PV QA 3 - Poster Viewing Q&A 3
Purpose/Objective(s): We investigated the verification accuracy of a novel patient positioning system with four X-ray tubes and flat panel detectors which was in first clinical use in world by comparing it with a kV-MV imager (2D-2D, 2D-3D, and CBCT) attached to the linear accelerator.
Materials/Methods: In May 2017, we started using the novel system and had recorded the shifts of the coordinates between the linear accelerator setups and the novel system-based setups in each translational directions (AP, SI, LR) and rotational directions (rotation, pitch, roll) until Jan 2018. The number of patients and image acquisition of linear accelerator were (A) 395 CBCT images for 20 patients of prostate cancer, (B) 73 CBCT images for 7 patients of brain tumor, (C) 231 2D-2D images for 16 patients and (D) 165 2D-3D images for 24 patients of pelvic tumor. The Spearman’s rank correlation coefficient was calculated in couch shifts between the CBCT or the kV-MV, and the novel system. The difference between the measured couch shifts was then plotted against the average couch shift using Bland-Altman analysis to assess the fixed bias.
Results: Spearman’s rank correlation coefficients were 0.96, 0.98, 0.96, 0.89, 0.76 and 0.87 in AP, SI, LR, rotation, pitch and roll, respectively between novel system and CBCT for prostate cancers, 0.91, 0.98, 0.95, 0.90, 0.95 and 0.94 between novel system and CBCT for brain tumors, and 0.85, 0.94, 0.97, 0.79, 0.88 and 0.88 between novel system and 2D-3D for pelvic tumors. The localization accuracy of novel system was very strongly correlated with that of CBCT and 2D-3D. In contrast, the correlation coefficients between the novel system and 2D-2D for pelvic tumors, it is thought most common, were 0.322, 0.689, 0.823, 0.474, 0.314 and 0.084 in AP, SI, LR, rotation, pitch and roll, respectively. They were less correlated than others. According to a Bland-Altman analysis, the 95% CI were (0.005, 0.024) in AP, (0.016, 0.028) in SI, (-0.026, 0.001) in LR, (-0.150, 0.032) in rotation, (-0.410, -0.283) in pitch, and (0.019, 0.211) in roll between the novel system and CBCT for brain tumors, indicating that there were fixed biases except for LR and rotation. There were no fixed biases between the novel system and 2D-2D, 2D-3D or CBCT in pelvic tumors.
Conclusion: Localization accuracy of this novel system was very strongly correlated with that of 2D-3D or CBCT system of linear accelerator. But the correlation between this system and 2D-2D was not confirmed. This implied that this novel system could be useful in bony structure based patient positioning even in radiation therapy that requires high accuracy of patient positioning such as stereotactic radiotherapy.
The asset you are trying to access is locked. Please enter your access key to unlock.