Robert Holt, PhD, Associate Director, Pharmaceutical Science, digiM Solution LLC

Robert Holt, PhD

AAPS PharmSci 360: Microstructure Intelligence in Predictive Drug Development, Part 2

 keynote on predictive drug development, 

, associate director, Pharmaceutical Science, 

, provides examples of how microstructure intelligence compresses development timelines and strengthens regulatory filing, particularly for modern, complex products such as long acting therapies, oral peptides, and combination products.

Holt explains that microstructure data delivers value at each development stage, such as feasibility and screening, baseline scans for calculating critical quality attributes from minimal material, and informing early formulation decisions. During optimization, these tools link particle structure to dissolution and release performance, enabling in silico formulation tuning. At scale-up and tech transfer, platforms like digiM’s dissoLab and ivisLab run mechanistic simulations that stress-test manufacturing variations in minutes, reducing batch-failure risk and generating physics-based data packages for regulators.

Providing an example of the value of microstructure data, Holt shares a case study where a major pharma client saw a dissolution slowdown between lab-scale and good manufacturing practice (GMP) batches of spray-dried dispersions, despite conventional characterization (laser diffraction, mercury injection, BET, XRPD) showing the batches as comparable. X-ray microscopy combined with artificial intelligence (AI) image analysis revealed the real cause: the GMP batch had stronger aggregation from droplet merging, cutting particle surface area by 40% versus the lab batch's looser, electrostatic aggregation.

When asked about this technology’s future, Holt notes that digiM holds an FDA IDIQ contract for regulatory assessment and guideline development on complex drug products, and its work informed product-specific guidelines for minocycline microspheres and dexamethasone ophthalmic implants, with the FDA encouraging microstructural characterization for generics. Holt calls digiM “the first FDA-recognized microstructure platform for drug product development,” positioning the technology as an emerging pillar of quality-by-design and real-time release strategies, with new AI tools that generate 3D drug substance and product models from 2D images.

Looking ahead to AAPS PharmSci 360 in New Orleans, Holt says the keynote his company is presenting will explore imaging science, predictive science, and AI applications for emerging modalities, such as long-acting injectables, high-concentration biologics, and combination products, where microstructure intelligence is poised to address next-generation manufacturing challenges.

AAPS PharmSci 360 is happening in New Orleans from October 25-28, 2026.

Click the video above to watch the interview.

Videos

Robert Holt, PhD, associate director, Pharmaceutical Science, digiM Solution, discusses microstructure intelligence in more detail in part 2 of this preview of the company’s keynote presentation at AAPS PharmSci 360.

AAPS PharmSci 360: Microstructure Intelligence in Predictive Drug Development, Part 1

, explains microstructure intelligence and how it can enhance how drug developers understand dosage forms.

According to Holt, “microstructure intelligence” is the science of understanding exactly where drug particles sit inside a dosage form, down to the microscopic level. Holt argues this spatial information is something traditional dissolution and content-uniformity testing simply can't capture. Two tablets can show identical drug content when tested, but if one has API concentrated at the surface and the other has it buried in the core, they will perform very differently. Conventional testing methods may miss this distinction, especially for extended-release products where uniformity matters most.

digiM employs multiple imaging modalities to understand the hidden architecture of a drug using an approach the company calls "correlative imaging." Based on the length scale specific to the given product and question, the approach will use electron microscopy, X-ray microscopy, polarized light microscopy and synchrotron CT. Its dissoLab and ivisLab tools then convert these real images into in silico microstructure models, applying physics-informed simulation kernels to predict dissolution and release behavior directly from the imaged sample. The company’s I2S tools use artificial intelligence (AI) for image segmentation, while a newer generative 

 lets scientists digitally alter traits like drug loading, particle size, porosity, and wall thickness to model how reformulated products would perform without manufacturing and testing new physical samples.

For manufacturers, microstructure intelligence may provide root-cause clarity: instead of just flagging that a batch dissolved faster or slower than expected, it can explain why. Holt notes the tools are browser-based and designed to be accessible regardless of company size, positioning this as a scalable diagnostic and formulation-optimization capability rather than a large-biopharma-only resource.

Click the video above to watch part 1 of the interview.

In part 1 of this 2-part preview of AAPS PharmSci 360’s keynote on predictive drug development, Robert Holt, PhD, associate director, Pharmaceutical Science, digiM Solution, explains microstructure intelligence and how it can enhance how drug developers understand the dosage form.

Robert Holt, PhD

Articles