Medicine of the near future

Vice-President of the Skolkovo Foundation: human cloning is a matter of ethics, not technology

Roman Balandin, TASS

There are 430 projects in the Skolkovo biomedical technology cluster. Some of them are located at the junction of information technology and medicine. Kirill Kayem, Vice president of the Skolkovo Foundation, executive director of the biomedical technologies cluster, told TASS in an interview about whether a computer will replace a doctor in the future, whether editing human genes can create a perfect being and what prevents human cloning.

– Kirill Vladimirovich, you are in charge of medical issues at Skolkovo.  To what extent are digital technologies in demand here?

– At the moment, there are 430 projects in the cluster of biological and medical technologies, of which a good half are related to pharmaceutical developments. And at least 20% of projects related to medical devices and diagnostics use digital technologies. "Digit" is everywhere now. The use of sufficiently good computing power allows you to make a diagnosis faster, more accurately and more efficiently, prescribe a more correct treatment. The technologies that are coming into practical and clinical medicine today are good assistants to the doctor.

– And who finally makes the diagnosis? Computer or human?

– The decision is still made by the doctor. We will not rely on the diagnosis and treatment prescribed by the robot for a long time. But taking into account the complexity of modern biology and medicine, a doctor, when making the right decision, must analyze such a large number of factors that the human brain is simply not able to cover it all instantly. A good example is translational medicine (a field of science that allows you to transfer theoretical aspects of biomedical research and knowledge from physics, mathematics and other disciplines to practical medicine - approx. TASS), related to genetics. Scientists decoded the human genome about 10 years ago. Now we understand better and better what each part of the genome is responsible for. There are genes that affect expression (the process of converting hereditary information into RNA or protein) and the functioning of the cells that we are made of.

To describe the process of transferring information from genes to ribonucleic acid (RNA) or protein, scientists need many, many pages of text. That's when computers come to the rescue, because the doctor, of course, is not able to analyze all this data. He needs computing power, the same digital medicine, to get clear recommendations.

Where can this digital medicine be useful? In the diagnosis of diseases, because it is necessary to understand quite clearly what kind of failure has occurred. This applies not only, for example, to genetic diseases, but also to oncological ones, since they are often associated with failures in the genetic code.

– That is, the machine uses all the knowledge and experience gained by science, including information on the genome, for accurate diagnosis?

– Yes, that's right. It analyzes the entire array of data encoded in the genome to understand their effect on the metabolism of a human cell. But not only that. The symptoms that appear in various patients, signs of the disease and deviations that occur during the course of the disease are also analyzed. Using these gigantic data, it is possible to more accurately diagnose and do it not after the fact, when the disease has already appeared, but to predict its course. The doctor can then give the patient, and sometimes even the future patient, still a healthy person, the necessary recommendations or prescribe corrective treatment that will avoid serious consequences in the future.

– Will the doctor be able to predict diseases?

– Predictive or predictive medicine is what we are counting on in the development of clinical medicine with the introduction of machine analysis of big data. And in the future we will change the understanding of what health is. After all, today a healthy person is the person who does not have a disease at the moment.

In the future, medicine will probably consider an individual to be a healthy person who, according to the predictive approach, the disease will not develop in the near future. The healthcare system will incur significant costs for the patient even before the onset of the disease. De facto, unfortunately, we are all potentially sick people with you…

– There is a sad joke: our health is undiagnosed diseases.

– There is an even tougher joke among oncologists who say that not every patient lives to see his cancer. One way or another, the volume of accumulated genetic errors increases with age. And if you understand, based on a person's predisposition, which mistakes will accumulate in the first place, you can really influence both lifestyle and corrective therapy. But this applies not only to diagnostics and predictive diagnostics.

Working with the data of patients who have already developed the disease makes it possible to prescribe much more accurate and effective therapy with fewer side effects. After all, the development of medicines is now proceeding on the principle of – a complex word – spelling.

– Could you explain?

– Orphanization is a narrowing of the population of patients to whom a particular drug is directed.

