Overall response rate (ORR) as a potential surrogate for progression-free survival (PFS): A meta-analysis of metastatic non-small cell lung cancer (mNSCLC) trials submitted to the U.S. Food and Drug Administration (FDA).

Lung Cancer - Non-Small Cell Metastatic
Session Type and Session Title: 
Clinical Science Symposium, Targeting EGFR: The Next 10 Years
Abstract Number: 
J Clin Oncol 32:5s, 2014 (suppl; abstr 8012)
Gideon Michael Blumenthal, Stella Karuri, Sean Khozin, Dickran Kazandjian, Hui Zhang, Lijun Zhang, Shenghui Tang, Rajeshwari Sridhara, Patricia Keegan, Richard Pazdur; U.S. Food and Drug Administration, Silver Spring, MD

Abstracts that were granted an exception in accordance with ASCO's Conflict of Interest Policy are designated with a caret symbol (^).

Abstract Disclosures


Background: Targeted therapies (TT) administered as single agents in molecularly defined mNSCLC subsets are yielding high ORR. Improvements in PFS of large magnitude with favorable benefit-risk have served as the basis of drug approval in mNSCLC. The relationship between ORR with PFS or Overall Survival (OS) in mNSCLC is not established. Therefore, we conducted a meta-analysis of mNSCLC trials submitted to the FDA, including 3 trials of TT in molecularly enriched populations with high ORR. Methods: We identified 15 trials of 12,534 patients (median N = 698) of 9 experimental agents (tyrosine kinase inhibitor = 5, chemotherapy = 2, monoclonal antibody = 2) submitted for treatment of mNSCLC in initial or supplemental New Drug or Biologics License Applications since 2003. Criteria for inclusion of the trials in this analysis were: randomized, active-controlled, multicenter, N ≥ 150. Three trials tested TT in defined populations (EGFR mutant = 2, ALK+ = 1). The estimated PFS hazard ratio (HR- ratio of hazard of treatment versus hazard of control group) and OS HR versus the estimated odds ratio (OR) of ORR (ratio of odds of response in controls to odds of response in treatment) on the log-scale was calculated. Weighted least square (WLS) regression analyses (weight equal to the number of patients) were performed on log-transformed effects. Results: For the PFS HR vs. ORR OR analysis, the R2 = 0.89, the slope of the WLS = 0.41 (95% CI: 0.32, 0.49). For the OS HR vs. ORR OR analysis, R2 = 0.12, slope = 0.05 (95% CI: -0.03, 0.13). Using trials ≥ 500 patients (n=12), R2 = 0.50, slope = 0.16 (95% CI: 0.04, 0.3). For the OS HR vs. PFS HR analysis, R2 = 0.1, slope = 0.1 (95% CI: -0.1, 0.3). Using trials with ≥ 500 patients, R2 = 0.39, slope = 0.32 (95% CI: 0.04, 0.6). Conclusions: On a trial level, the meta-analysis of randomized, active-controlled trials indicates a strong correlation between ORR and PFS. A correlation between ORR or PFS and OS is not established and may be confounded by cross-over in the TT trials. At the trial level, a TT in a molecularly defined subset of mNSCLC with a large magnitude of effect on ORR will likely have a large effect on PFS.