| Part 1:The value of mono-exponential,bi-exponential and stretched exponential based multiple b value diffusion weighted imaging in the diagnosis of pancreatic carcinomaObjective:To investigate the value of parameters of mono-exponential(ADC),bi-exponential(D?D*?f)and stretched exponential(DDC??)in the differential diagnosis of pancreatic carcinoma and normal pancreatic tissue,employing the magnetic resonance multiple b value diffusion weighted imaging technique.Materials and methods:41 patients with clinically suspected solid pancreatic lesions were collected from March 2018 to December 2018,and finally 30 patients with pancreatic cancer confirmed by pathology or clinical date meet the requirements of our research,including 17 males and 13 females,aged 35-87 years,average age 66.23±12.61 years,the median age 70 years.All subjects were performed by routine MRI sequences and multiple b value DWI sequences on GE Signa 3.0T magnetic resonance scanner.The multi-b value DWI sequence include 11 different b values(0,50,100,150,200,300,400,600,800,1000,1500s/mm2).the original images and data by scanning were inputted into the post-processing software,and data acquisition and analysis were performed by two radiologists respectively.In the pancreatic cancer group,we chose essential part of the three adjacent layers of the tumor for display three regions of interest.In control group,we selected three regions of interest(ROI)in different parts of the normal pancreatic tissue which were not invaded by tumor.The corresponding parameters(ADC,D,D*,f,DDC,a)were measured,and take it average.Intraclass correlation coefficient(ICC)analysis was used to analyze the consistency and Repeatability of parameters measured by the two radiologists respectively.The difference of the parameters between pancreatic carcinoma and normal pancreatic tissue were compared by t test or Mann-Whitney U test,and the ROC curve was used to evaluate the effictiveness of the parameters which were statistical significance;Pearson correlation test was used to analyze the correlation between ADC and D,ADC and DDC,D and DDC,DDC and a in pancreatic carcinoma and normal pancreatic tissue.Result:1.The ICC of ADC,D,D?,f,DDC,a in normal pancreatic tissue were 0.907,0.849,0.815,0.811,0.84,0.836,respectively,and the ICC of ADC,D,DDC,a in the pancreatic carcinoma were 0.915,0.833,0.914,0.849,respectively,suggesting a good consistency and Repeatability;the ICC values of D*and f in pancreatic carcinoma were 0.662 and 0.634,espectively,exhibiting a moderate consistency.2.ADC?D*?f and DDC in pancreatic carcinoma(1.32±0.19×10-3mm2/s?6.81×10-3mm2/s?0.329?1.45±0.26×10-3mm2/s)were signallylower than those in normal pancreatic tissue(1.66±0.24×10-3mm2/s?11.60×10-3mm2/s?0.390?2.45±0.80×10-3mm2/s),the differences were statistically significant(P<0.05);a(0.87±0.06)in pancreatic carcinoma was evidently higher than that in normal pancreatic tissue(0.70±0.06),the difference was statistically significant(P<0.05);There was no significant difference in D between pancreatic carcinoma(1.07±0.20×10-3mm2/s)and normal pancreatic tissue(1.13±0.21 ×10-3mm2/s)3.The area under the curve of ADC,D*,f,DDC and a in differentiating pancreatic carcinoma and normal pancreatic tissue were 0.853,0.883,0.718,0.935 and 0.988,respectively,and the optimally diagnostic threshold were 1.45 ×10-3mm2/s?9.45×10-3mm2/s?0.359?1.76×103mm2/s?0.794,respectively.On this base,the corresponding sensitivity were0.8?0.8?0.733?0.833?0.967,respectively,and the corresponding specificity were 0.733?0.867?0.7?0.867?0.933.4.ADC was positively correlated with D and DDC in pancreatic carcinoma and normal pancreatic tissue,and the correlation coefficient in pancreatic carcinoma(r=0.86,0.956)was higher than that in normal pancreatic tissue(r=0.55?0.756).D was significantly positively correlated with DDC in pancreatic carcinoma(r=0.694),while there was no correlated in normal pancreatic tissue(r=0.173),There was no correlation between DDC and a in pancreatic carcinoma(r=0.000),however DDC was negatively correlated with?(r=-0.446)in normal pancreatic tissue.Conclusion:1.ADC from the mono-exponential model,D*?f from the bi-exponential model and a?DDC from the stretched exponential model have certain value for differentiating pancreatic carcinoma from normal pancreatic tissue,and a has the highest diagnostic efficiency;2.The Stretched exponential model is superior to the mono-exponential and bi-exponential model in the identification of pancreatic cancer and normal pancreatic tissue,3.The parameters from the mono-exponential and the stretched exponential model are more repeatable than those from the bi-exponential model.Part 2::The value of mono-exponential,bi-exponential and stretched exponential based multiple b value diffusion weighted imaging in evaluating the pathological stage of pancreatic carcinomaObjective:To investigate the value of different exponential model in distinguishing moderately with poorly differentiated pancreatic carcinoma,employing the magnetic resonance multipleb value diffusion weighted imaging technique.Materials and methods:We retrospectively analyzed the date of 15 patients with pathologically proved pancreatic carcinoma and with definitely pathological stage from March 2018 to December 2018,8 patients with moderate differentiation and 7 patients with poor differentiation.All subjects underwent routine MRI and multiple b value DWI sequence scanning on GE Signa 3.0T magnetic resonance scanner.The original image and data obtained from the multiple b value DWI sequence were inputted into the post-processing software,and data acquisition and analysis were performed by two radiologists jointly.We chose essential part of the three adjacent layers of the tumor for display three regions of interest.The corresponding parameters(ADC,D,D*,f,DDC,?)were measured,and take it average.The difference of the parameters between moderately and poorly differentiated pancreatic carcinoma were compared by t test or Mann-Whitney U test,and the ROC curve was used to evaluate the effictiveness of the parameters which were statistical significance.Result:1.a(0.894)in moderately differentiated pancreatic carcinoma was higher than these in poorly differentiated pancreatic carcinoma(0.835),the difference was statistically significant(P<0.05);while there was no significant difference in ADC?D?D*?f?DDC between moderately and poorly differentiated pancreatic carcinoma2.The area under the curve of a in differentiating moderately and poorly differentiated pancreatic carcinoma were 0.893,and the optimally diagnostic threshold was 0.85.On this base,The corresponding sensitivity of diagnosis was 1,and the corresponding specificity was 0.857.Conclusion:In the differentiation of moderately and poorly differentiated pancreatic carcinoma,? from the Stretched exponential model has a certain value,and parameters from the mono-exponential and bi-exponential model have no value. |
PDF Full Text Request