The Clinical Factors Of Affecting Gd-EOB-DTPA-Enhanced Degree During Hepatobiliary Phase And The Optimization Strategy Of MRI Scanning Protocol

Part I:The correlation study between clinical factors and the enhancement degree of Gd-EOB-DTPA during hepatobiliary phaseObjective:To explore the related clinical factors correlating to enhancement degree of Gd-EOB-DTPA during hepatobiliary phase.Materials and methods:1.Patients:The patients with Gd-EOB-DTPA MRI study from November 2012 to May 2015were collected by PACS retrospectively.Inclusion criteria:（1）all patients were undergone Gd-EOB-DTPA enhanced study with hepatobiliary phase delaying 20 min;（2）The patients had no serious impairment of renal function（GFR<30mL/min/1.73m²）;（3）No spleen and liver surgery history.Exclusion criteria:（1）The images with serious motion artifact,which impaired the measurement;（2）Hepatobiliary phase imaging with incorrect delay time（over 5 min earlier or later than standard hepatobiliary delay phase）;（3）The patients had incomplete laboratory evaluation which is essential for the study;（4）Diffuse lesion（more than 5 lesions）or huge single lesion（diameter>3 cm）in liver impaired measurement;（5）Abnormal enhancement on arterial and portal vein phase in liver after contrast;（6）Focal or diffuse hepatic steatosis or hemochromatosis.Fat fraction and iron fraction was calculated based on in/out phase sequences,the formula was as follow:Fat fraction=(SI_in-SI_out)/(2×SI_in)×100%.Hepatic steatosis was defined as 5%or above fat fraction in liver.Moreover,Iron fraction was defined as(SI_in-SI_out)<-10.2.The protocol of Gd-EOB-DTPA study:All patients were performed MRI study on Philips Achieva 1.5T MR instrument and body coil.Supine position and respiratory suppressor were adopted for scanning.The scanning routine sequences were as follows:T1WI、T2WI、T1WI in/out phase、DWI（b=0,50,600 s/mm²）.The 3D-THRIVE were performed before and after contrast,the parameters were as follows:TR=3.36ms,TE=1.62ms,FA=10°,matrix=240×240,FOV=350×284mm,slice thickness=5mm.All patients received intravenous bolus injection of Gd-EOB-DTPA（Primovist,Bayer Schering）with dose of 0.1 ml/kg（0.025 mmol/kg）at a flow rate of 1.0 ml/s,and then flushed with 30 ml saline with same flow rate.Arterial phase,portal vein phase and delay phase were acquired by 25s,60s,3min delay.Delay-20min imaging was defined as hepatobiliary phase.3.The included clinical factors in the study:The included clinical factors were as follows:age,weight;the included laboratory indicators were as follows:alanine aminotransferase（ALT）,aspartate aminotransferase（AST）,gamma-glutamyl transpeptidase（GGTP）,alkaline phosphatase（ALP）,total protein（TP）,albumin（ALB）,total bilirubin（TBIL）,creatinine（Cr）,platelet count（PLT）,prothrombin time（PT）,activated partial thromboplastin time（APTT）,and thrombin time（TT）.All the laboratory tests were performed before or after MRI study within one week.4.The quantitative evaluation of hepatobiliary phase images of Gd-EOB-DTPA:All the measurement was performed by two radiologists with 11-and 9-year experience of abdominal imaging by PACS,they were blinded the clinical history and laboratory test result.Four region of interests（ROI）were placed at right anterior and posterior liver lobes,left lateral and medial lobe.The average of measurement was used as final data for statistical analysis.All ROIs were placed avoiding vessel,bile duct,lesions,calcification or artifact,and the largest ROI was preferred with area in the range of 1.6-2.7 cm².The pre-and post-contrast SI of hepatic parenchyma were measured.Copy and paste were used in order to ensure that the ROIs among different phase images were place at identical position of liver.Relative enhancement ratio（RER）were calculated based on the signal intense of hepatic parenchyma during pre-and post-contrast,and the formula was as follow:RER=(SI_hbp-SI_pre)/SI_pre.5.Correlation analysis between clinical factors and laboratory indicators with RER:The correlation analysis was performed between age,weight,ALT,AST,GGTP,ALP,TP,ALB,T.BIL,Cr,PLT,PT,APTT,TT with RER.6.Statistical analysis:Kolmogorov-Smirnov analysis was used to evaluate the data distribution.Pearson’s correlation coefficient was used for normal distribution data and Spearman’s rho correlation coefficient for abnormal distribution to compare the correlation between clinical and laboratory indicators with RER.All statistical analysis was performed using commercially available software SPSS 14.0（SPSS Inc.,Chicago,IL,USA）,P<0.05 was considered as statistical