In simple words, literally 40-50 years ago we just treated people for cancer. Then, with the development of science, we started treating people for specific cancers – lung cancer or breast cancer. Then, with the further development of science, there was an even greater split – we began to identify groups of patients with a certain type of lung cancer or breast cancer. And drugs are being developed for smaller and smaller populations of patients – for a specific genetic malfunction or for a specific biological target that causes this disease.

The more precisely we determine the mechanism of the disease, the more effective medicines we can use. Now the doctor prescribes the recommended line of therapy, the use of certain medications. With the improvement of the situation, everything is fine, we got there, the patient has a positive dynamic. But if the medicine does not work, second–line therapy is prescribed - other drugs are tried. Sometimes – third-row therapy. There is, for example, a disease – triple-negative breast cancer – triple negative breast cancer, when the patient does not have good dynamics for three lines of therapy. Because this is a very narrow branch of the reasons why cancer has arisen, there are currently no specific drugs for it. At the same time, each line of therapy has its own side effects that negatively affect health.

– So, not exactly the right tool can do harm?

– That's right. The patient's health suffers because cancer drugs are still toxic. What does digital medicine help to do? It is possible, having a large array of data of patients with similar diseases, including genetic and clinical information, to force a computer to make a rating of drugs that a particular patient, based on a combination of his clinical and genetic data, will be able to help.

This allows you not to go through several lines of therapy, but to immediately prescribe a drug that is most likely to be effective with minimal side effects. This is what is called personalized medicine. I cite oncology as an example, but, of course, this is a much broader concept. It's just that oncology is a very painful problem due to the high toxicity of drugs. Thanks to the existing technologies that are now on the border between science and clinical medicine, we are already implementing an individual approach to the patient.

However, there is a problem here. The data used by doctors are calculated on equipment, the principal purpose of which is scientific. And it is registered in the Russian Federation, first of all, as scientific equipment.

– In other words, experimental?

– We are talking about machines, about robots that allow you to read the DNA sequence, read the genome. But they are made for laboratories, first of all. And here, de facto, a medical service is provided. We, and this is absolutely fair, have Roszdravnadzor, which monitors this. What we use for clinical practice must be registered as equipment for medical services.

This problem exists all over the world. But countries where experimental medicine has developed faster have already proposed a mechanism called Laboratory Developed Test (LDT), which is the certification of laboratories using medical tests developed for scientific and clinical research. A special permit is obtained for such laboratories – in the United States, for example, such certification is called CLIA (Clinical Laboratory Improvement Amendments).

– And what exactly does this measure allow you to do?

– It allows the regulator to analyze the entire process of laboratory diagnostics, developed specific regulations, including the use of scientific instruments, and for this particular laboratory to confirm that "yes, we allow you to use such a device." It is not yet certified as laboratory diagnostic medical equipment, but it gives the doctor the opportunity to legally diagnose.

In order to change the situation in Russia, we have created a working group that includes representatives of Roszdravnadzor, as well as several companies, Skolkovo residents and partners who are engaged in developments on the border between science and practical medicine, including genetic diagnostics.

We are discussing the possibility of issuing such permits on the territory of Skolkovo so that our residents can help patients. It is very nice that the regulator understands the need for this. Colleagues at Roszdravnadzor think progressively and realize what a great help there will be for patients, how many human lives will be saved.

I believe that the introduction of rules for issuing permits by analogy with LDT will allow several times to increase the number of startup projects that are engaged in this kind of development. The number of patients receiving such care will also grow many times, and this is several thousand people a year. Only our residents help 2-3 thousand patients. We want to increase this figure to tens of thousands in Russia.

– We are not talking about the creation of new equipment, but about the development of data analysis systems?

– Yes, about regulations and data analysis systems mainly. At the moment there are several large equipment manufacturers. It is primarily supplied for scientific laboratories. What our developers are doing is just the use of neural networks, artificial intelligence (AI).

Ultimately, this will allow the analysis of "big data". In the meantime, the doctor will simply press the so-called "big green button", and the system will eventually give him recommendations on the selection of a drug or diagnosis without diving into the complex interrelationships of signaling pathways that affect the development of the disease or the choice of an effective priority drug.

