On May 13, the 3rd HealthTech Observer conference was held at La Paz University Hospital in the Community of Madrid. The event was organized by GMV and dedicated to “Artificial intelligence applied to radiological diagnostics.”

Fátima Matute, health councilor of the Community of Madrid, opened the event focused on two relevant topics: the interpretation of medical images with artificial intelligence (AI) in the field of primary care and the application of AI in healthcare in outer space.

In January 2021, GMV received the task of leading the ALISSE project consortium from the European Space Agency. The aim of this project is to help astronauts take care of their own health in an environment as aggressive as the one experienced during space missions. To do this, it developed AI technology based on deep learning, guiding and assisting astronauts to take high-quality diagnostic ultrasound images of various organs that may be affected by the conditions of crewed space travel. 

GMV led the consortium’s technological work, taking on engineering and project management tasks. Meanwhile, the Emergency Radiology Department of the La Paz Hospital was in charge of clinical consulting, labeling the images, and validating the technology. The consortium was rounded out by the Complutense University of Madrid’s (UCM) Nuclear Physics Group, which worked on creating synthetic images.

“We’re proud to have been a pioneering hospital in terms of implementing an artificial intelligence model for astronaut diagnoses in outer space,” said Rafael Pérez-Santamarina, the managing director of La Paz University Hospital. Since the start of the ALISSE project, led by the Radiology Department team, images of more than 50,000 patients per organ have been pre-selected from the more than 70,000 ultrasound studies performed by the Department each year. This will help astronauts in future space missions diagnose any conditions they may develop.”

In the complex context of life in outer space, ultrasound is an excellent diagnostic tool, since it involves lightweight, safe equipment that doesn’t take up much space or use much electricity and is capable of generating images in real time. For all of the above, this medical imaging equipment is also very useful in primary care visits. However, it’s an operator-dependent technology with a steep learning curve. 

Carlos Royo, GMV’s health strategy director, who is himself a primary care physician, moderated  the panel discussion on “Artificial intelligence for healthcare in outer space,”  and highlighted the current role of artificial intelligence in diagnostic radiology, paving the way for more accurate, efficient, and patient-centered medical care regardless of geographic location, even in outer space: “Ultrasound, an invaluable tool in modern medicine, has undergone a radical transformation thanks to breakthroughs in artificial intelligence. The combination of ultrasound technology with machine learning algorithms has transformed the way we diagnose diseases and medical conditions. We’re currently at the cutting edge of this revolution, exploring how artificial intelligence can improve the accuracy, efficiency, and accessibility of ultrasound diagnostics, from the early detection of disease to guidance during surgical procedures.” 

Primary care screening

Considering that the World Health Organization estimates that 80% of medical decisions are based on diagnostic imaging, incorporating artificial intelligence in ultrasounds is undoubtedly opening new frontiers in healthcare. In fact, AI can improve the training of healthcare workers and diagnostic accuracy in primary care services with the support of high-quality medical images,  facilitating earlier diagnoses and making it possible to rule out certain conditions without having to refer patients to a hospital. This increased efficiency has a direct impact on patient care, providing better results and greater resolution capacity and meaning patients don’t have to wait for the results of medical imaging performed by radiologists. In this way, not only do AI algorithms help non-radiologists use the equipment, they also make it easier for them to interpret the images. 

GMV’s AI solution can be extrapolated from spacecraft to primary care centers, including those far from hospitals or specialized radiology centers. In many cases, this saves patients from having to travel long distances and reduces uncertainty while waiting for results, freeing up hospital resources that do require specialized interventions.

Since primary care services address over 90% of the reasons for patients’ visits, we can consider them the gateway to the public healthcare system. Carlos Illana, GMV’s product manager and head of the team behind the ALISSE project, emphasized this point, explaining that “our technology is capable of guiding users with basic training in the acquisition of ultrasound images through a system offering multimedia information and artificial intelligence algorithms for guidance and the detection of diagnostic-quality images. The technical tests carried out have shown that the system is highly accurate, both in determining various healthy anatomical features, as well as in detecting pathological ones, and is compatible with several makes and models of ultrasound equipment.” In the clinical evaluation carried out by specialists from the Radiology Department of La Paz University Hospital, “it was found that staff with no medical experience or training produced images of sufficient diagnostic quality nine times out of ten, which is fairly close to how expert sonographers performed on the same tests. This means that applying ALISSE’s AI technology in primary care centers would make it possible for non-specialist healthcare personnel to carry out initial medical imaging tests, meaning patients wouldn’t have to be referred to radiology services at other facilities.”