A Fully Convolutional Network (FCN) was developed and applied to the task of contouring esophageal tumors on diffusion weighted images. After proper training, tumor classification by the FCN demonstrated excellent agreement with tumor contours from an inter-reader agreement study in the validation images. The FCN was able to achieve correct tumor classification in most cases with respect to different tumor position and shapes, and in the presence of intratumoral esophageal lumen.
This abstract and the presentation materials are available to members only; a login is required.