论文部分内容阅读
Objective: Part volume effect is a common phenomenon in nuclear medicine.It can cause incorrect estimation of radiopharmaceutical uptake.The aims of our study were to evaluate the influence of part volume on the measurement of lesion to background ratio (L/B) on SPECT and coincidence circuit SPECT; and to evaluate whether the wide peak acquisition of coincidence circuit SPECT affects the lesion detection.Methods: A dual-headed coincidence circuit SPECT/CT (Discovery VH, GE Medical Systems,United State) equipped with 1-inch-thick sodium iodide crystals was used.Using a simplified Jaszczak phantom with six cylindrical lesions having diameters of 2.8, 1.8, 1.3, 0.8, 0.8 and 0.8 cm, we performed three parts of experiments.The contrasts of radioactivity were varied from 2∶1 to 14∶1 in each part.In the first part, low energy high resolution collimator and 99mTcO4-solution were used.In the second and third parts, coincidence system and 18F-fluorodeoxyglucose (18F-FDG) solution were used.In the second one (routine acquisition), the energy window was 511keV ± 10%; the image acquisition time was 30 minutes.In the third one (wide peak acquisition), the energy window was 511keV± 15% and the image acquisition time was 20 minutes.Tomographic imaging was performed followed with CT scanning for attenuate correction in all.Then OSEM reconstruction was performed.Three kinds of lesion counts were recorded on the transitional slices.Firstly, the maximal counts were recorded; Secondly and thirdly, the average counts were recorded.The size of region of interest (ROI) was equal to the lesion size in the second one and 6 pixels in the third one.The background counts were average counts in all the three ones.Then three kinds of L/B values were measured.The correlations of the L/B values with the contrasts were studied and if they were statistics significant, the regression equations would be got.SPSS 11.0 was used to perform Paired Students t test, linear correlation and regression.Visual analysis was also used to compare the two FDG studies by two experienced doctors of nuclear medicine besides the L/B measurement.Results: When the contrasts were 2∶ 1, the 0.8-cm-lesions could not be detected in all the three parts of experiment.Except for the 2.8-cm-lesion on the two 18F-FDG experiments, the L/B values were all significantly underestimated and there were strong positive correlations between the L/B values and the contrasts.The smaller the lesion size, the greater the underestimation of the L/B values was.The regression equations were y2.8=0.086+0.689xcontrast, y1.8=0.350+0.436xcontrast,y1.3=0.837+0.307xcontrast, y0.8=0.564+0.199xcontrast for part one, and y2.8=0.221+0.941xcontrast,y1.8=0.426+0.647xcontrast, y1.3=0.484+0.528xcontrast, y0.8=0.429+0.318xcontrast for part two.The L/B of part three which had wider energy window and shorter scan time had no significant differences with part two.The findings of visual analysis of part two and part three were almost the same by the two observers.The L/B values using the maximum lesion radioactivity counts were the best one.Conclusions: Part volume effect significantly affects the measure of L/B values, especially to small lesions.The influence of part volume effect is more obvious on SPECT than on coincidence circuit SPECT.The regression equation can be used to calibrate the L/B measurement with the knowledge of lesion size.It is simple and accurate, and it can be repeated if we use the maximum lesion counts.Wide peak acquisition shortens the scan time without affect the lesion detecting.