Benefits of AI in preclinical studies

The preclinical phase of drug discovery involves extensive safety evaluations of the potential therapeutic intervention, often using cell and animal models of disease. These studies commonly include going through numerous histopathological samples. This contributes to the drug development process being time-consuming and labor-intensive for pharmaceutical and contract research organizations.

How can preclinical studies benefit from AI?

Preclinical studies can involve very difficult analysis tasks that need to be repeated over and over again. The difficulty stems from, for example, having to detect very subtle changes with high precision and accuracy. Manual quantification of small changes, or specific cell counting, is not only cumbersome but also time-consuming; therefore also often involving high costs.

One tangible example of how AI can benefit preclinical studies is Lundbeck pharmaceuticals performing neuroscientific histopathological analysis of alpha-synuclein, a neuronal protein linked to Parkinson’s disease. Quantifying alpha-synuclein positive neurons or the total pathological area with traditional methods is notoriously challenging and slow, but with the help of the Aiforia Create AI solution, this was accomplished with high speed and accuracy. In fact, the manual cell counts were compared to the AI model with a significant correlation (p-value <0.0001) demonstrating the AI-based quantification was as accurate as the manual method in quantifying α-synuclein pathology.

Aiforia Create is a cloud-based AI software that enables users to easily train their own unique AI models without any coding needed by the end user. This allows the automated analysis of any feature,  in any image, making the solution a huge benefit for preclinical studies since they commonly involve analyzing novel targets. Alternatively unique AI models can also be trained by Aiforia’s scientists through Aiforia Custom AI Services. 

AI-enabled increased precision and accuracy are something to emphasize when reflecting on the benefits of AI in preclinical studies. When detecting vague and unclear targets, as is commonly the case in histopathology, the risk of intra-observer variation and subjectivity increases. These targets often need to be detected in huge amounts, leading the work to be split among several pathologists and scientists, also increasing the risk of inter-observer variation. With AI-powered image analysis the risk of all these can be minimized as the AI model can be trained to detect even the slightest changes with high consistency and high reproducibility of results.

Research results validation is another section in preclinical studies that can benefit from AI. Scientific validation requires objective evidence that the methods for obtaining results are robust, reliable and reproducible. A well-trained AI model is exactly this. One demonstrated example of using Aiforia Create in the validation of challenging neurotoxicity studies in a preclinical toxicology study by a Finnish pharmaceutical organization Orion Pharma Ltd.

Finally, using AI models in preclinical studies can not only significantly improve and accelerate preclinical study workflows but also standardize analysis, help find hard-to-spot objects, and give important visual feedback that helps interpreting current and future results.