How AI is creating new opportunities in healthcare

Tim McCarty: When we think about … and when most people in this room think about artificial intelligence (AI), healthcare doesn’t immediately come to mind. It’s been a slow adopter of artificial intelligence. And this would lead some to conclude that healthcare and AI are mutually exclusive. And I’m going to try and make the case today that this is the wrong conclusion to make. And we think over the next three to five years, we’re going to see accelerated adoption of artificial intelligence in healthcare and healthcare is going to leapfrog other sectors in the market in terms of AI adoption.

AI applications in healthcare: Detecting cancer early

The opportunity for AI in cancer is earlier diagnosis to lead to better treatment outcomes. The issue with cancer is there’s only … a lot of times it’s too late to treat when symptoms show, and there’s only a handful of screening pathways today. Think mammography for breast cancer or colonoscopy for colon cancer.

Well, AI is being used in early stages to use blood-based screening of healthy or non-symptomatic individuals to screen for cancer to lead to earlier diagnosis to give people a better chance. And a real world example to bring this to life is what the Dallas Cowboys did. And with that, what quarterback Dak Prescott did is really truly inspirational.

So, unfortunately, cancer hits home to him. His mother passed away of cancer, and he has used his foundation to advocate for early screening, and he advocated the entire Cowboys organization get tested. And last summer, during training camp, a member of the media team took a blood-based screening test and was told he may have a signal for head and neck cancer. He then had a follow-up diagnostic test that confirmed he did, in fact, have stage 2 head or neck cancer. But the key is he was having no symptoms, was treated then during the bye week, and credits Dak Prescott for saving his life. So, this is a real world example to bring that to life.

Improving the patient, payor, & provider experience

So, there’s a condition called severe aortic stenosis, which has a two-year mortality rate of 50% if left untreated. However, the good news for patients is that there’s a minimally invasive approach as an alternative to open heart surgery called transcatheter aortic valve replacement that gives patients another lease on life . And artificial intelligence is better diagnosing this condition across all three areas – at the primary care doctor, at the general cardiologist, and then merging with health systems.

So, first, the normal patient pathway to this disease is done at the primary care doctor, where a primary care doctor through an annual physical must…needs to hear a heart murmur. However, it’s very imperfect. It’s a very imperfect exercise prone to errors. But AI is being incorporated into a stethoscope at the primary care doctor to reduce the error rate.

Then, a patient’s journey goes to the general cardiologist. Then at a general cardiologist, the patient will need to get a diagnostic test called an echocardiogram. However, like any diagnostic test, there are error rates, and AI is serving as an extra set of hands to the echocardiographer.

So, then leading academic medical centers in this country estimate that only 50% of their diagnosed patients get treated. So, what AI is doing is merging with electronic health records, going to health systems to proactively flag patients that have slipped through the cracks one way or another. And this is, this is a what I call the healthcare bullseye, where all three Ps in the healthcare ecosystem – the patient, the payor, and the provider – are all incented to adopt this technology.

Accelerating drug discovery

Drug discovery is a very inefficient process, and we think artificial intelligence can improve that efficiency. And then the “so what” of that is that we think longer term, this can rerate the multiples of biopharma companies. So, I like to talk about what I call a 10/10 rule to illustrate how inefficient drug discovery is. It takes, on average, 10 years from start to finish for a drug to make it to market, and then only 10% of drugs make it to market; 90% fail and that’s 5% when excluding reformulations, or me-to drugs. So, a very inefficient process.

Biopharma companies are replete with data, but didn’t really have the means to use this data for their benefit until now. But artificial intelligence is interrogating this data,
allowing them to better predict the structure and interaction of biomolecules. And we think this can lead to an improvement in the 10/10 rule. And then when you think across the healthcare space, biopharma companies have some of the best margins, but they trade at the lowest multiples because of this 10/10 rule or inefficiency. But if AI can improve that, we think it can lead to a longer-term rerating of biopharma…of biopharma multiples over the long term.