1 00:00:15,356 --> 00:00:15,796 Speaker 1: Pushkin. 2 00:00:20,196 --> 00:00:21,956 Speaker 2: The story of heart attacks used to be a story 3 00:00:22,076 --> 00:00:26,356 Speaker 2: basically about plumbing. You know, the arteries that feed the 4 00:00:26,396 --> 00:00:29,676 Speaker 2: heart with blood are like pipes, and over time and 5 00:00:29,756 --> 00:00:32,996 Speaker 2: some people, fatty plaque builds up on the inside of 6 00:00:33,036 --> 00:00:35,996 Speaker 2: those arteries, like gunk on the inside of a pipe. 7 00:00:36,436 --> 00:00:39,676 Speaker 2: And as the plaque builds up more and more, the 8 00:00:39,756 --> 00:00:43,436 Speaker 2: space where blood can flow through the artery gets narrower 9 00:00:43,476 --> 00:00:47,876 Speaker 2: and narrower, and eventually a blood clot blocks that narrow 10 00:00:48,036 --> 00:00:51,436 Speaker 2: artery all together. That's heart attack. And so a key 11 00:00:51,476 --> 00:00:54,196 Speaker 2: thing that doctors looked at when they were trying to 12 00:00:54,236 --> 00:00:58,556 Speaker 2: prevent heart attacks was how much narrowing is there, how 13 00:00:58,596 --> 00:01:02,396 Speaker 2: clogged is the pipe? Over time, though, doctors noticed a 14 00:01:02,396 --> 00:01:06,956 Speaker 2: surprising pattern. Lots of people were having heart attacks even 15 00:01:07,036 --> 00:01:09,796 Speaker 2: when they didn't have a ton of plats build up, 16 00:01:10,036 --> 00:01:12,996 Speaker 2: even when their pipes weren't particularly clogged to begin with. 17 00:01:13,636 --> 00:01:15,676 Speaker 2: It took a while to figure out what was going on, 18 00:01:16,116 --> 00:01:18,516 Speaker 2: and it's taken even longer to figure out what to 19 00:01:18,556 --> 00:01:21,476 Speaker 2: do about it. Heart disease is still the biggest killer 20 00:01:21,476 --> 00:01:24,996 Speaker 2: in the world, but in the past few years a 21 00:01:25,036 --> 00:01:28,596 Speaker 2: bunch of new technologies have come together, and it seems 22 00:01:28,716 --> 00:01:32,716 Speaker 2: like a new, better approach to preventing heart attacks may 23 00:01:32,796 --> 00:01:42,036 Speaker 2: be here very soon. I'm Jamie Goldstein and this is 24 00:01:42,036 --> 00:01:44,036 Speaker 2: What's Your Problem, the show where I talk to people 25 00:01:44,036 --> 00:01:47,636 Speaker 2: who are trying to make technological progress. My guest today 26 00:01:47,756 --> 00:01:51,556 Speaker 2: is Eric Topel. He's a cardiologist and the founder of 27 00:01:51,596 --> 00:01:56,036 Speaker 2: the Script's Research Translational Institute in San Diego. He's also 28 00:01:56,156 --> 00:01:59,116 Speaker 2: one of the most cited medical researchers in the world. 29 00:01:59,916 --> 00:02:03,276 Speaker 2: His problem is this, how can doctors do a better 30 00:02:03,356 --> 00:02:07,436 Speaker 2: job at preventing heart attacks. Topel was early to this 31 00:02:07,556 --> 00:02:11,076 Speaker 2: shift in thinking about heart age, this shift away from 32 00:02:11,116 --> 00:02:14,876 Speaker 2: the plumbing model. The new model, he told me, focus 33 00:02:14,996 --> 00:02:17,636 Speaker 2: is not on how much plaque has built up in 34 00:02:17,676 --> 00:02:20,556 Speaker 2: the arteries around the heart, but on what's going on 35 00:02:20,636 --> 00:02:24,436 Speaker 2: with the plaque that's there, because even a small amount 36 00:02:24,476 --> 00:02:26,596 Speaker 2: of plaque can put patients at high risk of a 37 00:02:26,596 --> 00:02:30,316 Speaker 2: heart attack if that plaque is inflamed and unstable and 38 00:02:30,476 --> 00:02:33,676 Speaker 2: likely to rupture. He recently wrote an essay about this 39 00:02:33,716 --> 00:02:38,116 Speaker 2: shift and about the emerging screenings and interventions that could 40 00:02:38,236 --> 00:02:42,356 Speaker 2: lead to fewer heart attacks. To start, we talked about 41 00:02:42,356 --> 00:02:44,876 Speaker 2: how doctors figure out who is at high risk in 42 00:02:44,916 --> 00:02:48,876 Speaker 2: the first place. Topel is particularly optimistic about a new 43 00:02:48,916 --> 00:02:52,636 Speaker 2: approach that combines a CT scan of the arteries around 44 00:02:52,636 --> 00:02:56,516 Speaker 2: the heart with an AI model that detects inflammation in 45 00:02:56,556 --> 00:02:59,716 Speaker 2: those arteries. And he told me about one recent study 46 00:02:59,796 --> 00:03:02,076 Speaker 2: in the United Kingdom that used that approach. 47 00:03:02,236 --> 00:03:06,916 Speaker 3: The results were pretty striking. Forty thousand consecutive people in 48 00:03:06,996 --> 00:03:10,716 Speaker 3: the UK they use this tech and they found that 49 00:03:10,756 --> 00:03:14,076 Speaker 3: people with inflamed arteries with no narring had a very 50 00:03:14,196 --> 00:03:19,596 Speaker 3: high risk of heart attack and death. And so now 51 00:03:19,596 --> 00:03:23,396 Speaker 3: we have a way without an invasive technology to map 52 00:03:23,436 --> 00:03:28,076 Speaker 3: out a person's arteries, which is exciting, and that is 53 00:03:28,356 --> 00:03:30,156 Speaker 3: you know, if you had all three arteries of flame, 54 00:03:30,436 --> 00:03:32,876 Speaker 3: the risk ratio for having a heart attack over the 55 00:03:32,916 --> 00:03:35,716 Speaker 3: next years is like thirty fold, I mean huge increase. 56 00:03:36,076 --> 00:03:38,636 Speaker 3: If one artery inflamed is still important. So we have 57 00:03:38,796 --> 00:03:44,436 Speaker 3: now moved in the whole dogma of cardiology from blockages 58 00:03:45,036 --> 00:03:46,516 Speaker 3: to inflammation. 59 00:03:47,476 --> 00:03:52,796 Speaker 2: And so there are several companies right that are doing 60 00:03:52,796 --> 00:03:55,516 Speaker 2: this now, that are doing CT scans of the heart 61 00:03:55,556 --> 00:03:59,196 Speaker 2: plus AI to assess somebody's risk of having a heart attack. 62 00:03:59,676 --> 00:04:01,476 Speaker 2: Just talk me through those companies kind of what they 63 00:04:01,556 --> 00:04:04,716 Speaker 2: figured out, what they're offering, like what's beenef to approved, 64 00:04:04,756 --> 00:04:06,716 Speaker 2: what's covered by insurance if you know that, Like what's 65 00:04:06,716 --> 00:04:07,356 Speaker 2: going on with that? 66 00:04:07,756 --> 00:04:11,316 Speaker 3: Yeah, there's four of them now and they use different 67 00:04:11,356 --> 00:04:17,636 Speaker 3: technology with AI, so you know, the one that is 68 00:04:18,076 --> 00:04:20,636 Speaker 3: soon to be approved by FDA is a University of 69 00:04:20,676 --> 00:04:24,436 Speaker 3: Oxford spin out from the UK. It's called Karista, and 70 00:04:24,476 --> 00:04:28,836 Speaker 3: it's the one that really directly uses AI to gauge inflammation, 71 00:04:28,956 --> 00:04:31,436 Speaker 3: like I described, with the big data sets of the 72 00:04:31,476 --> 00:04:34,636 Speaker 3: forty thousand people and the outcomes. 