Organ-At-Risk Motion Analysis for Pancreatic Tumor Stereotactic Radiotherapy
T Gray*, C Liu, B Guo, D LaHurd, E Balagamwala, A Vassil, P Xia, Cleveland Clinic Foundation, Cleveland, OH
Presentations
PO-GePV-T-390 (Sunday, 7/25/2021) [Eastern Time (GMT-4)]
Purpose: Using daily diagnositic CT to quantify inter-fractional anatomical changes and motions in the pancreas tumors and nearby organs-at-risk (OARs) in a course of radiation therapy for pancreatic cancer.
Methods: Thirteen daily CT scans were contoured and evaluated for interfractional motion of target volumes and surrounding OARs organs-at-risk. The GTV, PTV, duodenum, stomach, small bowel and colon were manually contoured on each of thirteen CT scans. Using an imaging analysis software program, contours from each daily CT scan were compared with those from the corresponding planning CTs with the Dice Similarity coefficient (DSC), mean slice-wise Hausdorff distance (MSHD), and average centroid motion.
Results: Median DSCs and Hausdorff coefficients for each target and OAR were plotted and compared. Mean DSCs and Hausdorff coefficients for each target and OAR were also compared in a table with standard deviations. Comparing DSC between the planning CT and daily CT for all thirteen CTs, the largest DSCs were observed for the PTV and stomach, with average values of 0.63 ± 0.07 and 0.65 ± 0.08, respectively. Median values were 0.64 and 0.66, respectively. For the small bowel and colon, the Hausdorff distance was greatest with average values of 13.5 ± 5.1 mm and 11.7 ± 6.1 mm, respectively. Average change in contour centroid was seen for the small bowel and stomach, with average values of 76.3 mm and 88.6 mm, respectively.
Conclusion: This study demonstrates a method to quickly and effectively evaluate target and organ-at-risk motion and that these structures can shift greatly during a patient’s radiation therapy treatment course. This study also shows the great importance in implementing better adaptive radiotherapy measures to treat pancreatic cancer more effectively.
ePosters
Keywords
Not Applicable / None Entered.
Taxonomy
IM/TH- Image Segmentation Techniques: Segmentation Method - other
Contact Email