The advent of IO has led to greater availability of effective subsequent treatments and extended survival of patients, increasingly complicating the evaluation of overall survival (OS) in adNSCLC trials. To expedite drug developments and allow timely approvals of potentially effective agents for previously treated adNSCLC, we assessed the validity of PFS and 1y survival as SE in 2nd-line adNSCLC trials, especially in those containing immunotherapies.
We conducted a systemic literature search of 2nd-line adNSCLC trials. A weighted linear regression analysis between post 2nd-line treatments and OS was performed to establish the necessity of SE in 2nd-line trials. Adopting Buyse’s criteria for surrogacy, a two-stage meta-analytic validation model was used to assess associations between SE and OS at patient-level and trial-level. The strength of association was quantified by the coefficient of determination (R2). R2 >0.6 was defined a priori as clinically relevant. Results were validated with leave-one-out cross validation methods.
Of 1680 studies identified, 85 trials with 146 arms and 25,698 patients were included. Data of 22,804 patients from 50 trials (103 arms) were used for surrogacy assessment. A significant correlation between percentages of post 2nd-line treatments and OS improvements was identified (R2 [95% confidence interval] =0.347 [0.345-0.351], PPearson <0.0001, PSpearman <0.0001). PFS showed weak patient-level (0.100 [0.098-0.101]) and trial-level (0.064 [0.059-0.069]) associations with OS, while 1y survival strongly correlated with OS at both levels (R2patient=0.707 [0.704-0.708]; R2trial=0.829 [0.828-0.831]). Subgroup analysis of IO trials yielded similar results (PFS: R2patient=0.177 [0.128-0.200], R2trial=0.835 [0.791-0.918]; 1y survival: R2patient=0.965 [0.960-0.968], R2trial=0.778 [0.734-0.856]).
A valid SE is needed in 2nd-line adNSCLC trials in the era of IO. PFS poorly correlated with OS at both patient-level and trial-level, while 1y survival showed the potential of being a valid SE in future 2nd-line adNSCLC trials.
Clinical trial identification
Legal entity responsible for the study
Has not received any funding.
All authors have declared no conflicts of interest.