| Objective:To explore the value of energy spectrum CT multi-parameter imaging in the diagnosis of multiple myeloma.Materials and Methods:Retrospective analysis was performed on 36 patients with multiple myeloma confirmed by clinical bone marrow aspiration and pathological results in our hospital from February 2018 to November 2019.The features of MRI images and energy spectrum CT images were compared,and the detection rate and diagnosis accuracy rate of lesions were compared.With MRI images as reference,GSI in AW 4.7 post-processing workstation was used in the viewer software,the MRI area is displayed as abnormal signal,but there is no obvious bone destruction area on the energy spectrum CT image to sketch the area of interest.The position and size of ROI in normal control group were basically the same as that in MM group,and the corresponding data file is exported.This file included the CT of each energy level of 40ke V?140 ke V(at 10 ke V intervals).The value(HU)and the concentration of each paired base pair(mg/cm~3).(1)Fisher's exact probability method was used to compare the detection rate and correct rate of lesions in MRI images and energy spectrum CT images;(2)The CT values and energy spectrum curve slopes of the two groups of lesions at 40 ke V?140 ke V single energy(K=CT(40 ke V)-CT(100 ke V)/(100 ke V-40 ke V),and the base substance for the comparison of concentration by using independent sample t test,an analysis of the ROC curve was statistically significant parameters and calculate the area under the ROC curve and best diagnostic threshold,P<0.05 was statistically significant.Results:(1)Compared with MRI,there was no significant difference in the detection rate and diagnostic accuracy of lesions by energy spectrum CT(P>0.05);(2)A total of110 lesions were delineated in the MM group and the normal control group,respectively.It was found that the CT value of ROI in the MM group and the normal control group was statistically significant in the 40 ke V?100 ke V single-energy image and the slope of the energy spectrum curve(P<0.05),but not in the 110 ke V?140 ke V single-energy image(P>0.05).The ROC curve area under the 40 ke V?100 ke V single-energy image decreased gradually with the increase of ke V,which were 0.774,0.746,0.711,0.680,0.648,0.621 and 0.600,respectively.The best diagnostic thresholds were 286.545HU,204.585HU,150.500HU,119.515HU,99.910HU,92.600HU and 92.390HU,respectively.The ROC area of the slope of the energy spectrum curve was the largest,which was 0.839.The optimal diagnostic threshold was4.011.According to the comparison of the concentration of base substances,the P values of HAP(Water),HAP(Fat),Calcium(Water),Calcium(Fat),Water(HAP),Fat(HAP),Water(Calcium),and Fat(Calcium)in 8 groups of the MM group and the normal control group were all less than 0.001,with the differences were statistically significant.The areas under the curves of HAP(Water),HAP(Fat),Calcium(Water)and Calcium(Fat)were 0.839,0.831,0.840 and 0.832,respectively.The optimal diagnostic thresholds were 95.220 mg/cm~3,127.720 mg/cm~3,44.820 mg/cm~3 and61.275 mg/cm~3,respectively.The areas under the curves of Water(HAP),Fat(HAP),Water(Calcium)and Fat(Calcium)were 0.838,0.838,0.641 and 0.641,respectively.The optimal diagnostic thresholds were 997.490 mg/cm~3,956.260 mg/cm~3,1049.455mg/cm~3 and 1025.660 mg/cm~3,respectively.Conclusion:(1)There is no difference between energy spectrum CT and MRI in the detection efficiency of MM.(2)Energy spectrum CT multi-parameter imaging can show bone changes before the occurrence of significant bone-melting lesions in MM. |
PDF Full Text Request