Science and Technology

ISR Researcher wins Google Award for developing Artificial Intelligence model to melanoma prognosis

Professor Catarina Barata’s research aims to help understand how does melanoma evolve and the various treatments.

Melanoma is the most dangerous type of skin cancer. According to the World Health Organization (WHO), there are about 200,000 new cases of melanoma and 55,000 melanoma-related deaths worldwide each year. In Portugal, there are 1320 new cases of melanoma and a mortality rate of 27% per year. Early diagnosis is key to reduce these numbers, as well as the ability to predict the progression of the disease and tailored therapies to the specific patient. It is with this aim in mind that Catarina Barata, professor at the Department of Electrical and Computer Engineering (DEEC) and researcher at the Institute for Systems and Robotics (ISR-Lisboa), develops her research for two years now, which has earned her a Google Research Scholar Award this year.

The Google Research Scholar Awards are delivered annually and aim to support early-career professors who are pursuing research in fields relevant to Google. The awards amount to around €50,000 and focus on developing collaborations with new professors and encouraging the formation of long-term relationships with the academic community. The Técnico professor was the only Portuguese researcher awarded from a total of 86 researchers, from more than 50 universities around the world.

The ISR researcher has been working in the field of medical image analysis using Artificial Intelligence (AI) for 10 years. Before doing her PhD, her work focused on melanoma diagnosis, but her background in Machine Learning (ML) has allowed her to make important contributions in other areas – security, robotics, etc.

Two years ago, she decided to focus her work on melanoma prognosis and started to develop a joint project with IPATIMUP and I3S, which is also led by Dr Helena Pópulo. “We contacted some hospitals across the country and asked for dermoscopy images to train the AI model and also the tumors. We are trying to understand if the extracted genetic information can be integrated into an artificial intelligence model to predict the prognosis”, highlights the professor.

Artificial Intelligence, a powerful ally for melanoma prognosis

As professor Catarina Barata points out, the diagnosis of melanoma is not the final step. “After diagnosis, the tumor is assessed in terms of genetic mutations in order to understand the best therapy to apply to patients”, she highlights. “Melanoma has a propensity to metastasize very quickly, so it is necessary to have tailored therapies to the specific patient, which is impossible right now”, she adds.

The ability of ML tools to detect key features from complex datasets can be a strong ally. “The machine is able to analyze many more cases than any doctor in an entire lifetime” and “combine different levels of information”, stresses professor Catarina Barata.

According to the ISR researcher, these technologies will be even more important in the future, due to new emerging therapies and hypothetical increase of melanoma incidence. “If I get sunburned yesterday it will not have an impact today, but 20 years from now, and so it is expected that the number of cases will increase”, she says.

“The reluctance that still exists on the part of doctors about this work makes me more willing to prove that it is possible”, says professor Catarina Barata. “I think their perspective changes when they realize that they can give some input and these technologies might be of some help”, she adds.

This work of persuading the medical community has been successful, as evidenced by the number of hospitals that are already collaborating with the project: Hospital de São João, Hospital de Santarém, Hospital de Santo António dos Capuchos, and the Portuguese Institute of Oncology (IPO).

“This project will take us much time”, stresses professor Catarina Barata. Even so, the ISR researcher believes that this award will allow her to take an important step forward: “predicting genetic alterations using dermoscopy images”. “We already have this result at the clinical level, and now we are going to try to obtain it using AI”, she adds. Although the project may take longer than desirable, one thing is certain: the ISR researcher will continue to work to make it possible.