AI Assisting Radiologists
The Israeli algorithm that saves lives in hospitals around the world has now started to work in Israeli clinics.
The artificial intelligence software identifies various urgent medical problems during CT scans, X-ray examinations and immediately sends alerts to doctors. This saves waiting time and eliminates delays that can sometimes result in serious damage and even death of the patient. The software, which is already installed in medical centers around the world, was recently put into operation in Israel. Expert: “The workload on radiologists is colossal; a doctor who has viewed 30 thousand photographs may miss a small pathology when interpreting. A computerized system never gets tired.”
Israeli hospitals have begun installing Aidoc software, an artificial intelligence that helps decipher CT and X-ray results and detects medical emergencies. The software moves urgent tests to the front of the queue and alerts the radiologist about a suspicious test. This helps reduce the number of errors made by on-call interns who have to perform thousands of tests.
Aidoc is actually a computer algorithm based on artificial intelligence. It is installed in the imaging departments of many Israeli hospitals and scans hundreds of thousands of X-rays and CT scans. Whenever the system detects a suspicious pathology in the images, the radiologist is immediately alerted about the suspicious results, and the patient’s data is moved to the front of the queue for priority. Meanwhile, the system marks the suspicious area on the photograph itself.
The system was developed at Sheba Medical Center, where it has been in use since 2016, through a lengthy “training” process that involved scanning hundreds of thousands of CT and X-ray images.
The system has learned to detect suspicious pathologies with high efficiency, which has also been proven in clinical studies published in the medical literature and presented at several conferences. To date, the system has been trained to detect strokes, cerebral hemorrhages, cerebral aneurysms, aortic ruptures, pneumothorax – a condition in which severe lung damage is observed, fractures of the cervical, dorsal and lumbar spine, rib fractures, pulmonary embolism, and to monitor the correct placement of an artificial respiration tube during resuscitation. These cases are defined as emergency situations, and any misdiagnosis or miss in them, as well as a delay in decoding the images, can lead to a danger to the life or even death of the patient. The company is working on developing additional algorithms that will allow diagnosing a huge list of diseases in adults and children.
The system was installed in the imaging department at Sheba Medical Center, and recently also at Bilinson, Ichilov and Soroka Medical Centers, and will soon be installed in other medical centers in Israel that have purchased it. Meanwhile, it has been successfully operating for several months in medical centers in the United States, South America and Europe. Other countries have already expressed interest in the development.
Never gets tired and reduces human errors
Director of Brain Catheterization Department at Sheba Medical Center Dr. Gal Yaniv ( also Aidoc co-founder) considers that it absolutely has proved its effectiveness. The scientist also points out that system determines the most important areas at image. Then mark them and inform the specialist about suspicious results. Thus, it reduces waiting times, improves the treatment process for patients arriving at the emergency room and hospital, reduces the likelihood of errors and misses for radiologists who sometimes have to review thousands of images in a single day, and, most importantly, it is a system that does not get tired, unlike a person.” Aidoc began operating at Soroka Medical Center about six months ago and has shown good results. Dr. Gal Ben Arie, Soroka Imaging Centre specialist recognizes that every radiologist can tired after a lot of conducted tests. “The system provides another layer of safety for patients, changes the order of priorities and speeds up the execution of urgent studies. According to studies published by the company, the device increases the potential for reducing human errors, and there is no doubt that this is a tool that can help save lives. An institution that does not have this tool will be left behind in the future.”
Is artificial intelligence expected to replace a doctor?
According to Professor Arik Blashar, Director of the Radiology Department at Ichilov Medical Center, the chances of this happening are slim: “We use artificial intelligence in radiology. The workload of radiologists is enormous, sometimes one intern has to check hundreds of tests in turn, each of which contains thousands of images. Artificial intelligence can only help, but it will certainly not replace a radiologist.”
According to Professor Blashar, the system increases the doctor’s confidence in the diagnosis, reduces the processing time of tests, since it focuses only on suspicious areas from the large number of images of each test, and helps to consider urgent cases. “People consider, that we are not allowed to make any mistake. The truth is that we are all human. People get tired, and a doctor who has looked at 30,000 photographs may miss a small pathology that a computerized system that does not get tired will recognize. This system has proven to be very accurate in its work, it definitely saves lives, and there is no doubt that this is the future.”