A computationally informed framework for lipid-based formulation developability: Integrating predictive models with...

The increasing prevalence of poorly water-soluble drugs (PWSDs) in development pipelines has intensified the need for bio-enabling strategies such as lipid-based formulations (LBFs). However, formulation selection for optimal bioavailability often relies on empirical, resource-intensive approaches. This study introduces a computationally informed developability framework that integrates predictive models to streamline LBF development. The framework combines in silico tools for assessing key determinants of LBF suitability-including predictions of food effect, lipid solubility, and biopharmaceutical dose number (Do)-with tailored strategies such as lipophilic salt synthesis for challenging molecules. A database of >200 FDA-approved oral drugs (2010-2023) was screened through the computational framework, and four case studies (dapagliflozin, mavacamten, ospemifene, flibanserin) illustrate the framework's ability to classify risk and guide formulation decisions. Predictions identified low-risk candidates for or type-I LBFs, highlighted medium-risk molecules requiring type-IIIa formulations, and demonstrated the utility of lipophilic salts for drugs which are high-risk with respect to dose-loading and dose number limitations. This integrated approach enables rapid, data-driven formulation decisions, reducing reliance on trial-and-error methods and supporting early-stage development of PWSDs. By coupling computational predictions with targeted experimental validation, the framework offers a practical roadmap for accelerating LBF adoption and improving R&D efficiency.