It’s only the size of a fist, but the human heart has 17 recognized segments used for classification, diagnosis and treatment. Before the advent of advanced cardiac MRI, each patient’s heart had to be segmented manually, eating up precious time for radiologists.

However, a wide array of automated MR segmentation techniques emerged in the new millennium, promising a new era of speed and accuracy. According to Arterys, a San Francisco-based startup promising to provide AI-driven cardiac segmentation, 2017 could represent yet another phase shift.

In what they are calling the first use of FDA-cleared, zero-footprint use of cloud computation in a clinical setting, Arterys’ Cardiac DL program uses a data set of over 3,000 patients to contour heart segments, providing valuable diagnostic information about heart function and size. The software received 510(k) approval in November.

"Arterys is committed to broadly accelerate the transformation of data-driven medicine through advanced cloud medical imaging analytics applications while always protecting patient data privacy," said Fabien Beckers, Arterys CEO. "Today's approval signifies a major regulatory milestone. This application demonstrates the power of deep learning combined with cloud supercomputation to aid physicians in interpreting medical images - eliminating tedious manual tasks carried out on a workstation by accurately automating those processes."

Arterys raised $12M during its Series A financing round in March, and Cardiac Dl became commercially available in late 2016. Primary investors include GE Ventures, Stanford Start-X, and Emergent. Other FDA-approved products created by Arterys include Visoworks, a cardiac MR suite designed to capture 2D, 3D, blood flow perfusion and function data.