Hacking Longevity: Dr David Karow at Singularity University Global Summit
00:03
I’m going to start with this slide. You know, everyone’s talking about extension of the healthy human lifespan and the so-called ‘health span’. But what does that actually mean? We have an objective for all the clients and patients who come through HLI. It’s that they live to the age of 128, and they die doing something they love, not of disease. And the question is, how are we going to get to that point? How do we get to the point where most of us are living healthily to the age of 100? I actually think we have that right in our hands. I think the new pill is you, as Dan said. I don't think it's necessarily going to be a new drug-eluting stent or a new chemotherapy regimen. I actually think it's going to be data-driven AI that gets us to that point. Where we are right now is we have sick care, right? If you think about bladder cancer, the way that we find out if you have bladder cancer is you find that you have blood in your urine or hematuria, you go to your doctor and you find out, often too late, that you have bladder cancer.
01:00
I would argue that the next version of sick care, but moving towards well care is pre-symptomatic diagnosis. The idea that using advanced data, MRI, and genetics, we can find out pre-symptomatically you have that tumor long before it's too late. The bottom line is you still had that tumor in the first place. I think where we're moving with AI, the promise of AI, is that we can tell you about your risk for bladder cancer, or Alzheimer's, or coronary artery disease, or diabetes, one or five or ten years before disease onset, and that if you do X, Y, and Z, you never get the disease in the first place. I think that's how we get to a healthy human lifespan of 100 or above. I would argue that the reason that, in general, AI has not moved the needle on the healthy human lifespan, because everyone's talking about AI, how it's going to change the practice of medicine, right? You go to any healthcare conference, JPMorgan included AI, right?
01:55
I would argue the reason that it hasn't made a big impression on the clinical practice is we're not measuring what matters. I love my Fitbit and I love my Apple Watch, but it turns out if you really want to know about your condition now, and your risk for chronic age related disease, you've got to go much deeper than that. You've got to measure what matters. At HLI, we actually started a science experiment four years ago, and that was to enroll presumed healthy individuals to come into what we now call the Health Nucleus. They underwent every test imaginable, whole genome sequencing at 30x, whole body MRI, CT calcium scoring, microbiome, EKG, a number of other tests as well. The initial reason for doing the testing, by the way, was to explore the normal human variation and phenotype as a better way of understanding genotype - our own whole genomes.
02:45
What we discovered from these presumed healthy individuals was that they weren't so healthy after all. In fact, 40% had clinically significant disease that wasn't previously known. One in fifty, a new high-grade early stage tumor. 2.5% - roughly one in fifty with brain or aortic aneurysms that weren't previously known. Almost one in ten with advanced coronary artery disease that wasn't previously known. 30% with elevated liver fat, evidence for metabolic syndrome. And get this - on whole genome sequencing, nearly one in four have a rare genetic mutation, a rare monogenic variant like a BRCA mutation, that affects the disease of that patient individually. And these are just some examples. And this speaks to measuring what matters. I wish we could get this data from a superficial sensor wearable, but we can't. So, two women in their 30s with papillary thyroid cancer that we caught early. Note that many of these findings are in folks under the age of 50.
03:45
Often we think of cancer as something that strikes fifties, sixties, seventies, and eighties, but not necessarily, right? Lung adenocarcinoma in a female patient, 54 year old. No risk factors for lung cancer, a non-smoker, this was resected a week later. Not just tumors. Brain aneurysms. Grape size aneurysms like this that normally present with sometimes fatal brain bleeds. This one was resect or coiled a week later on an outpatient basis, because we found it before it ruptured. Advanced coronary artery disease. Again, in an individual under the age of 50 at age 45, significant calcified plaque in the left anterior descending coronary artery that supplies the left ventricle. Turns out this individual had a LDL receptor mutation on whole genome sequencing, which explains in part why she, at the age of 45, has coronaries that look like an 80 year old. Again, it speaks to measuring what matters.
04:40
The overall summary, and by the way, this is in review and available in the pre-print bio-repository bio-archives. The overall summary from the first 1200 patients that we've had through and by the way, we've seen over 5000 now, is that 40% have clinically significant disease. 14% have clinically significant life altering findings that weren't previously known. New tumors, new aneurysms, new rare monogenic mutations like BRCA mutations. So, at the end of the day, what we like to say is that we're building the story of you through our assessment. We're doing that by measuring what matters. We're using data to drive decision support for nearly all other decisions in our lives, whether they're investment decisions, our car has a thousand sensors, our social media. Why are we not using deep data to drive decision support for health and wellness? I think that's where the whole field of precision health and personalized medicine is moving.
