In the past few years, artificial intelligence and deep learning are rapidly changing the world. In the field of medical health, artificial intelligence for disease diagnosis and pathological analysis is endless. As one of the pioneers and leaders of artificial intelligence, Google naturally will not miss this new direction that is booming. Recently, scientists from Google, Google Brain and Verily have developed an artificial intelligence that can be used to diagnose breast cancer. It has outperformed professional pathologists. â–²The first time on Google’s blog, this news was published (Source: Google) Limitations of manual diagnosis The diagnosis of many diseases relies on pathologists' analysis of tissue sections, which has become the gold standard for disease diagnosis. For breast cancer patients, the pathologist will use a microscopic examination to carefully look at the lymph nodes next to the breast to find traces of the tumor. Based on the results of the microscopic examination, the pathologist will tell the patient where the breast cancer is and whether the tumor has metastasized. These analyses directly determine the treatments and disease management methods that patients should take. It is estimated that there are 230,000 breast cancer patients in the United States each year who need to be diagnosed and listen to doctors' verdicts on their fate. â–² Can you find out which tumors are from this slice, which are normal tissues, and which are normal tissues that look like tumors? (Source: Google) However, such a link that affects life has a great congenital deficiency. First of all, manual diagnosis is very error-prone. Many researchers have found that even for the same patient, the diagnosis given by different pathologists tends to be very different: a 2015 paper found that the consensus rate of different pathologists for breast cancer diagnosis was only 75. 3%. In some heterotypic breast cancers, the consensus rate of diagnosis has dropped to 48%, less than half. It is conceivable that many patients are at risk of misdiagnosis, which undoubtedly makes the patients who are racing with death go round the curve, making the situation worse. Secondly, although the pathologist is not perfect, it is not an overnight success to cultivate such a talent. After studying in a basic medical school, these experts must undergo years of training in order to gain sufficient experience and learn to analyze the techniques of pathological sectioning. In areas with insufficient medical resources, it is a luxury to want to get a diagnosis. Scientists say that behind these two congenital deficiencies, there is a common reason that these pathologists need to process too much information. For a single patient, there are more than one slice, and each slice contains tens of billions of pixels under the microscope. What kind of concept is this? A Google report gave us an analogy. Suppose we have 1000 HD photos with tens of millions of pixels, and you have the ability to tell which of the 1000 photos may have a problem. â–²To correctly diagnose cancer, you have to find the wrong ones in tens of billions of pixels (Source: KingofWallpapers) Crt Monitoring,Crt Graphic Display Unit,Crt Device Fire Monitoring System,Crt Graphic Display Unit Control Panel LIAONING YINGKOU TIANCHENG FIRE PROTECTION EQUIPMENT CO.,LTD , https://www.tcfiretech.com
