Supercomputer Speeds Up Cancer Analysis

By marrying good science to a supercomputer, researchers have found a way to do the genomic analysis of a cancer tumor, a process that once took 8 weeks, in just 47 seconds per patient. The news could for the first time give oncologists critical information about the cancer they are going to treat before they begin treatment.

"We have quietly amassed the largest collection of cancer genomes in the nation," Dr. Patrick Soon-Shiong, chairman of NantHealth and the Chan Soon-Shiong Institute for Advanced Health, said at a press event held recently at The Newseum in Washington D.C.

Based in Phoenix, Ariz., NantHealth built the supercomputer-based high-speed fiber network with technology from Verizon and AT&T, as well as Intel and HP. Soon-Shiong told the audience that the supercomputer took approximately six years to develop and will provide thousands of oncology practices with life-saving information.

[ Is it time to re-engineer your clinical decision support system? See 10 Innovative Clinical Decision Support Programs. ]

Another partner is Eviti, Inc., which provides an oncology clinical decision support platform that connects and shares cancer patient information with physicians and payers. Using this platform, a scientific team from Nantworks streamed data into the supercomputer that consisted of 6,000 genomes representing over 3,000 patients, with 19 cancer tumor types.

The data took up 96,000 GB, and the network is now one of the fastest genomic platforms, with connectivity to over 8,000 practicing oncologists and nurses. Soon-Shiong estimates that this new infrastructure is capable of analyzing 5,000 patients per day.

It currently takes approximately two months and tens of thousands of dollars to conduct sequencing and analysis of a single cancer patient's genome.

Today, most oncologists prescribe a course of treatment for cancer patients based on the anatomical location of the cancer. However, a patient with breast cancer might benefit from positive results discovered from a patient with lung cancer if the underlying molecular pathways involving both cancers are similar. Part of the difficulty in applying genomic sequencing to guide treatment has been due to the inability to convert a patient's DNA into actionable information in a timely manner.

The supercomputer also tackles barriers such as matching patients with their medical records, which Dr. Peter Tippett, chief medical officer, Verizon, described as "a huge problem." During his remarks, Tippett mentioned that Verizon's recent initiatives around universal identity will help the healthcare ecosystem "get the right patient hooked up with the right medical record and the right electronic labels."

Tippett also said Verizon is focused on building an infrastructure for health delivery organizations that is HIPAA-compliant, and the company is developing a system that can handle 150 million concurrent patients and 400,000 transactions per second.

Looking ahead, Soon-Shiong said his team is focused on connecting genomic analysis from the supercomputer with electronic health records, which he described as disparate systems that are "basically medical bridges to nowhere."

Soon-Shiong asserts that if a grid is built that could connect EHRs to the supercomputer's platform, then up-to-date information on cancer patients, including the genomic analysis of their tumor and the treatments they've received could follow the patient wherever they reside.

A network of major cancer centers, including those at City of Hope, John Wayne Cancer Institute, and Methodist Hospital in Houston, have contributed to this collection of over 6,000 genomes.

InformationWeek Healthcare brought together eight top IT execs to discuss BYOD, Meaningful Use, accountable care, and other contentious issues. Also in the new, all-digital CIO Roundtable issue: Why use IT systems to help cut medical costs if physicians ignore the cost of the care they provide? (Free with registration.)