PET tumor volume assessment with different CT patient positions in lung cancer

1751 Objectives: The purpose of this study is to accurately estimate PET metabolic tumor volumes (MTV) using a deformable image registration, by investigating the effects of different patient positions on PET and CT image registration in lung cancer. FDG-PET imaging has been used to delineate target volume including functional information in CT–based radiation treatment planning for lung cancer. For that purpose using hybrid PET/CT system is desirable. However, when PET and CT images are obtained from stand alone modalities, the patient arm positions and phases of respiration are usually different. For diagnostic chest CT, the patient takes a deep inspiration breath-hold (DIBH), enlarging the thorax and lungs and lowering the diaphragm, whereas for CT simulation in radiation therapy either a DIBH or a free breathing (FB) method can be used. In PET imaging the patient breathes normally, so the thorax and lungs are smaller and the diaphragm is usually higher up compared to the DIBH method. Methods: PET and CT images were acquired for a total of 17 patients with a mean age of 73, including 13 lung cancer patients. Both arms were placed at sides for PET imaging, but arms were up for CT imaging. The images were registered using an iterative non-rigid algorithm (REVEAL-MVS) based on voxel intensities. The displacements after the deformable image registration were measured for four to six points in the lung boundaries of the central axial, coronal and sagittal planes. The MTV and its mean standard uptake value (SUV) in the PET images were measured before and after the image registration for lung cancer cases. Results: The mean displacements in the lung boundaries and diaphragm ranged from 1.42 to 2.57 cm for coronal, 1.57 to 1.83 cm for axial and 1.81 to 2.40 cm for sagittal images. In 7 out of 14 cases the PET MTV became greater and the mean SUV became smaller after the deformable image registration. The PET MTVs after the deformable image registration resulted in volume differences of -5.1 to +29.3 cm3 whereas the original MTVs ranged from 0.6 to 120.2 cm3. Conclusions: Different PET tumor volumes, extent and locations were produced by different patient positions (arms-down for PET with normal breath; arms-up for CT with deep breath) using stand alone PET and CT imaging modalities. The non-rigid image registration should be used to obtain more accurately coregistered PET MTV with CT-adapted location and extent. Thus, the adapted PET MTV obtained using this method may be utilized for PET diagnosis with corrected location and extent, and for CT-based radiation treatment planning in lung cancer.