Quantitative Radiology Solutions (QRS) has entered into the next phase of the clinical evaluation of its Automatic Anatomy Recognition (AAR) software – start of clinical acceptability and efficiency evaluation. Four academic medical centers have enrolled 214 studies into the evaluation whose endpoints will verify the accuracy, efficiency, and clinical acceptability of the AAR-generated contours for radiation treatment planning.
AAR uses deep learning based artificial intelligence algorithms to automatically identify and delineate (contour) anatomical structures that might be damaged by excessive radiation exposure. It enables physicians to increase the accuracy of treatment planning while reducing the time necessary to complete a plan. Radiation oncologists from the four participating sites will rate the acceptability of AAR-generated contours for use in treatment planning, and clinicians at those sites will document productivity improvements resulting from automatic contouring.
“AAR produces contours that are consistent with those published in radiation oncology guidelines,” said Gargi Pednekar, QRS Product Manager. “The algorithm routinely delivers accurate results even for difficult anatomical structures and in studies where image quality is compromised.”
Results of the evaluation are expected in October, 2020.