Sketch-Based Interactive Segmentation and Segmentation Editing for Oncological Therapy Monitoring

Cancer is one of the most common causes of death. Worldwide about 14 million new cancer cases and 8.2 million cancer deaths were reported for 2012 [1]. Even though minimally invasive treatment methods like radiofrequency ablation (RFA) become more and more popular, chemotherapy remains an important systemic treatment method, especially for metastatic cancer. An important criterion for assessing a tumor’s response to chemotherapy is its change in size. The current gold standard for measuring the size of solid tumors and their progression over time is the maximum diameter as defined in the RECIST guideline (Response Evaluation Criteria In Solid Tumors) [2]. A diameter is not always appropriate, though, since most tumors grow and shrink irregularly in 3D [3]. Therefore, the volume of a tumor has shown to be a more accurate and reproducible measure for tumor size in various studies (e.g., [4]). To measure a tumor’s volume based on computed tomography (CT), it needs to be delineated (segmented) in the 3D image. Segmentation in medical imaging is an intensively studied field and many algorithms have been developed during the past decades. For a long time, research has aimed at automatic segmentation methods. In recent years, however, interactive methods have grown in interest due to limitations of automatic methods in clinical routine. In real-world applications, control over the segmentation result is an essential requirement [5,6]. However, as CT images suffer from partial volume effects caused by their limited spatial resolution, the delineation of a tumor is an ill-defined problem. Particularly for interactive methods, this can result in significant differences in the volume measured by different readers or at different points of time. This thesis tackled the problems of interactive segmentation, the editing of a given segmentation result, and the accurate and reproducible measurement of a tumor’s volume based on a given segmentation mask in the context of oncological therapy response assessment. Each contribution is summarized in the following sections.