Google System can Improve Breast Cancer Detection

Researchers in the United States and Britain reported that the Google artificial intelligence system proved to be good as expert radiologists found that women have breast cancer based on screening mammograms and promised to reduce errors.

Google Breast Cancer Detection

The study, published Wednesday in Nature, is the latest to show that artificial intelligence (AI) has the potential to improve screening accuracy for breast cancer, affecting one in eight women globally.

The American Cancer Society states that radiologists miss about 20% of breast cancers in breast cancer, and that half of all women receiving screening over a 10-year period have a false positive result.

The findings of the study, developed with Alphabet Inc. (GOOGL.O) DeepMind AI unit, which merged with Google Health in September, represent a major advance in early detection capability of breast cancer, Mozziyar Etemadi , One of its co-authors of Northwestern Medicine in Chicago said.

The team, which included researchers from Imperial College London and Britain's National Health Service, trained the system to identify breast cancer on tens of thousands of mammals.

They then compared the performance of the system with a set of 25,856 mammograms in the United Kingdom and 3,097 actual results from the United States.

The study suggests that the AI ​​system can identify cancers with a similar degree of accuracy to specialist radiologists, while reducing the number of false positive results by 5.7% in the US-based group and 1.2 in the British-based group % is.

It also cut the number of false negatives, where tests are classified as normal by 9.4% in the US group and 2.7% in the British group.

These differences reflect the ways in which mammograms are read. In the United States, only one radiologist reads the results and the test is performed every one to two years. In the UK, tests are done every three years, and each is read by two radiologists. When they disagree, a third is consulted.

SUBTLE CUES

In a separate trial, the group formulated the AI ​​system against six radiologists and found that it did not make them better at accurately detecting breast cancer.

Connie Lehman, head of the breast imaging department at Harvard General Hospital in Harvard, said the results are consistent with findings from several groups using AIIMS to improve cancer detection in mammograms, including their own work Huh.

The notion of using computers to improve cancer diagnosis is decades old, and computer aided detection (CAD) systems are common in mammography clinics, yet CAD programs have not improved performance in clinical practice.

The issue, Lehman said, is that current CAD programs were trained to identify the things a human radiologist could see, while with AI, computers learn to spot cancer based on the actual results of thousands of mammograms.

"More than human ability is the ability to identify subtle signals that the human eye and brain are not able to see," Lehman said.

Although computers have not been "super helpful" until now, "the device we've shown in at least tens of thousands of mammograms can actually be a very well-informed decision," said Etemadi.

Google System can Improve Breast Cancer Detection

There are some limitations of the study. Most tests were performed using the same type of imaging devices, and the number of confirmed breast cancer patients was much higher in the US group.

Critically, the team has yet to show equipment to improve patient care, said CureMetrix Chief Medical Officer Dr. Lisa Watanabe, whose AI mammogram program received US approval last year.

"AI software is only helpful if it actually transfers the dial to the radiologist," she said.

Atemadi agreed that those studies are needed, as is regulatory approval, a process that could take several years.