The practical application of artificial intelligence (AI) in medicine 2022

In recent years, artificial intelligence in the context of health has been talked about a lot and often, but most people still see it as an element of the future, not the present. That is why I decided to ask several experts in the health industry operating at the interface between medicine and technology to indicate which applications of artificial intelligence already implemented in practice are the most important from their point of view. In addition, experts present which projects that are still at the R&D stage are worth following, because their dissemination will be another milestone in the development of medicine, as well as the biggest barriers to the wider adoption of AI solutions in medicine.

Ligia Kornowska, leader of the AI ​​in Zdrowie Coalition

Top three key current applications of AI in health, with particular emphasis on solutions used in Poland

Currently, the key application of AI in medicine is the use of algorithms for diagnostic support. Especially in radiology, but also in cardiology and other fields of medicine, many AI algorithms pride themselves on the sensitivity and specificity of a diagnosis that is several to several percent greater than the diagnosis proposed by a doctor or a team of doctors.

In my opinion, they have the greatest potential solutions supporting diagnostics in the early stages of the disease, for example, cancer or cardiovascular disease. There are more and more such algorithms, and thanks to hard scientific evidence confirming their effectiveness, they can be and are already used in many healthcare entities in Poland and around the world.

Examples of diagnostic algorithms include: In: Ultreon System. 1.0, which enables the detection of the most difficult to treat calcified atherosclerotic plaques. This system is implemented in 4 hospitals in Poland (Poznań, Wrocław, Kraków). The assessment of CT examinations in the field of tumors is possible thanks to the algorithms of the companies Radiato.ai and RSQ AI, and their solutions can be found in hospitals in Gdańsk, Poznań or Wejherowo. The system supporting the evaluation of histopathological preparations, implemented by the hospital in Zakopane, created by the CancerCenter company, is also extremely interesting.

The most important R&D projects worth observing in Poland and around the world

In addition to projects related to the creation of AI algorithms, it is definitely worth following the development of projects enabling obtaining high-quality medical data. To create each new medical technology, each AI algorithm or personalized treatment paths, access to data is needed – so projects that provide such access will have a high growth rate. In Poland, among others companies such as Upmedic or CliniNote help in structuring medical documentation, improving its quality. Upmedic solutions are implemented, among others in hospitals in Warsaw and Gdańsk.

I definitely recommend watching the development the idea of ​​data donationwhich is promoted by the Podaruj Dane Foundation. The idea is that the patient – just as he can donate blood or bone marrow to save the life and health of other patients – can also donate his data to accelerate the development of new drugs and therapies. Already this year, the Foundation will make it possible to record the patient’s first consent on an innovative internet portal using blockchain technology.

The biggest barrier, the overcoming of which will make AI much more widely used in medicine

There are two challenges: the first is about access to medical data for the development of AI algorithms. The second concerns education of medical personnel and management in the area of ​​implementation of algorithms in clinical practice.

We try to address the first challenge by developing a data donation scheme that will enable obtaining medical data while respecting patient rights.

Responsible for this second challenge, we create a White Book of AI application in clinical practice, which will not only present good practices related to the implementation of AI in a therapeutic entity, but also serve to educate medical personnel in the use of algorithms in an effective and safe way for the patient.

Piotr Orzechowski, CEO of Infermedica

Top three key current applications of AI in health, with particular emphasis on solutions used in Poland

One of the key uses of artificial intelligence in the medical industry is to use it for telemedicine support. In particular, these are devices that allow you to pre-assess a health condition, detect health events or speed up the diagnosis process. Solutions such as StethoMe, or the so-called wearable devices based on algorithms (Apple Watch, Samsung Watch, Sidly) allow doctors to access complete and precise data on the patient’s health.

Huge opportunities will also bring support for diagnostics based on photos and images. An example is the use of deep learning in creating models to identify dermatological changes (Google Health, Triage) or radiological images.

