Program Information
Application of Taguchi Methodology for Optimizing Parameters of Cone Beam CT for Optimal Image Quality
s alani1*, A Schlocker2 , N Honig3 , G Elikhis4 , M Levita5 , (1) Tel Aviv Sourasky Medical center, Tel Aviv, Isael, (2) Tel Aviv Medical Center, Modiin, ,(3) Tel Aviv medical center, Tel Aviv, Israel, (4) Tel Aviv medical center, Tel Aviv, ISRAEL, (5) Tel Aviv Medical Center, Tel Aviv, Israel
Presentations
PO-BPC-Exhibit Hall-20 (Saturday, March 5, 2016) Room: Exhibit Hall
Purpose:
Taguchi Method is a statistical approach to optimizing multiple process parameters to determine a combination resulting in the best output. This study utilizes the Taguchi Method to evaluate the image quality parameters of Varian OBI v2 CBCT for adaptive radiotherapy image sets.The CATPHAN 600 CT phantom, which includes several modules for measuring image quality indices for imaging protocols’ various parameters, was utilized.
Methods:
Taguchi developed a unique mathematical design of orthogonal arrays to study an entire parameter space with few experiments. The experimental results are transformed into what Taguchi referred to as signal-to-noise ratio (t-S/N) that is a measure of quality characteristics deviating from or nearing to a desired value. Analyses of variance (ANOVA) were employed to study the performance characteristics of the CBCT image quality. In this analysis, three factors of the acquisition protocol were considered: (a) CT slice thickness; (b) image reconstruction filter types; (c) matrix size. Respectively, a suitable orthogonal array was selected (3x3x3) to perform the experiments. After conducting the experiments the image quality parameters: pixel value stability, noise, contrast-to-noise ratio, and uniformity were measured, and t-S/N for each parameter was calculated.
Results:
The quantified t-S/N values analyzed in JMP software. The optimal combination of three factors was determined to be comprised of (a) 1 mm CT slice thickness; (b) “sharp” filter; (c) 512x512 matrix. The dominant factors influencing image quality are (b) filter type, and the cross interaction between the filter type and the CT slice thickness (aXb). Matrix size (c) plays a relatively minor role since it was incapable of yielding a 95% confidence level in the ANOVA test.
Conclusion:
The Taguchi Method provides simple, systematic, and efficient methodology for optimizing the process parameters for evaluating image quality for CBCT studysets and may allow us to improve adaptive radiotherapy as it matures.
Contact Email: