Deep Learning Triage

In the past few years, we witnessed a flood of deep learning (AI) algorithms aimed at dealing with medical imaging studies. Some of them are used for classification (i.e., identifying body parts or organs in scanned images) while others are used to highlight abnormalities in the images or used to specific structures like parts of an organ or tumor segmentation to be used by quantification and analysis applications. In an attempt to make all these “algorithms” more available to clinicians, we have also seen both frameworks and marketplaces introduced.

Introducing Deep Learning (AI) will bring great value to the medical and health industries and will change the way we do and receive medical analysis, treatment, and care. But, as of today, Deep Learning has yet to bring the amazing value it is expected to; it has yet to seamlessly help radiologists and physicians cope with the growing need to efficiently and accurately diagnose while still treating the most urgent patients.

We at Shina believe in Deep Learning Triage™, deep learning algorithms tightly integrated into clinical workflows. This is why we are working to augment our automatic, model-based segmentation and our semi-automatic tools with Deep Learning. Using Deep Learning to augment the workflow of our clinical analysis applications and to estimate a risk score will drive our future 3Di Web Patient Browser worklist prioritization.