Segmentation and Registration Based Automatic Cancer Proton Treatment Analysis
Abstract
Due to its low side effects, proton therapy is rapidly developing around the world. However, the time and labor it takes to deliver the treatment have prevented widespread use in the clinic setting. For this reason, an automatic proton treatment analysis system is needed to improve efficiency during the treatment process. The challenge lies in improving the accuracy and speed of the current proton treatment analysis system. The Range shifter correction factor is considered the same value under any given condition in our treatment system. This approximate algorithm limits the accuracy of the proton treatment analysis system. In addition, medical professionals spend a great deal of time checking for errors. Errors, such as contouring mistakes and unnecessary spots are manually reviewed during the treatment plans analysis process. At present, the dose calculation engine in the analysis system requires five hours to complete one treatment plan. The calculation speed limits our treatment capacity. Because the status of a tumor changes over the duration of the treatment process, monitoring its status is imperative for delivering an accurate dosage. Our research involves the development of an automatic proton treatment analysis system that uses methods based on segmentation and registration algorithms to solve these problems. The automatic proton treatment analysis system (and the accuracy and speed at which it operates) has been validated in a clinical setting. Therefore, this analysis system could potentially replace manual operation during the proton treatment plan quality assurance process. The amount of time it takes to deliver proton therapy will be significantly reduced.
Subject Area
Biomedical engineering|Oncology|Public health
Recommended Citation
Zhang, Yang, "Segmentation and Registration Based Automatic Cancer Proton Treatment Analysis" (2019). ETD Collection for University of Texas, El Paso. AAI10281194.
https://scholarworks.utep.edu/dissertations/AAI10281194