Latest / TeqTalk with Jas / Diabetes Reversed in 71% of Patients. Cardiac Surgeons Haven't Missed Once. Here's Why
Transcript
- Teqfocus: Imagine this. You walk into a hospital with chest pain. The doctor does tests. Tomorrow you have your results and they tell you you need a surgery. You ask, what if we don't want? They say, Let me show you. And they pull up your perfect virtual copy of your heart, not a generic heart, your heart. Then they run a simulation. They perform the surgery in the digital version first. They show you exactly what's gonna happen. They show you the risks, recovery time, what your life looks like in the next six months if you do the surgery, and what it looks like if you do not. This isn't a science fiction. Hospitals are already doing this now. And this is called as digital twin. But here is the real question: if healthcare can create a digital version of your heart, your brain, your disease, how much should we trust it? Because if digital twin is right, It can save time, reduce risk, and personalize the care. But if it's strong, incomplete, or poorly governed, it creates a new kind of healthcare risk. So in this video, we will understand where digital twins came from. We will deep dive into three clinical use cases where digital twins are no longer just an idea. They are already proving their values in the real world. And the future of digital toys. Let's get started. Most people think that digital twins are new in healthcare. A 2020s invention, something born after AI, cloud, gen AI. But that's not true. To understand digital twins, we have to go back to April 1970. Apollo 13. Three astronauts are almost 200,000 miles away from Earth. Suddenly, an oxygen tank explodes. The spacecraft is damaged. Oxygen is leaking. Power is dropping and the astronauts may have only hours to survive. And now think about it. NASA engineers cannot go to space and fix their spacecraft. They cannot afford to guess. They cannot say, Let's try this and see what happens. So what do they do? They recreate the problem on Earth using spacecraft simulators, engineering models, telemetry, and real-time mission data. Mission controls, tests, repair options before actually sending instructions to the astronauts. If we shut the system down, what happens? If we move power here, what happens? If we have this sequence instead, what fails? They simulated everything first. Only after testing the solution on Earth do they tell the astronauts what to do in their space. That decision-making approach helped bring Apollo 13 home safely. And this is the core idea behind any digital twin. When the real system is too risky, too expensive, or too far away to experiment on directly, you build a virtual version, test possibilities, and act with more confidence. So that was space. Then came jet engines. In the 1990s and 2000s, companies like GE and Siemens faced a similar problem. A jet engine was flying at 40,000 feet. Is not something you casually experiment on. One failure can really cost millions. One wrong adoption can risk lives. So they started using digital models to monitor engines, predict maintenance needs, and test failure scenarios before the real machine broke down. Then factories, turbines, power plants, and aircraft followed. And now healthcare is asking the same question: if we can simulate a spacecraft before we actually give instructions to astronauts. If we can simulate a jet engine before it fails midair, can we simulate a heart before surgery? Can we simulate a tumor before radiation? Can we simulate a patient's response before changing treatment? That is why digital twins matter in healthcare. Because in medicine, trial and error is not just expensive, it is personal. It is somebody's body, somebody's child, somebody's future. And that is where the story becomes very, very If you're still watching, you are exactly who TikTok is built for. Hit subscribe because what I am going to show you next, the three real clinical use cases, is more exciting than anything I have described so far. Let's start with one of the most time-critical situations in healthcare: sepsis. A patient comes into the ICU with a severe infection. Within hours, the body can start shutting down, blood pressure drops, oxygen level change. Kidneys may start failing. The immune system overall reacts. And here is the scary part. In sepsis, doctors don't always have the luxury of time. Sepsis is one of the leading causes of death in hospitals. Some estimates show that it's about one in three patients who die in a hospital had sepsis during that hospitalization. Now imagine the pressure on the ICO team. They have to decide quickly: should we give more fluid? Should we change the antibiotic? Should we wait or will waiting make the patient worse? For decades, this has involved clinical expertise, guidelines, and constant monitoring. And yes, some of the level of trial and error. But in 2020, Mayo Clinic researchers tested a different idea. So, what if instead of only watching the patient and reacting, we could create a digital twin of the sepsis patient? Data driven model. Built from patients' real ICU information, vital signs, blood tests, oxygen levels, medications, fluids, medical history, and treatment responses, the model asks a very practical question. If this patient receives this treatment now, what might happen to him in the next few hours? That's powerful because in sepsis, you cannot safely test five different treatment paths on the real patient. But if you can test possibilities inside the digital model, it's great. In the original 2020 Mayo linked study, researchers tested the model using 29 sepsis patients and 145 observations during the first 24 hours of ICO admission. That was not perfect AI. This is important. The result was promising early evidence. The model showed fair to good agreement depending on the treatment response being measured. And MAO did not stop there. Later work looked at larger sepsis datasets. Close to 20,000 sepsis patients admitted to MAO ICUs over the eight-year period, identifying different patients' trajectories such as fast recovery and slow recovery. Then in 2024, Mayo researchers presented newer work around a digital twin model trained on close to 30,000 critical care sepsis patients. At 24 hours, the model was reported to predict 14-day mortality with 82% accuracy on Mayo patients and 78% accuracy on the external validation groups. Now let's pause here. This is a real promise of digital twins in critical care. The goal is not to replace the ICU doctor, the goal is to give the doctor another layer of foresight. But here is the honest