Last March 13th the European Parliament approved the AI Actthe regulation on artificial intelligence which is a point of reference and at the same time a springboard for the use of AI while respecting the risks it entails and for the protection of fundamental rights.
A necessary debate which, however, can divert attention from the benefits of AI, often overshadowed by presumed cataclysms that these may represent for the world of work.
Mostly out of the spotlight, AI is reshaping the world of healthcare as a whole, from the doctor-patient relationship to treatment management, passing through diagnosis and also entering operating rooms.
Retina scans, prosthetics and voice recognition: how Google’s artificial intelligence helps us heal by Emanuele Capone 15 January 2020
AI and healthcare
AIs are revolutionizing the way medical data are examined, with benefits for both diagnoses and predictive investigations. What benefits are therefore the early detection of diseases, therapeutic paths (even remotely) but also the search for new drugs.
Artificial intelligence itself is a construct that si erige sul deep learning e sul machine learning and it is precisely by virtue of these two disciplines that research advances at a rapid pace.
Both assist analysis and research. To give an example, before going into the topic in depth, they read clinical tests identifying full-blown diseases and the risks of the onset of future diseases and, at the same time, they lend themselves to the discovery of new molecules which have a future purpose in the pharmacopoeia.
What does it mean in practice
When we talk about AI, we generally describe something with vague outlines that doesn’t help understanding what technologies are capable of doing. So let’s take a tour around the country to give some practical examples of what is happening and how AI is now part of healthcare, also clearly describing the risks that all this entails.
Starting from the diagnosis it is necessary stop in Rozzano (in the province of Milan) at the Humanitas Institute, in which the Radiology diagnostic unit uses 3D images for the diagnosis of the fetus which allow the development of the unborn child to be followed, facilitating any medical or pharmacological support. Similarly, 3D images are used in the oncology field, especially for research and personalized medicine to be applied to tumors of the biliary tract.
Il Gemelli Polyclinic in Rome uses advanced imaging techniques so that radiologists can be more precise in formulating diagnoses and, again thanks to the support of specific algorithms, the best treatments are identified by evaluating the patient’s response, thus moving towards predictive analysis. Something similar happens in Pesaro, near the hospital Ospedali Riuni Marche Nordwhich diagnoses with AI techniques that enable greater precision.
In each of these cases, the AIs are based on both the radiological images and the symptoms complained of by the patients to consult the relevant medical documentation, formulating the most precise diagnoses which will then be validated by the responsible doctors. AIs are able to compare data and images with those previously archived in special databases at a speed that man cannot emulate, discarding irrelevant information and focusing only on that which can lead to an accurate diagnosis.
AI in operating rooms
Three-dimensional imaging is not only useful for diagnosis but, as happens in Gemini, it is integral part of software that guide surgical interventions and allow medical devices to be placed inside the human body, increasing the safety of interventions and the likelihood that treatments will be more effective.
Returning to Rozzano, always Humanitas brings 3D technologies to surgery rooms orthopedic to reproduce bones or parts of bones (for example, a kneecap) customized and perfectly adhering to the anatomical characteristics of the patients. In addition to reducing any complications, this technique improves patient recovery and hospitalization times.
In the operating rooms of the Casa Sollievo hospital in San Giovanni Rotondo (Foggia) robots are used to cut, aspirate, irrigate and suture. The Vinci robot is a factotum that lends itself to various interventions, including those of gynecology, thoracic and abdominal surgery, but also for maxillofacial surgery. In the future it may also be used in cardiac surgery and breast surgery.
AI and machine learning
Machine learning is a subset of AI that allows machines to process data and learn without specific human intervention. The general context is that in which a machine learn to emulate human actions interacting with the surrounding environment (this is the case of robotics) and with data sets. In the field of healthcare, the issue of data, as we will see, plays a crucial role both for the advancement of technologies and for the particular needs of protection: it is difficult to find a more stringent and important form of privacy among all information that concern a person.
Still proceeding through examples (mentioning all the hospitals that use advanced technologies is prohibitive), the San Camillo Forlanini hospital in Rome and the San Raffaele in Milan they are Italian flagships. In particular, the Lombardy hospital uses AI for the treatment of non-small cell lung tumors to select patients to undergo immunotherapy. The success of therapies is stuck at 45% and AI is also used for offer more targeted ones and alternatives. These possibilities did not exist until recently.
Telemedicine
Among the many, the Bambino Gesù pediatric hospital and the Campus Bio-Medico Polyclinic, both in Rome, stand out. The term telemedicine includes a set of features that allow remote monitoring and communication between patients and doctors. The information collected is used both for diagnosis and for prevention and treatment of people.
These technologies they exploit sensors and algorithms to suggest to healthcare personnel how and when to act. To give an example, it is possible to release drugs remotely not so much by respecting an hourly administration schedule but when the drug is most needed. Likewise, telemedicine devices improve the therapeutic path by avoiding unnecessary medical visits and summoning the patient to a healthcare facility only when truly appropriate.
Telemedicine devices are taking hold everywhere even in Italy, the term you may come across is IoMT (Internet of Medical Things) and is made up of 3 large families, namely in-body, in-home and in-clinic. The former are devices that patients are equipped with and may not be invasive (for example smartwatches) that monitor specific parameters or, as in the case of pumps, that they release insulin when the patient needs it.
The ih-home devices, as the name suggests, they are placed where the patient lives and, for example, intelligent devices, beds equipped for sleep monitoring, fall into this category. When we talk about in-clinic devices we tend to refer to the same technologies which, for different reasons but often linked to the health conditions of patients, are applied within hospitals or for encourage hospitalizations or to examine data useful for formulating a diagnosis. In Italy, a structure particularly known for the use of IoMT devices is the Niguarda hospital in Milan.
Change! Fuse, with artificial intelligence in search of the stuff dreams are made of by Bruno Ruffilli 25 March 2024
Criticalities
The above are some examples of how AI they are part of Italian healthcare but, regardless, artificial intelligences as such work with data. Without quality data, AI can do little and, when it comes to health information, there should be specific rules for its protection. A possible loss of data would not only be a violation of privacy but could potentially cause harm to people and undermine the relationship between patients and technologies.
The European Union is addressing this issue, albeit partially: the European Health Data Space (EHDS) is a regulation in the process of being defined which wants to regulate the use of a common space for accessing and exchanging health data. The basic idea is to give European citizens greater control over their data which, at the same time, will favor research and health policies of the Member States.
This regulation, however important, does not resolve the issue of data security, which remains central and will continue to be the real challenge for all healthcare stakeholders.