Unencrypted login | home

Program Information

Improving Target Delineation by Using Deformably Registered Multi-Modality Images for Radiation Therapy of Pancreatic Cancer

C Yang

Y Cungeng*, I Moraru, E Dalah, V Hart, E Paulson, B Erickson, X Li, Medical College of Wisconsin, Milwaukee, WI

SU-E-J-112 Sunday 3:00PM - 6:00PM Room: Exhibit Hall

Purpose: Combined multi-modality images can provide unsurpassed target recognition, yet the large variation of contour volumes from different modalities prevents accurate contour transfer onto CT images for radiation therapy (RT) planning. This work demonstrates improved agreement between target contours from multi-modality images after deformable multi-modality image registration for pancreatic cancer RT.

Methods:PET, and various MRI including T1, T2, DWI and DCE, were rigidly registered with CT for representative patients with pancreatic cancer. Gross tumor volume (GTV) and organs at risk (OARs) were delineated on multi-modality images and were exported into a presently-developed deformable multi-modality image registration tool. The derived deformation fields were applied to deform the corresponding modality contours, which were later overlapped onto the planning CT.

Results:Substantial variations among contours from multi-modality images were observed for both GTV and OARs. When using T1 weighted contour volume as nominator, the delineated volumes vary from 0.5 to 1.8 for GTV, from 0.81 to 1.05 for OARs between the image modalities. The overlapping ratio between different modalities varies from 0.22 to 0.74 for GTV, and from 0.65 to 0.84 for OARs. After deformable image registration, the contour volumes are changed by deformation fields by 6% to 11%. These changes do not impact volume variation between different modalities, but improve significantly the overlapping ratio between contours from different modalities, for example, changing from 0.55 to 0.82 for GTV.

Conclusion:The deformable image registration of multi-modality images increases the agreement between contours from different image modalities, improving accuracy for target and normal structure delineation for radiation treatment planning of pancreatic cancer.

Contact Email: