Patrick Soon-Shiong is remaking the US healthcare system. Here's how
Science
09 August 13 by Joao Medeiros
Over the past five years, LA-based surgeon and businessman Patrick Soon-Shiong has ploughed over $700 million (£450 million) into about 50 companies to remake the US healthcare system. He heads NantWorks, an umbrella organisation created in 2011 that includes businesses conducting research in semiconductor technology, supercomputing, artificial intelligence and cancer-drug discovery.
It's the 60-year-old South-African-born Chinese-American CEO's latest in a long list of ambitious projects. In 1993, he performed the first human islet cell transplant in diabetic patients. He developed Abraxane, a cancer drug wrapped in a coating of nanoparticles that allow the medicine to be delivered inside the tumour. By 2008, he had already founded and sold two multibillion-dollar pharmaceutical companies. According to Soon-Shiong, medicine has made dramatic advances, but its delivery has lagged behind. The key to that change is digital technology. He talks to Wired's Start editor João Medeiros about the future of his groundbreaking work.
In January, you announced the foundation of yet another company, NantOmics (part of NantWorks), to investigate cancer treatments based on knowledge that you developed in the 90s.
Yes, I was working with stem cells as part of a Nasa programme. We realised that the science of stem-cell proliferation was also fundamental to cancer cells when cancer enters the phase of metastasis. I thought there might be a unifying biology between stem-cell differentiation and cancer metastasis: the same protein could be involved in both processes. The idea that all cancer cells, regardless of type, would absorb this protein led us to create a nanoparticle of the protein and place a drug inside it. It was very difficult to convince the oncologists that we would be able to treat all tumour types. It was a magic portal to the tumour cell. We've proven in clinical trials that this is the case. The drug Abraxane is now approved in breast cancer and lung cancer in the United States, China and Europe. It has also succeeded in treating melanoma and pancreatic cancer.
Your approach to cancer research is not just limited to drugs. You're also looking at the level of the genome and the proteome [the proteins expressed by the genome].
We need a genomic and proteinomic understanding of the disease in a personalised way. When I started calculating the amount of data and computer power needed, I discovered they were an impossibility at the time. We needed to accommodate yottabytes. You have in the US around two million new diagnoses of cancer a year, and 13 million survivors, so you have about 10,000 patients that require analysis every day. That's about five petabytes that need to be transmitted and computed on a daily basis.
That's an immense computational challenge.
It's a deluge of big data. You also need an algorithm that takes all the various analyses together - genome, proteome, known pathways and networks of protein systems and patient's clinical data. You also need a way to transport this in a secure cloud to the doctor and the patient in the clinic. That was our goal and we just announced the first complete computation for more than 3,000 patients. The largest analysis ever done. We completed the transmission and computation in 69 hours -- about 47 seconds per genome.
All this adds up to the concept of convergence that underpins your vision for healthcare.
Yes. It's the convergence of semiconductors, of machine vision, supercomputing algorithms, fibre networks and medical connectivity. Collecting all the clinical data and the vital signs of a single patient throughout the course of their life didn't exist before because there was no interoperability between electronic medical records and medical devices to capture your blood pressure, your heart rate and vital signs -- what I call the human signal engine. I funded a programme which developed ways of interacting with thousands of medical devices to capture the human signal engine. Our supercomputer for the human genome has been completed and we can now complete the bioinformatics of the cancer mutation in 47 seconds rather then the 11 weeks that it currently takes. What excites me is that these systems are in the cloud, connected by our dedicated secure fibre network spanning the country.
Artificial intelligence is also a big component of your project.
We are now building a human cognition engine in the form of a smartphone. We have developed the ability to recognise physical objects in the real world and also unlock information from the internet thereby creating what I call "augmented intelligence".
That's a description of your image-recognition app, iD.
Yes, iD is a machine vision and sensor browser for the physical world. That's what we have been working on with Coca-Cola, Verizon, Bank of America and Disney to launch content when an image is recognised. We've also released an app called LookTell Recognizer for the blind that recognises images and tells the user what it is. And we created an app, called DreamPlay, that recognises physical objects and creates an augmented reality based on that image.
Your vision with NantWorks isn't just limited to healthcare.
We're also interested in wellbeing. When we harness that power of convergence, we can achieve what was impossible before. Every patient is a consumer and every consumer is a potential patient. What NantWorks is doing is building the world the way Da Vinci saw it, and augmenting every frame a human being sees as they work, live and play.
AI
