• There are no suggestions because the search field is empty.

Charles River Laboratories case study: accelerating preclinical analysis with AI

Pathologists at Charles River Laboratories used the Aiforia AI cloud-based software globally to automate preclinical assessments.
Written by Aiforia

Pathologists and Scientists at multiple Charles River Laboratories’ locations use Aiforia Create 

Ten different pathologists and scientists spread around the world are all sharing an Artificial Intelligence (AI) solution for their work using one software subscription. From analyzing viral capsids in North Carolina to quantifying ovarian follicles in Colorado, AI through Aiforia Create is being used for a versatile range of image analysis tasks by pathologists and scientists at Charles River Laboratories (CRL). This CRL AI working group aim to innovate within their fields of Discovery and Safety Assessment and replace traditional, time-consuming methods by improving workflow efficiency with deep learning AI.

The CRL pathologists accomplish this virtual and global collaboration using different samples and disease models, impacting their research in idiopathic pulmonary fibrosis, colitis, non-alcoholic steatohepatitis, and hepatobiliary, bone marrow, and endocrine toxicity from multiple locations across the US, UK, and Canada. This global reach and ease of access is enabled by the fact that no coding or hardware is needed to use Aiforia Create. The Aiforia Platform’s cloud-based product empowers the user to build their own deep learning AI models by simply drawing annotations on an intuitive interface.

Enabling pathologists to work and collaborate remotely

Only an internet connection is essential to get started. As Michael Staup, a manager of specialty pathology at CRL, described: “The cloud-based system is a benefit. No GPU is needed, you can run the software from any computer, and it works very well.” This is particularly beneficial to the biotech corporation as CRL Senior Pathologist Dan Rudmann explained: “We have a lot of pathologists that work from remote locations and I am anticipating this will continue to grow. You could have a pathologist like myself, who would be working remotely, and I could access the Aiforia software just through the internet. A huge advantage.”

The ease of access to the software is pivotal but it was not what attracted these CRL pathologists to Aiforia in the first place. “We needed something to help advance innovation in our work and make our workflow more efficient. With the idea of increasing innovation and efficiency in mind, we looked to AI,” explained Dr. Rudmann. “Aiforia is so streamlined. You get good results without a lot of bells and whistles. You don’t have to be a computer scientist. The pathologist, they can actually just focus on the field they were trained and educated within while using Aiforia. They don’t have to learn coding. It is very user friendly,” Dr. Staup observed.

AI models created directly by pathologists

Aiforia Create allows the user to develop, or train, their own AI models for their own specific image analysis tasks through annotating and informing the AI what to look for. Dr. Staup described this process: “I got Aiforia up and running quickly, it was pretty intuitive. I was able to train it myself in a very short period of time with very limited data. I was really impressed.” The AI models can learn anything from a 2D image. This versatility is what allowed for the CRL pathologists to analyze such a wide breadth of sample and disease models with Aiforia Create.

Once training is complete, image analysis can be automated. “The AI model processed ovaries much faster than with manual counting, the reproducibility and concordance between the two methods was excellent,” described Dr. Rudmann of using the software to quantify ovarian follicles in rats for  reproductive toxicology safety assessment. Aiforia Create not only allows pathologists to work more efficiently but it also connects healthcare professionals around the world to focus on what matters the most, enhancing scientific progress.