Spreading education on health prevention as well as supporting patients in making decisions are other uses of AI. Examples include Infermedica solutions, as well as services such as Amazon Care, which has a voice interface for checking symptoms and quick consultation with a doctor.

The most important R&D projects worth observing in Poland and around the world

Digital health is a dynamically developing area with a huge number of projects worth observing, so I will focus only on a few of them.

As far as the Polish market is concerned, the following projects deserve special attention: StethoMe – a device detecting abnormal sounds in the respiratory system, AioCare – a personal system for monitoring and treating lung diseases, Saventic Health – a system supporting doctors in diagnosing rare diseases, or Consonance – a company supporting enterprises from the border of medicine and technology.

In the world, in addition to Google Health and Amazon Care, for example, AliveCor – a personal ECG system – is interesting.

The biggest barrier, the overcoming of which will make AI much more widely used in medicine

According to WHO data, by 2030 there will be a shortage of nearly 10 million doctors, nursing and obstetrics specialists worldwide. So we are dealing with the already visible and deepening problem of the lack of medical staff, which cannot be solved by the existing education systems. In the European Union, the average waiting time for an appointment with a general practitioner is already 4-5 days, which is a long and, in some cases, dangerous period.

I believe that they will be crucial for us solutions to relieve the health care system, facilitate the work of medical staff and automate tasks (especially administrative tasks). Currently, completing paperwork takes a lot of time, which often contributes to burnout among health workers.

It will be very important too creating solutions supporting and educating patients on proper disease prevention, a healthy lifestyle and, whenever possible, self-treatment of simple health problems.

Bartłomiej Lubiatowski, President of RSQ Technologies

Top three key current applications of AI in health, with particular emphasis on solutions used in Poland

The key applications of AI in Poland are image data analysis, body composition analysis and predicting how patients will respond to treatment.

The most important R&D projects worth observing in Poland and around the world

It is always worth observing the directions of the four main corporations: Amazon, Apple, Facebook and Google. If some projects are going very slowly for these great players, it means that something in the world is not going well. Then, space is created for smaller companies, about which it is not loud and it is not yet known whether they are worth observing. In addition, I see two trends in R&D projects dealing with pictorial artificial intelligence. One is the creators of large cameras who put AI on, for example, resonance. The second is makers who are camera-independent and create solutions that work with any set of photos based on a PACS database. Both have their pros and cons, but the latter, which works independently of the cameras, is more implementable – this is the approach we chose for the RSQ AI. We want to improve everything hospitals already have without spoiling their established processes, but only simplifying paths.

The biggest barrier, the overcoming of which will make AI much more widely used in medicine

For now, AI works mainly in scientific publications, not in a real medical environment. And it works on very narrow data available to universities, and not on the data of patients coming to the HED. It is crucial that the academic community learns to make the product. We have to realize that no one really needs artificial intelligence. You need smart reports and information on patients’ health so that you can take immediate action. AI is only one of the components that must run on PACS databases, and must function in a hospital infrastructure in which it can be deployed and then maintained. It is 10 times more difficult to do something in a hospital, in a process that is practical and convince medical staff to use, than to do the algorithm and model for evaluating photos alone. We also know from the implementation, e.g. the experience of RSQ AI with the HCP hospital in Poznań, how positive it is to take into account the realities of hospitals and enrich the current processes.

Kazimierz Cięćak, Consulting & Product Management Director, Comarch Healthcare

Top three key current applications of AI in health, with particular emphasis on solutions used in Poland

Looking at the current applications of AI, especially in the Polish health care system, one should first of all pay attention to:

1. Support for the diagnostic process / intelligent medical intelligence – the use of artificial intelligence in the processes of initial diagnostics already significantly contributes to the improvement of processes in this area.
2. Real-time systems supporting medical staff to identify patients at risk (analysis of previous medical data). One such area is the risk of re-hospitalization. Based on the information and data collected from home telemonitoring, patients are identified who are more likely to return to the hospital within 30 days from the date of discharge.
3. Improving logistics in health care – Artificial intelligence in health care and medicine is able to better organize the diagnosis of patients or their treatment plans, so that the organization time of the entire process is the most favorable from the point of view of the process.

