MR Imaging Of IVIM-DKI Model:Preliminary Study In Lung Cancer

PART I Comparison of Parameters from Diffusion-weighted Intravoxel Incoherent Motion Imaging in Different Histopathological Classifications of Lung CancerObjective To extend the intravoxel incoherent motion diffusion-weighted imaging(IVIM-DWI)model to assess different histopathological classifications of lung cancer.Materials and Methods A total of 105 patients with lung cancers were recruited in this study,including 68 patients with adenocarcinomas,22 patients with squamous carcinomas and the rest 15 with small cell carcinomas.All these patients underwent MR examination consisting of axial IVIM-DWI sequence on a 3.0T whole body scanner,then the standard ADC(sADC),diffusion coefficient(D),pseudo-diffusion coefficient(D?),perfusion fraction(f),distributed diffusion coefficient(DDC)and water diffusion heterogeneity index(a)were calculated for each lesion within the IVIM-DWI model.Results Mean sADC values were(1.45±0.26)× 10-3 mm2/s for adenocarcinomas,(1.36±0.48)× 10-3 mm-/s for squamous carcinomas,and(1.35±0.40)X 10-3 mm2/s for small cell carcinomas,respectively.Mean f values were(59.75±16.37)%for adenocarcinomas,(47.41±18.69)%for squamous carcinomas,and(48.96±19.88)%for small cell carcinomas,respectively.Mean a values were 0.72±0.13 for adenocarcinomas,0.62±0.12 for squamous carcinomas,and 0.63±0.11 for small cell carcinomas,respectively.Results from one-way ANOVA indicate that the sADC,f,a values among adenocarcinomas,squamous carcinomas and small cell carcinomas were significantly different(P<0.05),while there was no significant difference in D,D?,DDC values among them(P>0.05).And comparison between subgroups shows that the sADC,f.a values were significantly higher than that of squamous carcinomas or small cell carcinomas(P<0.05),but there was no significant difference between squamous carcinomas and small cell carcinomas in D,D? DDC values(P>0.05).Conclusion The sADC.f.a values derived from the IVIM-DWI model made comprehensive noninvasive evaluation on diffusion,perfusion and heterogeneity of micro-environment of lung cancer.The IVIM-DWI model shows promising applications in characterizing histopathological classifications of lung cancer.PART II Diffusion Kurtosis Imaging:Correlation Analysis of Quantitative Model Parameters with Molecular Features in Lung AdenocarcinomaObjective To preliminarily evaluate the correlation of non-Gaussian diffusion kurtosis imaging(DKI)parameters with expression of molecular markers(EGFR.ALK and Ki-67)in patients with lung adenocarcinomas,and diffusion-weighted imaging(DWI)was used as the ref-erence standard.Materials and Methods DKI sequences were included in the imaging protocol of a series of patients who underwent MRI examinations to evaluate lung tumors.The clinical and MRI materials of 96 patients with lung adenocarcinomas meeting inclusion criteria were analyzed retrospectively,within parameters Kapp,Dapp measured from the non-Gaussian kurtosis fit,ADC value from DWI model,and biomarkers EGFR,ALK,Ki-67 detected by immunohistochemistry(IHC)and/or by molecular biology(MB)techniques after biopsy or surgery.Quantitative parameters(ADC,Kapp,Dapp)of patients with different EGFR,ALK,Ki-67 status were compared to explore the relationships between noninvasive imaging parameters and pathologic outcomes in lung adenocarcinoma.One-way ANOVA was used to assess correlation of quantitative parameters with histopathologic results.Result Of the 96 lesions,the amount of EGFR positive,ALK positive,and Ki-67 high-expression lesions were 53,12,and 84,respectively.The Kapp values were significantly higher among patients with EGFR gene positive mutations,ALK rearrangements or high Ki-67 PI(p<0.0001).The Dapp values were significantly lower among patients with high Ki-67 PI and EGFR positive mutations(p<0.0001.p=0.002).No statistical significance was reached regarding the difference in mean Dapp or ADC value between the groups with or without ALK rearrangements.Conclusion DKI may be a potentially quantitative technique to analyze different status of molecular markers.There was a significant correlation between EGFR,ALK,Ki-67 expression levels and Kapp value,while little correlation was found between molecular markers expression and Dapp or ADC value.DKI model is more powerful than a standard mono-exponential model of DWI in pathological characterization evaluation of lung adenocarcinoma.