The artificial intelligence of Google is a step forward in the diagnosis of the tumor. To mark the new milestone is a study published in Nature and conducted by experts at Northwestern Medicine in Chicago, Imperial College London, and the uk national health Service. The To the Big G this time has been involved with about 26 thousand mammograms of patients in the United Kingdom and 3 thousand in the us. The software has been shown to have an accuracy in the diagnosis is comparable to that of experienced radiologists, and has made the best of the man by reducing the incidence of false positives and false negatives. It is one of the many projects that Google Health is bringing forward in the field of health, still under study and is not yet available for widespread use.
The algorithm can be ’trained’ to recognise patterns and interpret the images helping to detect lung cancers from Ct scans, to diagnose diseases of the eye in people with diabetes and to find the cancer on a microscope slides.
In this case, the artificial intelligence of Google applied to screening of breast cancer has reduced the false positive rate of 5.7% (on cases, Usa) and 1.2% (on cases Gb), and false-negative rates of 9.4% (Usa) and 2.7% (Gb). The diversity of the results is related to the fact that in the United States, mammograms are read by a single radiologist, and the examination is carried out every 1-2 years, while in the United Kingdom, every mammogram is examined by two radiologists, with a third party expert consulted in case of discordant opinions – and the test is done every 3 years. In a separate survey, the system based on The is compared with six radiologists, demonstrating a greater accuracy in the diagnosis.
“Many radiologists do make mistakes, some well outside of the acceptable margins of normal human error,” he told the New York Times Constance Lehman of the Massachusetts General Hospital in Boston, is not involved in the study.
In this case the reading of the mammography on the part of the neural systems have proven to be of help. “We have taken the mammograms, we have shown to the radiologists, and we asked to check for the presence or absence of the tumor. Then we submitted to artificial intelligence, and we performed the same test,” he explained Mozziyar Etemadi at Northwestern University. The Ai was found to be more accurate radiolog, because unlike humans the computer does not get tired and does not distract you towards the end of a long day of reading of mammograms, has clarified Etemadi.
“The Republic will fight always in defense of the freedom of information, to its readers and to all those who have at heart the principles of democracy and civil coexistence”
Carlo Verdelli SUBSCRIBERS TO REPUBLIC © Reproduction reserved Today on The face of bipartisan that wants to say goodbye to an Altitude of 100: here’s how you could change to another idea of Italy, The challenge for the Count: survive in January and accelerate on the program for The 2020 opens with a verification obstacles for the executive To Mattarella, the address that you never expect
the Republic
