Prediction of drug absorption: different modeling approaches from discovery to clinical development.

Drug absorption modeling is a task that requires different tools depending on the drug development stage. Scientific progress and increased computer power have initiated a veritable revolution of modeling and simulation. Technical developments have enabled a high model complexity that was unthinkable only a couple of years ago. Today, model simulations can be used to guide experimental planning and, in some cases, the extent of work can even be reduced. There are several success stories of pharmaceutical modeling and the prediction of a drug’s absorption, distribution, metabolism, and excretion (ADME) properties has certainly been of particular interest. Considering the advances in this area, the question can be raised, how close is the pharmaceutical industry to a ‘prediction paradise’ [1]. How wide are the gaps with respect to good predictions? The present editorial addresses this question, but with an exclusive focus on drug absorption modeling. The opportunities and limitations of drug absorption prediction are often viewed differently depending on the scientist asked. A discovery researcher, for example, tries to anticipate the extent of absorption based on the chemical structure alone. Approaches range from basic molecular property rules [2] to more recent approaches using advanced statistical procedures [3–7]. However, even more challenging than intestinal absorption is the prediction of oral bioavailability. In the latter case, additional drug clearance may lead to an increased scattering to the experimental data. Starting with the chemical structure, there are interesting approaches to quantitative structure–bioavailability relationship [8,9]. However, it is fair to raise the question of whether bioavailability can be well predicted at all, based on the chemical structure [10,11]. The answer to this question must, in any case, consider the various needs of prediction accuracy. In discovery, the goal is to select a promising drug candidate, thus the prediction of an absorption and bioavailability potential seems to be sufficient. However, once a drug candidate is chosen, more accurate and versatile tools are required.