Technologies based onartificial intelligence “already control large diagnostic imaging equipment (tomography
computerized, Tc o magnetic resonance), standardizing acquisition protocols and drastically reducing exam acquisition times, in order to improve patient compliance and comfort “. There are also algorithms” capable of supporting the radiologist in identifying pathologies (for example, nodules breast or lung, bone fractures, pneumothorax) by minimizing perception errors, or helping the specialist in the characterization of lesions (for example, the dermatologist with cutaneous melanoma), with the aim of improving the diagnosis “. The Higher Council of Health (Css) points out that it has released a document that outlines the proposals for the present and the future‘Artificial intelligence systems as a diagnostic support tool’.
“Artificial intelligence systems are rapidly entering the medical world, with the prospect, more or less in the short term, of significantly modifying diagnostic and therapeutic pathways, the decision-making methods of the doctor and, ultimately, also the doctor-patient relationship. Before they can be used in the healthcare sector – experts warn – these systems require rigorous scientific validation, based on methodologically sound studies (prospective, possibly randomized and conducted in real clinical environments) that demonstrate non-inferiority, or superiority, as well as the cost-effectiveness, compared to the conventional diagnostic and decision-making process “.
“It is also important to demonstrate the safety and reproducibility in the use of the software, as well as to consider the emerging ethical and legal problems inherent in the professional responsibility of the Doctor in the interaction with the algorithms”, specifies the Css.
According to the Css, “in order to introduce artificial intelligence systems (AI) safely into clinical practice and to compete internationally in their programming and development, it is desirable that the following interventions be implemented in our country: creation of organizational, computerized infrastructures, at local, regional or national level, of data stewardship and data governance; the creation of a governance structure of the AI systems by the Italian regulatory agencies, in particular the Ministry of Health as regards medical devices and AIFA for any therapeutic aspects, with the aim of establishing strict rules for the approval and registration of these systems – continues the document – the preparation of national guidelines regarding the methods of integration and the correct use of Ai systems in diagnostics, in agreement with the relevant scientific societies “.
“Furthermore – according to the Css – it is necessary to create a permanent national observatory at the Ministry of Health, to monitor the performance of the AI systems placed on the market (post-market analysis); the preparation of university and post-university training modules to improve AI knowledge and skills of medical staff and health professions integration of methodological elements on AI within upper secondary school programs and creation of information content, also through IT channels, at the service of the citizen ” .