73 00:04:35,116 --> 00:04:38,436 Speaker 2: I mean, I will say when you're just for me 74 00:04:38,556 --> 00:04:41,676 Speaker 2: knowing essentially nothing about the specifics of this case and 75 00:04:41,756 --> 00:04:44,796 Speaker 2: knowing a little bit about aim more generally, like there's 76 00:04:44,796 --> 00:04:47,436 Speaker 2: some level where what you really want is just the 77 00:04:47,476 --> 00:04:52,756 Speaker 2: biggest data set you can get and of coronary CT 78 00:04:52,916 --> 00:04:56,276 Speaker 2: scans and clinical outcomes and then just throw it at 79 00:04:56,276 --> 00:04:57,956 Speaker 2: the AI and say you figure it out. 80 00:04:58,196 --> 00:05:02,436 Speaker 3: Yeah, well that's basically what Charista, or well, Universe of 81 00:05:02,436 --> 00:05:04,276 Speaker 3: Oxford then they had a spin out company. 82 00:05:04,356 --> 00:05:05,276 Speaker 1: That's what they did. 83 00:05:05,516 --> 00:05:09,636 Speaker 3: So you're absolutely right, you want basically auto d idactic 84 00:05:09,716 --> 00:05:11,476 Speaker 3: to let the data see for itself. 85 00:05:11,556 --> 00:05:13,596 Speaker 1: It don't have any biases. 86 00:05:13,196 --> 00:05:16,156 Speaker 2: Like like the AI that beat go you know, like 87 00:05:16,316 --> 00:05:19,156 Speaker 2: they try and teach it the rules. They just said, 88 00:05:19,196 --> 00:05:21,636 Speaker 2: go play a billion games ago and figure it out 89 00:05:21,676 --> 00:05:28,076 Speaker 2: and it did right exactly. So you're saying the FDA 90 00:05:28,236 --> 00:05:32,036 Speaker 2: is likely to approve this test soon. Oh yeah, So 91 00:05:32,076 --> 00:05:36,396 Speaker 2: say the FDA approves this, like there's some universe of 92 00:05:36,476 --> 00:05:39,716 Speaker 2: people who maybe they haven't had a heart attack, but 93 00:05:39,756 --> 00:05:42,436 Speaker 2: they know they're at high risk for the classic reasons, right, 94 00:05:43,436 --> 00:05:45,036 Speaker 2: or they think they may be at high risk, and 95 00:05:45,076 --> 00:05:51,276 Speaker 2: this can give them a better sense, assuming that insurance 96 00:05:51,316 --> 00:05:53,516 Speaker 2: will pay for it, or people have had a pucket. 97 00:05:53,596 --> 00:05:55,116 Speaker 2: What do you do about it? What if I go 98 00:05:55,196 --> 00:05:57,556 Speaker 2: get this test and it says you're at high risk, 99 00:05:57,676 --> 00:05:59,876 Speaker 2: like I'm already taking a stat and I already exercise, 100 00:05:59,956 --> 00:06:01,556 Speaker 2: I don't smoke. What does it mean? 101 00:06:02,116 --> 00:06:03,996 Speaker 1: There's always more you can do? Right? 102 00:06:04,236 --> 00:06:08,156 Speaker 3: Well, So first, okay, firstly, on the LDL story, we 103 00:06:08,236 --> 00:06:11,156 Speaker 3: can get that down to very low levels. Yeah, and 104 00:06:11,196 --> 00:06:13,436 Speaker 3: that will help, you know, getting it down well below 105 00:06:13,756 --> 00:06:17,436 Speaker 3: to thirty twenty, you know. Okay, So we have so 106 00:06:17,556 --> 00:06:18,716 Speaker 3: many tools to do that. 107 00:06:19,716 --> 00:06:25,236 Speaker 2: So LDL is bad cholesterol basically and typically right, typically 108 00:06:25,236 --> 00:06:27,596 Speaker 2: the guideline is you wanted under one hundred. But you're 109 00:06:27,596 --> 00:06:30,316 Speaker 2: telling me, if a patient is at high risk, it's 110 00:06:30,316 --> 00:06:31,876 Speaker 2: good to get it much lower than that. 111 00:06:32,556 --> 00:06:35,596 Speaker 3: If your ct says you've got three inflamed guarteries. I 112 00:06:35,756 --> 00:06:39,076 Speaker 3: wanted to get your LDL down to you know, twenty 113 00:06:39,156 --> 00:06:41,596 Speaker 3: or ten. Yeah, okay now, and we have the way 114 00:06:41,636 --> 00:06:43,156 Speaker 3: to do that in almost everyone. 115 00:06:43,716 --> 00:06:45,076 Speaker 1: But that's just the beginning. 116 00:06:45,956 --> 00:06:49,436 Speaker 3: Then we check your LP LITTLEA because a lot of 117 00:06:49,436 --> 00:06:52,756 Speaker 3: people they're ldl's okay, but their LP little A lipoprotein 118 00:06:52,836 --> 00:06:53,876 Speaker 3: A is high. 119 00:06:53,956 --> 00:06:56,076 Speaker 1: And later this year we're going to have drugs to. 120 00:06:56,036 --> 00:06:59,036 Speaker 3: Treat that we've never had that before, okay, And that's 121 00:06:59,076 --> 00:07:01,556 Speaker 3: a high proportion of people that have that risk factor. 122 00:07:02,716 --> 00:07:07,436 Speaker 3: We would measure inflammation in your blood with things like 123 00:07:07,916 --> 00:07:12,076 Speaker 3: high sensitivity c P and cytostat and other emerging markers, 124 00:07:12,236 --> 00:07:14,396 Speaker 3: and we could use that to say, oh, well, we're 125 00:07:14,436 --> 00:07:16,516 Speaker 3: bringing it down and that's really good. 126 00:07:16,756 --> 00:07:19,036 Speaker 2: Let's just talk about inflammation for a seconds. That's where 127 00:07:19,036 --> 00:07:21,876 Speaker 2: we are, and like people have some kind of sense 128 00:07:21,916 --> 00:07:25,996 Speaker 2: of what inflammation is, but also not right. 129 00:07:26,076 --> 00:07:32,556 Speaker 3: So what is inflammation, Well, inflammation is a response to. 130 00:07:34,556 --> 00:07:35,076 Speaker 1: Injury. 131 00:07:37,076 --> 00:07:39,076 Speaker 3: You know, we've been talking about in the artery, so 132 00:07:39,116 --> 00:07:41,996 Speaker 3: it can use that. So let's say you're a smoke 133 00:07:42,116 --> 00:07:46,716 Speaker 3: or you have high blood pressure, your diabetes, you may 134 00:07:46,796 --> 00:07:51,116 Speaker 3: have some injury of the lining of your artery, and 135 00:07:51,316 --> 00:07:54,916 Speaker 3: the immune system comes along and picks that up because 136 00:07:54,956 --> 00:07:58,476 Speaker 3: it's trying to to seal it out, to heal it. 137 00:07:58,836 --> 00:08:03,716 Speaker 3: But unfortunately it backfires because cells from the immune system 138 00:08:04,076 --> 00:08:07,756 Speaker 3: get into the artery where there's been some injury, and 139 00:08:07,876 --> 00:08:11,756 Speaker 3: those cells have a mind their own and they produce 140 00:08:12,196 --> 00:08:17,156 Speaker 3: so called cytokines chemokind. These are proteins that then bring 141 00:08:17,236 --> 00:08:20,436 Speaker 3: in more self and it sets up a vicious cycle 142 00:08:20,796 --> 00:08:25,556 Speaker 3: of inflammation. So inflammation stems from the immune system. Of 143 00:08:26,316 --> 00:08:31,236 Speaker 3: those are white cells that include lymphocytes macrophages. But the 144 00:08:31,356 --> 00:08:36,836 Speaker 3: problem is that they can unleash inflammation. So it's part 145 00:08:36,876 --> 00:08:40,356 Speaker 3: of our healing inflammation is it can be a healing 146 00:08:40,396 --> 00:08:45,356 Speaker 3: response and it can be an untoward injury disease causing response. 