Charles River speeds digital pathology

Digital pathology is the business of turning microscope slides into software objects. That sounds boring, but it sits right in the middle of drug safety work — where delays in slide review can slow entire toxicology programs. The gap has never really been image capture. Labs can scan slides already. The slow part is everything around the scan — quality checks, routing, triage, and getting the right slides in front of the right pathologist. Charles River says it is now rolling out an AI-enabled workflow aimed at exactly that bottleneck. ### What actually changed? The new piece is not just “we use AI somewhere.” Charles River is packaging an end-to-end workflow for histopathology that uses AI for slide quality control, decision support, and anomaly detection before and during review. The company framed the launch around removing manual review steps that happen before a pathologist even starts interpreting tissue changes. (biospace.com) ### Why is slide quality control such a big deal? Because it is one of those annoying, load-bearing tasks that nobody notices until it slows everything down. In a traditional workflow, trained staff have to inspect slides one by one to make sure the tissue section, staining, and scan quality are good enough for interpretation. Charles River says that step is still highly manual and can create delays upstream of the actual pathology readout. (biospace.com) ### What does the AI do here? Basically, it acts like a fast first pass. It screens for slide issues, flags anomalies, and helps prioritize what needs a pathologist’s attention. Charles River has been talking for a while about AI-assisted review that can identify subtle patterns and support review prioritization, but this rollout matters because those tools are being embedded into the operational workflow instead of sitting off to the side as a demo. (biospace.com) ### How big is the claimed gain? Charles River says the AI-enabled workflow can improve pathologist efficiency by more than 30%. That is the headline number, and it is meaningful because toxicology studies generate a huge slide burden. On Charles River’s own digital pathology page, the company says an average toxicology program can involve roughly 30,000 slides. Even modest time savings per slide add up fast at that scale. (biospace.com) ### Is this built from scratch? Not really. This looks like the next layer in a multi-year buildout with Deciphex. Charles River and Deciphex partnered back in 2020 on digital pathology tools, expanded that relationship in 2024 around image management and workflow, and earlier launched Patholytix Foresight as an AI decision-support product for toxicologic pathology. So the current announcement is less a sudden invention than a stack finally getting integrated. (biospace.com) ### Why does GLP validation matter so much? Because in nonclinical safety studies, being useful is not enough — the workflow has to be defensible. Charles River has been emphasizing GLP-aligned and 21 CFR Part 11-compliant digital pathology systems, and more recent company materials describe digital pathology as validated and increasingly essential in nonclinical research. That is the real shift here: AI is moving from “interesting assistive tool” to something meant to live inside regulated study operations. (biospace.com) ### Does this mean pathologists get replaced? No — the pitch is augmentation, not substitution. Charles River keeps describing AI as a way to remove repetitive review work, improve consistency, and let pathologists focus on interpretation and peer collaboration. Think of it less like autopilot and more like a very aggressive pre-sorter that never gets tired. ### Why should anyone outside pathology care? (criver.com) Because this is where AI stops being a flashy model and starts becoming infrastructure. If Charles River can make digital pathology faster inside a regulated CRO workflow, that changes timelines for biotech clients running toxicology studies. And it raises the bar for competitors — not just to offer image analysis, but to offer validated, end-to-end systems that fit how real studies get done. (criver.com) ### Bottom line? The news is not that Charles River discovered AI in pathology. The news is that it is trying to operationalize AI at the ugliest, slowest parts of the workflow — and do it in a form clients can actually use for regulated drug development. If that holds up in practice, digital pathology stops being a nice add-on and starts looking like the default path. (biospace.com)