Integrated computational-experimental workflow for feasibility assessment of amorphous solid dispersion preparation by hot melt extrusion using

Abstract

Amorphous solid dispersions (ASDs) are widely used to enhance the solubility of poorly water-soluble drugs. Among commercial manufacturing approaches, hot melt extrusion (HME) is increasingly favored over spray drying due to its smaller environmental and physical footprint, low cost, continuous nature, and scalability. However, HME remains underutilized in early drug development due to the mismatch between minimum batch size requirements and availability of active pharmaceutical ingredient (API). To address this gap, we developed and evaluated a material-sparing workflow which could be applied in early development for HME feasibility using encorafenib (ENC) as a model compound, targeting API consumption below 1 g. Perturbed-chain statistical associating fluid theory (PC-SAFT) was first applied as an in silico technique to predict polymer compatibility, leading to the selection of appropriate polymer systems defined within a suitable process operating design space.

Vacuum compression molding (VCM) was then utilized as a material-sparing technique to prepare several formulations, which experimentally validated PC-SAFT predictions through solid state characterization (amorphous/crystalline), chemical purity (thermal degradation), and non-sink dissolution performance. Four polymers were evaluated: Soluplus (SP), polyvinylpyrrolidone-vinyl acetate (PVPVA), hydroxypropyl methylcellulose acetate succinate (HPMCAS), and Affinisol hydroxypropyl methylcellulose (HPMC). PC-SAFT and VCM screening identified PVPVA and SP as the most compatible polymers which enable ENC processing at lower temperatures, thereby reducing the risk of thermal degradation of the drug or polymer. Degradation analysis revealed that HPMCAS-based ASDs had the lowest thermal stability. Non-sink dissolution performance ranked the ASDs prepared from the corresponding polymer as PVPVA > HPMCAS > HPMC > SP. Based on overall performance, PVPVA was selected as the optimal polymer, consistent with the current commercial ENC ASD formulation. Importantly, this workflow required < 500 mg API, demonstrating its suitability for early-stage development. The integrated computational-experimental approach can enable API-sparing HME feasibility assessments, facilitating earlier adoption of HME in the development process and potentially accelerating timelines toward scalable commercial ASD manufacturing.

Introduction

Two leading technologies for manufacture of amorphous solid dispersions (ASDs) are spray drying and hot melt extrusion (HME) (Moseson et al. 2024; Vasconcelos et al. 2016). Spray drying has a distinct advantage in early development, as its material-sparing nature enables formulation feasibility to be easily conducted (Mosquera-Giraldo et al. 2021; Friesen et al. 2008; Anane-Adjei et al. 2022). However, in light of external drivers such as sustainability goals, accelerated clinical development timelines, and the push for co-location of manufacturing facilities of active pharmaceutical ingredients (API) and drug product, solvent-free manufacturing technologies such as HME become attractive alternatives (Ding 2018; Domokos et al. 2021; Fernando et al. 2022). Beyond being solvent-free, HME is inherently continuous, high-throughput, inexpensive, scalable, and occupies a small footprint on the manufacturing floor (Patil et al. 2024). Additionally, HME equipment is versatile, capable of performing other unit operations such as melt or wet granulation and in producing various drug delivery systems such as multiparticulates, implants, or powders/granules (Patros Zagaja et al. 2025; Kittikunakorn et al. 2020; Maniruzzaman et al. 2012).

The HME process to prepare an ASD involves mixing powdered components (e.g. API, polymer), and applying heat and shear to obtain a homogeneous molecular dispersion of amorphous drug and polymer (Crowley et al. 2007; Huang and Williams 2018). Two primary critical quality attributes of a successful HME process are the avoidance of thermal degradation and absence of residual crystallinity (Censi et al. 2018; Moseson and Taylor 2023). High temperatures may result in degradation of drug and/or polymer, and amorphous materials have accelerated kinetics of degradation in comparison to their corresponding crystalline forms (Moseson et al. 2020; Huang et al. 2017; Gupta et al. 2016, 2014; Meena et al. 2014; Parikh et al. 2014). Furthermore, polymers may be protective or destabilizing due to intermolecular interactions or reactivity (Alvarenga et al. 2022; Ben Osman et al. 2018; Jelić et al. 2019). To achieve the balance of minimal degradation and full crystalline-to-amorphous phase transformation, temperatures below the drug’s melting point may be employed, bounded by the formulation critical temperature (Tc), also called the solubility temperature, identified on the temperature-composition phase diagram (Moseson and Taylor 2018; Davis et al. 2021). Additionally, ternary formulation components such as plasticizers or surfactants may be required to facilitate adequate rheological properties of the molten material during the HME process or to improve in vitro dissolution performance (Saboo et al. 2021; Yang et al. 2023; Badruddoza et al. 2024).

