QSPR in oral bioavailability: specificity or integrality?

During the last decade the technological advances in drug discovery changed the absorption, distribution, metabolism, excretion and toxicity (ADMET) profiles of New Chemical Entities (NCEs). Among ADMET processes, absorption plays an important role in the research and development of more effective orally administered drugs. Although significant progress has been made in in vitro, in situ and in vivo experimental determinations of absorption, the development of in silico methodologies has emerged as a cheaper and fast alternative to predict them. Even though several in silico models have been described in the literature to predict oral bioavailability and related properties, the prediction accuracy and their potential use is still limited. The low precision and high variability of data, the lack of a complete experimental and theoretical validation of in silico approach, and above all, the multi-factorial nature of the oral absorption term, make the development of predictive in silico models a thorny task. The present review discusses several important advances regarding the QSPR approaches used in the development of predictive oral bioavailability models. The importance of fixing the problem associated with data resource, as well as improving the reliability of in silico results is highlighted. Optimization of individual properties along the absorption process must be integrated in a multi-objective scenario for studying oral bioavailability behavior in the early drug discovery and development.