Deep Learning Tool Identifies Disease Markers of Alzheimer's – Health IT Analytics

Deep learning tool identifies disease markers of Alzheimer's

Source: Thinkstock

By Jessica Kent

May 15, 2019 – Deep learning continues to demonstrate its ability to enhance providers’ expertise and analysis.

In a study published in Nature Communications, researchers from the University of California, Davis, and the University of California, San Francisco, developed a deep learning tool that accurately identified one of the markers of Alzheimer’s disease in human brain tissue.

People with Alzheimer’s have amyloid plaques, clumps of protein fragments in the brain that destroy nerve cell connections. The model was able to detect types of amyloid plaques in samples of brain tissue very quickly, enabling neuropathologists to analyze thousands of times more data.

Researchers designed a convolutional neural network (CNN) algorithm and trained the tool using tens of thousands of labeled example images. The model was able to process an entire whole-brain slice slide with 98.7 percent accuracy.

“CNNs have achieved expert-level performance in complex visual recognition tasks, including the diagnosis of skin and breast cancers,” the team said.

READ MORE: Deep Learning Can Support Personalized Predictions for Stress

“These flexible models learn to recognize intricate patterns directly from visual data without the need for manually-defined image features or expert-delineated templates, and can account for non-trivial variations in image quality and color.”

Although the model performed accurately and quickly, people shouldn’t expect computers to take over for physicians completely.

“In practice, such deep-phenotyping techniques will have limited utility if their underlying predictions cannot be interpreted, critiqued, and refined by expert neuropathologist supervision,” researchers said.

Deep learning and other forms of artificial intelligence will likely serve as support for providers, allowing them to diagnose and treat patients with greater precision.

“We still need the pathologist,” said Brittany N. Dugger, PhD, an assistant professor in the UC Davis Department of Pathology and Laboratory Medicine at UC Davis and lead author of the study.

READ MORE: Deep Learning Malware Can Fake Cancer on Medical Images

“This is a tool, like a keyboard is for writing. As keyboards have aided in writing workflows, digital pathology paired with machine learning can aid with neuropathology workflows.”

Experts across the care continuum have echoed these statements. At the 2019 World Medical Innovation Forum, hosted by Partners HealthCare, experts discussed how AI will play a critical supporting role in the field of medical imaging.

“Human interpretation of images is highly subjective. We don’t have enough people to read these images, and we don’t have enough people who can do it to the highest standards,” said Constance Lehman, MD, PhD, chief of the breast imaging division at Massachusetts General Hospital (MGH) and a professor of radiology at Harvard Medical School.

“Deep learning can use full-resolution mammogram images to predict the likelihood of a woman developing breast cancer accurately. Importantly, it is accurate across all races.”

In the Alzheimer’s study, researchers noted that deep learning technology will also allow providers to examine research areas that previously went unexplored.

READ MORE: Google Tries to Patent Healthcare Deep Learning, EHR Analytics

“The tool is tireless and scalable. It’s a co-pilot, a force multiplier that extends the scope of what we can accomplish and lets us ask questions we never would have attempted manually. For example, we can look for rare plaques in unexpected places that could give us important clues about the course of the disease,” explained Michael J. Keiser, PhD, an assistant professor in UCSF’s Institute for Neurodegenerative Diseases and Department of Pharmaceutical Chemistry.

The team made the tool and the study data publicly available online, which has led to researchers in other institutions using the model in their own labs. Going forward, the team expects that tools like these will become a standard part of neuropathology, helping clinicians to quickly identify disease and treat patients using a more targeted approach.

“If we can better characterize what we are seeing, this could provide further insights into the diversity of dementia. It opens the door to precision medicine for dementias,” said Dugger.  

“These projects are phenomenal examples of cross-disciplinary translational science; neuropathologists, a statistician, a clinician, and engineers coming together, forming a dialogue and working together to solve a problem.”

The study was limited in that it used data from a single brain bank, and all the data was collected and digitized under the same conditions. To develop more reliable models, the team noted that they will have to conduct research using more diverse datasets. Still, the group believes that the study shows the potential for deep learning and artificial intelligence to improve disease detection and image analysis.

“We anticipate collecting annotated datasets from multiple sources and experts will improve recall, sensitivity, and accuracy of the resulting neural network models and support training of more sophisticated model architectures,” the researchers concluded.

“We hope this proof of concept motivates further work in this field, where automated pathology classification could have far-reaching impact.”

Author: crh9e