| Part ? The clinical value of diffusion kurtosis imaging in the diagnosis and classification of lymphoma[Objective]To investigate the clinical value of diffusion kurtosis imaging?DKI?in the diagnosis and classification of lymphoma,comparing with the parameters of the traditional single exponential model.[Materials and methods]We prospectively collected 67 cases of lymphoma with cervical lymph node infiltration?lymphoma Group?,50 cases of cervical lymph node metastasis?metastasis group?and 35 healthy volunteers?benign group?.All subjects accepted the single exponential model Diffusion weighted imaging?DWI?and the non-Gaussian model diffusion kurtosis imaging?DKI?scans on a Philips 3.0T MRI.The b value for DWI was 0,800 s/mm2,for DKI was 0,1000,2000 s/mm2.All the scan data were imported into the Philips post-processing workstation,and calculated the DWI-related parameter ADC value and the DKI-related parameters D and K value.The parameters among the lymphoma,benign,metastatic group and the parameters among pathological subtypes of lymphoma were compared respectively.The receiver operating characteristic curves?ROC curves?were drawn to evaluate the diagnostic performance of DWI/DKI parameter for lymphoma.[Result]The ADC and D values of benign group were higher than the metastastic group(Benign:1.029±0.107×10-3mm2/s,1.370±0.183×10-3mm2/s;metastastic group:0.809±0.120×10-3mm2/s,1.102±0.162×10-3mm2/s),The lowest ADC and D values belonged to the lymphoma Group(0.735±0.228×10-3mm2/s,1.033±0.468×10-3mm2/s),the difference among three groups was statistically significant.The K value of the benign group?0.823±0.211?was lower than the metastatic and lymphoma Group?1.005±0.208,1.001±0.248?.But the K value between the metastatic and the Lymphoma Group was close.In the pathological subgroup comparation,the ADC value and D value of the HL group(1.006±0.244×10-3mm2/s,1.839±0.289×10-3mm2/s)were significantly higher than the NHL group(0.699±0.201×10-3mm2/s,0.923±0.371×10-3mm2/s).But there was no statistically significant difference of the K values between the two groups.In the NHL subgroup comparation,there was no statistically significant difference of ADC,D and K values between T cell-derived and B cell-derived group.With the ROC curve analysis for each parameter,the diagnostic performance of ADC value was slightly better than D value for lymphoma.While taking the optimal diagnostic threshold of 0.685×10-3mm2/s for ADC value,the sensitivity and specificity of the diagnosis of lymphoma were 96.5%and 47.8%.However,ADC and D value showed a better diagnostic efficacy for NHL than lymphoma,in which D value is higher than ADC value.The sensitivity and specificity for diagnosing NHL are 92.5%and 64.4%respectively when the best optimal diagnostic threshold of D value is0.9395×10-3mm2/s.K value was not effective in diagnosis of lymphoma and NHL lymphoma.[Conclusion]Both DWI and DKI model parameters can be used for differential diagnosis of lymphoma,in which the D value shows better diagnostic performance for NHL lymphoma.The K value can provide useful information for differential diagnosis of benign and malignant lesions of lymph node.Part ? Prognostic risk factors and invasive classification of DKI and lympho ma[Objective] To investigate whether the DKI related parameters correlate with the prognostic risk factors?Ki67 expression?and invasiveness of lymphoma,and to explore the value of DKI in the evaluation of the prognosis and invasiveness of lymphoma.[Materials and methods] 67 patients with lymphoma accepted DWI and DKI scans prospectively.The patients were divided into three groups?indolent,invasive and very invasive groups?accordingg to the 2000 WHO Lymphoma Classification.The Ki67 expression of lesions were determined by standard immunohistochemical staining?SP method?.The difference of DKI parameters among three different invasive groups and the correlation between Ki67 expression and DKI parameters were examined.[Results] The ADC values and D values decreased with the increased invasion of lymphoma,meanwhile,the K value and Ki67 expression increased.There were significant differences of ADC value,D value,K value and Ki67 expression between the indolent and invasive group,the indolent and very invasive groups.But there was no statistical difference of the DKI parameter and Ki67 expression between the invasive and very invasive group.The Spearman correlation analysis showed that ADC value and D value had significant negative correlation with Ki67 expression?r=-0.340,P<0.01,r=-0.478,P<0.001?,and K value was positively correlated with Ki67 expression?r=0.618,P<0.001?.[Conclusion] The quantitative parameters of DWI and DKI model are significantly correlated with the invasiveness and prognostic risk factors of lymphoma.DKI has the potential ability to assess the invasiveness and prognosis of lymphoma.Part ? Correlation between the DKI parameters and the characteristics of pathological images in Lymphoma[Objective] To explore the correlation between the parameters of DKI model and the characteristics of pathological images using computer-aided clustering analysis,and the correlative cytological mechanism of DKI parameters.[Materials and methods] 16 patients with lymphoma accepted DWI and DKI scans prospectively and their parameters were measured.The pathological characteristics of HE images including the cell density,nucleus portion,cytoplasm portion,extracellular space portion,nuclear/plasma and nucleus/extracellular space ratio were analyzed and calculated by computer-aided clustering analysis.In addition,the texture parameter?homogeneity?was calculated by GLCM texture analysis,and the correlation between the DWI and DKI parameters and the pathologic features were analyzed.[Results] The ADC value was negatively correlated with cell density?r=-0.847,P<0.001?and positively correlated with cytoplasm portion,nucleus/extracellular space ratio and homogeneity?r=0.539,P<0.05;r=0.582,P<0.05;r= 0.772,P<0.001?.The D value was negatively correlated with cell density and extracellular space portion?r=-0.788,P<0.001;r=-0.515,P<0.05?,and the nucleus portion was positively correlated with extracellular space portion and homogeneity?r=0.535,P<0.05;r=0.763,P<0.01?.The K value was positively correlated with the cell density?r=0.711,P<0.01?and negatively correlated with homogeneity?r=-0.598,P<0.05?.[Conclusion] There was significant correlation between the DWI/DKI parameters and pathological features.The parameters of DKI model,especially the K value,can be used as a unique index to evaluate the complexity and heterogeneity of tumor microstructure. |
PDF Full Text Request