147 00:08:45,436 --> 00:08:48,036 Speaker 2: Yeah, and when you get the disease causing response, say 148 00:08:48,076 --> 00:08:51,116 Speaker 2: in the arteries, like what is the what is the 149 00:08:51,156 --> 00:08:54,276 Speaker 2: sort of mechanism? Why is it? So I get that 150 00:08:54,356 --> 00:08:56,636 Speaker 2: there is some injury to the artery and the immune 151 00:08:56,636 --> 00:08:59,516 Speaker 2: system responds and that's like so far, so good. Yeah, 152 00:08:59,556 --> 00:09:01,636 Speaker 2: And if everything goes well, it fixes it and then 153 00:09:01,676 --> 00:09:04,476 Speaker 2: it's done and the artery goes on as morral. But 154 00:09:04,556 --> 00:09:06,996 Speaker 2: sometimes that doesn't happen, and you have the unhappy inflammation 155 00:09:07,036 --> 00:09:09,676 Speaker 2: story where it keeps feeding on itself and instead of 156 00:09:09,876 --> 00:09:13,556 Speaker 2: leaving well enough alone, it just keeps being inflamed. What 157 00:09:13,716 --> 00:09:16,236 Speaker 2: happens then? Why is that bad? Why does that cause 158 00:09:16,236 --> 00:09:17,276 Speaker 2: a heart attack? Eventually? 159 00:09:17,956 --> 00:09:20,716 Speaker 3: Oh? Well, I mean the common thread here is the 160 00:09:20,756 --> 00:09:24,636 Speaker 3: heart attack is a rupture. It's like a burst, you know. 161 00:09:24,836 --> 00:09:30,236 Speaker 3: Finally the artery is hot. It's just accumulating more of 162 00:09:30,276 --> 00:09:36,236 Speaker 3: this inflamed constituents and it just eventually it just leads 163 00:09:36,276 --> 00:09:42,276 Speaker 3: to that lining so called endothelium. It splits and then 164 00:09:42,316 --> 00:09:45,436 Speaker 3: the body says that, oh, we better go here and 165 00:09:45,436 --> 00:09:48,556 Speaker 3: clot that to seal it up, and unfortunately that clot 166 00:09:49,036 --> 00:09:51,916 Speaker 3: can cut off the BUTD supply and cause a heart attack. 167 00:09:52,116 --> 00:09:56,236 Speaker 3: And by the way, that inflammation and arteries is probably 168 00:09:57,276 --> 00:10:04,196 Speaker 3: that response. There's unquestionably there's some genetic story going on there. Okay, 169 00:10:04,476 --> 00:10:09,916 Speaker 3: So that why would some people be exubert inflammation response 170 00:10:09,916 --> 00:10:13,276 Speaker 3: whereas others you know that they don't really. 171 00:10:13,676 --> 00:10:14,356 Speaker 1: Get into that. 172 00:10:14,596 --> 00:10:18,796 Speaker 3: And so the point being is we have new drugs 173 00:10:18,836 --> 00:10:22,836 Speaker 3: and ways besides lifestyle, which you know, seventy five percent 174 00:10:22,876 --> 00:10:27,356 Speaker 3: of Americans don't even have the basal minimum exercise, physical activity. 175 00:10:27,436 --> 00:10:29,356 Speaker 1: Right, so you know there's. 176 00:10:29,156 --> 00:10:34,916 Speaker 3: Things things like sleep. Sleep health influences your body level 177 00:10:34,956 --> 00:10:37,836 Speaker 3: of inflammation. By the way, we have these clocks now, 178 00:10:38,436 --> 00:10:42,956 Speaker 3: artery clock, heart clock, immune system clock. These are collection 179 00:10:43,116 --> 00:10:46,356 Speaker 3: of proteins from your blood, from your plasma. They can 180 00:10:46,396 --> 00:10:50,676 Speaker 3: tell us if your artery arteries are aging faster than 181 00:10:50,676 --> 00:10:52,756 Speaker 3: the other organs in your body. That would be a 182 00:10:52,796 --> 00:10:55,396 Speaker 3: real tip off that something's not good here. So we 183 00:10:55,436 --> 00:10:58,596 Speaker 3: could find we could find the high risk people. We 184 00:10:58,636 --> 00:11:00,556 Speaker 3: can use not only find it, but then when you 185 00:11:00,556 --> 00:11:04,636 Speaker 3: can use these various blood tests to say, are we 186 00:11:04,716 --> 00:11:08,156 Speaker 3: making headway now that we got your sleep better, and 187 00:11:08,196 --> 00:11:10,316 Speaker 3: we got your exercise more and we have you on 188 00:11:10,356 --> 00:11:13,436 Speaker 3: a much better healthier diet. We got your LDL down 189 00:11:13,476 --> 00:11:17,556 Speaker 3: to twenty, what's your artery clock doing, what are your 190 00:11:17,596 --> 00:11:20,396 Speaker 3: inflammation markers doing, what's your immune system look like? 191 00:11:20,476 --> 00:11:24,356 Speaker 1: Now? We never had this stuff before. That's what's so extraordinary. 192 00:11:24,436 --> 00:11:28,356 Speaker 2: And everything you're talking about is available now. You're not 193 00:11:28,396 --> 00:11:29,916 Speaker 2: talking about what will be available. 194 00:11:30,276 --> 00:11:33,116 Speaker 3: Everything I've talked about except the LPA drugs which are 195 00:11:33,196 --> 00:11:38,636 Speaker 3: later this year, and also the multimodal AI packages that 196 00:11:39,036 --> 00:11:42,396 Speaker 3: take all your data. You're electronic health threat get your exposures, 197 00:11:42,676 --> 00:11:47,916 Speaker 3: your genetics, your proteins markers. That is, no doctor can 198 00:11:48,316 --> 00:11:51,476 Speaker 3: basically do this analysis in their head or even if 199 00:11:51,516 --> 00:11:54,556 Speaker 3: you if you lock them up in an office. 200 00:11:54,156 --> 00:11:56,556 Speaker 1: They couldn't do the analys You need help. 201 00:11:56,916 --> 00:11:59,836 Speaker 3: So this is another way AI kicks in. And then 202 00:11:59,876 --> 00:12:03,276 Speaker 3: there's other things like, for example, your retina. Did you 203 00:12:03,356 --> 00:12:07,996 Speaker 3: know your retina picture is incredible gateway to your risk. 204 00:12:07,876 --> 00:12:08,636 Speaker 1: For heart disease? 