Feasibility And Value Of Quantitative Study Of Pancreatic Fat Content In Healthy Population Based On MR Proton Density Fat Fraction And Quantitative CT

Part 1 Repetitive Study of Region of Interest Size in Quantitative Evaluation of Pancreatic Fat using MR m DIXON-QuantBackgroundPancreatic fat deposition,also known as nonalcoholic fatty pancreatic disease(NAFPD)refers to the accumulation of fat in pancreatic acinar or pancreatic islet cells,or replacement of pancreatic parenchyma by adipose tissue.Excessive fat depose in pancreas can leading to the decreased insulin secretion from the reduced number of pancreatic cells and the apoptosis of pancreatic islet cell,which plays a very important role in the pathogenesis of diabetes and metabolic syndrome.Until now,histopathology was considered the gold standard for quantifying pancreatic fat content,but has obvious limitations related to the complexity of biopsy of the pancreas due of its deep retroperitoneal location and the risk of sampling errors.Therefore,imaging plays a significant role in the diagnosis of NAFPD.Regions of interest(ROI)refers to the artificial drawing of regions to be processed during image processing,which is the basis of imaging research.Pancreatic has the following characteristics:(1)deep anatomical position and slender shape,which make it easier to be affected by respiratory movement;(2)internal blood vessels and pancreatic duct,which make it to be more heterogeneity;(3)fat around the abdominal cavity through the interlobular frequently infiltrate into the marginal lobule of the pancreas,which make it easier to overestimate the fat content.So,no matter what measurement,such as MRS,CSE-MRI,QCT and so on,the range and size of ROI may affect the measurement results.Therefore,the size of the ROI is the basis of quantitative fat study on pancreas.ObjectiveThe aim was to evaluate the repeatability of ROI of different sizes in quantifying pancreatic fat content by the proton density fat fraction which measured by MR m DIXON-Quant.MethodsA total of 52 volunteers were included for the m DIXON-Quant scan of the upper abdomen.Two physicians independently measured the pancreas head,pancreas body and pancreas tail three times under different ROI areas(70~90 mm2,120~140 mm2,170~190 mm2).Two weeks after the first measurement,measurement was conducted again without referring to the previous measurement results.The variance analysis was used to compare the differences of PDFF values measured by the same observers in the same location and different ROI areas.The intra group correlation coefficient(ICC)was used to evaluate the repeatability of PDFF values measured at the same site and ROI area by the same observer between the two weeks.Bland-Altman plot and the intergroup correlation coefficient were used to evaluate the consistency of PDFF values measured by two observers at the same site and the same ROI area.Results1.When the physician A measured the pancreas body,the difference in the PDFF values measured under the ROI area of 120~140 mm2 and 170~190 mm2 was statistically significant(p < 0.05).2.ICC is higher at the same location and the same ROI area for two weeks apart.3.The two observers were more consistent in measuring the pancreas at the same location and ROI: the Bland-Altman showed that the points outside the ±1.96 SD range were all less than 5%.For inter observer precision,the ICCs > 0.8.The maximum value of ICC was 0.993,the minimum value of ICC was 0.894.Conclusions1.Different ROI sizes affect the results of MR m DIXON-Quant on pancreatic fat quantification.2.When ROI areas are 70~90 mm2 and 120~140 mm2,the repeatability is good.Part 2 Quantitative Evaluate Pancreatic Fat Content in Healthy Population using MR Proton Density Fat Fraction and Its Influencing FactorsBackgroundPancreatic fat deposition is a kind of heterotopic deposition of visceral fat which can be caused by a variety of factors,such as age,obesity,diabetes,dietary ischemia,hormone therapy,and so on.It is associated with insulin resistance(IR),type 2 diabetes mellitus(T2DM)and metabolic syndrome.At present,the diagnostic criteria for non-alcoholic fatty liver disease(NAFLD)have been recognized,that is to say liver fat content greater than 5% can be diagnosed.However,less attention has been paid to pancreatic fat deposition,so there is no unified understanding of the criteria for diagnosing pancreatic fat content increase.Therefore,the purpose of our study was to select an accurate,feasible and simple method to study the normal range of pancreatic fat content in the general healthy population,exploring the related influencing factors and distribution trends of demography,which is also the basis for subsequent research.ObjectiveMR m Dixon-Quant technique was used to investigate the changes of pancreatic fat content in normal healthy people with age,and to analyze the effects of age,gender and BMI on pancreatic fat content in normal healthy people.MethodsMR m Dixon-Quant of the upper abdomen was performed on 237 subjects,including 107 males and 130 females,aged from 21 to 79 years.Fat content of pancreas(head,body,tail)was measured as PDFF.The Friedman test was used to verify the uniform distribution of fat in the head,body and tail of the pancreas.Different genders were divided into 5 groups according to age: group 1(21~30 years old),group 2(31~40 years old),group 3(41~50 years old),group 4(51~60 years old),group 5(61~79 years old),to investigate the distribution trend of pancreatic fat content in different age groups.Meanwhile,Pearson correlation analysis was performed on the pancreatic fat content and age.The multiple linear regression analysis was utilized to identify the factors including the age,gender and BMI related to the pancreatic fat content.Results1.The median pancreatic fat content was 3.33%(1.82~6.16)in general population.2.The pancreatic fat content did not vary significantly between head,body,and tail.3.There was a positive correlation between pancreatic fat content and age in men(r = 0.336,p < 0.05)and women(r = 0.433,p < 0.05).In men,the growth rate was-20.87%,-1.34%,62.61% and 94.41%(group 2~5 versus group 1),respectively.In women,the growth rate was 30.46%,58.02%,205.90% and 205.90%(group 2~5 versus group 1).The difference between different age groups was statistically significant(F value: 14.283 for males,33.012 for females,p < 0.05).The differences in pancreatic fat content between groups 4 and 5 in males and females compared with groups 1 to 3 were statistically significant.Among men and women of the same age group,the pancreatic fat content of men in each age group was higher than that of women,but there was no statistically significant difference in the pancreatic fat content of different genders in each group.4.A logistic regression model was used to access the association between pancreatic fat and demographic(age,gender and BMI).The results showed that the pancreatic fat content was association with gender.The PDFF in man was higher than that in women.Meanwhile,the age and BMI were associated positively with the pancreatic fat content.Conclusions1.The pancreatic fat content was about 3.33% in general population.2.Fat tends to be homogeneous evenly distributed in the head,body and tail of the pancreas.3.The fat content of pancreas increased with age in both men and women.After 50 years old,the rate of fat deposition in the pancreas was significantly accelerated.Women showed a faster deposition rate than men with age growing.4.PDFF of pancreas is correlate with age,gender and BMI.It is higher in elder people,obesity,and men.Part 3 Pancreas Fat Quantification with Quantitative CT: an MRI Correlation AnalysisBackgroundFatty infiltration of pancreas also as nonalcoholic fatty pancreatic disease(NAFPD)is associated with the occurrence and development of obesity,type 2 diabetes and metabolic syndrome.Considering the imitations of pancreatic biopsy,imaging plays an important role in the diagnosis.Magnetic Resonance(MR)has been advocated as the optimal technique to assess pancreatic fat content with the use of different approaches such as MR spectroscopy or chemical-shift-encoded MRI.And,CSE-MRI measurements of the pancreatic proton density fat fraction(PDFF)is the most accurate modality.In the CT domain,new approaches have been introduced in recent years,in particular the use of Quantitative CT(QCT).Although this technique is most commonly applied to the measurement of bone mineral density(BMD),with the same physical approach it is also possible to quantify the percentage fat content of soft tissue.However,there have been no studies using QCT to quantify the pancreatic fat content until recently.ObjectiveThe purpose of this paper is to assess the fat content of the pancreas using Quantitative Computed Tomography(QCT)and to correlate the results with MR m Dixon-Quant(CSEMRI)measurements of proton density fat fraction(PDFF).MethodsIRB approval for this research was obtained and fifty-two subjects(25 males,27 females;mean age 35.1 years;age range 22~50 years)who were enrolled in the Prospective Urban Rural Epidemiology(PURE)Study underwent QCT and MR m Dixon-Quant scans for quantification of fat content in the pancreas on the same day.First,two observers placed regions of interest(area of 100~140 mm2)in the head,body and tail of the pancreas as closely matched as possible on the two scans.The second,the equivalent density value of the pancreas was calculated,and the QCT measurement value of the fat content of the pancreas was calculated based on the same principle of QCT liver measurement.Pearson correlation and Bland-Altman analysis were performed to evaluate the correlation between the QCT and MR m Dixon-Quant measurements and the systematic difference between the two techniques.Results1.The equivalent density of pancreatic tissue was calculated and put into the correction equation.The average fat percentages in the MR m Dixon-Quant and QCT studies were 2.7% ± 2.6% and-3.6% ± 3.7% respectively with a statistically significant difference(p < 0.0001).2.For intra observer precision,the ICCs of QCT measurement in the head,body and tail of pancreas were 0.992,0.993 and 0.996,respectively.The ICCs of MR-PDFF in the head,body and tail of pancreas were 0.996,0.995 and 0.994,respectively.For inter observer precision,the ICC of QCT was 0.992,meanwhile,the ICC of MRI-PDFF was0.994.3.The QCT and MR m Dixon-Quant measurements of pancreatic fat content were well correlated(r = 0.805,p < 0.0001).4.Bland-Altman analysis showed that the QCT measurements were systematically lower by 6.3% compared with the MR m Dixon-Quant measurements.Conclusions1.This study is the first attempt to calculate the equivalent density value of pancreatic tissue,and converts the measured value of QCT into the pancreatic fat content through the correction equation.The results of QCT were significantly lower than those of MRIPDFF,but the repeatability and reproducibility were higher and the reliability was better.2.The results of this study suggest a good correlation between QCT and MR m DixonQuant measurements of pancreatic fat content.3.QCT measurements were systematically lower by 6.3% compared to MR m DixonQuant PDFF.Further studies are required to improve the numerical agreement of QCT measurements with PDFF.Innovation points1.The first part of this study evaluated the repeatability of the ROI to the quantification of pancreatic fat in order to optimize the selection of areas of interest for pancreatic measurement.Until now,about quantification of pancreatic fat in China,there have been no study evaluating the size of ROI,and few foreign reports have been reported.The purpose of part one is to make the follow-up research more rigorous,accurate and refined.2.It is well known that without data from the general population,direct diagnosis of pancreatic fat elevation is not warranted.Although the current diagnostic criteria for nonalcoholic fatty liver disease have been recognized,the diagnostic criteria for nonalcoholic fatty pancreas disease have not been reached due to the lack of previous studies on the pancreatic fat content of normal healthy people.Therefore,this study using MRPDFF to evaluate pancreatic fat content of general population and explore the related factors and distribution of the demographic trend.The sample size of our study is large.3.Quantitative CT was originally used to measure bone density.In recent years,based on the same physical principles,some scholars began to use it to quantify the fat content in soft tissue.However,no scholars have applied it to quantify the content of pancreatic fat.This study is the first attempt to use QCT to calculate the fat content in the pancreas by the same method which have been used to assess soft tissues,and compare the results with MR-PDFF to explore the feasibility of QCT to quantify the fat content of the pancreas.