Room: ePoster Forums
Purpose: To assess the dosimetric impact of inter-observer variation in localization of ring and tandem applicators in high-dose rate brachytherapy magnetic resonance planning images.
Methods: Five cervical cancer HDR brachytherapy cases treated with ring and tandem applicator and two interstitial catheters were studied. Applicator reconstruction on MR images was performed by four experts. Variability of applicator placement was evaluated by measuring mean distance between corresponding applicator points in the original clinical plan and that placed by each observer. Displacement and rotation of each applicator relative to the original was measured using rigid registration. Target and OAR EQDâ‚‚ parameters were determined using the clinical treatment plan parameters applied onto each reconstructed model.
Results: The mean distance between applicator points in the clinical case and those applied by different experts ranged 0.1-2.1 mm. Maximum translational parameters in Left-Right, Ant-Post, and Sup-Inf directions from rigid registration were 3.0, 2.8, and 2.0 mm, respectively. Pitch, roll, and yaw rotation angles were 0.0-1.4, 0.1-4.7, and 0.1-7.5 degrees, respectively. Needle tips were on average at a distance of 1.1Â±0.6 mm from clinical case. Systemic differences in needle registration were observed for two observers. Target and OAR EQDâ‚‚ parameters for the clinical plan were on average 0.3-3.3 Gy different from those re-calculated on the four independent reconstructions. Standard deviation of EQDâ‚‚ values across five datasets was 0.3-2.6 Gy. D98 for GTV, high-risk-CTV, and intermediate-risk-CTV were at most 7.0, 5.4 and 3.0 Gy lower than the clinical plan, respectively. D2cc for OARs increased a maximum of 3.9 Gy for the bladder relative to the clinical plan.
Conclusion: Variation of applicator and catheter reconstruction among four experts demonstrated small discrepancies within the image resolution. The dosimetric impacts for target and OAR were within acceptable tolerance. However, better image quality and more standardized protocols are recommended based on this study.