The most important R&D projects worth observing in Poland and around the world

It is very difficult to clearly identify the most important R&D projects. I think you should be aware that even the best research and development activities remain useless without commercialization. From my point of view, the activities of the giants should be watched with particular interest. For example, Amazon’s recently announced partnership with Teladoc shows the significant commitment of technology companies not only to existing, well-known processes, but to the potential they see in the health sector.

The second direction that must be followed is the activities in the Middle Kingdom. In 2017, China announced the creation of the New Generation Artificial Intelligence Development Plan – a strategy that aims to achieve the position of a global leader in the AI ​​sector by 2030. And China is making great strides in this regard. In 1997, Chinese scientists accounted for only 4% of scientific articles on AI, but in 2017, Chinese scientists conducted more than 27% of all research, ahead of the United States, which was previously the country with the most advanced AI development.

The biggest barrier, the overcoming of which will make AI much more widely used in medicine

In my opinion, the most important challenge will be creating appropriate legal regulationswhich, on the one hand, will allow for the definition of data accessibility rules for developers of AI-based solutions, and on the other hand will ensure adequate protection of patients’ rights.

Intensive work has been going on for some time, both at the European and national level, but being aware of how complicated issues require regulations and how unknown issues we touch, I am convinced that appropriate solutions in this area may contribute to a fuller development of artificial intelligence solutions in medical industry. Particularly noteworthy in this regard is the initiative of the “White Book of AI in Health”, which is currently being developed.

Łukasz Osowski, Founder & CEO of Vika Life

Top three key current applications of AI in health, with particular emphasis on solutions used in Poland

Currently, we are seeing significant benefits for patients and healthcare in a large scale of impact, among others for the following solutions using AI:

1. Diagnostics, incl picture diagnosis – supporting diagnostic doctors in diagnosing all types of images: X-ray, MRI, ultrasound, CT, etc.

1. benefit for the doctor: significant acceleration of work
2. benefit for the patient: reduction of misdiagnosis

2. Initial heart diagnostics on fitness trackers

1. benefit for doctors: capturing previously undetectable cases
2. benefit for the patient: diagnosis without visiting a doctor

3. Democratization of prevention. Extending healthy life and reducing the incidence of diseases thanks to the mass, cheap availability of diagnosis and prevention, incl. with the use of lifestyle medicine.

1. Benefit for the doctor: reduction of incurable diseases of affluence
2. Benefit for patients: reduction in the incidence of diabetes, cardiovascular diseases, cancer, stroke, hypertension

The most important R&D projects worth observing in Poland and around the world

1. Vika.life – Polish application and at the same time a game positively influencing a healthier lifestyle. Soon, it will be the first in the world to show “live” the long-term impact of any physical activity on health and the body. Physical activity is responsible for up to 12 years of health in our lives!
2. Attain – a pilot project of Apple and Aetna – one of the largest insurers in the world. It uses Apple Watch and gamification positively influencing the healthier lives of millions of people insured in Aetna in the USA.
3. Fitness trackers equipped with functions of monitoring ECG, movement, temperature and other vital parameters They allow hundreds of millions of people in the world to have an ongoing insight into life parameters and be warned about disorders that would be difficult to detect without the use of such devices.

The biggest barrier, the overcoming of which will make AI much more widely used in medicine

The biggest challenge is providing patients with easy access to the most effective solutions. The widely available Internet has democratized access to knowledge and contacts with people. Likewise, readily available health tools will democratize diagnostics and medicine because they are effective, cheap and scalable thanks to AI. Everyone will have access to the highest quality diagnostics, prophylaxis and treatment elements immediately, easily, cheaply and often for free. In this way, the average person should gain at least a few extra years in health, while reducing expenses and the amount of time spent on diagnosis and treatment.

