Font Size: a A A

Application Of Multi-b Value Diffusion Weighted Imaging In Rectal Cancer

Posted on:2019-06-05Degree:MasterType:Thesis
Country:ChinaCandidate:G W ZhangFull Text:PDF
GTID:2404330563955907Subject:Medical imaging and nuclear medicine
Abstract/Summary:PDF Full Text Request
Objectives:1.To identify the best mathematic model of multi-b value DWI in fitting the water signal attenuation curve between Gaussian(mono-exponential model)and non-Gaussian(IVIM and SEM)diffusion models in patients with rectal cancer,and evaluate the diagnosis efficacy of functional parameters in differentiating tumor from normal tissue.2.To explore the correlations between functional parameters and clinicopathological features(differentiation,T stage,N stage,AJCC stage and intra-mural lymphovascular invasion).Methods:1.The multi-b value DWI data of 59 patients with rectal cancer confirmed by pathology after curative resection were retrospectively analyzed,the range of b-value is 0to 2000 s/mm~2.Water signal attenuation curve at different b values was fitted according to Gaussian(mono-exponential model)and non-Gaussian(IVIM and SEM)diffusion models,the best mathematic model was evaluated with adjusted determination coefficient(R~2),and the diagnosis efficacy of functional parameters in differentiating tumor from normal tissue was tested using the t-test and ROC analysis.2.The multi-b value DWI data of 110 patients with rectal cancer confirmed by pathology after curative resection were retrospectively analyzed,the range of b-value is 0to 3500 s/mm~2.All functional parameter maps were acquired by using the FuctionTool processing program at AW4.6 workstation of GE 750 scanner.The maximum,minimum and mean value of functional parameters of whole tumor were extracted using the MATLAB.The correlations between functional parameters(Standard-ADC,slow-ADC,fast-ADC f,?,DDC and ADCaqp)and clinicopathological features(differentiation,T stage,N stage,AJCC stage and intra-mural lymphovascular invasion)were evaluated using t-test and ANOVA analysis.The correlation between T stage and N stage with differentiation were tested using Spearman rank correlation.The correlation between T stage and N stage with intra-mural lymphovascular invasion were tested using chi-square test.Results:1.The SEM exhibited the best fitting results of actual DWI signal in rectal cancer and the normal rectal wall(R~2=0.998,0.999,respectively).2.The functional parameters(DDC,?,slow-ADC-mono,f-Mono,slow-ADC-bi,fast-ADC-bi,standard-ADC)showed significant statistics in distinguishing the rectal cancer from normal tissue(P<0.05).The DDC,slow-ADC-mono,slow-ADC-bi and standard-ADC value had the better diagnosis efficacy(AUC>0.95)among of all functional parameters,but the fast-ADC-mono and f-bi showed the lower diagnosis efficacy(AUC=0.520 and 0.558).3.The CNR(tumor to normal tissue)was high at b=1000 s/mm~2,and there were no statistical significance between CNR at b=1000 s/mm~2 and CNR at b=800,1200,1500,2000 s/mm~2(P=0.389,0.876,0.443 and 0.085,respectively).4.ADCaqp_Max and ADCaqp_Mean value showed significant statistics in distinguishing the differentiation of rectal cancer(P<0.05),and the values of these two functional parameters were increasing with the lower tumor's differentiation.5.?_Min value showed significant statistics in distinguishing the differentiation of rectal cancer(ANOVA,P<0.05),and the?_Min value of well and moderate differentiated rectal cancer(0.361±0.096 and 0.378±0.139)were higher than that of poor differentiated group(0.262±0.220).6.There were no statistical significance of all functional parameters in differentiating the T stage,N stage,AJCC stage and intra-mural lymphovascular invasion(P>0.05).7.No correlation was found between T stage and tumor's differentiation(r=0.158,P=0.099),while the correlation was found between T stage and IMLVI(?~2=9.733,P=0.002),N stage and tumor's differentiation(r=0.265,P=0.005),N stage and IMLVI(?~2=13.364,P<0.001).Conclusions:1.The non-Gaussian diffusion models could evaluate the water diffusion characters more adequately in vivo than Gaussian diffusion model,and the SEM model was the best.?_Min value exhibited a correlation with tumor's differentiation.2.For the purpose of keeping the balance between high CNR(tumor versus normal tissue)and well image quality,the b vaule at 1000 s/mm~2 maybe a better choice for rectum routine DWI scanning.3.Considering the positive correlation between ADCaqp_Max,ADCaqp_Mean value and tumor's differentiation,we assumed that ultra-high b-value DWI might be a practicable and promising modality for evaluating the expression of AQP in vivo.4.Further researchers are needed for exploring the correlation between functional parameters and T stage,N stage,AJCC stage and intra-mural lymphovascular invasion.
Keywords/Search Tags:Rectal cancer, multi-b value DWI, Gaussian diffusion model, non-Gaussian diffusion model, AQP MR imaging, ultra-high b-value DWI
PDF Full Text Request
Related items