藉由田口設計法最佳化SPECT之影像品質

Optimization the SPECT imaging quality via the Taguchi methodology

黃亭硯(Ting-yan Huang)

指導教授 ： 潘榕光

中臺科技大學/健康科學院/放射科學研究所/碩士(2009年)

https://doi.org/10.6822/CTUST.2009.00004

摘要

以田口的穩健設計方式，對單光子電腦斷層(SPECT)影像造影參數作最佳化的調整。造影影像品質優劣可以以影像空間解析(spatial resolution)評估，唯空間解析會受到許多造影參數設定所影響。在此使用Siemens E-Cam單光子電腦斷層造影儀器，配合五個造影參數分別以不同之水準設定十八種組合，依照田口直交表方法排序，擷取Jaszczak phantom (Cold study)和Line sources (Hot study)影像。五個造影參數因子分別為Scan mode、Collimator type、Ｍatrix size、Zoom和Frame time。S/N( )值明確的定義出單光子電腦斷層在做資料擷取時擁有最佳空間解析度的參數設定。經由田口方式依據評分的實驗數據做參數組合的最佳化之調整，再利用ANOVA分析實驗結果並以F-test做檢定，驗證田口的實驗設計法的可信度。田口方法找出冷區影像與熱區影像的最佳造影參數組合分別為(1) Continuous scan mode, LEHR collimator, 128x128 matrix size, Zoom = 1.23, Frame time = 40 sec/frame和(2) Continuous scan mode, LEHR collimator, 256x256 matrix size, Zoom = 1.45, Frame time = 30 sec/frame。影響單光子電腦斷層影像擷取的主要控制參數為matrix size，調整參數為collimator type和frame time，皆達100%之信心水準。與主要控制參數的交互作用圖顯示matrix size與其他參數皆無強烈的的交互作用。與其他文獻相比，最佳化影像造影參數設定下的影像品質確實有較佳的表現。

關鍵字

最佳化 ； 田口方法 ； 空間解析 ； 單光子電腦斷層造影儀

並列摘要

This work demonstrated the optimization process for Single Photon Emission Computed Tomography (SPECT) image using the Taguchi methodology. The spatial resolution was a quality characteristics for the SPECT and it can be affected by many factors. The Siemens’s E-cam SPECT system was used for acquiring the Jaszczak phantom (Cold study) and line sources (Hot study) image herein. A total of eighteen combinations of five parameters were arranged. in the manner suggested by Taguchi. The five parameters were defined as Scan mode, Collimator type, Matrix size, Zoom and Frame time. The S/N( ) adopted herein was specifically defined to satisfy the requirements of good spatial resolution of the SPECT image acquisition in various parameters set. Moreover, a statistical analysis of variance (ANOVA) and F-test is utilized to determine which parameters are statistically significant. The optimized values in the Cold study and Hot study were (1) Continuous scan mode, LEHR collimator, 128x128 matrix size, Zoom = 1.23, 40 sec for frame time and (2) Continuous scan mode, LEHR collimator, 256x256 matrix size, Zoom = 1.45, 30 sec for frame time, respectively. Additionally, the parameters that dominated the SPECT image acquisition were matrix size. the adjust parameters were collimator type and frame time. confidence levels are exactly 100%. No strong cross interaction occurred between matrix size and other parameters. Compared with other papers, the optimization parameter setting for image quality of the SPECT has really a superior performance.

並列關鍵字

Taguchi methodology ； Optimization ； Spatial resolution ； Single-Photon-Emission-Computed-Tomography (SPEC

參考文獻

4. Powsner RA, Powsner ER. Essentials of Nuclear Medicine Physics. Malden, MA: Blackwell Science, 1998.

5. Sorenson JA, Phelps ME. Physics in Nuclear Medicine. 2nd ed. Philadelphia: WB Saunders, 1987.

8. Gopal B. Saha, Ph.D. Physics and Radiobiology of Nuclear Medicine. 2nd ed. Springer-Veriag New York, 2001.

9. Quality control of scintillation cameras. U.S. Dept of Health, Education and Welfare, HEW publications FDA 76-8046, 1976.

13. Ross PJ, Taguchi techniques for quality engineering Second ed. McGraw-Hill, 1996.