A Google algorithm has been shown to be able to distinguish between the two most common cancers with an accuracy of 97 percent, according to a study conducted by researchers at the University of New York published in Nature Medicine.
Inception v3 is an open source algorithm launched by the technological giant that was trained with the images of cancerous and healthy tissue from The Cancer Genome Atlas.
The deep learning tool distinguished cancer cells from healthy ones 99 percent of the time. After this stage, he was taught to differentiate between adenocarcinoma and squamous cell carcinoma.
Testing the algorithm with independent samples of patients with cancer, accuracy decreased but, still, correctly diagnosed between 83 and 97 percent of cases.
Researchers at the University of New York will continue to train the system with data from a greater number of sources. After that, they will seek approval from the FDA to standardize the sequencing of tumor samples in the United States.