According to the magazine specializing in new techniques in health, culture and other fields, that one in eight women will be diagnosed with breast cancer during her lifetime.
In an attempt to speed up the detection process, the researchers experimented with a deep learning algorithm to identify breast cancer in optical scanners with the same or better radiologist precision.
Although the study is still in its infancy, it could potentially help reduce the results of erroneous exams in the United States and alleviate the shortage of technicians and radiologists in Britain, according to the article published by the journal for journalist Nicole Kobe.
And according to the definition of the Wikipedia site, deep learning is part of a broader family of machine learning methods based on artificial neural networks. Learning can be supervised, semi-supervised or unsupervised.
Algorithms are a set of rules and orders that are implemented in a sequential and orderly fashion to solve a specific problem.
And since early detection of the disease is a key factor in treatment, women over the age of fifty are examined in the United States and Britain even if they show no signs of developing breast cancer.
Negative screening results may be incorrect and cancer may be fatal.
In cooperation with the British National Health Service, Google’s DeepMind Corporation is developing artificial intelligence so that it can read retinal exams and detect neck cancer.
Researchers from British Imperial College for Cancer Research, Northwestern University, Royal Surrey County Hospital, and Google Medical Service used the deep learning system developed by DeepMind on two different sets of breast exam data, one in the United States and the other in Great Britain.
The researchers concluded that artificial intelligence can help accurately read x-ray images of mammograms.
In her article, Nicole Kobe, citing Google director of medical services, Dominic King, said that what had been achieved was another step in trying to answer some of the essential questions to present the results in the field.
“This is another step that brings us closer to our efforts to apply this type of technology safely and effectively,” she said.
The study indicates that the AI model can predict breast cancer as accurately as an experienced radiologist. Compared to human experts, the system was able to reduce false-positive results rates by 5.7% in the United States and 1.2% in Great Britain, and by 9.4% compared to negative results in the United States and 2.7% in Great Britain.
However, these results do not necessarily reflect how these diagnoses are read.
The study indicates that the DeepMind algorithm works better than a radiologist alone, and it is not inferior to the performance of two of them.
Despite the study’s success in its conclusions, artificial intelligence cannot completely replace radiologists, but it can help them in their work, as affirmed by Caroline Rubin, deputy head of the department of clinical radiology at the Royal College of Radiologists.