part. This is still not routine bedside care in every ICU. It is still moving through validations, clinical workflow testing, physician, trust, and regulatory questions. And that is exactly how we should understand digital twins in healthcare today. Not magic, not hype, but a serious move from reactive medicine towards simulation-supported decisions. In sepsis, where every hour can matter, that shift. Be life-changing. Now let's move from emergency care to one of the most delicate areas in medicine: heart surgery on newborns and children with complex congenital heart defects. Imagine a baby born with a heart defect so complicated that there is no simple textbook answer. Holes between the chambers, blood flowing in the wrong direction, vessels positioned differently. Traditionally, surgeons study CT scans. MRI images, clinical reports. These tools are powerful, but the surgeons still have to mentally reconstruct a living, breathing, three-dimensional heart from the images on the screen. And in a complex case, the uncertainty is frightening. So this is where Boston Children's Hospital became one of the leading pediatric centers showing how patients. Specific 3D modeling and computational simulation can move from research into real surgical planning for complex heart disease. Instead of only looking at images, the team can take the child's scan and build a digital model of that specific heart. The model can help the surgical teams understand the blood flow, evaluate pressure changes, and compare possible repair strategies before entering the operating room. So in one of the reported cases from May 2019, Boston Children's worked on a child with holes involving all four chambers of the heart. A highly complex case with no easy answer. The team created a digital model, tested three different repair strategies, studied the expected flow and pressure outcomes, chose the most promising approach, and rehearsed the plan before surgery. Then they operated. The repair worked as predicted. And this is a real shift. Innovation does not have to happen during the operation. It can happen before the operation, inside the model, before the first cut is made. By 2023, this approach had grown into a broader program for complex hard cases. By 24-25, Boston children started using advanced modeling, simulation, and surgical planning across many complex cardiac cases, including training and planning workflows. And the outcomes show why this matters. For the arterial switch procedure used for babies born with reverse arteries, Boston children reported 60 cases with 100% survival. The success is because of the digital twin. Let's move on to our third case, which isn't just a crisis. It isn't a surgery. It's the world's most common chronic disease. Over 500 million people globally have type 2 diabetes. 1 in 10 adults in the US, and almost every single one of them gets the same advice on diagnosis day. Cut sugar, exercise, take this medication. You'll manage this for life. But here is the problem. Medicine has known for years and never solved. Two people, same age, same diagnosis, give them the exact same treatment and they respond completely different. One reverses it with diet, another one needs three medications. The third needs none, just sleep and stress management. Why? Because every human metabolism is different. And for decades, medicine has been treating them all the same. In 2018, a company called Twin Health set out to change that. They built what they call a whole-body digital twin, a real-time computational model of your individual metabolism. And hear it what it does. You wear a continuous glucose monitor, small sensor of your arm that track your blood sugar levels every minute around the clock without any needle pricks, and you pair it with fitness watch, scale, or a blood pressure cuff. Together The platform captures up to 3000 data points every single day across 174 health makers. That's 174 different measurements of how your body is functioning. From glucose to sleep quality to heart patterns. It builds a model for you, not the average diabetic patient, you specifically. The digital twin learns. When you eat rice, your blood pressure does this. When you sleep 6 hours instead of 8, Your body's ability to process sugar, what your doctors call insulin sensitivity, changes like this. When you walk for 20 minutes after a meal, your post-meal glucose spike drops by this much. Then it predicts, then it personalizes. Every recommendation calibrated to your biology updated in real time. So what happened when the clinic put this into a rigorous test? 150 adults with type 2 diabetes, randomized controlled trial. Meaning half got the digital twin, half got standard care randomly assigned so that the results can be explained by luck. Published in the New England Journal of Medicine Catalyst 2025 odds. And the results? 71% of patients using the digital twin achieved an HBA1C below 6.5. That's the clinical definition of reversal. 75% Cross that line while eliminating most of their medications. In the control group, the standard care, the way diabetes is treated everywhere right now, only 2.4% achieve the same result. Here's what else happened in the digital twin group. GLP1 medications, drugs like OZEMPIC, that have been taken by the world by storm and cost thousands of dollars a year, their use. dropped from 41 to 6%. Insulin, the daily injection that millions of diabetics depend on, dropped from 24% to 13%. Weight loss was twice as high compared to the standard care. Over 500 million people are living with a disease that medicine says you manage for life. The studies say for the majority of patients that's not inevitable. It's not a medication problem. It's a personalization problem. And a digital twin just solved it. And here is what gives me genuine hope. Digital twins are being actively built for cancer, brain disease, for conditions that have stumped medicine for decades. So digital twins aren't just a medical technology, it's a mirror. A mirror that shows you your biology before disease does, your risks before symptoms do, your future before it arrives. For the first time in human history, medicine isn't just reacting to what goes wrong, it's seeing it coming. So that shift is happening right now in real hospitals on real patients. But here is a question, nobody is asking aloud enough. When your digital twin predicts something you do not want to know, do you look forward? When it tells your insurance companies you are at high risk, who will protect you? When it is wrong, who is responsible? The technology is very much getting there. The harder question is are we?