ESMO 2014 - Abstract LBA43

ESMO 2014 Abstract LBA43: Antitumor activity of pembrolizumab (PEMBRO; MK-3475) and correlation with programmed death ligand 1 (PD-L1) expression in a pooled analysis of patients (pts) with adanced non-small cell lung carcinoma (NSCLC)

Garon, E.B. et al. 

Immunotherapy in Lung Cancer: The Hottest Topic! 

Introduction: Pembrolizumab (Pembro) is an anti-PD-1 antibody that has shown activity in advanced NSCLC. The correlation between tumor PD-L1 expression and antitumor continues to evolve. Most interesting is the duration of benefit in patients who respond. KEYNOTE -001 was presented in the oral session metastatic NSCLC as a late breaking abstract. With a large sample of 282 patients, it expands our understanding of immunotherapy in metastatic NSCLC. 

Methods: 282 pts with treatment-naive or previously treated advanced NSCLC were enrolled in randomized and nonrandomized expansion cohorts of the phase 1 study KEYNOTE-001. The biomarker of PD-L1 expression was assessed by a prototype IHC assay (PA) and re-confirmed by an independent clinical trial IHC assay (CTA). Pembro was given at 2 mg/kg every 3 weeks (Q3W), 10 mg/kg Q3W, or 10 mg/kg Q2W. Overall response rate (ORR) was assessed per RECIST by central review and per immune-related response criteria (irRC) by investigator review. 


Response Rate: As per methods, two different assessments of response rate were presented.  The overall response rate (ORR) by RECIST and irRC was 21%/23% overall. This was higher in treatment naïve (26%/47%) than those patients who had been previously treated (20%/18%). The response was similar in pts with squamous (18%/25%) and non-squamous (23%/23%) histology.

Dose: In regards to the antitumor activity by dose, the response rate was 33%/67% at 2 mg/kg Q3W (n = 6), 21%/22% at 10 mg/kg Q3W (n = 141), and 21%/22% at 10 mg/kg Q2W (n = 115).

Biomarker: As per methods, two different assays for the PDL1 biomarker were presented. In patients whose tumors had ≥1% PD-L1 staining by PA, the ORR was 23%/25%. In patients with no PD-L1 staining, the ORR was lower at 9%/13%. Data for PD-L1 staining using the CTA were available for one-half of patients. In these pts, ORR (RECIST/irRC) was 39%/47% in pts with strong PD-L1 expression (≥50% staining) and 16%/9% in pts with weak/negative PD-L1 expression.

Adverse Events: Pembro was well tolerated with Grade 3–5 drug-related AEs occurring in 24 (9%) patients, most commonly pneumonitis n = 5 (no Grade 5).

Efficacy: Progression free survival for the previously treated patients was 10 weeks versus 27 weeks for treatment naive patients. The overall survival in previously treated patients was 8.2 months and has not been reached for treatment naïve. Durable responses were seen for both groups

Conclusions: Pembro is tolerable and provides antitumor activity in treatment-naive or previously treated advanced NSCLC. Patients with strong PD-L1 tumor expression may derive particular benefit.

CARE Faculty Perspective: This is a complicated area that we are moving quickly into with metastatic NSCLC. The KEYNOTE-001 had three different doses, two different ways to access response and two different assays used for the PDL1 biomarker. Although the number of patients treated at 2mg/kg were few, their responses seem as robust and durable as the higher dose so this is the dose moving forward. The response rate ranges between 20 and 26% for previously treated and treatment naïve respectively. Responses are durable. Side effects are few mostly fatigue.  Overall survival is impressive at 8.2 months for the heavily pretreated group and not even reached for the treatment naïve. What biomarker will be used? Likely the CTA assay with a cut off of >50% staining for pembrolizumab. We await Phase lll data!


Barb Melosky, MD