الفهرس 
الغلافOnly 14 pages are availabe for public view
 |  | from | 123 |  |  |
| | | from | 123 | | |
Abstract
Abstract
Positron emission tomography/computed tomography (PET/CT) is a diagnostic imaging tool that plays an important role in the management of cancer patients. PET/CT is used to diagnose, stage cancers and to monitor response to treatment. The most common used radio-isotopes is 2-(18F)-fluoro-2-deoxy-D-glucose (18F-FDG). The basis of PET/CT is the physiological 18F-FDG radiotracer uptake that has been administered to the patient at a cellular level. Therefore, it is important to quantify the uptake of a radiotracer and to have a reliable method to assess 18F-FDG uptake in a lesion detected by this scan.
There are two common methods to make quantification of 18F-FDG uptake; the first is the visual qualitative assessment that compares the uptake in diseased tissue relative to the normal surrounding tissue, and the other method is a semi-quantitative assessment is known as a standardized uptake value (SUV). The value of SUV is dependent on various factors including the patient’s body habitus, the dose of the injected radioisotope, and point spread functions (PSF). SUV is being used for staging, diagnosis, and monitoring in clinical oncology.
Normalization of lesion SUV to background activity is an important way to minimize the variation and ensure the quality of the scans. The liver is identified as a potential source for background normalization, however, there is a limited number of studies compared the liver to other background sites and less studies that compared the normalization of lesion SUV to different body weight indices.
The purpose of this study was to study the impact of different SUV variants in terms of mean and maximum measures as well as various normalization methods with respect to body weight, body mass index, body surface area, and lean body mass in patients with lymphoma and to study the effect of point spread function (PSF) iteration as well as the timing of PET/CT acquisition assessment of SUV measures including patients and phantom study.
In the case of SUV normalization, sixty-nine patients were retrospectively selected. All patients had undergone F18-FDG PET/CT using the standard imaging protocol. In the first part of this study, SUVmean and SUVmax of patients’ lesions and three background sites including liver, aorta, and muscle were determined. Then the normalization of lesion SUV to body weight and body background sites was performed. The ratio of lesion SUVmax to body background sites (muscle, aorta, and liver) SUVmax was computed in addition to the ratio of lesion SUVmean to body background sites SUVmean.
The second part of the study included the calculations of the body mass index (BMI), body surface area (BSA), and lean body mass (LBM). The normalization of lesion, liver, aorta, and muscle SUV to BMI, BSA, and LBM was calculated and compared to each other. For studying the point spread function effect (PSF) and timing effect, clinical examinations, as well as phantom methods, were carried out. For the clinical studies, 10 patients have been examined and the resulting lesions images were reconstructed to a different number of PSF iterations including PSF (1,6), PSF (3,6), and PSF (6,6). The % difference was then determined for every iteration number.
In the case of phantom studies, four contrast ratios have been performed including 3:1, 5:1, 7:1, and 10:1. The data was then reconstructed to different times 1 min, 3 min, 5 min, 7 min, and 10 min including different spheres sizes. For each study, SUVmean and SUV max were identified.
To support the phantom study, one patient was scanned for 10 min and the lesion image was reconstructed to 3 min, 5 min, and 7 min. Results showed that there is a significant difference in SUV measurements between the three background sites.
Lesions normalized to the liver were significantly lower than those normalized to aorta and muscle and the results also showed a higher magnitude of lesions normalized to muscle in comparison to the aorta.
The SUVmax and SUVmean normalized to different body weight indices showed the lowest variation with BSA and BMI while being increasingly higher with lean body mass using the two methods James and Janmahasatian respectively and then highest with body weight. Also, the results showed that the % difference for the lower number of PSF iteration was lower than the higher number of iterations. For the four contrast ratios 3:1, 5:1, 7:1, and 10:1, the SUV measurements showed a stable profile over time including the different phantom spheres sizes.
The present work recommends normalization of SUVmean to BSA and Janma lean mass and also the normalization of SUVmax to James’ lean body mass. We also recommend the normalization of SUVmax to lean (i.e. James), SUVmean to lean (i.e. Janma), and BSA which showed significant independence to body weight so as to reduce the systematic errors associated with elevated SUV values in patients with increased body weights.