Abstract Oral bioavailability (%F) is a key determinant of drug efficacy and a critical decision-making parameter in early-stage drug development. Current prediction of human oral %F often relies on preclinical animal studies, which are costly, ethically disputed and do not necessarily correlate with human %F. Machine learning (ML), an emerging in silico technology, has been leveraged to predict human %F. However, previously published ML approaches have often relied on binary classification or achieved limited regression performance, with R2 values ranging from -0.36 to 0.50. In this study, we demonstrate that an ML pipeline can indeed be used to predict accurately human oral %F using preclinical data, potentially reducing the need for animal experiments. A total of 192 FDA-approved compounds were curated, which were discovered to span a diverse chemical space. Model inputs included molecular descriptors, structural fingerprints, published animal %F and cytochrome P450-aware metabolic feature, with the latter designed to account for metabolism-driven interspecies variability. Five-fold cross-validation, repeated 100 times, yielded an R2 of 0.766 (RMSE = 16.13%), substantially outperforming all previously published models and individual animal species univariate correlation, including non-human primates (R2 = 0.691). Notably, high performance was maintained without non-human primate data, with rat and dog %F combined achieving R2 = 0.723 ± 0.016, reducing dependence on ethically and logistically demanding primate studies. Permutation feature importance analysis identified lipophilicity, surface area and CYP features as informative predictors, providing mechanistic interpretability alongside predictive accuracy. These results demonstrate that a combination of feature engineering and a comprehensive ML pipeline, can achieve precision and reproducible prediction of human oral %F from routinely collected preclinical data. The impact of this study includes the potential to reduce late-stage attrition and accelerate the translation of drug candidates to the clinic. Graphical Abstract
Subcutaneous (SC) dosing has become an increasingly popular non-oral route of administration for biotherapeutics due to its user-friendly nature and suitability for self-administration when compared to intravenous (IV) dosing. However, comprehensive methods for characterizing the post-injection depot within the complex SC tissue remain limited. In this study, we developed an ex vivo porcine skin model with magnetic resonance imaging (MRI) as a non-invasive and anatomically relevant tool to visualize injected solution depots formed after SC administration. Building upon this, a computational quantification approach using open-source Python libraries was established to calculate depot characteristics, such as effective depot volume, surface area, and surface-to-volume ratio. Using this integrated experimental-computational approach, we evaluated different injection parameters from both formulation and device perspectives with two model drugs, semaglutide and alirocumab. As the result, this method demonstrated its applicability to two different SC therapeutic modalities (peptide and mAb) with their range of injection volumes (0.05 to 1 mL), formulation concentrations (1 mg/mL to 10 mg/mL), and injection devices (prefilled syringe (PFS) and autoinjector (AI)). Notably, the measured effective post-injection depot volumes were consistently 1.2- to 1.6-fold larger than the injection volumes. Some differences in depot morphology were observed among injections of semaglutide with different formulation concentrations, while only minor variations were noted between injection devices. By integrating an ex vivo porcine skin model with MRI and open-source computational tools, this study presents a versatile and reproducible platform for 3D visualization and quantitative characterization of SC drug depots.
Large molecule therapeutics, including monoclonal antibodies, recombinant proteins, and nucleic acids, offer precision and efficacy that have transformed the treatment of complex diseases such as autoimmune disorders, cancers, and genetic disorders. Despite their potential, their large size, structural complexity, and inherent biological instability create significant barriers to effective delivery, frequently necessitating invasive parenteral administration (e.g., intravenous, subcutaneous, or intramuscular routes). While effective, these traditional administration routes often require frequent healthcare visits, limiting accessibility, increasing the risk of administration errors, and affecting patient adherence. Recent advancements in formulation science and drug delivery device technologies—such as prefilled syringes, autoinjectors, wearable injectors, transdermal patches, and smart delivery systems—are enabling a paradigm shift toward patient-centric care. These innovations facilitate self-administration and home-based treatment, offering improved dose accuracy, greater convenience, and enhanced patient adherence. Furthermore, progress in formulation science, nanotechnology, innovative device technologies, and digital health is pioneering new minimally to non-invasive and direct delivery routes, such as pulmonary, transdermal, ophthalmic, and neurological approaches. Emerging trends like smart devices and sustainable practices are redefining therapeutic strategies, emphasizing effectiveness, accessibility, and personalized care. This mini-review explores the challenges of large molecule drug delivery, highlights advancements in formulation and device technology, and discusses emerging trends—including sustainable practices and digital health integration—that are shaping the future of patient-centric large molecule therapies.
