| Objective Collimation is one of the important parameters of spectral CT scan,and its influence on imaging and radiation dose has been widely discussed.Studies have discussed the effect of imaging quality and radiation dose in coronary artery and neck and chest scans,while there are few studies on bone density measurement for different collimations.The study aims to assess the influence of different collimations on bone mineral density(BMD)measurements by gemstone CT energy spectrum imaging(GSI)using the European spine phantom(ESP).Methods To simulating abdominal total adipose tissue(TAT)of the patients as normal,overweight and obese,ESP was wrapped by fresh subcutaneous fat from pigs of200cm~2,320cm2 and 400cm2,respectively.The TAT of 200cm~2,320cm2 and 400cm2 were measured by QCT and adjusted manually,respectively.The ESP was scanned by GSI for 10 times with two kinds of collimations(group 40(64×0.625mm)and group 80(128×0.625mm),respectively),and CTDI was recorded.In these reconstructions,the image section thickness was 1.25 mm.The ASIR-V level was set to 0% to avoid potential bias in the iterative reconstruction algorithm.Then the reconstructed images were transmitted to the GE AW4.7 workstation,and the MD Analysis in the GSI Viewer software was selected.The region of interest(ROI,15mm×15mm)was set at the central level of L1-L3 vertebrae,avoiding the high BMD areas such as bone cortex and pedicle.The HAP(water)values of L1-L3 vertebrae were measured and recorded.The HAP(water)values between the two groups was compared by paired t-test;The relative measurement error(RME,%)and the root mean square error(RMSE)of HAP(water)values were calculated to assess the accuracy.The correlation analyses were determined by the Pearson correlation test.Bland-Altman analysis was conducted to assess agreement between the measurements of two groups.Linear regression was used to calculate formulas to make the measurement closer to the true BMD value.The CTDI was used to assess the radiation dose of the two groups of scans.Results There were good correlation and agreement between the measurements of two collimations(r>0.996;p<0.05),when the TAT area was 200cm~2,320cm~2,and400cm~2,respectively.The Bland-Altman plots show the agreement between the 40mm and 80 mm groups.When TAT was 200cm~2,for L1 to L3,RMEs were 2.12,1.17 and2.59 in group 40,respectively,and RMSEs were 0.44,0.65 and 1.64,respectively.In group 80,RMEs were 7.57,3.70 and 2.56,respectively,and RMSEs were 1.29,1.30 and1.64,respectively.The HAP(water)values between two groups were significantly different in L1 and L2(P<0.01),except for L3(P>0.05).For L1 and L2,The accuracy of40 mm group was better than that of 80 mm group.Linear regression analysis was performed on the HAP(water)values of group 40 and 80 for reducing deviations.For group 40,the regression equation was y=1.0431x-3.1983 with R2=0.9995;for group 80,the regression equation was y=1.0671x-8.6992 with R2=0.9986.Using this equation,RMSEs decreased significantly.When TAT was 320cm2 and 400cm2,there were no statistically significant differences for these values between the two groups(P>0.05).The values of CTDI were 9.09 m Gy and 8.83 m Gy for group 40 and 80,respectively.Conclusion For the patients as normal,especially for the patients with bone loss and osteoporosis,40 mm collimation shows higher accuracy than 80 mm collimation on BMD measurements by GSI.And the corrected HAP(water)measurement value can be made closer to true BMD value by the linear regression analysis.For the overweight and obese patients,there was no significant difference between the two collimations,and 80 mm collimation is recommended to reduce the radiation dose. |
