IC ≥ 25 % was associated with survival advantage in second-line (HR = 0.56, 95 % CI: 0.36-0.86) and third-line (0.69, 0.49-0.97) but not first-line (1.00, 0.70-1.41) therapy. CONCLUSION: No association was observed between PD-L1 TC ≥ 25 % and OS in any therapy line. PD-L1 IC ≥ 25 % may confer survival benefit among some patients who reach second-line therapy.
BACKGROUND:PD-L1 expression on tumor cells (TCs) or immune cells (ICs) may be used as a prognostic marker for survival in patients with NSCLC. We characterized PD-L1 expression on TCs or ICs in a patient cohort with NSCLC to determine associations between PD-L1 expression and overall survival (OS), according to EGFR and KRAS mutation status. METHODS: Danish patients aged >18 years diagnosed with NSCLC before 2014 on first- (N = 491), second- (N = 368), or third-line (N = 498) therapy were included. Data were extracted from population-based medical registries. Tumor samples from pathology archives were tested for biomarkers. High PD-L1 expression was defined as expression on ≥25 % of TCs or ICs based on first diagnostic biopsy or surgical resection. KRAS and EGFR mutation status were tested using PCR-based assays. Cox regression analysis was used to compute adjusted HRs and associated 95 % CIs. RESULTS:PD-L1 TC and IC ≥ 25 % were observed in 24.3 %-31.0 % and 11.7-14.7 % of patients, respectively. EGFR and KRAS mutations were detected in 4.7 %-8.8 % and 26.5 %-30.7 % of patients, respectively. PD-L1 TC ≥ 25 % was not associated with survival advantage in first- (HR = 0.96, 95 % CI: 0.75-1.22), second- (1.08, 0.81-1.42), or third-line (0.94, 0.74-1.20) therapy. PD-L1 IC ≥ 25 % was associated with survival advantage in second-line (HR = 0.56, 95 % CI: 0.36-0.86) and third-line (0.69, 0.49-0.97) but not first-line (1.00, 0.70-1.41) therapy. CONCLUSION: No association was observed between PD-L1 TC ≥ 25 % and OS in any therapy line. PD-L1 IC ≥ 25 % may confer survival benefit among some patients who reach second-line therapy.