Cancer is, in Slovenia as well as worldwide, a great public health problem, and among the key challenges of modern society in the contexts of health management and quality of life. In addition to surgery and chemotherapy, radiation and lately proton therapy are important and widespread cancer treatment options. In order to generate optimal radiation dose distributions, it is important to have a precise knowledge of the location and shape of the tumors as well as of nearby healthy organs at risk (OARs). For this purpose, three-dimensional (3D) images of the patient are acquired with the computed tomography (CT) and/or magnetic resonance (MR) imaging technique, and accurate and precise delineation of target volumes from such images represents the basis for the inverse planning of optimal radiation dose distribution. Our research hypothesis is that the delineation of target volumes in the head and neck area can be improved in the CT as well as MR image modality, and by exploiting the fused information from both imaging modalities, so as to reduce the subjectivity and time spent for manual delineation. Within the proposed research project we will therefore aim to (1) setup an annotated database of CT and MR images of the head and neck area with OARs precisely delineated by multiple observers on multiple occasions, (2) evaluate the generated image database from the perspective of intra-observer and inter-observer variability to generate reference delineation masks, (3) design and develop computer-assisted techniques based on state-of-the art technologies for automated segmentation of target volumes from 3D CT and/or MR images that are used for radiotherapy planning for the purpose of improving the accuracy and reliability of the resulting segmentation while reducing the time and effort required by the observer to verify or eventually modify the results, (4) design and setup a platform for an objective evaluation and comparison of the segmentation results by means of organizing a computational challenge, and (5) explore the possibilities of applying the developed methodology to related problems in the field of radiotherapy planning. It is expected that the results of the proposed research project will represent a break-through in the corresponding research field, and stimulate new research guidelines.
