12 April 2017 | News | By BioSpectrum Bureau
Cancer can be diagnosed by Google's artificial intelligence
An artificial intelligence (AI) program owned by Alphabet, Google's parent company, may be able to diagnose cancer much faster. Google is working hard to tell the difference between healthy and cancerous tissue as well as discover if metastasis has occurred.
Metastasis detection is currently performed by pathologists reviewing large expanses of biological tissues. This process is labour intensive and error-prone. Google presents a framework to automatically detect and localise tumours as small as 100 Ã—100 pixels in gigapixel microscopy images sized 100,000Ã—100,000 pixels. This method leverages a convolutional neural network (CNN) architecture and obtains state-of-the-art results on the Camelyon16 dataset in the challenging lesion-level tumour detection task.
Such high-level image recognition was first developed for Google's driverless car program, in order to help the vehicles scan for road obstructions.
The software is a "deep learning" program, a rudimentary artificial intelligence application that is capable of drawing conclusions from a large set of data.
The technology is impressive but unlikely to replace human pathologists. Still, the Google team is optimistic that this method could improve accuracy and consistency of evaluating breast cancer cases, and potentially improve patient outcomes.