05:38
Everyone knows this book. If you have a company, a start-up company, you're using ‘Measure What Matters’, you're using ‘OKR System’. Why are we not measuring what matters for our own health and wellness? We should be. I would argue that the fountain of youth is a reality. And it starts with data-driven AI. It starts with a fountain of data. At the end of the day, a deep, data-driven, precision health assessment allows you to know your condition now. Do I have that tumor or aneurysm? Do I have that rare monogenic variant? More importantly, based on everything that I know about you, what's your risk - your one year, five year, ten year risk for chronic age related disease? And by the way, if that risk is elevated, what's the one, two, or three risk factors that you can modify to get that risk down?
06:24
This is an example of an imaging biomarker. Imaging biomarkers that we derive from our whole body MRI. When I socialize whole body MRI, most folks think about whole body CT that was popular so many years ago. Actually, the point of whole body MRI is not screening. It's to derive imaging biomarkers, continuous measurements of chronic age related disease that we can integrate with genomic and blood biomarkers to give you a readout on your risk, long before disease onset. So take diabetes or metabolic syndrome. It's important to know your liver fat, your visceral adipose tissue, because these are going to be elevated long before your hemoglobin, A1C, or fasting blood glucose. It turns out that the biomarkers that are important for longevity are also important for your performance right now. If your liver fat is elevated, if your visceral adipose tissue is elevated, believe me, you're not jumping out of bed in the morning. You're not as energetic at work or for your family as you could be. Human performance, high performance, and longevity are definitely tied.
07:25
At the end of the day we’re giving actionable health intelligence like this to our patients and clients. This really sums up not only our value proposition, but where we think the field of precision health and personalized medicine is moving. Yes, you want to know about your risk - your condition now, but more importantly, thankfully most of the clients that come see us, they don’t have an aneurysm or a tumor, but everyone leaves with this health intelligence. Which is - for dementia, can I show you what your risk is a year from now, 5 years from now, 10 years from now, for mild cognitive impairment or dementia, long before you have disease onset. It turns out that 1/3 of the risk for Alzheimer’s and most dementias is modifiable - it’s in your control. So folks that say “I don’t want to know about my risk for dementia because there’s nothing I can do about it” - that’s just simply not true.
08:12
1/3 of the risk is modifiable. So, what this allows us to do is take imaging biomarkers, genetic biomarkers, like your APOE status, blood biomarkers, and put it all together to give you your 8 year pre-symptomatic readout. And if your risk is elevated, what’s the number one modifiable risk factor that we can use to get your risk down. We know there are six modifiable risk factors for dementia and mild cognitive impairment, that includes blood pressure, hypertension, BMI, visceral adipose tissue elevation, alcohol use, and so on. It’s a very different conversation if I say to someone “You should lose weight or lower your cholesterol” vs “Hey, we looked at everything we know about you, including your imaging and genetic biomarkers and we show that you have elevated risk for dementia, and the number one modifiable risk factor is cholesterol”, that’s a whole different motivation and incentive for lowering that cholesterol.
09:04
The Health Nucleus is really a solution to this problem, which is that, unfortunately, in this country, and it's really a global problem, that if you live to the age of 50, one in three of us won't make it to the age of 75. That's due to largely cardiovascular disease and cancer, but a number of others as well. That's largely because what's been driving decision support over the last 30 years hasn't changed. It's a stethoscope, a reflex hammer, and a tongue depressor. If we want to know about our family history, your physician asks you, what did mom and dad have? We need to move that to the next phase, the next generation, data-driven, measuring what matters, actually looking at all 6 billion base pairs of your whole genome, actually looking at your core with whole body MRI biomarkers and blood biomarkers and helping the physician with AI and machine learning. Our health nuclear assessment is based on three data silos, imaging, genetics, and blood biomarkers. The important concept is integrating all of that data together with AI, enabling the physician to offer a data-driven, precision health assessment.
10:06
I'll end with this slide. And it almost seems cliche, right? Because we can all debate about the 10 tenets of longevity. You all probably have different tenets than I have on this list, but I just want to focus on the top three. One is sleep. I would argue that all of these other tenets serve that one, because if you're not getting good sleep, you're not going to live to 100. You're not going to be a high-performing individual. Number two, Dan alluded to it, you are the new pill, right? If pills were working so great, why is our life expectancy for the first time in the last 40 years declining, over the last three years? We've given up our power to the drug industry, largely. Now, I'm a physician, so I know that there are extreme examples where pharmacotherapy is necessary. But for most of us, we're just at mildly increased or decreased risk for chronic age related disease. Most of us need a program of action, not a pill.
10:58
But number three is know your number. Because if you're doing all the right things and you're eating kale and practicing yoga, and you've got the right sleep regimen and the right diet and exercise, if you don't know that you have a BRCA mutation that's going to predispose you to breast or ovarian cancer in your 40’s, it doesn't matter if you're doing all those other things. We've got to measure what matters and know your number. I'll close with this slide, which is that medicine has historically been a clinical practice only sometimes supported by data. I think it's moving pretty quickly to a data science supported by clinicians, and I'll end with that. Thank you.