By Marissa Jones, PhD | March 2026
"We don't want to gamble with the success of our program, and we don't want to gamble with the safety of patients."
It's a simple statement from Greg Chrimes, Vice President of Analytical and Formulation Sciences and Site Head for the Louisville facility at KBI Biopharma, but it captures why particle forensics has become essential in biologics development. Industry approaches to particle testing are evolving rapidly because when your HIAC test can't see translucent protein aggregates, when it can't tell you whether those particles are dangerous or benign, you're not managing risk. You're rolling the dice.
When Chrimes talks about particles, he doesn't just talk about numbers. He talks about stories.
"A particle count, a number like 25, isn't an answer," he says. "There's a deeper understanding that we need to have. We need to know what those particles are, not just that they exist."
This distinction matters more than ever as the biopharmaceutical industry shifts toward increasingly complex modalities. Protein therapeutics, gene therapies, high-concentration formulations: none of these are the small molecules that compendial particle testing was designed for. Relying on traditional methods alone is exactly the gamble Chrimes warns against.
For decades, light obscuration testing (HIAC) has been the industry standard for particle counting. It was designed with a straightforward goal: catch particles large enough to block blood vessels during IV injection, typically 10 or 25 microns and larger.
However, today's biologics contain particles that traditional methods cannot see.
"Translucent protein aggregates have a refractive index very close to the buffer solution," Chrimes explains. "Light obscuration misses them entirely. You might have significant particle formation happening, but your HIAC test says everything's fine."
It's not just translucent particles that are overlooked. High-viscosity formulations, increasingly common in subcutaneous delivery, create their own challenges. "We're seeing more and more high-concentration formulations, over 200 mg/mL in some cases," Chrimes says. "Viscosity and refractive index combined can result in underreporting particles or even missing them entirely."
The consequences can be severe, such as immunogenicity issues that compromise efficacy, clinical failures, and product recalls, all from particles you never knew were there.
"That's why we recommend orthogonal methods early and often," Chrimes emphasizes. "You want to catch these problems in development, not when you're already in the clinic."
Unlike light obscuration, which essentially counts shadows, micro-flow imaging (MFI) captures actual photographs of every particle flowing through the system.
"With MFI, you get the morphology, the character of the particles," Chrimes explains. "You can detect the difference between fibrous, proteinaceous, silicone oil, etc. You're not just getting a count; you're getting identification."
The data speaks for itself. When KBI runs HIAC and MFI side by side, MFI consistently detects two- to 10-fold increases in particle counts, sometimes even greater, particularly in the subvisible range where protein aggregates live.
"Those particles were always there," Chrimes notes. "Light obscuration just couldn't see them."
But here's where it gets interesting, hundreds or thousands of particle images create a new problem. How do you analyze them all without spending weeks manually classifying those images?
The answer is PICTS, KBI's Particle Image Classification Tool Suite.
"We released the first version of PICTS back in 2018, " Chrimes recalls. "We were training machine learning models on terabytes of particle image data long before AI became a buzzword."
PICTS uses a constantly expanding library of classified particle images to automatically identify particle types based on morphology. The software can distinguish proteinaceous aggregates, silicone oil droplets, glass flakes, cellulose fibers, and more. It can process thousands of images and bin them into categories, giving scientists an immediate picture of what's actually in solution.
"An analyst could spend hours looking through images, get fatigued, [and] miss things," Chrimes says. "PICTS takes that same MFI dataset and classifies it in a fraction of the time. Then our scientists can interpret and apply the findings to remediate any issues that arise."
But here's the critical difference that sets PICTS apart from the AI tools everyone's rushing to adopt today: it runs locally.
"PICTS operates on our secure, validated network," Chrimes emphasizes. "Your data never leaves KBI. That means it's fully secure and you can use it in support of regulatory filings characterization data: IND, BLA, whatever you need. With a lot of cloud-based AI tools today, you can't do that. PICTS was purpose-built for data integrity compliance from day one."
After nearly eight years in production, continuously learning with every study, PICTS represents something competitors can't easily replicate: data traceable, filing-ready machine learning that actually works in pharmaceutical development.
Chrimes tells the story of a client whose low-dose formulation showed higher particle counts than the high-dose product. It was counterintuitive.
"Traditional testing would have flagged the issue based on counts alone," he explains. "The team would have assumed product instability, spent months investigating formulation problems. They might have even scrapped the low-dose product."
