“…The curve of logit transformation of the relative survival between the index cohort and the reference population would follow an approximately straight line if the excess hazard of the patients remains constant over time (Fang, Chang, Hsu, et al, ). This method has been successfully validated with simulation data and has been used to analyze real data in several studies (e.g., Cucchetti, Trevisani, Bucci, et al, , Hung et al, ‐, , H. Y. Lee et al, , Yang, Lee, Hsu, et al, ; Yang, Lai, Chang, et al, ). The assumption of an excess constant hazard, however, is usually untestable with a limited observed time period, because how closely the curve of logit‐transformed relative survival approaches a straight line is uncertain.…”