Drug development in oncology field suffers high failure rate since limited learning from early clinical trials to inform late studies [1].Using quantitative approach(a.k.a.modeling approach) to predict late phase endpoints from early surrogates(e.g.tumor size changes in phase Ⅱ trial or tumor inhibition matrices) has been reported in many types of cancers,including Non-small cell lung cancer(NSCLC) [2-5].