Published: Wed, October 17, 2018
Tech | By

New DeepMind AI-Augmented Cancer Detection Study Detailed

New DeepMind AI-Augmented Cancer Detection Study Detailed

USA tech giant Google has created a tool that could potentially help doctors to diagnose breast cancer using artificial intelligence (AI).

The new Google AI tool is called the LYmph Node Assistant, or LYNA. Pathologists study the tissue samples from breast cancer patients' lymph nodes to confirm how much the tumor has spread and how unsafe it can be. Such tumors are known to be hard to detect.

Cancer, a disease caused by the growth of abnormal cells beyond their usual boundaries, is the second leading cause of death worldwide, according to the World Health Organization. Finding them is a time-intensive and hard task for pathologists. The papers state that the Google AI tool has an accuracy rate of 99% when it comes to identifying which slides show metastatic cancer. Further, LYNA was able to accurately pinpoint the location of both cancers and other suspicious regions within each slide, some of which were too small to be consistently detected by pathologists. "These techniques may improve the pathologist's productivity and reduce the number of false negatives associated with morphologic detection of tumor cells".

LYNA wasn't flawless - it occasionally misidentified giant cells, germinal cancers, and bone marrow-derived white blood cells known as histiocytes - but managed to perform better than a pathologist. The results were remarkable when LYNA worked as a companion to pathologists. In fact, the company says it cut the review time by around half for each slide used.


The scientific report published on this issue explains that: "Artificial Intelligence Algorithms can exhaustively evaluate each occurrence, individually and in detail". This deep learning model achieved a 78.1% accuracy rate on Stanford's ImageNet data set. As the investigators explained, it examines a 299-pixel image (Inception-v3 default image size), then analyzes volumes at the pixel level and then exports them as labels - that is, predictions and classifies them as benign or malignant.

Introduced past year, the LYmph Node Assistant (LYNA) algorithm was trained to recognize characteristics of tumors in metastatic cancer. It also has yet to be used in real-life clinical situations.

When and if LYNA will become ready for practical use, the benefits of the AI would be huge - not only it would allow doctors more to time care for their patients but it would also lead to more reliable and swift diagnoses that could save hundreds if not thousands of lives.

Luckily, Google AI and researchers at the Naval Medical Center San Diego have a solution.

Like this: