MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

SSIM Index for the Quantitative Evaluation of Radiotherapy Dose Distribution

C Shi*, S Lim , M Chan , Memorial Sloan Kettering Cancer Center, Basking Ridge, NJ

Presentations

(Saturday, 4/7/2018)  

Room: Foyer

Purpose: To evaluate using SSIM index into plan evaluation and dose map comparison with various QA devices in radiotherapy.

Methods: Structural Similarity Index (SSIM), which measures the degree of similarity of an image to a defined reference image, is defined as the product of image luminance (L), contrast (C), and structure similarity (S) terms. In radiotherapy, luminance, contrast and structural similarity are defined as dose deposition, device specific characteristic dosimetric noise, and the patient specific two dimensional dose distribution patterns respectively. A set of IMRT QA distributions, measured with Varian EPID imager, Sun Nuclear MapCheck, and IBA Dolphin, is compared with the reference calculation generated by Varian Eclipse treatment planning System using SSIM. The same measurements are also compared with the reference distribution using gamma index. The exponent terms of L, C, S are varied systematic to assess the sensitivity of SSIM. A test pattern with 1 mm distance shift and 2% dose difference has been benchmarked using SSIM index and gamma index respectively for the following three scenarios: (1) distance shift 1 mm alone; (2) dose difference 2% alone; and (3) background 2% alone. SSIM indexes have been calculated to show the comparison of those images, as compared to those used with gamma index.

Results: The test pattern shows that SSIM index could detect the minor changes in distance, intensity, and contrast rather than gamma index. The comparison also indicates different qualities of the dose map from different QA devices. Though a clinically relevant tolerance for the SSIM index has not yet been determined, the results show that the SSIM index is more sensitive to image/fluence/dose from different QA modalities.

Conclusion: We have investigated a new evaluation metric for dose distribution and device comparison - SSIM index. SSIM index has the potential to be implemented into radiotherapy QA.

ePosters

Contact Email