Dan Klevisah, CEO, Pion
There’s no getting around it in drug development: what happens at the bench doesn’t always predict what will happen in the human body. This challenge is amplified with poorly soluble compounds such as those classified as Biopharmaceutics Classification System (BCS) Class II. While formulation scientists can improve dissolution by reducing particle size or creating amorphous solid dispersions and nanocrystal formulations, these approaches don’t always translate to better absorption or predictable performance in patients.
The reason is that oral absorption is a dynamic process. The drug must dissolve, remain available in solution, and cross the intestinal membrane. Depending on the compound and its formulation, precipitation, solubility limitations, or membrane permeability can all influence this process. Better formulation decisions begin with understanding these factors and leads us to the key question:
Can we use laboratory data to predict drug absorption more reliably, before committing to more extensive and costly development work?
At Pion, we answer that question by combining in vitro dissolution–permeation testing with mechanistic in silico modeling.
Connecting laboratory measurements to real-world performance
Conventional dissolution testing is an important tool, but it doesn’t fully capture the relationship between dissolution and permeation. Dissolution–permeation testing provides the necessary layer of information by measuring how much drug dissolves and how much crosses a membrane under biorelevant conditions.
When those data are integrated with mechanistic absorption modeling, we can take leap forward and translate laboratory measurements into quantitative predictions of how a drug may behave in vivo.
Our recent application note describes this approach using Pion’s MicroFLUX™ and BioFLUX™ systems alongside Predictor™ software. Our team first evaluated the model using compounds with established human absorption data, then applied the approach to formulations of aripiprazole and piroxicam.
Across the model validation set, predicted and reported fractions of drug absorbed showed a strong linear relationship, with an R² of0.98 and a slope of 0.99. These results support the ability of the integrated approach to translate dissolution–permeation measurements into meaningful predictions of absorption across multiple Biopharmaceutics Classification System (BCS) classes.
The aripiprazole study provides a real-world example; aripiprazole is a poorly soluble BCS Class II drug. Predictions for five commercially available generic formulations and the innovator product fell within the study’s predefined acceptance ranges relative to clinical absorption data. The evaluation included whole-tablet experiments, demonstrating the potential to assess formulations under conditions that reflect how products are administered.
Looking beyond how much drug is absorbed
The piroxicam case study highlights the fact that the extent of absorption is only part of the story. The rate at which a drug becomes available also must be considered.
The second study compared a nanocrystal tablet with a marketed reference tablet. In the clinical data, both formulations had a reported peak plasma concentration (Cmax) of 2,230 ng/mL, while the nanocrystal formulation reached its peak sooner: a median time to peak (tmax) of 1.75hours, compared with 2.75 hours for the reference.
The in vitro and in silico workflow reproduced the overall pattern, predicting tmax values of 2.23 and 3.52 hours, respectively, alongside comparable peak concentrations and total exposure.
These findings reinforce how an integrated approach can distinguish between formulation effects on dissolution, the timing of absorption, and overall drug exposure. It can also provide insight into the factors limiting absorption in a particular formulation.
Make formulation decisions with greater confidence
Better prediction capabilities can help teams understand why a formulation behaves the way it does and compare alternatives more effectively and efficiently. While no laboratory model eliminates the need for clinical evidence, improving the connection between in vitro testing and in vivo performance provides a stronger foundation for development decisions and bioequivalence assessment.
Our goal is to bridge that gap. By bringing dissolution, permeation, and mechanistic modeling together, we can move beyond asking whether a drug dissolves and begin answering the more consequential question: What does that tell us about its likely performance in the body?