205 00:12:09,476 --> 00:12:13,756 Speaker 2: I didn't know it until I read your essay. We're discussing, Yes, it's. 206 00:12:13,676 --> 00:12:16,516 Speaker 3: A remarkable and we're not using you and someday we'll 207 00:12:16,516 --> 00:12:19,156 Speaker 3: be taking the selfies of our retina and getting AI 208 00:12:19,236 --> 00:12:23,156 Speaker 3: interpretation of that. The key here is identifying who is 209 00:12:23,196 --> 00:12:27,356 Speaker 3: the vulnerable patient and then getting all over that all 210 00:12:27,396 --> 00:12:28,116 Speaker 3: the ways we know. 211 00:12:28,636 --> 00:12:35,236 Speaker 2: So here is an interesting coincidence. My father lived in 212 00:12:35,276 --> 00:12:37,316 Speaker 2: San Diego for a long time, and one time I 213 00:12:37,396 --> 00:12:39,636 Speaker 2: was visiting him there and he happened to be going 214 00:12:39,636 --> 00:12:43,676 Speaker 2: to visit his cardiologist at Scripts where you work. His 215 00:12:43,796 --> 00:12:48,356 Speaker 2: name was Evan Muse, and I was just chatting. I 216 00:12:48,396 --> 00:12:50,756 Speaker 2: was sitting in the exam room chatting with the doctor 217 00:12:50,796 --> 00:12:53,996 Speaker 2: and he mentioned that he was running this was an 218 00:12:53,996 --> 00:12:58,836 Speaker 2: experimental project at the time to assess your genetic risk 219 00:12:58,996 --> 00:13:03,396 Speaker 2: for heart attack, and essentially every man in my family forever, 220 00:13:03,476 --> 00:13:04,956 Speaker 2: as far as I know, has had a heart attack, 221 00:13:05,516 --> 00:13:07,356 Speaker 2: many of them relatively hung. So I said I should 222 00:13:07,396 --> 00:13:10,636 Speaker 2: do that, and I took the test, and unsurprisingly, I 223 00:13:10,676 --> 00:13:14,396 Speaker 2: have an extremely high genetic risk for heart attack. So 224 00:13:14,436 --> 00:13:16,876 Speaker 2: I started taking a statin. It's basically the end of 225 00:13:16,876 --> 00:13:20,676 Speaker 2: that story. So anyways, I'm glad that I learned about 226 00:13:20,716 --> 00:13:23,916 Speaker 2: that test. It still seems pretty, at least at that time, 227 00:13:23,996 --> 00:13:26,436 Speaker 2: kind of fringy. My primary care doctor didn't tell me 228 00:13:26,436 --> 00:13:30,796 Speaker 2: about it, Like what's happening with genetic tests for heart 229 00:13:30,836 --> 00:13:31,396 Speaker 2: attack risk? 230 00:13:32,316 --> 00:13:33,836 Speaker 1: I'm so glad you're asked me about that. 231 00:13:34,836 --> 00:13:37,996 Speaker 3: By the way, Evan Muse and I share offices at 232 00:13:37,996 --> 00:13:41,916 Speaker 3: Scripts Clinic. I trained Evan as a cardiologist and really 233 00:13:41,956 --> 00:13:46,876 Speaker 3: proud of him. Now, polygenic risk scores should be universal. 234 00:13:47,516 --> 00:13:52,636 Speaker 3: They're very low cost, they're very accurate now to help 235 00:13:52,996 --> 00:13:55,716 Speaker 3: gauge risks, as they did in you. So you are 236 00:13:55,756 --> 00:13:56,796 Speaker 3: a vulnerable patient. 237 00:13:57,436 --> 00:13:57,636 Speaker 1: Now. 238 00:13:57,716 --> 00:14:01,716 Speaker 3: The difference polygenic risk is this is now looking at 239 00:14:01,796 --> 00:14:06,036 Speaker 3: hundreds of gene variants that collectively, even if you know 240 00:14:06,076 --> 00:14:09,716 Speaker 3: what's really important is you knew it from your family tree. 241 00:14:09,796 --> 00:14:11,596 Speaker 3: But you might not know it from your family tree 242 00:14:11,636 --> 00:14:14,636 Speaker 3: because there's a unique combination of variants from both mother 243 00:14:14,716 --> 00:14:17,596 Speaker 3: and father, so it may not be so simple as 244 00:14:17,636 --> 00:14:22,516 Speaker 3: a father or mother family lineage. Now, the reason why 245 00:14:22,516 --> 00:14:25,516 Speaker 3: it's so important is, Okay, it tells you, yes, you're 246 00:14:25,556 --> 00:14:28,796 Speaker 3: at high risk or not all right, and if you're 247 00:14:28,796 --> 00:14:31,316 Speaker 3: at low risk or even moderate, you can feel a 248 00:14:31,316 --> 00:14:34,516 Speaker 3: lot more assured that that's It doesn't mean you should 249 00:14:34,516 --> 00:14:37,156 Speaker 3: go out and eat bad things and smoke and whatnot, 250 00:14:37,156 --> 00:14:40,276 Speaker 3: but at least you're not in this really high risk group. Now, 251 00:14:41,156 --> 00:14:43,636 Speaker 3: the problem with it is it doesn't tell you when. 252 00:14:44,956 --> 00:14:48,636 Speaker 3: So just as yes, okay, so it doesn't mean you're 253 00:14:48,756 --> 00:14:51,676 Speaker 3: have a heart attack, it's probably coming, but we don't 254 00:14:51,716 --> 00:14:54,236 Speaker 3: know when. It could be when you're ninety, and then 255 00:14:54,316 --> 00:14:57,076 Speaker 3: you know, maybe you're not even going to live to ninety. 256 00:14:57,116 --> 00:15:00,636 Speaker 3: You may have something else. So the reason it's different 257 00:15:00,716 --> 00:15:04,516 Speaker 3: now is we have new models and AI that cannot 258 00:15:04,596 --> 00:15:07,196 Speaker 3: just tell us like polygenetic risk or added to poly 259 00:15:07,556 --> 00:15:11,796 Speaker 3: risk or but when that's we never had this before. 260 00:15:12,036 --> 00:15:14,436 Speaker 2: So you're saying the new model is not just based 261 00:15:14,436 --> 00:15:17,196 Speaker 2: on the genetics or it's all the things you're talking about. 262 00:15:17,236 --> 00:15:18,956 Speaker 2: You put in that you put in your scans, you 263 00:15:18,956 --> 00:15:20,876 Speaker 2: put in everything you can. 264 00:15:21,196 --> 00:15:24,636 Speaker 3: Put in the whole kitchen sink of data. We call 265 00:15:24,636 --> 00:15:28,236 Speaker 3: it the stack, right, the stack of data with your proteins, 266 00:15:28,276 --> 00:15:32,796 Speaker 3: your genetics, your biomarkers, your scans, the retina, the works. 267 00:15:33,156 --> 00:15:35,436 Speaker 3: And you say, okay, you know what, You're going to 268 00:15:35,516 --> 00:15:39,116 Speaker 3: have a heart attack at age sixty seven plus or 269 00:15:39,196 --> 00:15:44,036 Speaker 3: minus a year unless we do some big things change 270 00:15:44,076 --> 00:15:44,636 Speaker 3: to change that. 271 00:15:44,956 --> 00:15:47,756 Speaker 2: Well, you just said about that. We put it in 272 00:15:47,796 --> 00:15:50,716 Speaker 2: and they I said sixty seven plus or whatever. How 273 00:15:50,796 --> 00:15:53,796 Speaker 2: much how true is that? How much are you kind 274 00:15:53,836 --> 00:15:55,636 Speaker 2: of making that up? And how much is that actually 275 00:15:55,676 --> 00:15:55,956 Speaker 2: a thing? 276 00:15:56,036 --> 00:15:58,756 Speaker 1: Right now? I'm not making it up. This was a shock. 277 00:15:58,876 --> 00:15:59,876 Speaker 1: This was a big shock. 278 00:15:59,956 --> 00:16:03,196 Speaker 3: Okay, this is the most exciting thing about AI is 279 00:16:04,116 --> 00:16:08,956 Speaker 3: predicting preventing disease. Yeah, people talk about accuracy and whatnot. Now, 280 00:16:09,236 --> 00:16:12,196 Speaker 3: the paper I'm referring to was in the journal Nature. 281 00:16:12,556 --> 00:16:17,316 Speaker 3: It was done from Morris Gerstung from Germany and his colleagues. 