Marius Røthe Arnesen1, Kjersti Bruheim2, Eirik Malinen3, Taran Paulsen Hellebust3. 1. Department of Medical Physics, Oslo University Hospital, Norway; Department of Physics, University of Oslo, Norway. Electronic address: mariua@ous-hf.no. 2. Department of Oncology, Oslo University Hospital, Norway. 3. Department of Medical Physics, Oslo University Hospital, Norway; Department of Physics, University of Oslo, Norway.
Abstract
BACKGROUND AND PURPOSE: This study aims to analyze subsections of the target volume that are sensitive to delineation uncertainties with respect to underdosage (spatial dosimetric uncertainty) in MRI-based brachytherapy of cervical cancer. MATERIAL AND METHODS: A methodology was developed to simulate delineation uncertainties by shifting an angular segment of the contour perpendicular to the original HR-CTV. For shifts of 3, 6 and 9mm resulting D90 and D98 were calculated for the modified contour. The sensitivity of the dose plan to the locally introduced error was estimated by linear regression of D90 or D98 against the magnitude of the shift. The methodology was employed on 20 patients treated with tandem ring brachytherapy. RESULTS: Topographic maps resulting from the dosimetric sensitivity analysis showed both large spatial variations and substantial inter-patient variations. For all plans included the spatial sensitivity in D90 ranged from 0.0 to -1.6%/mm, correspondingly sensitivity in D98 ranged from 0 to -4.6%/mm. A significantly increased dosimetric sensitivity was found in anterior direction and the cranial part of the tumor (p<0.05). CONCLUSIONS: The developed methodology identifies specific tumor regions and patients with increased risk of underdosage from delineation uncertainties in brachytherapy of cervical cancer.
BACKGROUND AND PURPOSE: This study aims to analyze subsections of the target volume that are sensitive to delineation uncertainties with respect to underdosage (spatial dosimetric uncertainty) in MRI-based brachytherapy of cervical cancer. MATERIAL AND METHODS: A methodology was developed to simulate delineation uncertainties by shifting an angular segment of the contour perpendicular to the original HR-CTV. For shifts of 3, 6 and 9mm resulting D90 and D98 were calculated for the modified contour. The sensitivity of the dose plan to the locally introduced error was estimated by linear regression of D90 or D98 against the magnitude of the shift. The methodology was employed on 20 patients treated with tandem ring brachytherapy. RESULTS: Topographic maps resulting from the dosimetric sensitivity analysis showed both large spatial variations and substantial inter-patient variations. For all plans included the spatial sensitivity in D90 ranged from 0.0 to -1.6%/mm, correspondingly sensitivity in D98 ranged from 0 to -4.6%/mm. A significantly increased dosimetric sensitivity was found in anterior direction and the cranial part of the tumor (p<0.05). CONCLUSIONS: The developed methodology identifies specific tumor regions and patients with increased risk of underdosage from delineation uncertainties in brachytherapy of cervical cancer.