Artificial Intelligence can help initiate cancer treatment quickly. For this reason, seven Greek scientists established DeepMed IO, along with an innovative system that accelerates and supports the diagnosis of cancerous tumours.
Observing a global problem—the lack of pathologists—which causes delays in the diagnosis of cancer, the UK-based DeepMed IO team developed an algorithm that helps buy time for doctors and patients - DeepPath—LYDIA.
“DeepPath-LYDIA leverages deep intelligence technologies to provide diagnostic assistance in cancer staging. It is a product that has received the CE-Mark and can be installed in any hospital or private organisation with digitised pathology laboratories," says Konstantinos Vougas, the founder of the start-up company.
The algorithm analyses digitised lymph node sections with particular sensitivity and precision, identifying cancerous areas in melanoma, breast, lung, and colon cancer patients. The system is designed to notify pathologists that cancer foci have been detected and provide specific measurements on those areas to help the pathologists complete the disease diagnosis faster and more accurately.
According to Mr Vougas, the algorithm is not designed to work autonomously and does not eliminate the necessity of the pathologist.
Experienced pathologists have nearly 100% accuracy in detecting metastasis. However, when people work under pressure and the hours are long and exhausting, performance is bound to decline.
“Unfortunately, stress and overtime are a daily reality in the pathology field worldwide due to the lack of workforce. Systems that are as accurate as an experienced and rested pathologist, like DeepPath – LYDIA, 24 hours a day, 365 days a year, will help dramatically reduce any problems of inaccurate diagnosis that exist at the moment," he points out.
Benefits for patients and Health systems
Using the system as a support tool reduces the waiting time for the final diagnosis and improves its accuracy, thus offering immediate benefits for patients in terms of quality of life and the final outcome of the disease.
"In addition to the patients, the above benefits for the health systems translate into an improvement in the quality of the services provided by them," explains Mr Vougas.
"In addition, it is a money-saving benefit as the system makes pathologists at least twice as fast in this diagnostic examination, which forms a very important part of their daily routine. Also, using the system saves money by reducing the use of immunohistochemistry to detect metastasis, which is particularly expensive as a technique since it involves antibodies."
Does the system apply to hospitals?
DeepPath—LYDIA is currently the only system in the world to detect melanoma metastasis and, furthermore, the only system to detect three additional types (four in total) simultaneously with the same AI model without the pathologist having to define the cancer type.
The system is being tested in a pilot phase in hospitals in the UK, the Netherlands, and Israel. The company has officially joined the providers of software solutions with Artificial Intelligence to the British health system (NHS) and, in fact, participated in a competition to install the system in PathLAKE-Plus, a consortium of 26 hospitals in the south of England, in 2024.
"We are currently seeking funding to expand the system to all cancer types and bring a pan-cancer detection system to market," said the founder of DeepMed IO.
When asked if there is interest in implementing the system in Greece, he notes that a digitised pathology laboratory is required to operate the system. Unfortunately, in Greece, hospitals are far behind in this area and need to have these services digitised.
The team behind the innovative algorithm, in addition to the startup's founder, consists of Akis Manolis, Christos Spyropoulos, Nikos Niotis, Thomas Papathymiopoulos, Konstantinos Gyftodimos, and George Kalikakis.
Maria-Niki Georgantas is a columnist for Ygeia Mou. Translated by Paul Antonopoulos.
READ MORE: Gallstones: A Greek expert explains the complications and treatment.