282 00:16:17,876 --> 00:16:20,196 Speaker 3: And what they did was they took four hundred thousand 283 00:16:20,236 --> 00:16:24,956 Speaker 3: people in the UK biobank and they basically got their 284 00:16:24,996 --> 00:16:28,956 Speaker 3: electronic health records and they also looked at polygenic risk 285 00:16:29,036 --> 00:16:33,676 Speaker 3: for and they could predict twelve hundred different diseases, including 286 00:16:33,716 --> 00:16:38,236 Speaker 3: heart attack, not just yes or no, but when Yeah, 287 00:16:38,276 --> 00:16:39,836 Speaker 3: and it works out of sample. 288 00:16:39,956 --> 00:16:40,956 Speaker 2: It works out of sample. 289 00:16:41,516 --> 00:16:43,076 Speaker 1: Now you see you had all the right question. 290 00:16:43,436 --> 00:16:43,716 Speaker 2: Yeah. 291 00:16:43,996 --> 00:16:47,556 Speaker 3: Then they took it to one point nine million people 292 00:16:47,676 --> 00:16:51,156 Speaker 3: in Denmark and it was as good or better than 293 00:16:51,156 --> 00:16:54,836 Speaker 3: in the four hundred thousand UK buy. Okay, okay, So 294 00:16:54,996 --> 00:16:57,636 Speaker 3: we have it's called it's called the del Phi two 295 00:16:57,876 --> 00:16:58,516 Speaker 3: M model. 296 00:16:59,196 --> 00:17:01,556 Speaker 2: Is there enough data on me, just as a normal 297 00:17:01,636 --> 00:17:03,996 Speaker 2: patient who goes to the doctor every year and is 298 00:17:04,036 --> 00:17:06,556 Speaker 2: basically healthy to put it into this model? Can I? 299 00:17:06,636 --> 00:17:08,116 Speaker 2: Can I go use it right now? 300 00:17:08,636 --> 00:17:12,196 Speaker 3: Well, they're not making it yet available for everyone, but 301 00:17:12,476 --> 00:17:13,156 Speaker 3: it's imminent. 302 00:17:13,556 --> 00:17:14,876 Speaker 2: That seems like a big deal. 303 00:17:15,356 --> 00:17:15,676 Speaker 1: It is. 304 00:17:16,236 --> 00:17:18,036 Speaker 3: I thought it was one of the most important papers 305 00:17:18,036 --> 00:17:19,556 Speaker 3: ever published in medicine. 306 00:17:19,796 --> 00:17:22,116 Speaker 1: I talked, I talked to more. It's about it. 307 00:17:23,356 --> 00:17:27,036 Speaker 3: He's he's a top flight researcher. There's no hype here. 308 00:17:27,076 --> 00:17:30,236 Speaker 3: And his team that did this they never expected. 309 00:17:30,196 --> 00:17:33,076 Speaker 2: It's not even a CT scan, right, it's you just 310 00:17:33,316 --> 00:17:35,876 Speaker 2: it's data I have already. It's just a smart way 311 00:17:35,916 --> 00:17:37,876 Speaker 2: to parse that data. That's part of what's so exciting 312 00:17:37,916 --> 00:17:38,756 Speaker 2: about it, right. 313 00:17:38,876 --> 00:17:42,716 Speaker 3: Yeah, it's just because as we humans can't see stuff 314 00:17:43,236 --> 00:17:47,676 Speaker 3: that multimodal AI can see. Yeah, interpret, and that was 315 00:17:47,956 --> 00:17:49,876 Speaker 3: not even with some of these other things that we've 316 00:17:49,916 --> 00:17:51,316 Speaker 3: been discussing, right. 317 00:17:51,276 --> 00:17:53,636 Speaker 2: Right, these are just normal patients who go to the 318 00:17:53,676 --> 00:17:57,716 Speaker 2: doctor in the UK. They're not getting weird fancy blood 319 00:17:57,756 --> 00:17:59,036 Speaker 2: tests or scans or whatever. 320 00:17:59,156 --> 00:17:59,756 Speaker 1: Right, Yeah. 321 00:17:59,796 --> 00:18:03,356 Speaker 3: And then another major paper in Nature medicine, James Zoe 322 00:18:03,516 --> 00:18:08,516 Speaker 3: from Stanford did a study that added sleep data to 323 00:18:08,596 --> 00:18:12,756 Speaker 3: everything electronic record and he did showed the predictive ability 324 00:18:12,876 --> 00:18:16,076 Speaker 3: was incredible in sixty five thousand people and replicated in 325 00:18:16,116 --> 00:18:19,836 Speaker 3: another co wort. So our ability to predict heart disease 326 00:18:20,636 --> 00:18:23,276 Speaker 3: is just all these things that we never knew When 327 00:18:23,276 --> 00:18:27,476 Speaker 3: you could combine them, it's so formidable. And that's what 328 00:18:27,796 --> 00:18:32,316 Speaker 3: Then then we have this you know, aggressive ability, Like 329 00:18:32,396 --> 00:18:35,196 Speaker 3: in your case, we know you're at some higher rest, 330 00:18:35,596 --> 00:18:38,476 Speaker 3: so we never let you have a heart attack in 331 00:18:38,516 --> 00:18:38,956 Speaker 3: your life. 332 00:18:38,996 --> 00:18:40,996 Speaker 1: It's not going to happen. We're going to have to 333 00:18:40,996 --> 00:18:42,076 Speaker 1: get over. 334 00:18:41,876 --> 00:18:43,476 Speaker 2: Your lips to God's ears. 335 00:18:43,516 --> 00:18:47,636 Speaker 1: That we're going to do that in the imminent time ahead. 336 00:18:48,516 --> 00:18:52,116 Speaker 2: Well, for now, I'm healthy, so I appreciate your sense 337 00:18:52,156 --> 00:19:09,556 Speaker 2: of urgency. We'll be back in just a minute. Let's 338 00:19:09,556 --> 00:19:12,796 Speaker 2: talk about drugs, and we're generally about interventions, right, medical 339 00:19:12,796 --> 00:19:15,316 Speaker 2: and non medical intervention. So clearly we have this incredible 340 00:19:15,356 --> 00:19:19,116 Speaker 2: new power, very exciting to predict not only who's very 341 00:19:19,196 --> 00:19:20,556 Speaker 2: likely to have our heart attack. 342 00:19:20,316 --> 00:19:25,316 Speaker 3: But when and monitor them with those markers. Yes, yes, 343 00:19:25,636 --> 00:19:28,396 Speaker 3: and even putting them back into the model to see 344 00:19:28,516 --> 00:19:30,476 Speaker 3: have we changed that predictive Yeah? 345 00:19:30,556 --> 00:19:36,556 Speaker 2: Yeah? Goodh So aside from like what are the interventions 346 00:19:36,716 --> 00:19:39,876 Speaker 2: and like what interventions do we have and what interventions 347 00:19:39,916 --> 00:19:42,156 Speaker 2: do we need that we don't have but we might 348 00:19:42,196 --> 00:19:42,556 Speaker 2: have soon. 349 00:19:43,236 --> 00:19:47,716 Speaker 3: Oh well, we start off the statins you mentioned, but 350 00:19:48,076 --> 00:19:49,956 Speaker 3: other LDL cholesterol drugs. 351 00:19:50,196 --> 00:19:53,756 Speaker 2: So basically you want to get LDL like so far down. 352 00:19:54,076 --> 00:19:56,636 Speaker 2: You want to get LDL to like crazy low levels. 353 00:19:56,556 --> 00:20:00,116 Speaker 3: Yeah, which we can in high risk people. 354 00:20:00,516 --> 00:20:03,116 Speaker 2: Is it just linear? Is it the lower LDL goes, 355 00:20:03,196 --> 00:20:04,596 Speaker 2: the lower your risk of heart attack? 356 00:20:06,196 --> 00:20:07,036 Speaker 1: It looks that way. 357 00:20:07,236 --> 00:20:11,876 Speaker 3: Yes, I every study when you put them all together, 358 00:20:12,196 --> 00:20:15,036 Speaker 3: it's a pretty linear result. 