OBJECTIVES:

Selumetinib is approved for children aged ≥2 (US) or ≥3 (EU) years with neurofibromatosis type 1 and symptomatic, inoperable plexiform neurofibromas (NF1-PN) based on data from the single-arm SPRINT trial (NCT01362803) and age-matched patients in the NH trial (NCT00924196). This PSM analysis estimates progression risk with selumetinib in SPRINT versus NH (external control arm), accounting for residual differences in baseline characteristics.

METHODS:

Balance in baseline characteristics between SPRINT and NH was assessed by standardized difference. Propensity scores (PS), accounting for age, sex, race, weight, height, and PN parameters (location, volume, progression status), were used in a 1:1 PS match without replacement (base case analysis), stabilized inverse probability of treatment weighting (sIPTW), and 1:2 PS match with replacement as sensitivity analyses. Selumetinib’s effect (maximum duration 5.6 years) on progression risk was evaluated using univariate and multivariable Cox models (adjusted for aforementioned baseline characteristics) of progression-free survival (PFS) and reported as hazard ratios (HRs). Kaplan–Meier method was used for nonparametric PFS estimates.

RESULTS:

Fifty SPRINT and 75 NH patients were included. Before application of PSM, SPRINT versus NH baseline characteristics were generally balanced (standardized difference ≤20%), except for PN location (51%) and PN progression status (60%). Following a 1:1 PS match (n=37), balance was achieved across baseline characteristics, except for PN status (standardized difference 25%). Balance across baseline characteristics was achieved with sIPTW and 1:2 PS match (n=46:43).

PFS HRs (95% confidence interval) for the direct comparison, 1:1 PS match, sIPTW and 1:2 PS match were 0.11 (0.05, 0.25), 0.11 (0.04, 0.29), 0.12 (0.06, 0.25), and 0.11 (0.06, 0.24), respectively, with p-values consistently <0.001 across all methods.

CONCLUSIONS:

Residual baseline differences in prognostic factors between SPRINT and NH were adequately accounted for. NF1-PN progression risk reduction with selumetinib was consistent with direct comparison and was statistically significant, robust, and comparable across methods.