Positional and Volumetric Changes in Parotid Glands During Head and Neck Radiation Therapy Assessed Using Deformable Image Registration

Abstract

Purpose/Objective(s)To assess the validity of using a deformable image registration algorithm for the assessment of position and volume changes of parotid glands of head and neck patients during the course of radiation therapy.Materials/MethodsA total of 296 daily MVCT images from 8 head and neck patients were acquired throughout the course of radiation treatments using a Hi∗Art II helical tomotherapy unit (TomoTherapy Inc., Madison, WI). These images were processed retrospectively using a deformable image registration algorithm1. The algorithm was applied to register daily MVCT images to the original planning kVCT image. The registered CT images then automatically provide the deformed parotid contours on the daily MVCT. Therefore, no manual manipulation was involved in this process. The daily contours were visually reviewed for their anatomical appropriateness by a head and neck specialty team. The reviewed contours were quantitatively analyzed for their position and volume changes during the course of treatment.ResultsThe anatomical shape and position of the automatically generated parotid contours were judged as acceptable. Figure 1(a) shows positional changes, while Figure 1(b) shows parotid volume changes throughout the course of the treatments. The trend was for parotid position to migrate towards midline, therefore migrating towards the target areas treated to high doses, i.e. in the 50 to 70 Gy range. The average distance change between two center of masses of left and right parotid glands was −6.8 mm. The parotid volumes tended to decrease during treatment by an average of 0.7%/day.Conclusions Purpose/Objective(s)To assess the validity of using a deformable image registration algorithm for the assessment of position and volume changes of parotid glands of head and neck patients during the course of radiation therapy. To assess the validity of using a deformable image registration algorithm for the assessment of position and volume changes of parotid glands of head and neck patients during the course of radiation therapy. Materials/MethodsA total of 296 daily MVCT images from 8 head and neck patients were acquired throughout the course of radiation treatments using a Hi∗Art II helical tomotherapy unit (TomoTherapy Inc., Madison, WI). These images were processed retrospectively using a deformable image registration algorithm1. The algorithm was applied to register daily MVCT images to the original planning kVCT image. The registered CT images then automatically provide the deformed parotid contours on the daily MVCT. Therefore, no manual manipulation was involved in this process. The daily contours were visually reviewed for their anatomical appropriateness by a head and neck specialty team. The reviewed contours were quantitatively analyzed for their position and volume changes during the course of treatment. A total of 296 daily MVCT images from 8 head and neck patients were acquired throughout the course of radiation treatments using a Hi∗Art II helical tomotherapy unit (TomoTherapy Inc., Madison, WI). These images were processed retrospectively using a deformable image registration algorithm1. The algorithm was applied to register daily MVCT images to the original planning kVCT image. The registered CT images then automatically provide the deformed parotid contours on the daily MVCT. Therefore, no manual manipulation was involved in this process. The daily contours were visually reviewed for their anatomical appropriateness by a head and neck specialty team. The reviewed contours were quantitatively analyzed for their position and volume changes during the course of treatment. ResultsThe anatomical shape and position of the automatically generated parotid contours were judged as acceptable. Figure 1(a) shows positional changes, while Figure 1(b) shows parotid volume changes throughout the course of the treatments. The trend was for parotid position to migrate towards midline, therefore migrating towards the target areas treated to high doses, i.e. in the 50 to 70 Gy range. The average distance change between two center of masses of left and right parotid glands was −6.8 mm. The parotid volumes tended to decrease during treatment by an average of 0.7%/day. The anatomical shape and position of the automatically generated parotid contours were judged as acceptable. Figure 1(a) shows positional changes, while Figure 1(b) shows parotid volume changes throughout the course of the treatments. The trend was for parotid position to migrate towards midline, therefore migrating towards the target areas treated to high doses, i.e. in the 50 to 70 Gy range. The average distance change between two center of masses of left and right parotid glands was −6.8 mm. The parotid volumes tended to decrease during treatment by an average of 0.7%/day. Conclusions