359 00:20:15,316 --> 00:20:15,556 Speaker 1: Yes. 360 00:20:16,156 --> 00:20:19,156 Speaker 3: But you know, I don't think that it's necessary to 361 00:20:19,236 --> 00:20:23,356 Speaker 3: be so aggressive with LDL unless you're at very high risk. Right, 362 00:20:23,716 --> 00:20:28,996 Speaker 3: that's where you get this big bang benefit. But we're 363 00:20:29,036 --> 00:20:32,676 Speaker 3: too much LDL centric and not enough you know a 364 00:20:32,756 --> 00:20:35,756 Speaker 3: broader like I mentioned LPA and inflammation. 365 00:20:36,356 --> 00:20:39,436 Speaker 2: Talk about inflammation, right, I mean inflammation is central to 366 00:20:39,516 --> 00:20:42,756 Speaker 2: the story you're telling of the mechanism of disease. Right, 367 00:20:42,796 --> 00:20:46,436 Speaker 2: that's kind of the big shift over these decades that 368 00:20:46,516 --> 00:20:52,436 Speaker 2: you're describing there. What do we do? How do you 369 00:20:52,476 --> 00:20:53,276 Speaker 2: lower inflammation? 370 00:20:53,956 --> 00:20:58,236 Speaker 3: Yeah, so there was a trial called Cantos. It was 371 00:20:58,276 --> 00:21:02,756 Speaker 3: done with this interluken blocker, all right, and it lowered 372 00:21:03,356 --> 00:21:07,236 Speaker 3: heart attacks in people who had a high marker CRP 373 00:21:07,996 --> 00:21:08,836 Speaker 3: for inflammation. 374 00:21:09,436 --> 00:21:11,676 Speaker 1: There's also these coachesine studies. 375 00:21:11,996 --> 00:21:14,996 Speaker 3: A lot of people know about coach zine for treatment 376 00:21:15,036 --> 00:21:18,196 Speaker 3: for gout, but it's an anti inflammatory and there have 377 00:21:18,196 --> 00:21:22,116 Speaker 3: been some very interesting trials over five randomized trials that 378 00:21:22,156 --> 00:21:26,116 Speaker 3: show with benefits people with inflammation taking low dose coach 379 00:21:26,276 --> 00:21:28,396 Speaker 3: zin and a lot of people don't even know that. 380 00:21:28,636 --> 00:21:30,556 Speaker 3: Doctors don't even know it because they just still think 381 00:21:30,556 --> 00:21:34,676 Speaker 3: coachuzinees for gout, not for deventing heart disease. Now, what 382 00:21:34,796 --> 00:21:37,956 Speaker 3: are we going to add to that? Well, what looks promising, 383 00:21:38,636 --> 00:21:43,956 Speaker 3: what's more impressive than statins, which works through LDL largely 384 00:21:44,756 --> 00:21:50,716 Speaker 3: is the golp one drugs, Yes, like to zeppethide, zepbound. 385 00:21:50,596 --> 00:21:53,076 Speaker 2: Weight loss drugs that suddenly everybody is taking. This is 386 00:21:53,116 --> 00:21:55,796 Speaker 2: what you were talking about. Yes, so let's talk about right. 387 00:21:55,836 --> 00:21:59,316 Speaker 2: So there's this line in your piece, which is golp 388 00:21:59,396 --> 00:22:02,956 Speaker 2: one drugs appear to lower the risk of heart attacks 389 00:22:02,956 --> 00:22:05,676 Speaker 2: a lot. That was unsurprising. The part that was surprising 390 00:22:05,716 --> 00:22:08,276 Speaker 2: to me was most of the benefit is not from 391 00:22:08,356 --> 00:22:08,916 Speaker 2: weight loss. 392 00:22:09,236 --> 00:22:09,636 Speaker 1: I know. 393 00:22:09,956 --> 00:22:13,636 Speaker 2: That was like, what that's another shocker, right, shocker? So 394 00:22:13,716 --> 00:22:16,436 Speaker 2: what's going on there? Two thirds of the benefit is 395 00:22:16,476 --> 00:22:17,356 Speaker 2: not from weight loss? 396 00:22:17,436 --> 00:22:17,636 Speaker 1: Yeah. 397 00:22:17,676 --> 00:22:20,516 Speaker 3: So they had the paper when it first came out, 398 00:22:20,756 --> 00:22:23,756 Speaker 3: the big trial, you know, tens of thousands of people 399 00:22:24,556 --> 00:22:28,716 Speaker 3: randomized to ozempic or placebo. There was this big benefit 400 00:22:28,756 --> 00:22:32,036 Speaker 3: of reduction of heart attacks right now. Then they did 401 00:22:32,076 --> 00:22:34,996 Speaker 3: an analysis because ten percent of people don't lose any 402 00:22:34,996 --> 00:22:38,156 Speaker 3: weight on ozempic, people don't realize that, and then some 403 00:22:38,236 --> 00:22:41,516 Speaker 3: people they lose a little weight but not much. Anyway, 404 00:22:41,676 --> 00:22:44,756 Speaker 3: they found that two thirds of the benefit had nothing 405 00:22:44,756 --> 00:22:45,076 Speaker 3: to do. 406 00:22:45,076 --> 00:22:45,836 Speaker 1: With weight loss. 407 00:22:46,596 --> 00:22:49,516 Speaker 2: That's unbelievable. Yeah, what's going on? 408 00:22:50,036 --> 00:22:50,276 Speaker 1: Right? 409 00:22:50,396 --> 00:22:54,676 Speaker 3: So that's that's basically the biggest story about this family 410 00:22:54,756 --> 00:22:58,916 Speaker 3: of drugs is that their main mechanism of action whell 411 00:22:58,956 --> 00:23:02,996 Speaker 3: the cosmetic and other health benefits of carrying lesser weight, 412 00:23:03,436 --> 00:23:08,916 Speaker 3: that's great, But the big impact is to lower total 413 00:23:09,276 --> 00:23:11,716 Speaker 3: body and brain inflammation. 414 00:23:12,356 --> 00:23:13,276 Speaker 2: Huh. 415 00:23:13,316 --> 00:23:15,596 Speaker 1: And they are much more potent stats. 416 00:23:15,196 --> 00:23:19,076 Speaker 2: And it's clear that the reduction in inflammation is independent 417 00:23:19,236 --> 00:23:23,556 Speaker 2: of weights less. Presumably weightless alone would drive down inflammation, right. 418 00:23:23,476 --> 00:23:26,236 Speaker 1: Yeah, but forget about that. 419 00:23:26,396 --> 00:23:28,836 Speaker 3: If you just have the same amount of belly fat 420 00:23:28,916 --> 00:23:32,476 Speaker 3: and you're on these drugs and you're seeing big reductions 421 00:23:32,716 --> 00:23:35,676 Speaker 3: of inflammation, huh, does anybody know why? 422 00:23:36,476 --> 00:23:36,836 Speaker 1: Well? 423 00:23:37,436 --> 00:23:41,396 Speaker 3: Yeah, in part, the receptors of the immune cells are 424 00:23:41,556 --> 00:23:46,436 Speaker 3: rich with GLP receptors, so when you take the drug, 425 00:23:46,516 --> 00:23:50,396 Speaker 3: it's affecting your immune system, which drives inflammation. So all 426 00:23:50,436 --> 00:23:54,676 Speaker 3: these drugs work directly through the immune system and also 427 00:23:54,756 --> 00:23:55,236 Speaker 3: the brain. 428 00:23:55,476 --> 00:23:58,916 Speaker 1: The brain our command center for inflammation. Right. 429 00:23:59,436 --> 00:24:04,676 Speaker 3: The brain has GOLP receptors as well. So anyway, we 430 00:24:04,916 --> 00:24:08,716 Speaker 3: knocked down the immune system, knocked down inflammation. That's something 431 00:24:08,756 --> 00:24:12,236 Speaker 3: we've never been able to do with a drug safely 432 00:24:12,316 --> 00:24:17,556 Speaker 3: before statins included. So this is the beginning of a 433 00:24:17,636 --> 00:24:20,756 Speaker 3: new era of anti inflammatory drugs. 434 00:24:21,036 --> 00:24:24,436 Speaker 2: Okay, one question, and maybe it's a dumb question, but 435 00:24:24,436 --> 00:24:28,756 Speaker 2: when you're talking about this sort of immune system being 436 00:24:28,796 --> 00:24:32,956 Speaker 2: what's modulating the inflammation response, Like, is there any concern 437 00:24:33,236 --> 00:24:36,036 Speaker 2: about that? Is there are there instances where it's bad 438 00:24:36,076 --> 00:24:39,716 Speaker 2: that you're taking GLP one drugs and it's suppressing your 439 00:24:39,716 --> 00:24:40,996 Speaker 2: immune response at some level. 440 00:24:41,556 --> 00:24:42,636 Speaker 1: Well, that's the difference. 441 00:24:42,676 --> 00:24:45,596 Speaker 3: In that earlier trial I mentioned Cantos with the endo 442 00:24:45,716 --> 00:24:49,876 Speaker 3: lukeing of blocking. The problem was they reduced heart attacks, 443 00:24:49,876 --> 00:24:54,236 Speaker 3: but it increased infections. Right here, we don't see any 444 00:24:54,596 --> 00:24:58,556 Speaker 3: increased infection. And that's what's so exciting is that we're 445 00:24:58,916 --> 00:25:01,836 Speaker 3: we're so more selective. We're not blocking the immune system 446 00:25:01,996 --> 00:25:05,436 Speaker 3: that defends us. We're blocking the immune system that hurts us. 447 00:25:05,956 --> 00:25:08,436 Speaker 3: But they're not the end of anti inflammatory drugs. 448 00:25:09,196 --> 00:25:09,916 Speaker 2: Another one more. 449 00:25:10,436 --> 00:25:14,676 Speaker 3: Oh gosh, there's the n l RP inflammazome. There's the 450 00:25:14,796 --> 00:25:19,036 Speaker 3: gas sting pathway. I mean, these are your listeners, you 451 00:25:19,076 --> 00:25:22,276 Speaker 3: know they All I can say is there's so much 452 00:25:22,316 --> 00:25:26,876 Speaker 3: more going on right now to reduce inflammation with more 453 00:25:26,916 --> 00:25:32,396 Speaker 3: gut hormones, with new drugs, new pathways, with gut microbiome manipulations, 454 00:25:32,796 --> 00:25:35,956 Speaker 3: changing the gut microbiome or it's metabolis. We are on 455 00:25:36,156 --> 00:25:41,796 Speaker 3: the inflammation express Okay, anti inflammation. You're going to see 456 00:25:41,796 --> 00:25:46,036 Speaker 3: a blossoming of strategies to block untoward inflammation like never 457 00:25:46,116 --> 00:25:47,796 Speaker 3: before in medicine. 458 00:25:48,476 --> 00:25:50,796 Speaker 2: So let's talk about the like I don't know the 459 00:25:50,836 --> 00:25:53,596 Speaker 2: medium to long term future. What does the field look like. 460 00:25:53,676 --> 00:25:54,916 Speaker 2: What does your job look like? 461 00:25:55,556 --> 00:25:58,836 Speaker 3: Well, it ought to be revamped today or you know 462 00:25:58,996 --> 00:26:04,476 Speaker 3: in the next year. As you know, medical community doesn't 463 00:26:04,556 --> 00:26:09,076 Speaker 3: change very quickly, but eventually it will be incorporating all 464 00:26:09,276 --> 00:26:12,316 Speaker 3: the things we've discussed today. You are head of the 465 00:26:12,356 --> 00:26:14,476 Speaker 3: curve by getting apologenic risk, and I give you a 466 00:26:14,516 --> 00:26:15,476 Speaker 3: lot of credit for that. 467 00:26:15,916 --> 00:26:18,036 Speaker 2: It's just a lucky break. I just happened to walk 468 00:26:18,076 --> 00:26:19,956 Speaker 2: into the right doctor's office at the right time. 469 00:26:20,356 --> 00:26:22,956 Speaker 3: It should be in everyone with a family history or 470 00:26:22,956 --> 00:26:25,356 Speaker 3: other risk factors. That should be you know, it's cheap, 471 00:26:25,396 --> 00:26:27,236 Speaker 3: it could be done. We can do it for like 472 00:26:27,316 --> 00:26:28,116 Speaker 3: twenty dollars. 473 00:26:28,156 --> 00:26:28,356 Speaker 1: You know. 474 00:26:29,316 --> 00:26:34,276 Speaker 3: You can actually now if you have your ancestry DNA 475 00:26:34,516 --> 00:26:38,356 Speaker 3: or twenty three and meters, you can use one of 476 00:26:38,356 --> 00:26:43,956 Speaker 3: the AI platforms like a cloud or open AI and 477 00:26:44,036 --> 00:26:46,076 Speaker 3: can analyze and give you your apology. 478 00:26:46,716 --> 00:26:47,236 Speaker 2: Interesting. 479 00:26:48,836 --> 00:26:52,676 Speaker 3: Yeah, so anyway we should have that. It will become accepted. 480 00:26:53,196 --> 00:26:56,676 Speaker 3: So will be the artery score, the heart score, the 481 00:26:56,676 --> 00:27:01,836 Speaker 3: brain organ clocks, so will be these better inflammation markers 482 00:27:02,316 --> 00:27:04,916 Speaker 3: and the models we talked about, they will become routine 483 00:27:05,636 --> 00:27:10,236 Speaker 3: and each you could even have a virtual coach and 484 00:27:10,276 --> 00:27:14,676 Speaker 3: you say, your virtual coach will say your new assessment 485 00:27:14,756 --> 00:27:17,796 Speaker 3: with your sleep and your exercise, and you know, whatever 486 00:27:17,916 --> 00:27:21,276 Speaker 3: test has changed, the chance of you having a heart 487 00:27:21,276 --> 00:27:24,196 Speaker 3: attack at age sixty seven is now changed to age 488 00:27:24,236 --> 00:27:29,596 Speaker 3: seventy nine. Yeah, wouldn't that be nice. 489 00:27:30,836 --> 00:27:32,916 Speaker 2: We'll be back in a minute with the lightning rip. 490 00:27:43,356 --> 00:27:46,916 Speaker 2: I read an interview with you recently and you said 491 00:27:47,076 --> 00:27:50,756 Speaker 2: chatbots are more empathetic than doctors. Yeah, which I found 492 00:27:50,796 --> 00:27:53,516 Speaker 2: extremely interesting coming from a doctor. Tell me more. 493 00:27:54,556 --> 00:27:59,516 Speaker 3: AI has no clue what empathy is, no clue, but 494 00:27:59,636 --> 00:28:05,996 Speaker 3: it channels all of its language training to basically be 495 00:28:06,796 --> 00:28:14,996 Speaker 3: a great way to support empathy. It's a far more impressive, 496 00:28:15,796 --> 00:28:21,076 Speaker 3: at large empathetic channel than the average doctor. The average 497 00:28:21,116 --> 00:28:25,036 Speaker 3: doctor doesn't have time. And it turns out there's been 498 00:28:25,196 --> 00:28:30,036 Speaker 3: thirteen studies and twelve of them have shown that by 499 00:28:30,516 --> 00:28:34,636 Speaker 3: doctors making the judgment that the AI was more empathetic 500 00:28:34,716 --> 00:28:35,516 Speaker 3: than the doctors. 501 00:28:36,196 --> 00:28:36,796 Speaker 1: Say WHOA. 502 00:28:37,236 --> 00:28:41,836 Speaker 3: Now, the other thing that's surprising here is that we're 503 00:28:41,876 --> 00:28:45,756 Speaker 3: going to see AI coaching doctors to be more empathetic. 