Personalized Cardiology

Medicine has been “personalized” (that is, patient-centered) since its inception. More than a century ago, Sir William Osler (1848-1919) noted: “It is more important to know what man has a disease than to know what disease that man has,” thus being among the first to draw attention to the importance of patient-specific features in the manifestations of a given disease [1].

The stages of the diagnostic process (history, physical examination, investigation plan, stage of disease, treatment plan, and monitoring the effectiveness of the treatment) are completed individually, not in a group. It is true that this way of developing the diagnosis has variable degrees of precision (doctors who had higher precision were sought more often by patients and achieved fame). Moreover, reaching an accurate diagnosis requires a lot of time, and here we are not talking only about the many years needed for the doctor’s training. Medicine and cardiology are practiced today by gathering data related to the particularities of the patient, data that is analyzed in the smallest details (as it should be!) in order to arrive at a precise diagnosis and to deliver treatment appropriate for the patient’s condition. True, meaningful personalized medicine would involve delivering treatment that is specifically suitable to that patient only. Currently, important support is provided by technology, data from clinical trials, medical practice guidelines, and, last but not least, the skill and experience of the doctor.

Clinical trials provide evidence derived from large study populations (tens of thousands of patients), assessed by expert panels. Today, one rarely hears – in multi-disciplinary case meetings – “my opinion is…”; rather, statements like “Guideline X, or article Y recommend…” are prevalent. Today, final treatment decisions are based on the clinical characteristics and investigations of each individual patient, yet the decisions we make are based on general principles and on evidence from large randomized clinical trials. We can say that the examination is individual – and the treatment is based on general principles. Here are a few examples, but there are many more: the choice of a coronary stent takes into account the anatomical particularities of the patient’s coronary artery, but we do not know anything about the patient-specific response; the choice of a valve prosthesis also takes into account the age of the patient and the size of the valve ring relative to that of the patient, but we have no way to predict the life-span of a specific valve model in a specific patient; even the choice of a specific drug takes into account some individual-specific characteristics (glycemia, creatinine levels, allergies, etc., etc.).

In the near future, precision medicine will be much more accurate (we hope) and will be influenced by genetics and artificial intelligence (AI), and then we will be able to talk about individual data-gathering mirrored by practice also based on individual characteristics [2]. Genetics has an important role, and great hopes were placed, but not many fulfilled, after the full ‘reading’ of the human genome. (The human genome was 92% deciphered in 2003, but research continues to elucidate the unknown sequences, and hope for further insights is rife). Genetic tests can identify a subject’s risk of developing a disease, define the personalized medicine needed to correct pathology and determine how the patient will respond to the treatment prescribed [3,4].

In cardiology, many genetic diseases are caused by modification/ mutations of structural proteins (hypertrophic cardiomyopathy, dilated cardiomyopathy, arrhythmogenic cardiomyopathy, dilated cardiomyopathy with hypertrabeculation, Takotsubo syndrome, Marfan syndrome) or through modification/mutations of proteins essential for the electrical activity of the heart (long and short QT syndrome, Brugada syndrome, catecholaminergic tachycardia, sick sinus node syndrome, atrial fibrillation). We had high hopes for clinical progress based on data provided by genetics, but few were fulfilled. For example, hypertrophic cardiomyopathy, a fascinating and intensively studied disease, still retains many mysteries. When I learned about the completion of the human genome, I hoped that the mysteries of this disease could be unraveled (the complexity of genetic mutations was not known then). We still do not know how to accurately identify patients at risk of unfavorable outcomes such as sudden death, and we do not yet know how a specific phenotype associates with a certain clinical course, so treatment decisions (such as implanting a defibrillator) cannot be made on genetic grounds, and we cannot prevent disease [5].

High hopes and great disappointments regarding genetic data are also related to other cardiovascular diseases (arterial hypertension, atherosclerosis, ischemic heart disease, etc.). A potential explanation is that most diseases have polygenic causes, and that the same gene can be found in many diseases [6]. We can, therefore, conclude that genetic data are not sufficient to allow us to practice precision cardiology. Other important factors have to be taken into account, as they undoubtedly modulate gene expression: individual risk (family history, lifestyle, exposure to environmental pollutants), adherence of the patient and of the family to the proposed treatment plan (the patient and their family are active factors in precision medicine and their preferences and values are paramount), genetic consultation (how we communicate the data to the patient and the impact of this data on the patient and family [7].