difference.Result:1.The study population based on inclusion and exclusion criteria:According to the inclusion and exclusion criteria,156 patients were included in the study of 127 males and 29 females,age 17-79 year-old,mean age 54.77±13.99year-old.2.The correlation between clinical and laboratory indicators with RER:A negative correlation with RER was found in ALT（r=-0.747）,T.BIL（r=-0.715）,PT（r=-0.678）,AST（r=-0.671）,GGTP（r=-0.455）,ALP（r=-0.390）and TT（r=-0.185）,and the positive correlation was found in PLT（r=0.493）and ALB（r=0.610）with RER,P<0.05 in all groups.However,there were no statistical difference in APTT,Cr,TP,age and weight,P>0.05.Conclusion:The factors of APTT,Cr,TP,age and weight had no correlation with enhanced degree of Gd-EOB-DTPA during hepatobiliary phase in adult,but the laboratory indicators of ALT,T.BIL,PT,AST,GGTP,ALP,TT,PLT and ALB had significant correlation with enhanced degree of Gd-EOB-DTPA during hepatobiliary phase.Part II:Development and validation of a semiquantitative test of insufficient enhancement during hepatobiliary phase of Gd-EOB-DTPA-enhanced magnetic resonance imagingObjective:To establish the semiquantitative test system（HBP enhancement test,HBP-ET）to predict the enhanced degree during hepatobiliary phase of Gd-EOB-DTPA,in order to avoid the invalid hepatobiliary imaging and improve the poor enhancement.Materials and methods:1.Patients:The Ethics Committee of our hospital approved this retrospective study.The inclusion and exclusion criteria were same as study Part I.According to the examination date,patients were divided into development data and validation data.The development data was used to establish the HBP-ET,while the validation data was used to validate the HBP-ET in predicting poor enhancement during hepatobiliary phase image of Gd-EOB-DTPA.2.MRI protocols:The MRI instrument and parameter in the study were same as study Part I.3.The quantitative analysis of hepatobiliary phase images of Gd-EOB-DTPA:The methods and criteria were same as study Part I.4.The semiquantitative analysis of enhanced degree during hepatobiliary phase of Gd-EOb-DTPA:The liver-to-portal vein contrast（LPVC）,liver-to-kidney contrast（LKC）and bile duct-to-panceras contrast（BPC）were measured.The semiquantitative scoring criteria were as follows:1:hyperintense;2:slight hyperintense;3:isointense;4:slight hypointense;5:hypointense.The invalid enhancement was defined as LPVC,LKC or BPC≥3.The measurement and judgement was performed independently by two radiologists with 11-and 9-year experience of abdominal imaging.Two radiologists were blinded the patients clinical and laboratory information.5.Establishment of HBP-ET for predicting invalid hepatobiliary phase imaging:Based on the result from study Part I,ALT,T.BIL,PT,AST,GGTP,ALP,TT,PLT and ALB were involved for multiple stepwise regression analysis and screened to establish HBP-ET.The indicators derived from multiple stepwise regression analysis were semiquantitative in terms of treatment guidelines and clinical experience.The total score from HBP-ET was used as final result to further statistical analysis.6.The evaluation and validation of HBP-ET:The correlation analysis between HBP-ET and RER was performed in both development and validation dataset.The difference between valid and invalid Gd-EOB-DTPA-enhance degree groups was compare in RER and HBP-ET using development and validation dataset.7.The utility of HBP-ET and traditional indicators in evaluating Gd-EOB-DTPA-enhanced degree during hepatobiliary phase by development and validation dataset:Child-Pugh classification and T.BIL were involved as traditional predictors in evaluating Gd-EOB-DTPA-enhanced degree during hepatobiliary phase.The comparison of utility in predicting Gd-EOB-DTPA-enhanced degree was performed between HBP-ET,Child-Pugh classification and T.BIL.8.Statistical analysis:Kappa-test was used to compare the coherence of semiquantitative scoring of Gd-EOB-DTPA-enhanced degree between two radiologists.Receiver operating characteristic curve（ROC curve）analysis was used to compare the utility of HBP-ET,Child-Pugh classification and T.BIL in predicting Gd-EOB-DTPA-enhanced degree during hepatobiliary phase.Kappa-test,multiple stepwise regression was performed using commercially available software SPSS 14.0（SPSS Inc.,Chicago,IL,USA）,and ROC curve analysis was performed using MedCalc 11.4（MedCalc Software,Ostend,Belgium）.P<0.05 was considered as significant difference.Result:1.Patients population:311 patients were included in the study according to inclusion criteria,and