PICTS revealed the truth in days: the particles were silicone oil droplets, not protein aggregates.
The root cause was straightforward: the low-dose compounding process used an extra syringe extraction step, introducing additional silicone oil. The high-dose process used a closed system transfer device with no extra exposure.
"That's why identification matters so much," Chrimes emphasizes. "Because the remediation is completely different. Silicone oil sends you down a container-closure review path: device compatibility, handling procedures. Protein particles point you toward formulation, shipping stress, or clinical use conditions. You don't want to go down the wrong path."
The problem was identified, the process was optimized, and the program continued.
"Standard particle counting showed the problem," Chrimes says. "Only particle forensics revealed the solution."
What makes KBI's approach different from standalone particle testing labs is integration with our other services and how we integrate with your team.
"When you work with KBI as a development partner, particle forensics is integrated with your development context," Chrimes explains. "We understand your formulation history, your process, your container closure, your manufacturing realities. The output is actionable for your program, not just a reported count of particles."
The particle core at KBI's Boulder facility includes 12 specialized scientists with expertise in particles, forensics, and in use studies. They partner with our broader Colorado characterization team include 120+ expert scientists in mass spectrometry, biochemistry, formulation science, and biophysical characterization."
We have access to a huge network of capabilities," Chrimes says. "If we need to understand colloidal stability, how your protein behaves in solution, we have scientists who specialize in that. If we need chemical fingerprinting with Raman spectroscopy to identify a mystery contaminant, we have that too."
It's this depth that allows KBI to solve particle mysteries quickly. The team doesn't just identify particles; they understand what those particles mean in the context of your entire program.
For programs on tight timelines, KBI offers its Standalone Services Portal (STAT portal), a platform where clients can request quotes, order testing, submit samples, track progress, and receive results.
"We've executed within 24 hours when it was critical for a program," Chrimes notes. "Our standard timeline is six weeks…but we have the ability to truncate the timeline, and we'll really work with our clients to define the urgency and react to that."
The portal reflects KBI's philosophy to make particle forensics as accessible and efficient as possible, so it becomes part of routine development rather than a special investigation reserved for crises.
Ask Chrimes when companies should start thinking about particle characterization, and his answer is immediate: "When they have an idea."
"I have a target, a molecule that I'd like to develop — how can we work together to devise that control strategy, that quality target product profile, from beginning to end? If we think about it early and often, we're going to save ourselves time and money later."
For a CMC leader about to file an IND, Chrimes recommends the following approach:
"The guidance on immunogenicity means you need to credibly define your particle and aggregate profile," Chrimes explains. "Regulators expect you to know this. And if you're moving into syringes or other devices, being able to differentiate silicone oil from protein aggregates changes your entire mitigation strategy."
As biologics become more complex (AAVs, lipid nanoparticles, personalized medicines), particle characterization evolves with them.
"A lot of gene therapy products and advanced therapies are basically particles by nature," Chrimes notes. "We want to know their distribution. With nucleic acid delivery systems, is the DNA under-packed or over-packed? But those same problem statements exist: do we have subvisible, translucent particles that aren't being counted?"
KBI continuously evaluates new instrumentation: mass photometry, analytical ultracentrifugation, and nanoparticle tracking, staying ahead of what emerging modalities will demand.
"What keeps me up at night?" Chrimes reflects. "Those invisible particles. If we're just using HIAC and just looking at those bins at 10 and 25 microns, are we missing data and trends that might cause problems later?"
It's why the industry is evolving toward greater expectations for particle characterization in general: smaller size bins, orthogonal techniques, and morphological identification.
"We need a comprehensive approach," Chrimes says. "Not just checking a box that says we meet specifications."
In an era where biologics represent years of development and millions in investment, particle mysteries can derail programs or compromise patient safety.
"Particle counting tells you there's a problem," Chrimes summarizes. "Particle forensics tells you what to do about it."
For KBI, it's about more than numbers. It's about understanding colloidal stability, identifying root causes, integrating particle data with formulation and process knowledge, and moving quickly when problems arise.
"We're collaborative by nature," Chrimes says. "We want to work as part of your team and not just provide a number but provide the depth of understanding that's necessary to make [your program] successful."
Because in the end, it comes down to a simple principle: a particle count isn't an answer. It's just the beginning of the question.
Ready to go beyond particle counting?
Schedule a consultation with KBI's particle characterization team to discuss your program's specific needs or explore particle forensics capabilities through our online STAT portal.
Related Content: