|
|
||||||||
|
|
|||||||||
|
|
Oncology-Basic Science: Therapy, Metrics & InterventionTherapy, Metrics & Intervention Posters |
1 Diagnostic Imaging, MD Anderson Cancer Center, Houston, Texas; 2 Electrical and Computer Engineering, Rice University, Houston, Texas
1376
Objectives: To compare the internal target volume (ITV) of tumors obtained from: 4D CT (10phase); volume encompassing end inspiration and expiration CT, mid inspiration CT with a 1.5 cm margin to those of various PET segmentations based on % max tumor activity concentration (AC).
Methods: 4 patients diagnosed with solitary pulmonary lung nodules (nodule volume range: 0.8 to 13 cc) had a PET/CT and a cine CT on a GE DST scanner. The cine CT was rebinned to 10 different phases of the patient breathing cycle. The 3D tumor volume on each CT phase was segmented using a threshold of -300 HU and the volume of the total tumor trajectory was calculated and constituted the measurement gold standard. The same process was repeated to measure the volume encompassing the end expiration and inspiration CT phases (2 phase CT), the volume encompassing the mid inspiration CT with a 1.5 cm margin (1 phase CT). Tumor ITV was also measured on PET images by using 10, 15, 20, 30 and 40% of the tumors maximum AC. The centroid and ITVs of all PET and CT volumes were then compared to the gold standard.
Results: The ITV of the 2 phase CT was on average 86% similar to the 10 phase CT. The 1 phase CT was on average 3740% larger than the 10 phase CT. The PET ITV corresponding to 30% threshold was on average 92% (range 50-150%) similar to the 10 phase CT. The centroid of all ITVs was within 1.5 pixels in X, Y and Z.
Conclusions: 2 phase CT reproduces the 10 phase ITV with good accuracy, while the 1 phase CT with margin is the worst predictor of ITV. A PET tumor segmentation using 30% max threshold might potentially be used to determine the tumor ITV.
| ||||||||||||||||||||||||||||||||||||||