Year: 2025
Type: article
Abstract: The best way to predict the future is to create it! Abraham Lincoln This August, the European Society of Cardiology (ESC) celebrates its 75th birthday. On 2 September 1950, the ESC was founded by its first President Karl G.W. Nylin, Charles Laubry as Honorary President, and by three Vice-Presidents, i.e. D.E. Belford, Eduardo de Arayjo Coehlo, and Jean Lenègre. The first ESC Congress was organized in London in September 1952 by Sir John Parkinson—75 years is quite an achievement for a learning society such as the ESC and demonstrates that it is here to stay. Within 75 years, cardiology was transformed from an observing to a therapeutic specialty. In 1950, there was no treatment for myocardial infarction, no defibrillator for ventricular fibrillation, no secondary prevention, nor any remedy for heart failure except digoxin. Even 5 years later when Dwight D. Eisenhower experienced his first heart attack, his doctor had only amyl nitrate and morphine at hand.1 An electrocardiogram (ECG) recorded an anterolateral ST-segment elevation myocardial infarction, but nothing could be done about it. There was no aspirin, its antiplatelet effects were described in 1971 by Sir John Vane and its life-saving effects together with streptokinase shown in the ISIS-2 trial only in 1988. There was no balloon angioplasty or stenting. Andreas Grüntzig performed his coronary balloon angioplasty in 19772 and Ulrich Sigwart3 and Jacques Puel implanted the first stent in 1988. Today, a myocardial infarction treated timely with aspirin, a balloon, and a stent is an acute event with a mortality of around 3% in stable patients, but remains a serious condition in those with cardiogenic shock (Figure 1)4—what an achievement! Change in in-hospital mortality of acute myocardial infarction over time (Lüscher and Obeid,4 by permission) Against all odds, Eisenhower survived, but there was no second prevention, e.g. no anti-hypertensives, lipid-lowering agents, or antidiabetics. Remodelling of the ventricle occurred uncontrolled as ACE inhibitors were not discovered yet. For beta-blockers, patients with heart failure had to wait for their discovery by James Black in the 1960s. It took decades until we recognized atrial fibrillation as a cause of stroke and anticoagulation to prevent it. In 1998, Michel Haïssaguerre showed that it can be cured by radiofrequency ablation.5 Today, with pulsed wave ablation, this has become an easy intervention with most clinical benefit in heart failure.6 Finally, valvular heart disease. The first mechanical aortic valve replacement by Dwight Harken in 1960 was a breakthrough for an otherwise incurable condition. It remained a surgical domain until Alain Cribier, inspired by Henning Andersen, performed the first transarterial valve implantation (TAVI) in 2002.7 Today, more TAVI than surgical aortic valve replacements are performed even in low-risk patients. The Alfieri stitch inspired MitraClip as a new percutaneous procedure for mitral regirgitation,8 now expanded to the tricuspid valve.9 Thus, in line with the ESC mission, members of our society transformed cardiology from an observing to a therapeutic specialty, thereby improving and saving lives. What is coming next? Cardiology will transform into a digital specialty providing precision medicine, and the ESC is on it!10 So far, once a diagnosis has been established, we have treated all patients more or less alike—this has made medical practice feasible and effective, but not precise. Artificial intelligence and machine learning (AI/ML) analyses large datasets and defines complex interplays of its components in an unforeseen fast and precise manner;11 it provides the basis for precision medicine by characterizing phenotypes and prediction of individual clinical outcomes. Artificial intelligence and machine learning integrates data such as demographics, clinical findings, ECG, imaging, biomarker signatures, routine laboratory parameters, and genetics for deep phenotyping and a digital twin of an individual patient.12 This sets the stage for precise risk prediction and individualized treatment strategies. For instance, after acute coronary syndromes, personalized prevention differs in phenotypes in whom the cardiovascular (CV) risk is lipid-driven, inflammatory, renal, or diabetic in nature. Similarly, treatment strategies for heart failure differ on whether the patient is of the obese, atria-typel, hypertensive, or any other phenotype. Algorithms further improve outcome prediction in males and females,13 as well as in ethnicities. What is the phenotype of the cardiologist in the digital age? When René Laennec created the stethoscope in 1821, he revolutionized physical examination. Today, it remains the symbol of the physician, but is no more than a screening tool. The advent of echocardiography quantifies speed and direction of blood flow within the heart, and AI/ML got as good at it as any cardiologist. When Willem Einthoven developed the ECG in 1895, it became another important tool. Mastering the ECG was a must