Why I Believe AI Will Break New Ground in Healthcare

Dr. John Kan
Chief Integration Officer of A*STAR
(Agency for Science Technology and Research)

Why I Believe AI Will Break New Ground in Healthcare

Dr. John Kan
Chief Integration Officer of A*STAR
(Agency for Science Technology and Research)

A few weeks ago, Historian Adrianne Mayor wrote a column in The Strait Times titled: AI and the Original Story of Pandora’s Box.

You know what’s coming. It argued how AI is an impending danger (that Prometheus or foresight warned us about) and how it will open up a Pandora’s box that threatens the very survival of humankind. A very compelling argument.

But ask any CIO excited about AI and its immense potential, and he or she’d vehemently shake their heads. Not a chance, they’d say.

That’s why when we asked Dr. John Kan, it didn’t come as a surprise: “If you say AI equals the human brain, I’d say AI is not there yet,” he said in his characteristic style.

As the CIO of Singapore’s A*STAR (Agency for Science Technology and Research)—and Deputy Chairman of A*STAR’s Computational Resource Centre—Dr. Kan should know.

“AI is not at a stage where it can replace a human fully yet. But it may change the way certain job roles are performed. As a term, AI is too broad, like IT. It’s a matter of context and how you define it,” he says.

As the debate around AI and its reputation as a job-snatcher, humankind-ender continue to hound CIOs and their organizations, AI has been steadily gaining ground in the mainstream. Take for instance, how it’s made a significant impact on industries like manufacturing, automotive, e-commerce, retail, and healthcare. Especially healthcare.

AI governance issues are important and policies need to evolve with time. If we don’t change then we will not transform ourselves to exploit its benefits.

In fact, an Accenture study noted that the healthcare AI market may hit $6.6 billion in the next three years. In 2014, that number was just $600 million. There’s no doubt that AI has well and truly arrived, despite the skepticism surrounding it. It has spoken for itself.

For someone who’s spent a significant part of his career in ICT and research—and someone who’s keen on ensuring Singapore sets an example, Dr. Kan believes AI has the potential to break new ground in three areas: Cancer detection, clinical trials, and elder care. His reading of—and about—the AI market and his experience supporting research institutes provide immense insight on how AI is aiding healthcare.

Cancer Detection

Cancer detection has been a poster child for AI. A recent study showed that a computer could detect melanoma (Skin Cancer) with nearly 10% more accuracy than dermatologists.

Scientists at A*STAR’s Genome Institute of Singapore have recently used AI to help pinpoint roots of Gastric Cancer. The project used two new AI methods to scan the entire genome of 212 Gastric Cancer tumors in a few months. According to the researchers, such an analysis would have taken 30 years to complete on a standard modern computer.

Researchers from Beth Israel Deaconess Medical Centre and Harvard Medical School have also used deep learning to train an algorithm to integrate speech recognition and image recognition to diagnose tumors.

An article in Tech Emergence explains how the researchers began with thousands of images that labelled regions showing cancerous and non-cancerous cells and how the labelled regions were extracted, which provided a multitude of examples that can be used to train the algorithm.

The study results showed that compared to human pathologists, the algorithm achieved a diagnostic success rate of 92 percent, lower than the human rate of 96 percent. However, when the algorithm and human results are combined, an accuracy rate of 99.5 percent was achieved.

It’s evident that AI is making many strides in Cancer research. At A*STAR, apart from diagnosing Gastric Cancer, researchers are also working on ways to study genome sequencing—determining the DNA sequence of an organism—of ethnic groups across Asia, says Dr. Kan.

And its projects such as these that’re guiding researchers who are looking for different ways to leverage AI to speed up the process of detection and diagnosis.

Clinical Trials and Diagnosis

Apart from Cancer detection and diagnosis, Dr. Kan believes AI and machine learning can help identify the right drugs for patients. “According to an article in Tech Emergence, through clinical trials and associated patterns, AI can help identify the right drugs for patients. This process shortens waiting time in diagnosing and prescribing the appropriate drug,” he says.

Take, for example, how AI-chatbots and speech recognition “are identifying patterns in patient symptoms to form a potential diagnosis, prevent disease and recommend an appropriate course of action,” says the article.

A majority of healthcare organizations, according to Tech Emergence, “are also combining facial recognition with machine learning to help clinicians diagnose rare diseases. Patient photos are analyzed using facial analysis and deep learning to detect phenotypes that correlate with rare genetic diseases,” reads the article.

Another article in Forbes says healthcare companies such as Johnson & Johnson are investing in advanced AI programs for competitive advantage. The article says that these technologies have driven significant scientific discoveries, including the correlation between fish oil and Raynaud’s disease.

AI is a boon for researchers looking for answers to treat rare diseases, as it helps combine different types of data from different sources to identify patterns that can lead to potential causes, symptoms, and their treatment.

Improved Elder Care

AI has lent a hand to another crucial but often neglected aspect of healthcare: elder care.

For the older generation, daily life is a struggle, and many have to rely on either family members or outside help. AI is at a stage where this dependency can be partially done away with. “A report in Smithsonian.com says home robots could help seniors with everyday tasks and allow them to stay independent and in their homes for as long as possible, which improves their overall well-being. That’s a great way to improve elder care,” says Dr. Kan.

The potential benefits of AI for elder care are immense. Elders no longer need to travel to receive care, for example.

Today, AI-enabled robot companions are being used to remind elders to take their medicines or even arrange cabs. Dr. Kan gives an example of the Toyohashi University of Technology, which has developed a way to estimate various human poses using deep learning with twin cameras placed in elderly care robots. Human pose detection is key to ensuring fast responses when accidents or falls occur, which is very common when it comes to the elderly.

That’s another area in which AI can play a significant role in reducing dependency on humans.

The Governance Puzzle

But for any of that to happen, organizations must first take steps to solve the AI governance puzzle. Like with any new technology, new threats are always part of the package. And governance and privacy could be major stumbling blocks.

“So we’ve been struggling with these governance concepts, ideas, and issues all the time. In a simplistic way, I could say that we always thought about people, process, technology, and now we need to also consider governance and policy issues too. So in each of these four areas, operational issues need to be addressed as an integrated whole. For example, do you allow end users to use scientific open-source applications in a secure corporate IT environment, or allow staff to use less secure mobile application QR codes or not? Who should have access to data? Answering these simple things could be a challenge,” says Dr. Kan.

And then, he says, there are the ethical questions which are also a struggle in the biomedical research field. “How do you do clinical trials? Will AI replace humans? I don’t think we have all the answers to these things, but, the point is that I don’t think AI can replace the human brain to decide on such issues,” he says.

Dr. Kan believes governance issues are important but policies need to evolve with time due to technological advances. “If we don’t change then we will not transform ourselves,” he says.

After all, AI makes everything good possible, doesn’t it? Pandora’s box can wait.

Dr. John Kan
Chief Integration Officer of A*STAR (Agency for Science Technology and Research)

Dr. John Kan is the CIO of A*STAR (Agency for Science Technology and Research), Deputy Chairman of A*STAR’s Computational Resource Centre, President SingAREN, and Board Member of APAN. He has more than 20 years of experience in the Infocom and Telecom sector. Prior to joining A*STAR, Dr. Kan held senior management positions in Teletech Asia, AT&T, Opentext-Vignette, and Siemens, to name a few. He is also a member of the ITMA, Singapore International Chamber of Commerce, and the Singapore IT Federation.