การประเมินการลดสิ่งแปลกปลอมโลหะโดยใช้วิธีการประมาณค่าแบบผสม สำหรับการรักษามะเร็งศีรษะและลำคอ ด้วยเทคนิคการฉายรังสีแบบปรับความเข้มลำรังสี (ศึกษาในหุ่นจำลอง)
Year: 2020
Type: article
Abstract: Backgrounds: Radiation therapy for head and neck cancer treatment requires an accurate treatment plan to obtain a consistent dose and avoid the surrounding organs. Dental prosthetics or other artifacts, commonly found in computed tomography simulation images of the head and neck, produce metal artifact. Metal artifact obstructs the images of the surrounding organs and tissues. These images are unclear and cannot be effectively used in treatment planning, however, the artifacts can be reduced by an algorithm. Objective: To develop a method for reducing metal artifacts in computed tomography simulation images of the head and neck with dental prosthetics before planning treatment with Intensity Modulated Radiation Therapy (IMRT) Materials and methods: The efficiency of these three methods of reducing metal artifacts, (linear interpolation method (Li), total variation inpainting method (TV) and hybrid interpolation method (Hybrid)) were compared on the head and neck phantom images of an inserted metal screw and two metal screws. The time of the metal artifact reduction, line profiles, image data values and the shape of metal screws of both initial and reduced artifact images were evaluated. Results: The average times of the hybrid interpolation method were 59.93 seconds and 148.71 seconds for a metal screw and two metal screws, respectively. The evaluated quality of the line profile, image data value and shape of the metal screws were similar to the TV method. Moreover, the hybrid method reduced metal artifact on the head and neck phantom images of two metal screws better than Li method. Conclusion: Hybrid interpolation used less time than TV methods. The evaluated quality of the line profile, image data value and shape of the metal screws were similar to the TV method.
Language: Thai
Subfield: Biomedical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7219251019