Recent research by Google Health has revealed that artificial intelligence can now assist doctors in detecting early signs of Breast cancer, earlier than radiologists. This publication is also vouching for another similar announcement that AI can predict death via EKG results. Is AI really strengthening its grounds in the medical field and can be considered as a replacement to medical experts?
If you’ve ever read about the diagnosis of cancer, you might be aware of the term mammogram and how these images are then studied by radiologists to detect the presence of abnormal cells. If the test results positive, further diagnosis is done to find whether the type is malignant or benign. In a similar fashion, substituting for the medical staff, an algorithm has been developed to identify cancer at similar or even fastest rates to these radiologists. The database of mammogram images is called as OPTINAM and has been made available for others to alter and make similar versions.
“We show an absolute reduction of 5.7% and 1.2% (USA and UK) in false positives and 9.4% and 2.7% in false negatives,” the study says.
We are still doubtful of how early-stage cancer shall be detected accurately as even today, over or false diagnoses result in unneeded treatments that can prove to be harmful to the patient. The experiment conducted included reading the mammogram’s results by a noted radiologist, and the same by Google Health’s trained algorithm. A well-made algorithm could further reduce the need for revisiting the results and could save a lot of time, resources and costs involved at the patient’s end.
Though repetitive scans and tests are recommended for women falling in certain age groups, the intervals have to be set by medical practitioners varying from case to case, as widespread exposure to rays can also affect negatively. In 2014, a similar study conducted in Canada found that women given annual scans felt difficulty surviving breast cancer in the longer run. Here again, an efficient system with further improvements could add to the intervals between these tests.
Another test on record was done comparing results from six individual radiologists in parallel to the AI-based algorithm. The latter gave 11.5 percent better results than the radiologists and confirmed a promising healthcare future. The same is also being developed for the diagnosis and screening of lung cancer. The success of this algorithm shall go a long way saving several females around the globe from one of the deadliest forms of cancer. But, studies are still under process to learn the trends of other types and how those interact with their surroundings. Hopefully, the pace with which AI is growing, we are quite hopeful that human critical thinking shall soon add more to medical sciences.