504 00:28:46,556 --> 00:28:48,916 Speaker 3: They're going to go over their notes. This is the 505 00:28:48,956 --> 00:28:52,876 Speaker 3: conversations and they say, well, doctor Smith, why did you 506 00:28:52,916 --> 00:28:56,316 Speaker 3: interrupt missus Jones after nine seconds? She never even had 507 00:28:56,356 --> 00:28:59,676 Speaker 3: a chance to express what was her concern and that 508 00:28:59,756 --> 00:29:00,476 Speaker 3: stuff like that. 509 00:29:00,956 --> 00:29:03,396 Speaker 2: How do you think doctors will take that? I could 510 00:29:03,436 --> 00:29:05,676 Speaker 2: have matter a doctor not wanting to hear from an AI. 511 00:29:05,876 --> 00:29:08,956 Speaker 3: Yeah, well, you know, the data is accruing and it's 512 00:29:09,196 --> 00:29:10,676 Speaker 3: not good for doctors. 513 00:29:10,876 --> 00:29:12,516 Speaker 1: It's not their doctor's fault. 514 00:29:12,596 --> 00:29:15,356 Speaker 3: It's because there's squeezed so much to see so many 515 00:29:15,396 --> 00:29:17,436 Speaker 3: patients in such short periods of time. 516 00:29:17,836 --> 00:29:23,156 Speaker 2: Yeah, how do you use AI in your in your 517 00:29:23,276 --> 00:29:25,676 Speaker 2: non work life? How do you use it when you're 518 00:29:25,716 --> 00:29:26,196 Speaker 2: not working? 519 00:29:26,356 --> 00:29:30,436 Speaker 1: If you use it, I use it a lot. I 520 00:29:30,556 --> 00:29:30,916 Speaker 1: use it. 521 00:29:31,076 --> 00:29:35,276 Speaker 3: I use deep research a lot when I'm curious about 522 00:29:35,316 --> 00:29:38,476 Speaker 3: something and I just want to produce a report. 523 00:29:38,796 --> 00:29:41,276 Speaker 2: What was a recent thing you used deep research for 524 00:29:41,396 --> 00:29:42,196 Speaker 2: outside of work? 525 00:29:42,916 --> 00:29:46,036 Speaker 3: Well, like today, there was a publication about the real 526 00:29:46,116 --> 00:29:48,156 Speaker 3: da Vinci code of Leonardo da Vinci. 527 00:29:48,236 --> 00:29:50,596 Speaker 2: Okay, I didn't know that was a real DaVinci code. 528 00:29:50,756 --> 00:29:52,916 Speaker 3: Well, it's in the works, So I did. 529 00:29:53,036 --> 00:29:53,516 Speaker 1: I did a. 530 00:29:55,756 --> 00:29:59,956 Speaker 3: Query about what do we know about Leonardo da Vinci's 531 00:29:59,956 --> 00:30:03,076 Speaker 3: geno and the answer was, you know very little actually, 532 00:30:03,436 --> 00:30:07,636 Speaker 3: But I mean that this curiosity, I wasn't like writing 533 00:30:07,676 --> 00:30:10,876 Speaker 3: any paper about it. It's just, you know, it's just 534 00:30:10,916 --> 00:30:12,796 Speaker 3: the way my crazy mind works. 535 00:30:12,876 --> 00:30:14,156 Speaker 1: Yeah. 536 00:30:14,196 --> 00:30:16,076 Speaker 2: What are you worried about with respect to AI? 537 00:30:16,516 --> 00:30:20,636 Speaker 1: Oh? Lots of things. Okay, I'm worried about. 538 00:30:20,756 --> 00:30:25,876 Speaker 3: There's lack of transparency, there's lack of compelling data in 539 00:30:25,916 --> 00:30:31,156 Speaker 3: many respects that have accrued. There's worries about implementing it 540 00:30:31,236 --> 00:30:35,356 Speaker 3: where you could have adversarial attacks. There's privacy, security issues, 541 00:30:35,396 --> 00:30:37,036 Speaker 3: there's inequity issues. 542 00:30:36,716 --> 00:30:37,156 Speaker 1: Worth thing. 543 00:30:37,516 --> 00:30:38,996 Speaker 3: I mean, there's a long list here. 544 00:30:39,076 --> 00:30:39,556 Speaker 1: Okay. 545 00:30:40,556 --> 00:30:43,316 Speaker 3: Of course there's errors and hallucinations, but we don't acknowledge 546 00:30:43,396 --> 00:30:46,436 Speaker 3: human errors enough. We think that AI has to be 547 00:30:46,516 --> 00:30:50,156 Speaker 3: perfect and it's never going to be perfect. But no, 548 00:30:50,276 --> 00:30:53,996 Speaker 3: so there's issues with AI that we have to confront 549 00:30:54,196 --> 00:30:57,956 Speaker 3: head on to make it better and be more pure positive. 550 00:30:58,436 --> 00:31:02,636 Speaker 2: Okay, last thing, forget about AI. Let's just talk about 551 00:31:02,676 --> 00:31:06,116 Speaker 2: the heart of cardiovascular system. What's one thing you wish 552 00:31:06,196 --> 00:31:09,236 Speaker 2: we understood that we don't understand about how it works. 553 00:31:10,516 --> 00:31:13,436 Speaker 1: That's a good one. 554 00:31:13,876 --> 00:31:19,796 Speaker 3: Well, I mean, I the inflammation story that we discussed, 555 00:31:20,676 --> 00:31:26,116 Speaker 3: I've given you the reductionist view. Yeah, I've oversimplified it. 556 00:31:26,636 --> 00:31:29,876 Speaker 3: I'd like to know all the intricacies of that. I'd 557 00:31:29,916 --> 00:31:33,796 Speaker 3: like to know why people that do everything right and 558 00:31:33,916 --> 00:31:37,236 Speaker 3: have low LDL and low everything that we can measure 559 00:31:37,276 --> 00:31:40,356 Speaker 3: still can get inflamed arteries and a heart attack. 560 00:31:40,636 --> 00:31:42,996 Speaker 1: We don't know why. I'd like to know exactly what's 561 00:31:43,036 --> 00:31:43,836 Speaker 1: going on there. 562 00:31:44,796 --> 00:31:47,716 Speaker 3: And so there's things that we don't understand that I 563 00:31:47,756 --> 00:31:53,356 Speaker 3: hope eventually will unravel. Anything else you want to say, Well, 564 00:31:53,716 --> 00:31:56,476 Speaker 3: I've really enjoyed this discussion. You're the first one that's 565 00:31:56,516 --> 00:32:01,236 Speaker 3: really you know, come to discuss this cardiovascular change of 566 00:32:01,796 --> 00:32:03,516 Speaker 3: the big miss or you know, the. 567 00:32:03,756 --> 00:32:07,676 Speaker 2: Completely talking about it for twenty years. Nobody ever interviewed 568 00:32:07,716 --> 00:32:08,276 Speaker 2: you about it. 569 00:32:09,356 --> 00:32:11,436 Speaker 3: In recent time, not with all the things that we 570 00:32:11,516 --> 00:32:14,076 Speaker 3: discussed today. So I thank you for that. Thanks for 571 00:32:14,316 --> 00:32:16,636 Speaker 3: queueing into what I think is really important. 572 00:32:23,836 --> 00:32:26,516 Speaker 2: Eric Topel is a cardiologist and the founder of the 573 00:32:26,556 --> 00:32:31,756 Speaker 2: Scripts Research Translational Institute. Please email us at problem at 574 00:32:31,796 --> 00:32:35,036 Speaker 2: Pushkin dot fm. We are always looking for new guests 575 00:32:35,156 --> 00:32:39,036 Speaker 2: for the show. Today's show was produced by Trinamanino and 576 00:32:39,076 --> 00:32:42,956 Speaker 2: Gabriel Hunter Chang. It was edited by Alexander Garreton. And 577 00:32:43,116 --> 00:32:46,276 Speaker 2: engineered by Sarah Brigett. I'm Jacob Goldstein and we'll be 578 00:32:46,316 --> 00:33:03,236 Speaker 2: back next week with another episode of What's Your Problem.