LAUSR

Dear Editor, In my original paper of the bridging study, the estimates of mean and variance for the treatment efficacy for CDx+ patient population are assumed to be independent, a typical assumption for an independent variable which follows a normal or asymptotical normal distribution. In the original paper, the estimates for mean and variance are first derived, separately, and then were pluged into the formula assumed values for the efficacy and variance of CDxpatient population for mean and variance, separately and independently. In the commentary paper, the authors first plugged in the mean estimate assuming the efficacy of CDxis a proportion of the efficacy of CDx+, and then the authors redrived the formula of the variance estimate based on updated mean plug-in formula. This approach assumes mean and variance are correlated in some degree, which may not be appropriate. Under this approach, the variance will be smaller than what was presented in the original paper. The authors may want to use “alternative” instead of the “weakness” in describing the variance estimate method in the original paper and provides some justification why this new approach may be used.