Researchers at Washington State University have unveiled a deep learning AI model that can identify disease in tissue images with unprecedented speed and accuracy. This breakthrough promises to revolutionize both medical diagnostics and disease-related research.
Researchers at Washington State University have developed a groundbreaking "deep learning" artificial intelligence model that can identify pathologies, or signs of disease, in animal and human tissue images much faster and often more accurately than human experts. This development holds the promise of revolutionizing both the pace of disease research and medical diagnostics.
"This AI-based deep learning program was very, very accurate at looking at these tissues," co-corresponding author Michael Skinner, a professor in the School of Biological Sciences at WSU, said in a news release. "It could revolutionize this type of medicine for both animals and humans, essentially better facilitating these kinds of analysis."
Developed by computer scientists Colin Greeley, a former WSU graduate student, and his advising professor Lawrence Holder, the AI model was trained using images from past epigenetic studies conducted in Skinner's laboratory.
These studies focused on molecular-level signs of disease in various tissues from rats and mice. Subsequently, the AI was tested with images from other studies, including those identifying breast cancer and lymph node metastasis.
Impressively, the deep learning model not only identified pathologies quickly but also found instances that human pathologists had missed.
"I think we now have a way to identify disease and tissue that is faster and more accurate than humans," Holder, a co-corresponding author, said in the news release.
Traditionally, analyzing tissue slides has been a painstaking process performed by teams of specially trained people. These experts meticulously examine and annotate the slides under a microscope, often spending hours on just one image to reduce human error.
In the realm of epigenetics research, where Skinner's team studies molecular processes influencing gene behavior without changing the DNA itself, such analysis can take a year or more for large studies. With the new AI model, the same data can be acquired within a few weeks.
Deep learning is an advanced AI method that emulates the human brain's neural networks. Unlike traditional machine learning, which relies on predefined algorithms, deep learning adjusts its processes over time through a network of neurons and synapses. If the model makes an error, it "learns" from it using backpropagation, which fine-tunes the network to improve its performance.
The WSU AI model is designed to handle high-resolution, gigapixel images, each containing billions of pixels. To manage these large files, the researchers configured the AI to analyze smaller, individual tiles while still considering their context within larger image sections, akin to zooming in and out on a microscope.
Other researchers have already taken interest in this deep learning model. Holder's team is currently collaborating with WSU veterinary medicine researchers to diagnose diseases in deer and elk tissue samples.
The implications of this technology extend beyond animal health. The AI model could significantly enhance human medical research and diagnostics, particularly for cancer and other gene-related diseases.
According to Holder, as long as there is annotated data identifying diseases within tissue samples, the AI can be trained to perform those analyses.
"The network that we've designed is state-of-the-art," Holder added. "We did comparisons to several other systems and other data sets for this paper, and it beat them all."
Published in the journal Scientific Reports, this groundbreaking research outlines a bright future for integrating AI into medical diagnoses and research.