Application Of Multi B-value Magnetic Resonance Diffusion Weighted Imaging In Non-small Cell Lung Cancer

Part Ⅰ Diagnostic value of Monoexponential,Biexponential,and stretched exponential Diffusion-weighted MR imaging and Diffusion Kurtosis MR imaging for solitary pulmonary lessionsObjective Quantitative comparison of Monoexponential,Biexponential,and stretched exponential Diffusion-weighted MR imaging and Diffusion Kurtosis MR imaging in differential diagnosis of benign and malignant solitary pulmonary lesions(SPL).Methods 44 SPL subjects according to inclusion criteria were selected,multi-b value DWI data acquisition,fitting linear fit with single exponential and DKI models and LevenbergMarquardt nonlinear fitting double exponential and tensile index models using least squares method,Describe the Region of Interest(ROI).Each parameter between benign and malignant SPL was compared using an independent sample t test or the Mann-Whitney U test.Assess correlations between all parameters by using Spearman correlation.Rusults The ADC,ADCslow,α,DDC,and Dapp values were significantly lower in benign SPL than in benign SPL(P < 0.001).Kapp was significantly higher in benign SPL than benign SPL(P = 0.001).For the ADCfast and f values reflecting the perfusion parameters,no difference was found between the two.In all subjects,ADCslow was significantly smaller than ADC(P < 0.05),and DDC and Dapp were significantly greater than ADC(P < 0.05).When observing the receiver operating characteristic curve identifying the benign and malignant SPL(Fig.3),the AUC value was 0.904 for ADCs,0.815 for ADCslow,0.942 for DDC,0.93 for Dapp,and 0.815 for Kapp.The DDC value has the largest area under the curve.De Long analysis showed no statistical difference between the areas under the ADC,DDC,and Dapp curves.There is a strong correlation between ADC and ADCslow,ADCfast,f,α,DDC,Dapp and Kapp(P < 0.001).Conclusions Compared to traditional diffusion parameters,the DWI extension model does not show significant advantages,but alpha and Kapp can provide additional information for the benign and malignant identification of SPL,which may help to improve treatment strategies and prognosis.Part Ⅱ Correlation between DKI whole-volume histogram analysis and histopathology of non-small cell lung cancerObjective Using Diffusion Kurtosis Imaging(DKI)quantitatively assesse Non-small cell lung cancer(NSCLC),and compare the correlation between DKI histogram parameters and tumor cell structure in a group of surgically confirmed NSCLC patients.Methods 64 patients with NSCLC were screened according to the inclusion criteria,and routine MRI and DKI images were collected.The DKI data was processed using internally written software,the relevant parameter maps were obtained,the degree of differentiation of each sample was evaluated by histology,and the nuclear count(NTCN)and nuclear-to-plasma ratio(N/C)of each sample were evaluated using Image Pro Plus software.The expression of Ki-67 in the specimen was examined.Histogram analysis was performed using the voxel-based raw data for each ROI,and then the following parameters for each ROI were calculated: mean,standard deviation(sd),skew,kurtosis,and 10,25 The percentile of the cumulative frequency distribution of 50,75,and 90.One-way ANOVA was used to assess the differences in NTCN count,N/C ratio,Ki-67 index,and ADC and DKI histogram parameters for different degrees of differentiation.Spearman correlation was used to analyze the relationship between ADC and DKI histogram parameters and tumor grade,NTCN and N/C ratio,and Ki-67 index.Receiver operating characteristic curve(ROC)analysis was used to assess the ability of ADC and DKI histogram parameters to distinguish between different grades of NSCLC.Rusults As the tumor grade increases,the patient’s N/C ratio gradually increases,the Ki-67 index increases,and the ADC’s 10 th percentile(10th)-90 th and mean are significantly lower than the lower grade.Group,high-grade group Dapp10th-90 th and mean were significantly lower than the low-level group,Kapp 10th-90 th,Kapp mean,Kapp sd and Kapp histogram kurtosis values were significantly higher than the low-level group(P<0.05).In identifying highly differentiated and moderately differentiated NSCLC,Kapp 75 th has the largest area under the curve(0.936),a cut-off value of 0.83,a sensitivity of 95.74%,a specificity of 80%,and a Youden index of 0.8.In the identification of moderately differentiated and poorly differentiated NSCLC,ADCmean has the largest area under the curve(0.923),a cutoff value of 1.25,a sensitivity of 92.33%,a specificity of 86.67%,and a Youden index of 0.77.There was a moderate negative correlation between ADC mean and tumor grade and N/C ratio,and a weak negative correlation with Ki-67 index.Dapp mean showed a moderate negative correlation with tumor grade and N/C ratio.Kapp mean and tumor The grade and N/C ratio showed a moderate positive correlation and a weak positive correlation with the Ki-67 index.Conclusions DKI histogram analysis showed significant differences in DKI parameters between different levels of NSCLC.DKI is feasible for assessing non-Gaussian behavior of water molecule diffusion and shows a certain correlation with tumor cell structure.Part Ⅲ Correlation between DKI and clinical pathological prognostic factors in nonsmall cell lung cancerObjective The multi-b-value diffusion-weighted imaging technique was used to evaluate the potential relationship between DKI-derived parameters and clinical pathological prognostic factors of NSCLC,and to explore imaging biomarkers sensitive to NSCLC prognostic factors.Methods 252 NSCLC patients were screened according to inclusion criteria and routine MRI and DKI images were collected.Sampling independent samples t-test or μ-test to analyze differences in parameters between different groups.The ROC curve was used to analyze the diagnostic efficacy of different prognostic factors.Spearman correlation was used to assess the relationship between DWI and DKI parameters and CEA levels,differentiation,T stage,N stage,WHO classification,and EGFR mutation status.Rusults ADC values were moderately negatively correlated with tumor T stage(r=-0.519),clinical stage(r=-0.544),and degree of differentiation(r=-0.599).Dapp values showed a negative correlation with tumor T stage(r =-0.519),clinical stage(r =-0.544),and degree of differentiation(r =-0.599).Kapp value was negatively correlated with CEA level(r=-0.283),and positively correlated with tumor T stage(r=0.429),clinical stage(r=0.378)and degree of differentiation(r=0.437).However,for the EGFR mutation status,no statistical differences and correlations were found between DKI and ADC parameters.The ADC,Dapp and Kapp values showed diagnostic significance(P<0.05)in different T stages,clinical stages and differentiation,and ADC values had the highest AUC.Conclusions ADC and DKI parameters are associated with important prognostic factors for NSCLC.DKI can quantitatively characterize NSCLC by detecting non-Gaussian diffusion characteristics associated with tumor microenvironment,which not only reflects tissue diffusion information,but also reflects the complexity and heterogeneity of tissue microstructure to some extent.Preoperative evaluation of NSCLC has important guiding significance.