NYU Langone Validates Automatic Segmentation For Clinical Adoption
NYU Langone was ready to adopt an AI workflow in favor of manual methods, which are time-consuming and prone to inter-observer variability. According to physicist and lead author Noah Bice, “Contouring is subjective. With any mix of individuals, there are inherently varying levels of expertise and personal preference involved.” But for all the promises of workflow improvement, how does the quality of the outputs measure up?
A team of three certified medical dosimetrists with 41 years of combined experience scored the quality of each individual contour on a scale from 1-5. A 1 score indicated the output required no edits before planning; a score of 3 indicated that the structure required some modifications but that no time was saved or lost using the software, including all edits. A score of 5 indicates the contour was unusable.
Download the case study to see their results.
“'[The dosimetrists are] getting contours through much faster and with less effort now with AutoContour. Overall, they’re much more efficient. It’s been a huge time saver.”
— David Barbee, Medical Physicist at NYU
Reduced Contouring Time
Minimal Installation Overhead