Automatic Parotid Gland Segmentation in MVCT Using Deep Convolutional Neural Networks

Radiation-induced xerostomia, as a major problem in radiation treatment of the head and neck cancer, is mainly due to the overdose irradiation injury to the parotid glands. Helical Tomotherapy-based megavoltage computed tomography (MVCT) imaging during the Tomotherapy treatment can be applied to monitor the successive variations in the parotid glands. While manual segmentation is time consuming, laborious, and subjective, automatic segmentation is quite challenging due to the complicated anatomical environment of head and neck as well as noises in MVCT images. In this article, we propose a localization-refinement scheme to segment the parotid gland in MVCT. After data pre-processing we use mask region convolutional neural network (Mask R-CNN) in the localization stage after data pre-processing, and design a modified U-Net in the following fine segmentation stage. To the best of our knowledge, this study is a pioneering work of deep learning on MVCT segmentation. Comprehensive experiments based on different data distribution of head and neck MVCTs and different segmentation models have demonstrated the superiority of our approach in terms of accuracy, effectiveness, flexibility, and practicability. Our method can be adopted as a powerful tool for radiation-induced injury studies, where accurate organ segmentation is crucial.

Evaluation and improvement the safety of total marrow irradiation with helical tomotherapy using repeat failure mode and effects analysis

Background & purpose Helical tomotherapy has been applied to total marrow irradiation (HT-TMI). Our objective was to apply failure mode and effects analysis (FMEA) two times separated by 1 year to evaluate and improve the safety of HT-TMI. Materials and methods A multidisciplinary team was created. FMEA consists of 4 main steps: (1) Creation of a process map; (2) Identification of all potential failure mode (FM) in the process; (3) Evaluation of the occurrence (O), detectability (D) and severity of impact (S) of each FM according to a scoring criteria (1–10), with the subsequent calculation of the risk priority number (RPN=O*D*S) and (4) Identification of the feasible and effective quality control (QC) methods for the highest risks. A second FMEA was performed for the high-risk FMs based on the same risk analysis team in 1 year later. Results A total of 39 subprocesses and 122 FMs were derived. First time RPN ranged from 3 to 264.3. Twenty-five FMs were defined as being high-risk, with the top 5 FMs (first RPN/ second RPN): (1) treatment couch movement failure (264.3/102.8); (2) section plan dose junction error in delivery (236.7/110.4); (3) setup check by megavoltage computed tomography (MVCT) failure (216.8/94.6); (4) patient immobilization error (212.5/90.2) and (5) treatment interruption (204.8/134.2). A total of 20 staff members participated in the study. The second RPN value of the top 5 high-risk FMs were all decreased. Conclusion QC interventions were implemented based on the FMEA results. HT-TMI specific treatment couch tests; the arms immobilization methods and strategy of section plan dose junction in delivery were proved to be effective in the improvement of the safety.

Testy kontroli jakości obrazowania megawoltowego akceleratora TomoTherapy oraz ich ocena w oprogramowaniu ARTISCAN do półautomatycznej analizy obrazów

Stosowanie megawoltowej tomografii komputerowej (Megavoltage Computed Tomography - MVCT) w aparatach TomoTherapy wymaga testowania poprawności działania systemu detekcyjnego. Prawidłowa praca detektora warunkuje precyzyjne ułożenie chorego na stole terapeutycznym oraz dokładne dostarczenie zaplanowanej dawki do napromienianego obszaru poprzez precyzyjne odwzorowanie kształtu planowanej objętości terapeutycznej (Planning Target Volume - PTV) oraz narządów krytycznych. Kontrola jakości (Quality Assurance - QA) pracy detektora jest zatem niezbędna dla zapewnienia poprawności realizacji planów leczenia. Do analizy danych otrzymanych z testów coraz częściej wykorzystuje się programy komputerowe. Celem niniejszej pracy było uzasadnienie wdrożenia oprogramowania ARTISCAN do półautomatycznej analizy parametrów MVCT oraz wskazanie sposobu wyznaczenia tolerancji i ich ustalenie dla wybranych własności fizycznych opisujących jakość obrazów tomograficznych oraz określenie ich precyzji.