The physiological relevance of single-phase (aqueous only) and 2-phase (aqueous and organic phase) in vitro dissolution experiments was compared by mechanistic modeling. For orally dosed dipyridamole, stepwise, sequential estimation/confirmation of biopharmaceutical parameters from in vitro solubilitydissolution data was followed, before applying them within a physiologically based pharmacokinetic (PBPK) model. The PBPK model predicted clinical dipyridamole luminal and plasma concentration profiles reasonably well for a range of doses only where the precipitation rate constant was derived from the 2-phase experiment. The population model predicted a distribution of maximal precipitated fractions from 0% to 45% of the 90 mg dose (mean 7.6%). Such population information cannot be obtained directly from a few in vitro experiments; however well they may represent an "average" and several extreme subjects (those with low-high luminal fluid volumes, pH, etc.) because there is no indication of outcome likelihood. For this purpose, direct input of in vitro dissolution/precipitation profiles to a PBPK model is insufficientdmechanistic modeling is required. Biopharmaceutical in vitro-in vivo extrapolation tools can also simulate the effect of key experimental parameters (dissolution volumes, pH, paddle speed, etc.) on dissolution/precipitation behavior, thereby helping to identify critical variables, which may impact the number or design of in vitro experiments.