Researchers at The Institute of Cancer Research (ICR) in London have developed an AI technology that can predict how cancer cells will respond to new drugs based on changing cell shapes. This innovative approach makes it possible to test drugs faster and more efficiently, meaning new treatments could be available to patients years earlier.
The technology uses 3D imaging and deep learning to analyse the characteristic 'fingerprint' of a cell. Previous methods have only been able to use 'flat' 2D images. The new AI tool can create a much more realistic picture of how cells behave in the body. By analysing almost 100,000 3D images of melanoma cells, the technology was able to predict which drug was being used with 99.3 percent accuracy. Even subtle differences between drugs with similar effects were detected.
Targeted treatments
In addition to accelerating drug development, the AI tool can help to develop targeted treatments. This will make clinical trials more efficient and effective. This reduces the chance of failed trials and saves considerable costs. In addition, the technology was found to be applicable not only to melanoma cells, but also to red blood cells, brain vessels and stem cells. This indicates that the technology has the potential to be widely applicable.
The development of a new drug usually takes ten to twelve years. The researchers estimate that the new AI approach can shorten this process by six years. In particular, the preclinical phase, which normally takes three years, can be reduced to just three months using this technology. This means that effective drugs can be available to patients more quickly, while side effects and effectiveness can be better predicted early in the process. The results of the research have been published in Cell Systems.
Collaboration
To implement the technology more broadly, the researchers are collaborating with the Center for Cancer Drug Discovery at the ICR. In addition, a spin-off, Sentinal4D, has been established to further develop the innovation and apply it in the drug development process.
“3D cell shape acts as a fingerprint of cellular condition. AI can decode this information and develop effective drugs faster. This saves time and money, and more importantly, it brings life-saving treatments to patients faster,” says Professor Chris Bakal, one of the principal investigators. With the creation of Sentinal4D, the team hopes that this technology will fundamentally change drug development and contribute to new treatments for cancer and other diseases.
AI helps find new drugs
Artificial intelligence (AI) is playing an increasingly important role in drug development by speeding up the process and making it more effective. AI can analyze existing drugs and discover new applications, enabling the reuse of medication. By recognizing patterns in large datasets, AI helps find promising substances and can even contribute to personalized treatments.
A good example of the added value of AI is the study in which scientists tried to identify the most effective new glaucoma medication using AI. To predict which new candidate drugs would be the most effective, millions of compounds had to be analyzed. AI, machine learning and big data have allowed researchers to select potential drugs much faster than traditional methods.