Currently and in the near future, an important (decisive?) factor is, and will be, artificial intelligence (AI), which will be able to analyze and integrate all the complex information obtained from the patient. AI has already transformed many domains of human activity (manufacturing, drug design, finance, etc.), but medicine is different from all these because of the immediate and profound impact on the individual of diagnostic and therapeutic decisions, and because of the complexity of the field (research, new discoveries, utility, safety, special regulations regarding application in practice, multiple unresolved ethical questions). AI, and especially machine learning and deep learning, will be transformative in this field, even more than we can imagine. Today, we can talk about four major applications of AI in medicine:

Clinical management of diseases - diagnosis and treatment, patient monitoring, imaging, telemedicine, preventive medicine;

However, AI is rapidly expanding in multiple areas. ‘Feeding’ the patient’s genetic profile to AI in order to make decisions about prevention, diagnosis, and treatment of the disease is - most likely - the domain of the near future [3,4].

We should never forget, as mentioned before, that multiple other factors matter, apart from the genetic profile and AI applications. Each patient has a different lifestyle, risk factors, history, features that influence the course of the disease, and a response to treatment. For the practice of precision cardiology, the following requirements will need to be fulfilled: advanced technology, suitable infrastructure, genetic research, and the training of doctors to understand and use the new technology [8].

A common question is whether precision medicine/cardiology is real progress or just hype [9]. We should notice that precision medicine also has potentially less attractive features. We still do not know very well whether the ‘precise’ treatment of a certain disease does not have negative effects elsewhere. Here is an example. Modern, precision, and personalized cancer therapy (successful for certain types of cancers), has led to new, hitherto unseen cardiovascular cytotoxic effects and to an increased burden of cardiovascular morbidity and mortality [10]. There are many other aspects related to AI and medicine/cardiology that have not yet been fully explored: who bears the ultimate responsibility for implementing AI-suggested measures? Is medicine at risk of becoming dehumanized? Is there any remaining role for humans in medicine and cardiology? Related to this, I note the pessimistic comments recently published in the Romanian Journal of Cardiology by Majd Protty and Adrian Ionescu under the title: The devastating impact of artificial intelligence (AI) on the traditional ways of practicing imaging and cardiology. Are ‘imagers’ an endangered species, and should trainees rethink their careers? [11]. The authors believe, in fact, they fear that medical imaging will be dominated/taken over by AI and young cardiologists should turn to other areas of cardiology where manual, procedural skills will be more difficult to replace. In the same issue of the Romanian Journal of Cardiology, Alan Fraser is analyzing ‘the good and the bad’ contributed by AI in imaging (echocardiography) and concluded: “The need for (truly) intelligent cardiologists will never be replaced by (falsely) “intelligent” machines. And patients want to be cared for by fellow human beings with whom they can talk and whom they can trust” [12]. We should have the answer to this debate soon, I hope, once we have the opportunity to work with these systems. For now, we are just making theoretical considerations, confined to the domain of ‘what if’…

More than ten years ago, Eric Topol wrote, in a 336-page book, about the “creative destruction” of medicine and referred to the digital revolution in medicine that opened a new era in medical practice [13]. Creative destruction refers to the transition from medicine based on rules extracted from population studies to medicine based on the individual characteristics of the subject to be treated, which means destroying an old model and replacing it with a new and superior model. There is also a question about whether the results generated by AI can be applied in medical practice and whether we trust them. We see that a lot of published work on medical uses of AI produces unexpected results, some difficult to understand and reconcile with clinical common sense, and which seem to be from the field of science fiction, but we still do not see the certification of AI products, their clinical impact is not obvious, no systems have appeared that can be purchased and used in daily practice, or if this is done, it is quite limited [15]. In the USA, the Food and Drug Administration had approved by the end of 2022 a staggering 500 (!) AI-based systems for medical use [16]. but these systems are nowhere to be seen in most hospitals [17].

I don’t want to end this brief presentation on a pessimistic note, so I will say that the personalization of treatment in cardiology holds great hope, but I am also asking myself whether personalized cardiology is ‘near us’ or whether we still have a long way to go. Our hope is that by adapting the treatment to the specifics of each patient, the effectiveness of the treatment can be increased, and the cardiovascular disease burden can be reduced.