Segmenting MRI scans is a crucial aspect of any neuroimaging workflow, with manual segmentation widely considered to be the most accurate and efficient technique currently available.
NEW YORK, August 6, 2018 (Newswire.com) - In recent years, clinicians have started turning toward automated segmentation in an effort to increase efficiency and consistency. According to Dov Bechhofer, however, all currently available automatic segmentation software is notoriously imprecise, and as a result, still requires that any output is manually checked and corrected.
Looking to establish a more efficient way to correct these errors, work has now begun on utilizing virtual reality technology to expedite the process.
Early studies have shown that when completed in a virtual reality environment, users are able to complete the task of correcting segmentation errors within MRI scans in a shorter time and with higher rates of accuracy. "This is compared with using conventional 2D interfaces," explains Bechhofer.
One such study tested the viability of virtual reality by enlisting 30 individuals to correct an error on a previously segmented MRI scan. "Half of the participants used the traditional 2D interface, while the other half used a prototype 3D, virtual reality system," says Dov Bechhofer.
The two groups were then switched and tasked with completing the process once again, this time using the alternate solution.
"The task," the radiologist continues, "was to identify and eliminate voxels representing the optic nerve from a sample MRI scan, which is something that's commonly missed during automated segmentation."
When using the traditional 2D interface, most individuals were around one minute slower than when using the prototype 3D system. "The new technology allowed for the task to be completed, on average, 68 seconds faster than when using the current solution," Dov points out. "This, we understand, is primarily owed to the greater precision afforded by the virtual reality environment."
According to Dov Bechhofer, a follow-up survey revealed that all 30 participants stated that they preferred the virtual reality environment over the current technique and 2D interface. "Most suggested that using the virtual reality software was more engaging, while around a third also claimed that the interface was more intuitive than the existing 2D alternative," he adds.
Research and development into several virtual reality MRI segmentation software solutions and interfaces are ongoing. "One example currently in the pipeline will include sound effects and feedback, as well as real-time accuracy tracking systems," Bechhofer explains. "Inbuilt training modes are also likely to be incorporated across future versions which will allow users to refine their technique via a library of sample data and scans."
"I see great potential," adds the New York City-based radiologist, wrapping up, "and I look forward to seeing virtual reality utilized more widely in all manner of clinical settings in years to come."
Web Presence, LLC
Source: Web Presence LLC