Google AI Can Now Detect Breast Cancer Better Than Humans

Left slide contains lymph nodes. Right slide shows LYNA is identifying the tumor region.

Google has once again shown why it is among the largest tech corporations in the world as it has developed an artificial intelligence program called Google AI that is claimed to be better than pathologists at detecting late-stage breast cancer.

The Google AI, named Lymph Node Assistant aka LYNA, was developed to recognize the characteristics of tumors by studying the scans from cancer patients. This AI can detect secondary cancer cells in the pathological slides at 99 percent accuracy while it is only 81 percent on average for pathologists under time constraints.

Secondary cancer cells can spread from original cancer and grow in different parts of the body. Therefore, finding them can be a hard task for pathologists.

LYNA was proven to be more effective when pairing with pathologists. In an LYNA’s trial, pathologists using LYNA were being able to inspect cancer scans with better accuracy in only half of the normal inspection time.

However, this AI can only be effective in detecting last-stage breast cancer which is usually pretty late for patients who have them. It also has not been applied in real situations. But the scientists said that metastasis is a common factor in most known forms of cancer.  Therefore, it should not be long until someone finds out ways to adapt the systems to detect other tumors. 

Google is contributing a lot to the health care AI. Google's Medical Brain claimed in June that they had managed to create an AI system that was able to forecast the mortality rate with the accuracy of 90%. Verily of Google has also made inroads with determining the risk of a person getting heart disease with retinal scans.

Anyway, hopefully Google's LYNA will be ready for practical use soon as it could boost the accuracy of cancer diagnoses and also help doctors save time to focus more on caring their patients.