For a small-molecule drug candidate, a compound must dissolve, reach the required exposure, remain stable and be manufactured consistently at progressively larger scales, but these requirements are becoming harder to reconcile as development pipelines include increasingly complex molecules.
Adding to the challenge, the consequences often emerge only after formulation work or pharmacokinetic studies have begun. At that point, the remedy may involve a new formulation, additional clinical or preclinical work, or redevelopment of the manufacturing process.
Avoiding this cycle requires selecting technology based on the molecule’s dominant limitation and staying informed by the requirements of the eventual product. An approach that improves exposure but creates an unstable intermediate, an expensive process or an impractical dosage form can simply transfer risk to a later development stage, but predictive development can help teams make these decisions earlier.
Fewer, more informed experiments allow teams to test the most promising pathways before committing large amounts of active pharmaceutical ingredient (API) and significant development time. The goal is to reduce the formulation options while ensuring stability, manufacturability, and downstream needs are considered.
Identifying developability problems
The first step is to build a detailed picture of the API against the target product profile. The intended dose, solubility under relevant intestinal conditions, and permeability help establish whether dissolution rate, intestinal solubility or membrane transport is the main barrier. Stability and solid-state behaviour then inform which interventions are feasible.
The developability classification system (DCS) can help determine whether absorption is primarily limited by dissolution, solubility or permeability. That distinction is significant because compounds with seemingly similar bioavailability problems may require very different interventions.
The key issue is what prevents the intended dose from reaching the necessary exposure levels: Does the compound dissolve too slowly, is it insufficiently soluble under intestinal conditions, or is transport across the intestinal membrane the main barrier? Dissolution and solubility measurements under relevant conditions, permeability assessment and solid-state characterisation can help distinguish these mechanisms. The precipitation following dissolution also requires careful consideration, as an early rise in dissolved drug levels may not last long enough to facilitate absorption. Poor solubility does not automatically mean an amorphous solid dispersion (ASD) is the solution, but selecting a technology without identifying the rate-limiting mechanism risks solving the wrong problem.
Choosing a formulation when several constraints apply
A molecule may simultaneously exhibit poor solubility, limited permeability, and sensitivity to heat or moisture. Teams should determine which of these factors most significantly restricts exposure at the planned dose, then evaluate if solving that issue would leave other barriers still preventing the product from hitting its target. The primary limitation serves as a hypothesis to test and update as new evidence emerges.
Candidate approaches should be evaluated based on the same product requirements. Improving dissolution alone offers limited benefit if permeability continues to restrict absorption. Likewise, a positive exposure profile is insufficient if achieving the necessary drug loading results in an impractically large tablet or the formulation fails to stay stable during storage. The best approach is to target the key limiting mechanism within a feasible development and manufacturing process.
For example, micronisation or nanomilling may be appropriate when the dissolution rate is the main constraint, as smaller particles provide greater surface area for dissolution. If the amount that can dissolve remains as inadequate at the intended dose, particle-size reduction alone may be insufficient. Salt selection or ASD may then merit investigation, depending on the compound and its solid-state properties.
By maintaining the API in a higher-energy amorphous state within a polymer matrix, an ASD can improve apparent solubility and dissolution. That benefit must be weighed against physical stability and the ability to maintain the required performance. Lipid-based formulations offer another option for suitable lipophilic compounds, but their capacity to carry the dose and maintain drug in solution after dilution or digestion needs evaluation.
The benefits of predictive development
Thermo Fisher Scientific is increasingly using predictive modelling to support this diagnostic stage. Its OSDPredict™ digital toolbox incorporates Quadrant 2, an artificial intelligence (AI) and machine learning-enabled platform designed to predict formulation behaviour using information about a compound’s molecular structure and chemical characteristics alongside the target product profile. Quadrant 2 applies several computational approaches, including molecular dynamics, quantitative structure-activity relationship (QSAR) modelling and absorption, distribution, metabolism, excretion and toxicity (ADMET) analysis. The aim is to narrow the range of potentially viable formulation strategies before scarce API is consumed in extensive experimental screening.
Predictive tools can also inform the wider development strategy. Absorption modelling can assess whether an anticipated improvement in dissolution is likely to translate into exposure, while stability prediction and process modelling can flag risks that affect a formulation’s usefulness beyond its initial performance.
However, these tools do not eliminate experimental development. Predictions depend on available inputs and model applicability, so shortlisted approaches still require experimental confirmation. The resulting studies should resolve specific uncertainties, including whether a formulation sustains its performance and whether the proposed process is feasible. New data can then refine the next development decision.
Assessing stability and scale
Selecting an ASD does not settle the manufacturing route. Spray drying, for example, can be attractive early in development because feasibility studies can be conducted with relatively small API quantities. However, it requires the drug and polymer to be soluble in an appropriate solvent system. A process built around very low solids content can also become expensive at commercial volumes, potentially creating a cost-of-goods problem that was not apparent during early development. Teams should investigate that cost and capacity burden before it becomes embedded in the process.
Hot-melt extrusion (HME) offers an alternative for compatible compounds. It is typically a solvent-free continuous process and can offer commercial advantages, but the API and polymer must tolerate the heat and shear involved.
The decision needs to account for drug-polymer compatibility, achievable drug loading, chemical and physical stability, solvent requirements, downstream processing and eventual scale. The intended final dosage form should also be considered, as an intermediate that performs well immediately after spray drying or extrusion still needs to be developed into a robust tablet, capsule or other finished product.
Thinking through those downstream steps at the outset can prevent a Phase I solution from becoming a Phase III manufacturing bottleneck.
Making early decisions with the final product in view
The goal should be a development strategy that is phase-appropriate without becoming phase-limited. A powder-in-bottle presentation may be perfectly reasonable for an early clinical study, for example, but teams should already understand how the formulation and process could evolve towards a scalable commercial product. Early decisions should document why the route was selected, which constraints remain unresolved, and what evidence would trigger reconsideration.
For a promising laboratory formulation, this means identifying the next studies needed to establish stability, practical drug loading, and process feasibility. It also means understanding the implications of a later formulation change before clinical supply, and manufacturing plans depend on the initial choice.
ThermoFisher’s OSDPredict platform is intended to connect these predictive tools across areas including bioperformance, stability and manufacturability, and its value is not simply in the availability of multiple modelling tools. Instead, it is its ability to narrow the formulation space earlier, prioritise more promising approaches, and focus experimental work where it is most informative.
