Software is an essential component of radiotherapy clinic, which supports numerous functions ranging from image processing to treatment planning. We have been developing novel software to translate in-house developed computational algorithms and tools to routine clinical practice.
One example is the AutoBrachy system to support treatment planning and quality assurance (QA) of high dose rate brachytherapy (HDRBT) for gynecological cancer. HDRBT treatment planning is conventionally performed manually in a high time-pressure environment, suffering from limitations such as being time-consuming, error-prone, and sub-optimal and variable plan quality depending on the planner’s experience. Widespread use of HDRBT is limited by the complexity of the treatment planning process. To address these problems, we developed the AutoBrachy system and implemented the system at UTSW clinic since 2015, aiming at a streamlined fully-automated planning process. The system applies image processing techniques to CT images to segment the applicator and place dwell positions. Source dwell times are determined via inverse optimization. The final plan is imported to the clinical treatment planning system and then manually inspected and adjusted as needed. After the plan is approved, the QA module of the AutoBrachy system performs a comprehensive check on its geometric and dosimetric components. A report is generated for manual inspection, which highlights potential issues requiring human inspection.
To enhance the effectiveness of our system, we have applied deep-learning techniques for several individual steps of the treatment planning process, such as interstitial needle segmentation, organ segmentation, and human-like inverse treatment planning. More details about our deep-learning work can be found here. It is also our on-going process to continuously develop novel modules in AutoBrachy to improve its functionality and to translate them into routine clinic to improve HDRBT healthcare.
- Hyunuk Jung, Yesenia Gonzalez, Chenyang Shen, Peter Klages, Kevin Albuquerque, Xun Jia, "Deep learning assisted automatic digitization of applicators in 3D CT image based high dose rate brachytherapy of gynecological cancer," to appear in Brachytherapy (2019).
- Chenyang Shen, Yesenia Gonzalez, Peter Klages, Nan Qin, Hyunuk Jung, Liyuan Chen, Dan Nguyen, Steve B. Jiang, Xun Jia, “Intelligent Inverse Treatment Planning via Deep Reinforcement Learning, a Proof-of-Principle Study in High Dose-rate Brachytherapy for Cervical Cancer,” Phys. Med. Biol., 64, 115013 (2019).
- Yuhong Zhou, Peter Klages, Jun Tan, Yujie Chi, Strahinja Stojadinovic, Ming Yang, BrianHrycushko, Paul Medin, Arnold Pompos, Steve Jiang, Kevin Albuquerque,Xun Jia, “Auto-mated High-dose Rate Brachytherapy Treatment Planning for a Single-channel Vaginal CylinderApplicator,”Phys. Med. Biol., 62, 4361, (2017). Special issue on ”Recent Progress in Applica-tions of Computing to Radiotherapy.”
- “Next generation high dose-rate brachytherapy treatment planning for gynecological cancer,” Varian Master Research Agreement, 12/2015 - 11/2018.