– That is, in the future, with the development of these systems, the doctor's work will be reduced only to listening to the patient, correctly entering his indicators and…

– No, not really. This is only one of the factors. That is, of course, such systems are designed to help the doctor make a decision based on the array of data that the human brain is unable to process.

I will now name terms that are strange to the layman's ear – genomics, proteomics, transcriptomics, metabolomics. These are the names of branches of knowledge describing the steps that a cell goes through in order to develop, produce a certain protein, and divide. Artificial intelligence identifies critical points at each stage that can strengthen, weaken, kill a cell or make it "go crazy" and turn into a tumor. Unfortunately, without computer help, it would take weeks to analyze such information.

– Accordingly, does this lead us to the fact that we will need to develop a certain database of all patients?

– Such databases are constantly being developed, but at the moment they are scattered. Access to a part is open, but not to a part. There are both Russian and international databases, but many of them are scattered across fairly small groups of bioinformatics researchers.

– Is such a system being created in Russia?

– In Russia at the moment it is in a fragmented state. In some countries, so-called "National Genome" projects have been adopted. The United States and Great Britain have done this in the form of government programs. An array of data is collected together, and the state provides access to these databases to groups of researchers for the development of intelligent systems based on them.

We don't have such a unified national program yet. Although it would be extremely useful, because it would increase the number of people who are engaged in these developments, and the developments themselves would become more efficient.

The larger the database, the less artificial intelligence makes mistakes. In this regard, we have made a proposal within the framework of the working group, which was included in the government program on the digital economy. Its essence is to gradually start genotyping the population according to a certain schedule. The beginning will be the collection of the genetic database of newborns, followed by monitoring of ongoing changes in health status. This will help us better and better understand how certain deviations in the genotype affect the future state of health.

– At what stage is this issue in Russia?

– He entered the digital economy program. There are certain key points spelled out when the work will begin. We really lack a single center – a project office that would coordinate the maintenance of this database, organize it according to uniform standards and allow integrating existing arrays of information from different sources into the database. Plus, he would provide equal rights to all developers and be responsible for depersonalizing information, hiding the identity of a particular patient. After all, it is not so important to the developer what a person's name is, it is important to him that all attributes match.

At the same time, if the doctor wants to get information on a specific patient, it will be necessary to attach his personal data back. The project office should also be responsible for this personalization and depersonalization. The Minister of Health of the Russian Federation Veronika Skvortsova said at the WEF that this can be done using blockchain technology (a continuous sequential chain of blocks containing information – approx. TASS), because we are talking about the security of patient data. And this kind of project office would speed up the work in many ways, because it would become a center around which bioinformatics could build their projects, and would coordinate standards between them.

Similar associations exist at the level of each of the countries. There is an International Association for Working with Genetic Data, Skolkovo is represented there as one of the participants. And, probably, it is now the only participant from Russia.

Skolkovo could be a good base for such a project office, but to be honest, I will be happy regardless of where it appears, because it is important for me to give bioinformatics developers access to this data so that they can start developing new intelligent algorithms.

– Does the legal system allow us to do this?

– In principle, yes. In the law on personal data, we see mechanisms that allow them to be personalized and depersonalized. Plus, no serious financial investments are required, since the number of sequencers in the Russian Federation in various scientific institutions, with proper coordination, will be sufficient.

A sequencer, to simplify it a lot, is a machine that just reads the genome data, after which they can be placed in a common database. Some large injections into the purchase of heavy equipment are not needed - you can collect what you have. Funds may be required to coordinate this process, to obtain biological samples from the population, to maintain a database and purchase reagents. But this is not so much money for the first wave.

I think that the initiative we have proposed will be adopted, and next year we will start acting either as a coordinator or through residents who will replenish this database and work with data.

– What do you mean by coordination? Will there be a supercomputer in Skolkovo that will control the process?

– The computer can be anywhere right now. Rather, we are talking about coordination from the point of view of the project office – the development of common standards and interfaces, work with personalization and depersonalization, ensuring access rights to this database.

– Will artificial intelligence technologies be used?