Mikołaj Gurdała, Account Director, Public Health, IQVIA

Top three key current applications of AI in health, with particular emphasis on solutions used in Poland

From the analyzes carried out by EIT Health (https://eithealth.eu/wp-content/uploads/2020/04/EIT-Health-Mckinsey-Infographic_FINAL.pdf), in which representatives of companies offering AI solutions and specialists in the field of medicine participated , almost 1/4 of the respondents indicated diagnostics, including imagingas the main area of ​​application of artificial intelligence. This is not only a trend in Poland, but all over the world. Imaging data is the most “grateful” range of data that AI algorithms can learn from easily, and by supporting diagnostic staff, they can reduce diagnostic time in areas such as oncology and neurology.

In the same report, almost 1/5 of the respondents noted that the widespread use of artificial intelligence in medicine is systems supporting decision making by clinical teams. Algorithms are used both in diagnostics and planning the therapeutic process. However, AI tools of this class must undergo a complicated procedure of registration as medical devices, because while supporting medical personnel in making decisions, they must guarantee the highest possible level of security, not only in terms of data processing, but also be precise and specific in presenting decision scenarios.

Data management in turn, this is an area of ​​AI use that is significant for 15% of respondents. Data management relates naturally to clinical trials, but also to the more and more intensively developing, also in Poland, use of data for the management of medical entities. For health service providers, planning the activities of the individual based on historical data as well as those collected in real time (real world data) is a real source of valuable information about costs, resources and patient movement. The most sophisticated algorithms can, for example, analyze the patient flow based on city traffic or weather conditions. This use of AI affects not only the efficiency of hospital management in terms of costs, but also enables better planning of treatments and other medical procedures (the phenomenon of no-show-ups).

The most important R&D projects worth observing in Poland and around the world

BrainScan – a very interesting Polish start-up developing a solution based on AI in computed tomography of the brain. The company develops a unique machine learning technology that allows not only to shorten the diagnosis time, but also translates into a change in patient management paths. BrainScan is also a very interesting set of business and technological competences, which allows the company to actively cooperate with clinical centers (private and public) and strategic partnerships, e.g. with PZU Zdrowie.

Prosome – the only Polish start-up with an advanced solution supporting oncological patients, which is in the process of obtaining certification in the DiGA mechanism in Germany. A dedicated mobile application allows patients to manage adaptation in the environment during the therapeutic process. Prosoma implements its project in cooperation with centers in Poland, Europe and the United States, having an extensive network of opinion leaders supporting the application. The Prosoma solution is a very interesting example of AI implementation in the area compliance in chronic diseases.

Patch AI – an Italian start-up that has developed a tool to engage patients in hybrid and decentralized clinical trials. The PatchAI algorithm in the form of a mobile application allows clinical teams, CROs and pharmaceutical companies to more effectively involve patients in the research process, strengthen their role in clinical activities and better management of data in real time. In 2021, PatchAI signed a partnership agreement with Roche, the basis of which is the wider use of the application for oncological patients, although it is already known that the company, by conducting an active business activity, will expand the use of the solution in other diseases, e.g. in diabetes.

The biggest barrier, the overcoming of which will make AI much more widely used in medicine

Access to the data – this issue is crucial for the wide application of artificial intelligence in medicine and its further development as a technology. In this respect, Poland and Europe must perform a lot of legislative and procedural work. The new agenda of the European Commission in the field of European Health Data Space and the announced regulations of the European Commission covering the use of artificial intelligence give hope. Both companies and clinical centers clearly indicate that data access, interoperability and transparent data management processes are the biggest barriers. If we create a safe and transparent system of general data collection and sharing in Poland in the next 3 to 5 years, serving the development of AI solutions, or more broadly, digital health, we have a chance to make our health services market an attractive space for investments.