A team of researchers at the Universitat Politècnica de València and the Instituto de Física Corpuscular, joint center between the Spanish National Research Council (CSIC, in Spanish) and the University of València have participated in the development of a system that helps diagnose breast cancer and is able to reduce the number of false positives. This new method provides 90% detection reliability, the highest of this kind of systems, and it will be quite useful in clinical practice. Scientific groups from seven other international centers are taking part in this project.
According to the researchers, the methods used today by radiologists are limited in detecting potentially suspicious areas in the image. However, this device is able to reduce the number of suspicious areas and false alarms and still provide information about the presence of a cancer. To do so, it is based on artificial intelligence techniques such as neural networks and the use of predictive algorithms.
Mammograms are diagnostic tests that for years have proved to be efficient for the early detection of breast cancer, one of the most prevalent tumors in developed countries. This new system can reduce false positives in every age range and minimize false alarms, thus preventing the need to do more damaging tests for women. In addition, it will make it possible to reduce clinical costs, which will help to incorporate new risk groups into the detection campaigns.
“Moreover, if there are other clinical signs that make the professional suspect there is a positive cancer diagnosis but it’s not obvious, the professional can expand areas where there is a greater suspicion of a tumor but is not detectable by the expert human eye. The aim is to locate future areas for biopsy,” explains Francisco Albiol, researcher at CSIC in the Instituto de Física Corpuscular.
“A patient’s life expectancy increases 20% for every year in advance that the breast cancer is detected. Thus, the algorithm we have designed can be a highly useful tool for the early detection of this kind of cancer, providing clinical professionals an additional expert system,” says Francisco Albiol.
Now, the partners of the project are studying how can they translate this method into clinical practice. “One of the possibilities might be using it to screen for the easiest cases in order to reduce radiologists’ fatigue,” adds Alberto Albiol, researcher at the Institute of Telecommunications and Multimedia Applications (iTEAM, in Spanish) of the Universitat Politècnica de València.
Digital Mammography DREAM Challenges
DREAM Challenges proposes challenges to the scientific community that are related to biology and medicine. The researchers submit their projects and those selected take part in the project, whose objective is to improve life for society in different aspects.
This system that helps diagnose breast cancer is the result of Digital Mammography DREAM Challenges, a global project promoted by the main American institutions fighting cancer, along with multinational companies such as IBM and Amazon. The aim is to improve breast cancer detection by interpreting the mammography with artificial intelligence techniques. 120 multidisciplinary teams have participated in the project. The Instituto de Física Corpuscular and the Universitat Politècnica de València are the only Spanish partners.
In this scientific study, patients’ data provided by USA medical institutions were analyzed. The results were recently presented at the International Society for Computational Biology Conference, held in New York.
“In order to use this kind of technology on a large scale, we need to generate and maintain local data collections from patients that generally represent the ethnic, nutritional and economical composition of a health system,” explains Francisco Albiol.