– Elements of artificial intelligence are already being used here. But in addition to working with genome data, it can also be used to work with medical images. And this is also a serious and great diagnostic support for the doctor. We already have a fairly large material base of scanned medical images, such as magnetic resonance imaging (MRI) or computed tomography (CT), we have wonderful radio diagnostics doctors. But the number of these expert centers is limited, and often in remote regions where there are not enough qualified personnel, doctors can make mistakes, miss alarm signals in medical images that speak of a particular disease.

On the other hand, Russian developers who work with artificial intelligence in the field of image recognition have already advanced far enough in this. For example, Skolkovo startups take first places in competitions held by Google and international independent organizations.

If we put these two competencies together, we will really seriously affect missed diagnoses for tuberculosis or oncology. There are such developments, they are actively underway, but we can manage to become trendsetters in this trend.

– Which doctors are currently analyzing medical images?

– There are two groups of them. The former work with radiology (X–ray, CT, MRI), the latter are so-called morphologists-pathologists, that is, doctors who look at sections of tumors and clarify the diagnosis. At the moment, this method is the "gold standard" in healthcare. It is these specialists who make the final diagnosis – determine the benign or malignant type of tumor, the specifics of the cells that are predominant or cause concern in this tumor.

The number of such highly qualified doctors is limited worldwide. Skolkovo residents are now working hard in terms of creating educational programs for them, creating a system that allows them to hold consultations of these doctors, discuss the most difficult cases online.

And here artificial intelligence can also help a lot. Images indicating atypical cells are entered into a computer, the neural network analyzes them and learns in the future to identify such atypical cells itself in order to give advice to a doctor. This will seriously reduce the volume of medical errors and increase the accuracy of diagnoses.

– Are there any examples of such companies in Skolkovo?

– There is. This is a company called "UNIM", it is now opening a laboratory here. They gathered the best morphologists and made a database consisting of high-resolution images obtained from digital microscopes.

Artificial intelligence is currently being trained to work with these images. It is carried out, alas, by the hands of the doctors themselves, who upload the image, and then check whether the grid has started to identify them correctly. That's why it takes so long.

– That is, we are talking about the fact that sooner or later the doctor can just say: "Siri, what does this image mean?"

– More like "where are the atypical cells here"? And Siri in the picture will prompt: "Yes, here are atypical cells here and here, I draw your attention, check." And the doctor will act.

– So who makes the diagnosis in this case?

– Anyway, the doctor makes a decision, he can specifically look at the selected areas and check if the artificial intelligence was wrong and write: "Yes, great. I agree with that." Or write "disagree" and find something else.

But if he points out an error, next time artificial intelligence will determine such cells more accurately. As I have already said, we have a good position in the field of image recognition. Some foreign clinics are already using the development of our guys.

– That is, these developments can be used in remote medicine?

– They are already used in it. For example, the laboratory may be located in Vladivostok. If they have a high-quality microscope and scanner, the image is scanned, and then transmitted to the network in a special format. Morphologists in America, Germany and Russia can already look at a complex case together and say: "Yes, we agree. Yes, the diagnosis is exactly like that. It is necessary to treat so-and-so." It is already working as a distance medicine. Colleagues from "UNIM" still want to add an intelligent hint from the computer.

– But the doctor still has the final word. That is, the machine does not make a diagnosis, only assumes?

– The doctor will have the final word for a long time. Because the machine can give a hint, but the person who prescribed the treatment is responsible.

– If possible, name the hero companies that are doing all this?

– I have already said about the company "YUNIM". This company is a resident of Skolkovo. There are examples from another area – the selection of therapy for patients with cancer based on genome analytics with an understanding of how the disease developed. The company PONCC (the First Oncological Research and Consulting Center) works here, which has already helped to prescribe the right treatment for several thousand patients.

Representatives of the company are working together with IBM Watson (an IBM supercomputer equipped with artificial intelligence) to test their algorithms. The algorithm analyzes the patient's medical history and genotype for prescribing treatment.

– You said that a working group on LDT has been created. Are there any forecasts, estimates or deadlines?

– We would very much like to start work in 2018 and move the laboratories from the gray zone to the white one, so that doctors from all over the country could contact them, use their data officially. But this requires legislative changes. The question is how fast we can do it.