for any cardiologist. Today, AI/ML algorithms surpass any cardiologist in reading ECGs; based on a few beats in sinus rhythm, they predict age and sex, left ventricular dysfunction, and episodes of prior atrial fibrillation, as well as suspect aortic stenosis and amyloid heart disease. Will the cardiologist of the future still spend hours to learn how to read an echo or an ECG? Most likely not: it will be readily analysed for him. Any visit starts by obtaining history. How are you today? Is this the last resort for the physician? The Articulate Medical Intelligence Explorer algorithm talks to the patient, while he is waiting to be seen.14 When he meets his physician, he has the most likely diagnosis at hand. Large language models (LLM) will further provide a summary of referral letters and any other document relevant to a working diagnosis; indeed, avatars are highly precise in providing a diagnosis. We will shortly have LLM within electronic health records recommending a diagnosis such as heart fialure with preserved ejection fraction (HFpEF) based on available clinical findings, an NT-proBNP and an echocardiogram15—a friendly co-pilot to make us faster, better, and more precise (Figure 2). Machine learning and precision medicine (modiefied from Asselbergs and Lüscher,10 by permission) In pictures analysis, AI/ML is particularly fast, reproducible, and precise. Algorithms match or surpass the eyes of imagers in the analysis of echocardiograms, coronary CTs, or cardiac magnetic resonance—and they learn as we speak. Will imagers be replaced by AI/ML? A certain number may still be required to approve and sign reports, but cardiac imaging will become an automated service. CV prevention will also be run by AI/ML, providing precise and personalized risk predictions and, with the help of the digital twins, recommendations for primary and secondary prevention and adaptations of medication. Introducing and positioning catheters in the aorta, ventricles, atria and coronary arteries will remain the competence of trained interventionalists and electrophysiologists for long. However, algorithms—by analysing ECGs and providing image integration into intuitive cardiac holograms—will make their job easier, more efficient, fast, and precise. Artificial intelligence and machine learning is as transformative as the industrial revolution in the 19th century. With the steam engine of James Watt, physical force became less important; machines did the hard work (Figure 3). Engines were used in factories, locomotives, and cars, thereby revolutionizing production, mobility, and human life. The discovery of electricity by Michael Faraday led to another source of power and light. Did that revolution lead to unemployment? Certainly not: novel professions evolved such as factory workers, engineers, electricians, drivers, and others. The question is: Does the AI similarly provide novel professions overall and in medicine in particular? We do not know yet, but this time the change will likely be massive. Change in human skills over time (Lüscher and Wenzl,16 by permission) So, what will be the cardiologist’s phenotype in the future? Classical skills such as auscultation, clinical examination, and reading ECGs and images will become less important in training and clinical practice.16 In contrast, AI/ML literacy, assessing the quality and proper use of algorithms will be essential: digital cardiology will become a new subspecialty. Manual skills will for a long time be the privilege of trained interventional cardiologists, electrophysiologists, and surgeons because robot technology—although it rapidly progresses—will still not mimic the fine-tuning of human fingers. In summary, AI/ML will not replace clinicians who use AI. However, AI will replace clinicians who do not use it. As AI/ML prepares data effectively and fast, it provides physicians with more time for the patient, for listening, showing empathy, and to explain its condition. And, indeed, there is need here as avatars give better answers and are more empathic than physicians17—so let’s revitalize that very virtue of our profession! What are the future major issues for CV scientists and cardiologists? There are three major issues: environmental hazards, the cardiometabolic pandemic, and the ageing society. The ESC is addressing environmental hazards at the political level through the EU CV Health Plan and its task force as it fosters cardiometabolic medicine as the task of the new generalist.18 Ageing is the final common pathway of most CV conditions and a new therapeutic target to be addressed. Artificial intelligence and machine learning will facilitate solving these issues by its analytic power and in silico experiments and trials. The future is bright—let’s shape it together! The author declares no disclosure of interest for this contribution.
Source: European Heart Journal
Author: Thomas F Lüscher
Institution: European Shippers' Council
Language: English
Subfield: History and Philosophy of Science
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W4413772349