| Background:Diffusion kurtosis imaging (DKI) is a new diffusion-weighted MRI technique and an extension of the widely used diffusion tensor imaging (DTI) method that with a genuine diffusion technique accounting for diffusional non-Gaussianity which have more ensitivity in reflecting the complexity of tissue. Recently,DKI was applied for evaluation of lesions of cerebral, liver, lung and prostate with the incomparable advantage compared with traditional diffusional and diffusional tensor imaging. DKI could describe the complicated diffusion behavior in tumor tissues and reveal more tissue information about microstructure such as the heterogeneity, vascularity and cellularity etc and provide quantitative information for tumors. But its application in diagnosis and differentiation of parotid tumors was rarely reported.Objective:Parotid gland tumors are increasing incidence and tend to be younger. For parotid gland, the inspection methods are mainly dependent on ultrasound and CT examination. Ultrasound for parotid gland deep lobe tumors showed poor ability. CT examination had radioactive damage and the side effects of contrast agent, and patients with high allergic constitution can't have enhancement CT examination. Routine MRI scans have less clinical value in distinguishing benign with malignant parotid tumors. Diffusion kurtosis imaging (DKI) is a new diffusion-weighted MRI technique and DKI is an extension of the widely used diffusion tensor imaging method (DTI). With DKI, all the standard DTI diffusion metrics,such as the mean diffusivity (MD) and the fractional anisotropy (FA) could be evaluated, and also for several additional metrics related to the diffusional kurtosis.In this way, DKI provides for a more complete characterization of water diffusion in tissue.This study was aimed to explore the magnetic resonance diffusion kurtosis imaging as a non-invasive checking method to identify its clinical value in diagnosis of benign and malignant parotid gland tumors.Materials and methods:There were 32 cases with 34 parotid gland tumors confirmed by pathology were included in this study. We retrospective analyze the conventional magnetic resonance imaging (MRI) and diffusion imaging (DKI) findings of all lesions. All patients without false teeth were examined with conventional MRI and DKI on a MAGNETOM Skyra 3. OT MRI scanner (Siemens AG,Erlangen, Germany) before operative three days. The inclusion criteria were as follows:1) patients with confirmed pathology by surgery;2) patients without any invasive examinations and any treatments including biopsy, radiation and chemotherapy before MRI examination. Conventional MRI scans including T1WI, T2WI,or T2WI fat suppression sequence and diffusion weighted imaging (DWI) with scanning position including transverse, sagittal and coronal position.MRI scanning was used to not only observe the pathological changes of size,shape, signal strength, boundary, but also show the details of pathological changes: 1, identification of single or multiple parotid tumors (involving multiple lesions in the same glands or lateral glands);2, the location of the clear tumors in parotid gland (deep lobe,anterior lobe of the parotid gland or spondylolysis);3, the relationship between tumor and facial nerve;4,the relationship between tumor and blood vessels within carotid sheath;5, clearing whether the tumor has a gland outside invasion, such as violation of skull base, etc.6, specific situation to the cervical lymph nodes. The DKI parameters were as follows:TR=3700ms, TE=95ms, FOV=210mm x 210mm,slice thickness=3mm with 0mm gap, b-values=0,1000 and 2000mm2/s,20 orthogonal directions. Ensure the patients not to make swallowing moventment during scanning as it may influence the imaging quality or accuracy of the measurements. The data of DKI was analyzed using in-house software (Diffusion and Kurtosis Estimator) in Matlab,Version Release 2012a /7.14 (MathWorks, Natick, Massachusetts) to calculate mean kurtosis (MK) map, axial kurtosis(Ka) map, radial kurtosis (Kr) map, mean diffusivity (MD) map,axial diffusivity(Da) map,radial diffusivity (Dr) map and fractional anisotropy (FA) map. A standardized ROI size about 20 voxels was drawn in the solid part of the tumor, and mean values of MK,Ka,Kr, MD, Da, Dr and FA was evaluated with measuring average three times on the basis of T2WI. With histopathological diagnosis as the gold standard, the sample was divided into benign and malignant groups. All statistical analysis was performed with the Statistical Package for Medical Statistics (Medcalc 15.6). At first, independent samples Wilcoxon Mann-Whitney test was conducted after Kolmogorov-Smirnov test for data normal distribution. And then Analysis of variance (ANOVA) was used for testing data homogeneity. With histopathological diagnosis as the standard and the maximum sum of the sensitivity and specificity, the cut-off values of MK, MD and FA for differentiating malignant from benign were obtained with Receiver Operating Characteristic (ROC) curve analysis,sensitivity and specificity were also obtained with ROC analysis.Results:32 patients with 34 parotid tumors were involved in this study. One patient with.adenolymphoma had two bilateral masses, and one patient with mucoepidermoid carcinoma had two masses in right parotid gland. In 23 benign parotid tumors, there were 10 pleomorphic adenoma,9 adenolymphoma,2 basal cell adenoma,1 multiple nodular acidophil adenoma,1 myoepithelioma. In 11 parotid malignant tumors, there were 4 mucoepidermoid carcinoma,3 acinic cell carcinoma,2 basal cell carcinoma;1 adenoid cystic carcinoma,1 carcinoma ex pleomorphic adenoma. There was a significant difference of benign and malignant parotid tumors in the values of MK, FA and MD (P=0.003,0.019 and 0.047). There was no difference of benign and malignant parotid tumors in the values of Ka, Kr and Da, Dr Values. The mean MK value of benign parotid tumors was significantly lower than malignant parotid tumors(0.728±0.263 vs 1.0911±0.253, respectively). The mean FA value of benign parotid tumors was significantly lower than malignant parotid tumors (0.198±0.047 vs 0.251±0.043), and the mean MD value of benign tumors was significantly higher than malignant tumors (1.048 ±0.323(10-3mm2/s) vs 0.738±0.330 (10-3mm2/s)).The diagnosis of cut-off point between benign and malignant parotid tumors for MK was 1.053. The diagnosis of cut-off point between benign and malignant parotid tumors for FA and MD was 0.240 and 0.647 (10 W/s, respectively;The AUC,Yuden index J, sensitivity, specificity.positive predictive value and negative predictive value for MK were 0.853, 0.663,75.000%,91.300%,75.000%,91.300%; The AUC for FA and MD in differentiating benign and malignant parotid tumor was 0.783 and 0.739. The Yuden index J, sensitivity, specificity, positive predictive value and negative predictive value for FA and MD were 0.576,75.000%, 82.610%,60.000%,90.500% and 0.457,50%,95.650%,46.700%,93.700%. ROC curve analysis showed that MK has the largest AUC and Yuden index J, and MD has the smallest AUC and Yuden index J, but has the largest specificity.Conclusion:DKI showed higher specificity and sensitivity than conventional diffusion-weighted and tensor imaging for assessment of benign and malignant parotid tumors. MK enables differentiation and characterisation of parotid tumors, providing quantitative information for the parotid tumors.DKI is a noninvasive biopsy method preoperative for parotid gland tumors. |
PDF Full Text Request