In particular, it will be necessary, for example, to introduce certain amendments to the law "On the Skolkovo Innovation Center". To solve this problem, we are currently consulting with our lawyers.

– Is Roszdravnadzor compromising?

– Roszdravnadzor understands that such risks will be reduced, since there is a single project office that will be easier to control. The department wants this to be created first of all on one site, and then spread from there to the entire country.

– Probably, this problem is universal for a number of countries. What are the global trends?

– Skolkovo, like all the centers supporting developments, is at the forefront, and regulators are slightly behind. But in some countries, the implementation process is moving faster and, interestingly, now the pole of progressive states is moving from West to East.

If 10-15 years ago we said that the United States was ahead, they allowed the use of methods much earlier than around the world, now Chinese colleagues are starting to get ahead of them. Today, the Chinese government is taking on quite high risks, allowing some projects that have not been allowed anywhere before.

For example, there are methods to correct the embryo's genome in such a way as to eliminate the cause of a congenital disease, and this change can even be inherited. China has become the first country to allow, so far for scientific purposes, to work with human embryos. That is, de facto they stepped over this ethical problem.

Another example is the so-called "Child of Three Parents" project. Its author Shukhrat Mitalipov is a native of the post–Soviet space, now works in England and in the USA. He deals with the so–called mitochondrial insufficiency - this is when the cell does not work quite normally. Mitalipov was able to combine the cell nucleus with someone else's shell. In other words, he took the nucleus of a fertilized egg, which contains the genetic material of two parents, and placed it in another egg…

– So that the genetic material is copied?

– To preserve the genetic material of the parents and at the same time remove the problem associated with cellular metabolism. The fetus develops from the material of three parents. Successful tests have already been carried out on great apes.

– Does Skolkovo somehow touch on such topics?

– We cannot act on projects that are not allowed in our country, therefore – no.

– But is the experience being studied?

- of course. That's what we're talking about. We even discussed the possibility of inviting Shukhrat Mitalipov to Skolkovo. But the Ministry of Health has a very cautious attitude to this, which is understandable.

– Do you mean ethical issues?

– Including ethical issues. For example, new methods of editing the CRISPR/Cas9 genome have appeared (a method of manipulating the genome, potentially can be used to treat hereditary human diseases – approx. TASS). Previously, scientists used to take a viral vector (that is, a piece of virus) to correct the genome. On this viral vector, they injected a useful part of the genome into the genome they were looking for. Here GMOs (genetically modified organisms) are mostly made on the viral vector in the world. But this part of the viral vector still remained alien CRISPR/Cas9 allows you to cut DNA faster and insert the desired section into it with fewer negative consequences.

This method, by the way, has already been confirmed on large animals. It can already be used technologically, many successful experiments have been conducted at the scientific level. The method can be used in the clinic only after serious validation, security checks, and this will be possible only in ten years.

However, there are also ethical problems. Potentially, the method can be used to improve the human breed by embedding useful features.

– We're talking about embryos. Accordingly, is this a correction of the genetic material of an unborn person?

– Yes, it is meant. The ethical question is, is it possible to make a perfect person by these methods or is it impossible? The answer to it is very complicated.

Here's a look at the cloning mechanisms – they are used right and left in agriculture. Now there is a giant trend for cloning pets. For example, in the USA, many owners of cats and dogs, after their death, donate genetic material to make them the same pet, one in one. Seriously. This is a huge market.

We were in San Diego for BIO, which is the world's largest exhibition in the field of biotechnology, and there were a number of companies that deal with this issue. Cloning dogs and kitties is, of course, a bit curious, but you look at it from a different angle.

For example, the production of breeding animals for agriculture, which cost hundreds of thousands of dollars. Now this is also already being done, although from the point of view of biology it is a complex organism, no less complex than a person. By the way, the issue of human cloning is primarily ethical. Technologically, this can already be done.

– This is already the plot of some kind of horror movie, in fact.

– Yes, sometimes I'm afraid of what I see.

– Can I call you an optimist in this regard?

– You can. Ethical topics are a very difficult question. For example, do we need another Einstein or not? Is it right to do this? But this is a completely different conversation.

Portal "Eternal youth" http://vechnayamolodost.ru  22.09.2017