There is a practical difficulty in assessing formulation attributes relevant to HME processing during early clinical development due to the limited availability of API and minimum batch size requirements of processing equipment, or the need for specialized equipment (Pluntze et al. 2024). For this reason, formulation design during early clinical phases may be biased toward solvent-based processing techniques (Moseson et al. 2024). In such cases, it is only later that manufacturing process selection can be performed, which may result in a need to change the formulation (such as switching to a more HME-friendly polymer, adjusting the drug loading or adding a plasticizer/surfactant). Furthermore, particle properties are distinctly different when formulations are prepared by different processing methods, which may result in different dissolution or crystallization kinetics (Trenkenschuh et al. 2024). Each of these changes may lead to differences in in vitro dissolution performance or physical stability, and ultimately to differences in in vivo bioperformance (Trenkenschuh et al. 2024; Schönfeld et al. 2021; Agrawal et al. 2013). To avoid later significant formulation or process changes, manufacturing process selection in early clinical development may be advantageous for the long-term acceleration of clinical development activities.

Use of computational formulation screening methods can help guide material-sparing formulation and process design (DeBoyace and Wildfong 2018). Models to predict drug-polymer miscibility, like the Hansen Solubility Parameter (Greenhalgh et al. 1999; Forster et al. 2001) and Flory–Huggins Interaction Parameter (Huggins 1942; Marsac et al. 2006; Flory 1942) have been used for decades to predict drug-polymer compatibility; however, these models can fail to account for intermolecular interactions, such as hydrogen bonding. More advanced computational tools utilizing molecular simulations (Turpin et al. 2018), quantum mechanics (Maniruzzaman et al. 2013), machine learning (Han et al. 2019), or perturbed-chain statistical associating fluid theory (PC-SAFT) (Gross and Sadowski 2001; Prudic et al. 2015) have recently become popular for their ability to encompass more complicated systems and better prediction for ASD miscibility and long-term stability. PC-SAFT, in comparison to the other techniques, is computationally simple for the user while maintaining the tunability through many input parameters (Trenkenschuh et al. 2024; Pavliš et al. 2023; Deac et al. 2024; Dohrn et al. 2020, 2021; Kerkhoff et al. 2025; Gottschalk et al. 2022; Castro et al. 2025). This model has been successful in ranking polymers for selection in ASDs (Prudic et al. 2014) and generating temperature-composition diagrams (Luebbert et al. 2021) with applications to formulation and process development for HME.

Vacuum compression molding (VCM) has been identified as an HME-surrogate technique with material-sparing API requirements. The VCM device applies even heating to a powder while under vacuum, causing the piston to apply pressure and create a melt devoid of air that is subsequently cooled. Polytetrafluoroethylene (PTFE) foils provide a non-stick surface preventing material loss (Treffer et al. 2015). Unlike HME, the VCM does not apply shear to the system, relying solely on diffusion for mixing, necessitating pre-treatment of samples to enable the preparation of a homogeneous system. For high viscosity polymers where diffusion is limited, VCM may not be able to form homogeneous ASDs (Pöstges et al. 2023). Several studies have been conducted in literature, highlighting the predictive quality of the VCM technique for ASD formulation and characterization, as well as HME feasibility studies (Eggenreich et al. 2016; Shadambikar et al. 2020; Evans et al. 2019; Kayser et al. 2022; Hörmann et al. 2018); however, a comprehensive workflow covering all critical quality attributes (CQAs) of a hot melt extruded ASD has yet to be developed.

In this study, several material-sparing methods were used to perform feasibility assessment of ASD preparation by HME, with the goal of developing a workflow which uses less than 1 g of API. Two critical quality attributes (thermal stability and crystalline-to-amorphous phase transformation) were assessed through material-sparing preformulation tests, as well as processing studies using VCM. Non-sink dissolution testing was conducted to assess formulation performance. Encorafenib (ENC) was selected as the model compound and the four HME-amenable polymers selected for this study were polyvinylpyrrolidone-vinyl acetate (PVPVA), Soluplus (SP), hydroxypropyl methylcellulose acetate succinate (HPMCAS), and hydroxypropyl methylcellulose (HPMC). PVPVA and HPMCAS are the most commonly used polymers for commercialized ASD formulations (Moseson et al. 2024), while SP and the Affinisol grade of HPMC are more recently developed polymers with improved stability and rheological properties suitable for HME processing (Gupta et al. 2016, 2014).

Download the full article as PDF here Integrated computational-experimental workflow for feasibility assessment of amorphous solid dispersion preparation by hot melt extrusion using <1 g API

or continue reading here

Materials

Encorafenib (ENC) was provided by Pfizer (Groton, CT). Kollidon VA64 (PVPVA) and Soluplus (SP) were gifts from BASF (Ludwigshafen, Germany). Affinisol 15LV (HPMC) was a gift from IFF (New York, NY). AQOAT LMP (HPMCAS) was a gift from Shin-Etsu (Tokyo, Japan). All other materials used were of reagent grade.

Benson, E.G., Ortiz, J.M., Buckle, E.L. et al. Integrated computational-experimental workflow for feasibility assessment of amorphous solid dispersion preparation by hot melt extrusion using <1 g API. AAPS Open 12, 26 (2026). https://doi.org/10.1186/s41120-026-00167-1