Building a Better Chest X-Ray

By medGadget

Using machine-learning algorithms and pattern recognition, Riverain's Temporal Comparison software can spot early lung cancer by looking at and comparing chest X-rays over time.

ChestXRays-Post.jpg

Riverain Technologies has received FDA clearance for its Temporal Comparison software for chest X-rays. The software compares current and previous chest X-rays of the same patient and highlights the differences in order to improve detection of new nodules which may be early lung cancer.

The software uses pattern recognition and machine-learning algorithms (including its existing bone suppression algorithms) to normalize each image to make it consistent with a standard format. The lung field is then identified for comparison. The current and prior image are then aligned and processed by subtracting the current image from the prior image. The system integrates seamlessly with the PACS system, saving the resulting difference image as an additional image in the patient's folder ready for the radiologist's interpretation.

In a study performed by the company as part of the FDA approval process, on 422 pairs of chest X-rays, 15 radiologists were able to demonstrate an average 12.4 percent improvement in sensitivity for actionable solitary pulmonary nodules using Riverain's Temporal Comparison software as compared to results based on comparing the X-rays side by side.

In addition to the United States, Temporal Comparison is also available in Canada and Europe. It requires the customer to have purchased Riverain's bone suppression software SoftView, functioning as an add-on to it. The software can also be used in combination with Riverain's On-Guard Computer-Aided Detection software to improve early detection of lung disease.


This post also appears on medGadget, an Atlantic partner site.

This article available online at:

http://www.theatlantic.com/health/archive/2012/01/building-a-better-chest-x-ray/251413/