Experts Find That New AI Is Officially Better Than Humans In Diagnosing Breast Cancer

4The results of the AI was more impressive in the US

In the USA the results were even more impressive. The false positives showed a reduction of 5.7% and the false negatives a reduction of 9.4%. What was incredible was the data available for the AI to study and make its diagnosis. The AI processed the information only form the most recent mammograms available without relying on extra information whereas the human specialists had access to patient histories and even earlier mammograms done on them.

Image Source: langimg.com

5It can detect breast cancer with greater accuracy

The research team involved with the study said that the latest study which was published in the Journal Nature. The report stated that the findings set the stage for the AI model to now support radiologists and specialists make more accurate cancer screenings and diagnosis. The UK lead researcher of Google Health Dominic King said that “Our team is really proud of these research findings, which suggest that we are on our way to developing a tool that can help clinicians spot breast cancer with greater accuracy.”

Image Source: www.medicalnewstoday.com

6Further testing is required before the method can be used by the public

He went on to elaborate that there was further testing also required including clinical validation, and regulation approvals before the AI system could be implemented to make a difference in breast cancer screening for patients. However, the team remains committed to working with its partners towards this common goal.

Image Source: cnbc.com

7Mammography is the most common form of breast cancer screening today

The most widely used breast cancer screening tool is mammography which is an x-ray taken of the breast from different multiple views. An analysis of the images is done by a radiologist but there are tie when the reports may be inaccurate such as false negatives and false positives.

Image Source: www.mayoclinic.org

You may also like...