University of Hawaiʻi Cancer Center, Western Digital Leverage AI and Big Data in Fight Against Breast Cancer

Print Friendly, PDF & Email

Western Digital today said it has partnered with the University of Hawaiʻi Cancer Center to provide the storage infrastructure for from mammography scans cancer detection. The center’s AI Precision Health Institute (AI-PHI) has deployed the company’s Ultrastar Data60 Hybrid Storage Platform  with 720TB of  hard disk drives (HDDs)  in support of artificial intelligence (AI), machine learning and deep learning to predict the risk of disease and relieve the increasing prevalence of cancer in the Pacific region.

AI-PHI’s mammogram database holds the Hawaiʻi Pacific Island Mammography Registry, specifically focused on women in Hawaiʻi, including native Hawaiian women who statistically fare worse with breast cancer compared to nearly every other ethnic group. With improved performance and hundreds of terabytes of storage, the goal is to provides researchers with access to large amounts of data needed for AI workflows. This helps them  develop, curate and share the massive and diverse medical datasets for cancer research to solve this unique problem for women in Hawaiʻi, and other women around the world.

“AI has changed everything in terms of how we addressed problems with big data,” said John Shepherd, Ph.D., founder of AI-PHI and professor of epidemiology and population sciences at the cancer center. “When you’re going to examine 6 million images, it takes a lot of time. With fast and reliable access to large amounts of data and the power of AI, you can say ‘Okay, there’s an outcome difference between these two images, which may ultimately help us pick out the women who may develop cancer versus those who don’t.’ We are extremely grateful for our partnership with Western Digital, which helps make this all possible with its highly reliable and scalable storage solutions.”

“Because of the uniqueness of Hawaiʻi, we have this really broad demographic of ethnicities, races, BMIs, and cultures, that if you train an AI model with that broad amount of data, it should deliver models that are useful anywhere in the world,” continued Shepherd.

Shepherd’s latest AI-based research studies biomarker patterns in mammograms. He began a career in medical imaging work before joining the University of California / San Francisco Medical Center, where he got started with mammography. He began breast cancer research around the same time AlexNet, a convolutional neural network, won the ImageNet contest in 2012, proving AI’s power to analyze images.

Dr. John Shepherd

“When AlexNet won that contest and I became of aware of [AlexNet], it changed everything in terms of how we addressed problems with big data,” said Shepherd. AI can recognize millions of images with immense precision only if coupled with a deep training model, which requires a high computational power. AlexNet solved this problem by using GPUs in combination with AI, so long as there were enough images to feed the network. This worked effectively for Shepherd, who had access to a database of 6 million mammograms at UCSF.

The cancer center said it plans to extend its Western Digital storage foundation with the addition of an OpenFlex composable infrastructure solution and Ultrastar NVMe solid state drives (SSDs). More efficient workflows and flexible collaboration provided by the new solutions will aid AI-PHI in its research on validating new AI cancer risk and detection models and developing advanced health analysis AI models through non-invasive methods. These include 3D full body-scans, dual-energy X-ray absorptiometry scans for body composition, blood tests for metabolic markers and strength assessments.

“The work Dr. Shepherd is spearheading at the University of Hawaiʻi Cancer Center’s AI Precision Health Institute is a powerful example of how data is being used to drive insights that lead to breakthrough discoveries, enabling the world to solve its biggest challenges,” said Kurt Chan, vice president and general manager, Data Center Platforms, Western Digital. “We are thrilled to contribute the storage foundation that provides fast and easy access to the data that helps enable these researchers to glean insights and understanding, and ultimately help save lives.”

source: Western Digital