How the cloud helps fight cancer

The man with stage 4 bladder cancer thought he'd run out of options. He had tried multiple treatments to fight the cancer, which had spread throughout his body. Every treatment was unsuccessful. Hope dwindled.

In a last attempt to find something that would help, the elderly man went to Dr. Lincoln Nadauld, an oncologist and director of Cancer Genomics at Intermountain Healthcare based in St. George, Utah.

Nadauld sequenced the patient's entire genome, looking for something in his DNA that would offer clues about the specific makeup of his disease and what medications might best treat it. He found a particular gene and began treating the man with medication that targets it specifically.

"His cancer had progressed through previous treatments very rapidly," Nadauld told Computerworld. "He's been on this new drug for 20 months and he feels great. His cancer stopped progressing and it started to shrink. He actually goes to the gym a couple of times a week. I hope I'm doing that when I'm 77."

Just what enabled the oncologist to sequence his patient's genetic makeup quickly and far less expensively than he could have just several years ago?

The cloud.

A new way to fight disease

Doctors are using genetics to find better, personalized treatments for cancer patients, while, on a larger scale, researchers are working to find genetic causes of diseases that could lead to better treatments -- or even cures.

This genetic work, involving doctors and researchers from companies, universities and medical centers around the world, could change the way we fight diseases from cancer to heart disease to autism. Using the cloud allows them to work with massive amounts of data and complicated analyses.

"The only way to do big research is with large population sets," said James Hirmas, CEO of GenomeNext, a Columbus, Ohio-based genomic analysis company. "Enabling this mass data processing...is ultimately going to lead to a lot of discovery. We can do that because of the cloud."

Doctors have been working to create more personalized medicine, using patients' individual genetics along with personal data like age and other health factors, to figure out if one chemotherapy treatment would work better than others. Doctors also can sequence the genetic makeup of a tumor to find a specific treatment strategy.

The result is a new world of patient care that gives doctors a better shot at success against some of the worst diseases they battle.

Genome sequencing is a complicated endeavor; the human genome contains more than 3 billion genetic characters. However, inside that sequencing, doctors can find clues about where genes are and how they work together to direct a person's growth and body maintenance.

Sequencing also enables doctors or researchers to spot mutations that could lead to cardiovascular disease, dementia or cancer.

This kind of work not only takes a lot of computing power, it needs a system that can scale to hold massive data loads and enable doctors and researchers from around the globe to feed data into it and access it for their own work.

The cloud is a perfect answer for that.

"The cloud opens the door for this," said Lauren Nelson, an analyst with Forrester. "This works with any type of big data problem.... This is a good example for really every organization that is trying to leverage information from big data."

Workloads and expense

For instance, Dr. Peter White, a genomics scientist with Nationwide Children's Hospital in Columbus, Ohio, said just a few years ago he didn't think he would be able to analyze an entire human genome sequence routinely. (A single human genome involves about 100 GB of data; sequencing the first human genome took about 15 years and cost approximately $3 billion.)

Now, it can be done in about 40 hours for between $1,500 and $2,500.

Using the cloud and a new algorithm, White and his team can analyze thousands of individual genomes at the same time. That not only allows him to help individual patients, but it also helps large-scale research efforts.

"This is just not possible to do without using the cloud," said White, who doctors call on to sequence patients' genomes. "We can analyze 1,000 genomes in a week using the cloud. If I had run that in house, it would have taken several years. Our in-house resources are just not able to scale to that level.... Being able to understand the genetic cause of disease will allow us to better treat that disease. That enhances our understanding of how to cure it."

Given the in-house constraints, White turned to Amazon Web Services' public cloud system.

"We struggled because we couldn't keep up with the machines spitting out these huge volumes of data," he said. "It was filling up our hard drives. We had about 1 petabyte reserved for our genomics work in our hospital. It was 90% full. We just can't keep building up this IT infrastructure internally."

Being able to sequence thousands of genomes at one time allows a doctor to move beyond finding a better treatment for one patient to exploring the genetic mutations that cause diseases.

"For something like heart disease, it requires sequencing thousands of individuals to find a genetic cause," he said. "If you look at just your genome, we'd typically find that you have about 4 million differences from the reference genome that we compare everything to. Then the challenge becomes how do we know which of those genetic differences causes the disease. If we can look at thousands of patients with a disease, we can see what genetic changes they have in common."

Turning to the cloud also enables doctors and researchers worldewide to share information and collaborate more easily, Nadauld said.

Doctors must worry about security and HIPAA issues when transmitting information back and forth via a currier or the mail. That process can also be cumbersome and slow.

"Before, we had tried a few times shipping information on encrypted hard drives," Nadauld said. "It just didn't work. Our partners weren't sure what they were receiving. There was difficulty getting all the data we needed into single sources. It was really slow. We need our turnaround time measured in hours, not days or weeks."

Researchers often use the public cloud, despite tight HIPAA regulations, because patient names and identifying information are not connected to the data being stored and manipulated.

And GenomeNext's Hirmas, who also uses the AWS platform, said using the cloud also dramatically cuts costs. In some cases, it's reduced the cost of genome analysis by 50% to 70%.

For exmaple, it used to cost GenomeNext about $5,000 just to run an individual's sequence. Interpreting the results added another $2,000 to $3,000. All together, according to Hirmas, the rpcoess could top $10,000.

Today, with the cloud, sequencing costs about $1,500 to $2,000, the analysis adds $800 and the entire process runs between $3,000 and $5,000.

We can't fail

In addition to saving money, cloud use delivers a higher level of reliability and scalability, according to Jonathan Hirsch, founder and president of Palo Alto, Calif.-based Syapse Inc., a software company focused on precision medicine. Intermountain is a customer of Syapse.

"Syapse can't fail," he said, noting that the company uses AWS. "The patient comes in and we need to be able to get that information for them. You need a platform behind you that you know is incredibly robust and will not go down.... Cloud is the preferred option at this point for running our service and that's a sea change. Without the cloud, it would be very hard to impossible -- especially for a company of our size."

Hirsch also pointed out that AWS has focused on its cloud services for the health care field over the past few years - even offering up a HIPAA compliance roadmap to use.

Steve Halliwell, director of healthcare and life sciences for AWS, said his operation has been building out infrastructure to take on the heavy lifting for medical companies and organizations so doctors and researchers can focus less on IT and more on medicine.

"The fax machine is still used for sharing data for physicians," said Halliwell. "You've got this atmosphere where there's plenty of data, but it's not connected. Rather than standing up servers..., these companies are able to get right to the work of working with the practitioners and doing things that will make changes in the medical process."