J Nucl Med. 2007; 48 (Supplement 2):144P
Oncology: Therapy, Metrics & Intervention
Technical Issues and Quantification |
Quantification of bone metastasis in whole body images of 18 F-FDG PET/CT
Mohei Abouzied1,
Omer Demirkaya2 and
Ayman Rifai2
1 Nuclear Medicine, SUNY at Buffalo, Buffalo, New York; ;
2 Biomedical Physics, King Faisal Specialist Hospital and Research Center, Riyadh, Saudi Arabia
492
Objectives: To develop a method to quantify the metabolic changes induced by the different types of bone metastasis in cancer patients using whole-body PET/CT images. The quantitative parameters along with the structural changes seen by CT bone window may serve as a useful tool in assessing the tumor response of bone metastases to therapy. Methods: Twenty five cancer patients with no prior history of chemo or radiotherapy who had definite bone metastases documented by PET/CT using an 8 slice PET/CT system (Discovery ST, GE) and other conventional modalities were selected for the study. PET and CT images were resampled to the same pixel size. Then the bone structure was segmented simply using a threshold of 150 Hounsfield unit. After the segmentation, the maximum SUV in the bone region in each slice was computed. Then the ROIs in PET slices were identified using the 50% of the corresponding maximum SUVs. The ROIs only within the bone were identified using the segmented bone mask. The lesion statistics including the average CT values were computed from the PET and CT images using these ROIs. The final results were saved into an excel file. The identified lesions were subjected to the visual confirmation by a nuclear medicine physician experienced in PET/CT who also described the structural changes in the CT bone window whether lytic, sclerotic or mixed type. Results: 149 bony lesions have been processed using our method in 25 patients (8 females, 17 males, mean age of 42.2 years, 13 head and neck carcinoma, 5 with lymphoma, 3 with sarcoma and 4 other types of carcinoma). Quantitatively, the mean SUV for the lytic, sclerotic, mixed and lesions with no structural changes were 6.88, 6.13, 6.79 and 6.27, respectively. On the other hand, the corresponding average CT values in Hounsfield units for the same lesions were 156.9, 266.7, 230.4, and 251.2, respectively. The automated method depicted all the lesions defined as bone metastases by the nuclear medicine physician. Conclusions: The automated method (with minimal user interference) provides a convenient way to process images and give a valuable functional and structural description of the different types of bone metastases that might serve as a useful tool in monitoring and assessing the therapy response.
