Comparative Study Of Monoexponential、Biexponential、Stretched Exponential Models Of Diffusion-Weighted Imaging And Diffusion Kurtosis Imaging In Diagnostic Value Of Solid Soft Tissue Tumors

Objectives:To compare the superiority of monoexponential、biexponential、stretched exponential models of diffusion-weighted imaging and diffusion kurtosis imaging in diagnostic value of solid soft tissue tumors.Materials and Methods:1.Clinical data:From June 2018 to Apirl 2019,30 patients who were initially diagnosed in the extremities,trunk and abdomen were diagnosed by clinical or routine imaging at the Second Hospital of DaLian Medical University.All cases were divided into benign and malignant groups according to the 2013 WHO soft tissue tumor classification criteria.2.MR scanning technique,image post-processing and Observation method:（1）Scanning technique:All examinations were examined by GE Discovery MR750w 3.0T MR scanner.The corresponding coils were selected according to the location and size of the lesions for routine MRI,DWI,DKI and routine enhancement examinations.（2）Image post-processing and Observation method:The MRI images were independently observed by two radiologists.The results were the same when the two were consistent,and the third person decided when there was a dispute.The DWI and DKI images obtained after the inspection are transmitted to the workstation of the MRI（GE Advantage Workstation 4.6）for image processing by MADC and DKI software in Functiontool.After selecting the optimal threshold and eliminating background noise,Find the image layer that contains the surface marker in accordance with the normal scanning level,and select the area where the most obvious enhancement and the most diffused part overlap,and draw a sense in the parenchyma region by hand-painting on the corresponding fitted image.In the area of interest,the area of the ROI ranges from about 50 to 70 mm²,and the values of the parameters are repeatedly measured three times to calculate the average value,generate:monoexponential DWI Standard ADC value;biexponential DWI D-mono,D^*-mono,f-mono,D-Bi,D^*-Bi,f-Bi value;stretched exponential models DWI DDC,αvalue;DKI MK,MD,FA value.3.MRI-pathology control method（1）Compare the image surface with the pathological section.（2）Compare ROI of the image with the pathological sampling spot.4.Histopathological observation methodThe HE section was observed under the microscope by two pathologists.The results were the same when the two were consistent.The third person decided when there was a dispute.The observation area is drawn on the finished HE slice,and the low power microscope and the high power microscope are sequentially observed.The histopathological information reflected by the microscopic image corresponds to the imaging information reflected in the image level ROI.5.Observation indicators（1）Conventional MRI findings:location,morphology,size,border,signal,homogeneity,and enhancement.（2）Various parameters obtained under the DWI and DKI:monoexponential DWI Standard ADC value;biexponential DWI D-mono,D^*-mono,f-mono,D-Bi,D^*-Bi,f-Bi value;stretched exponential model DWI DDC,alpha value;DKI MK,MD,FA values.（3）The parameters of DWI and DKI are used to identify the threshold,sensitivity and specificity of benign and malignant soft tissue tumors and comparison of the diagnostic efficacy.5.Statistical analysisAll data were analyzed using SPSS24.0 software.The normal distribution measurement data is expressed as mean±standard deviation,and the non-normal distribution measurement data is expressed as median（P25-P75）.（1）Firstly,the non-parametric single-sample Kolmogorov-Smirnov test is used to test the normal distribution of each parameter value.If the data is normally distributed and the variance is equal,the two groups are compared using the independent sample t test.If the data is non-normally distributed.A nonparametric Wilcoxon rank sum test was used.（2）Generate a receiver operating characteristic（ROC）curve for each parameter generated,analyze the area under the curve（AUC）,evaluate the diagnostic performance of each parameter,and obtain the optimal diagnostic threshold and corresponding sensitivity according to the Youden index.Results:1.MRI performance of Solid soft tissue tumorsMRI findings of 30 cases of soft tissue tumors:11 cases of benign soft tissue tumors showed high signal in 6 cases,slightly higher signal in 4 cases,and equal signal in 1 case.19 cases of malignant soft tissue tumors showed high and mixed high signals.With the increase of b value,the contrast of the low signal background between the tumor and the surrounding tissue is more obvious,but the image resolution is significantly reduced.As shown by the b-signal curve,the diffusion signal in the tissue decreases with increasing b value.2.Solid soft tissue tumors MRI-pathological control and its histopathological resultsAmong 11 cases of benign solid soft tissue tumors,1 case of hemangiomatosis,1case of hemangioma,2 cases of neurofibroma,3 cases of schwannomas,1 case of fibroid fibroma,2 cases of giant cell tumor of tendon sheath,1 case of nodular fasciitis.Among 19 cases of malignant solid soft tissue tumors,7 case were liposarcoma,2case were leiomyosarcoma,3 case were synovial sarcoma,3 case were gastrointestinal stromal tumors,2 case were malignant mucinous sarcoma,1 case was malignant peripheral schwannomas and 1 case of extramedullary mucinous chondrosarcoma.3.All data were consistent with normal distribution and variance.The results of independent sample t test showed that the monoexponential DWI Standard ADC value,biexponential DWI D-mono value,stretched exponential model DWI DDC value,and DKI MD value in the benign and malignant differences of solid soft tissue tumors were significant（t=2.415,3.678,3.467,4.113,P=0.005,0.001,0.002,0.003）.4.Monoexponential Standard ADC,biexponential D-mono,stretched exponential model DDC and DKI MD、MK showed that the AUC of each parameter is 0.804,0.844,0.852,0.919 and 0.878,and the AUC of MK is the largest.According to the Youden index,the optimal diagnostic thresholds for each parameter are 1.416×10^-3mm²/s,1.290×10^-3mm²/s,1.452×10^-3mm²/s,0.518,1.803×mm^-3/s.The sensitivity were 100%,91%,100%,79%,and 100%,and the specificities were 63%,73%,68%,100%,and73%.Conclusion:The following conclusions were obtained by comparing 30 cases of solid soft tissue tumors monoexponential,biexponential,stretched exponential model DWI and DKI:1.The monoexponential,biexponential,stretched exponential model DWI and DKI can identify benign and malignant solid soft tissue tumors,and the DKI may be superior to the monoexponential,biexponential,stretched exponential model DWI.2.In the parameters obtained by monoexponential,biexponential,stretched exponential model DWI and DKI,DKI MK value has the highest diagnostic value in the diagnosis of benign and malignant solid soft tissue tumors.