Technology

Reliable Deep Computation Using Machine Learning

ClearRead is a modern approach utilizing the latest advances in machine learning, such as deep learning. ClearRead has surpassed the state-of-the-art by a significant margin based on a combination of frameworks, modeling, and computational ingenuity.


Acquisition Independence

ClearRead handles a broad range of acquisition protocols, a difficult problem for automatic analysis algorithms. Riverain Technologies developed adaptive algorithms, so each scan is normalized for factors
such as:
• Noise
• Reconstruction kernels
• Slice sampling effects

Conventional approaches collect data from different sensors to adjust component algorithms. This leaves them vulnerable to changes in hardware, protocols, and reconstruction methods.
Our adaptive process allows our software to be vendor neutral. ClearRead provides enterprise imaging without compromise, while also enabling fast and simple installation.

The Riverain Technologies Difference

The standard approach to building large, complex models requires large measured training sets. These high-quality medical data sets are both time consuming and expensive, to collect. Many cases look similar, and do not include rare cases.

Riverain developed the capability to create synthetic nodules automatically and place them into relevant anatomical contexts – such as next to the pleura wall or attached to a vessel. ClearRead was built on thousands of simulated, diverse nodules. By doing this, our software has been trained on more complete cases (including more rare cases), and tested on full training sets.

Slice showing simulated nodules
Creating Adequate Amounts of Data Through Simulation

Building large complex models, such as deep neural networks, requires large training sets. Collecting large, high-quality medical datasets is both time-consuming and expensive, making it impractical in any realistic sense. To circumvent this problem, Riverain developed the capability to create synthetic nodules automatically, and place them into relevant anatomical contexts – such as next to the pleura wall or attached to a vessel.

Creating synthetic nodules was an essential capability to build vessel suppression. Vessel suppression, and other algorithms within ClearRead CT, were developed on thousands of simulated nodules. The simulated nodules are instrumental in realizing a robust clinical solution. The figure above illustrates a slice with several simulated nodules embedded.

ClearRead CT Vessel Suppressed Chip
Reliable Quantification and Unique Access to Clinically Important Quantities

Vessel suppression enables improved nodule detection, but its benefit continues throughout the processing chain. Suppressing vessels and surrounding structures allows for highly precise segmentation of nodules; which, provides an accurate assessment of size, volume, and other general nodule characteristics.

Vessel suppression aids the radiologist in interpretation through the removal of vascular structure within ground glass nodules, critical to the assessment of density and tissue composition. This allows a radiologist – or analysis software – to rapidly determine the relative amount of solid tissue, which is a critical aspect for clinical decision making. The figure above shows a zoomed-in chip of a ground glass nodule with and without vessel suppression.

xray machine
Reliable Deep Computation Using Machine Learning

ClearRead CT is a modern approach utilizing the latest advances in machine learning, such as deep learning. ClearRead CT has surpassed the state-of-the-art by a significant margin based on a combination of frameworks, modeling, and computational ingenuity.

Achieving high reliability and significant performance usually uses substantial amounts of processing. ClearRead CT system can run on commodity hardware, without the need for special computer cards (GPUs) or large memory systems.