The group, led by Dr. Savvas Nicolaou from Vancouver General Hospital, assessed the potential value of using cinematic rendering instead of traditional volume rendering to generate 3D images of patient anatomy. Unlike the single raycasting model of volume rendering, cinematic rendering makes use of a global illumination model to enhance the photorealism of medical images.
Nicolaou and colleagues processed an imaging dataset from several patients who presented with multiple types of trauma and who underwent CT exams at their institution's level I trauma center. They used computer software (syngo.via Frontier, Siemens Healthineers) to convert the CT data into cinematically rendered images.
Upon evaluating the resulting 3D reconstructions, the researchers found that the cinematically rendered images provided exquisite anatomical details of acute injuries -- allowing for simple delineation of bony, vascular, and soft tissue.
A team of trauma surgeons also assessed the 3D reconstructions and reported that the cinematically rendered images were much more useful for helping them make a clinical decision and educating trainees than were the conventional volume-rendered images.
"Cinematic rendering is a promising novel technique to display visually receptive 3D photorealistic high-definition images ... [and] provides remarkable details relative to [volume-rendered] reconstructions in context of complex acute trauma," the researchers concluded.
This paper received a Roadie 2018 award for the most popular abstract by page views in this Road to RSNA section.