then103 patients were excluded according to exclusion criteria.Finally,208 patients were included for the study.These patients were divided into development and validation datasets.The patients with MR study from November 2012 to May 2015 were considered as development dataset,and the patients with MR study from June 2015 to October 2015 were considered validation dataset.Finally,156 patients（127 males,29females,age 17-79 year-old,mean age 54.77±13.99）were included to development dataset,and 52 patients（41 males,11 females,age 27-81 year-old,mean54.69±12.21）were included to validation dataset.In development dataset was consist of 50 invalid and 106 valid HBP-enhancement patients,and 18 invalid and 34 valid HBP-enhancment patients in validation dataset.In both development and validation dataset,there were significantly statistical difference between invalid and valid HBP-enhancement group in RER,P<0.001。2.The significant laboratory indicators in predicting invalid enhancement during hepatobiliary phase of Gd-EOB-DTPA:RER was considered as variable,multiple stepwise regression analysis showed that increasing ALT,T.BIL,GGTP,PT or decreasing ALB,PLT indicated invalid enhancement hepatobiliary imaging of Gd-EOB-DTPA（R=0.861;adjusted R²=0.731;F=71.354;P<0.001）3.Establishment of HBP-ET:According to the treatment guideline,the abnormal severity of laboratory indicators was semiquantitative scoring as follows:ALT,T.BIL,GGTP,PT score from0 to 3,ALB score from 0 to 2,PLT score from 0 to 1.The score of all parameter was summed as the final predictive score of each patient which is in the range of 0 to 15.The higher scores indicating lower enhancement of liver parenchyma during hepatobiliary phase.4.The evaluation and validation of HBP-ET:In both development and validation dataset,HBP-ET scores of valid enhancement group were higher than that of invalide enhancement group with significantly statistical difference,P<0.001.In development dataset,the total score of HBP-ET was negative correlation with RER（r=-0.787,P<0.001）,and also there was negative correlation between total score of HBP-ET and RER in validation dataset（r=-0.727,P<0.001）.5.The utility of HBP-ET and traditional indicators in predicting Gd-EOB-DTPA-enhanced degree during hepatobiliary phase with validation:In development dataset,the utility of HBP-ET（>4）was highest in predicting Gd-EOB-DTPA-enhanced degree during hepatobiliary phase with sensitivity,specificity and AUC of 80.00%,95.28%and 0.90（95%CI=0.84-0.94）,as compared to Child-Pugh classification with sensitivity,specificity and AUC of56.00%,83.96%and 0.71（95%CI=0.63-0.78）,and T.BIL with sensitivity,specificity and AUC of 50.00%,95.28%and 0.73（95%CI=0.65-0.80）.According to z-test,there was significant difference between HBP-ET and Child-Pugh classification（z=3.43,P=0.0006）,T.BIL（z=4.37,P<0.0001）respectively.In validation dataset,the utility of HBP-ET（>3）was also highest in predicting Gd-EOB-DTPA-enhanced degree during hepatobiliary phase with sensitivity,specificity and AUC of 83.33%,88.24%and 0.88（95%CI=0.78-0.95）,as compared to Child-Pugh classification with sensitivity,specificity and AUC of50.00%,94.12%and 0.67（95%CI=0.53-0.79）,and T.BIL with sensitivity,specificity and AUC of 77.78%,70.59%and 0.75（95%CI=0.61-0.86）.According to z-test,there was significant difference between HBP-ET and Child-Pugh classification（z=2.00,P=0.0459）,T.BIL（z=2.54,P=0.0112）respectively.Conclusion:HBP-ET can predict the invalid Gd-EOB-DTPA-enhanced degree during hepatobiliary phase,and the utility was higher than traditional indicators,such as Child-Pugh classification and T.BIL.Therefore,HBP-ET was helpful to avoid the invalid enhancement and improve the poor enhancement of Gd-EOB-DTPA during hepatobiliary phase.Part III:The influence of increasing flip angle to the image contrast on hepatobiliary phase of Gd-EOB-DTPAObjective:To study the flip angle influence the image contrast of hepatic parenchyma,hepatic lesion and bile duct on hepatobiliary phase of Gd-EOB-DTPA.Materials and methods:1.Patients:The Ethics Committee of our hospital approved this prospective study.Inclusion criteria:（1）No serious impairment of renal function（GFR<30mL/min/1.73 m²）;（2）No liver and spleen surgery history.Exclusion criteria:（1）Incomplete Gd-EOB-DTPA contrast MRI study;（2）Serious motion artifact impair measurement;（3）Incorrect delay time of hepatobiliary phase（5 min earlier or later than standard hepatobiliary phase of delaying 20 min）;（4）Huge（size>3cm）lesion in liver impair measurement.（5）Focal or diffuse hepatic steatosis or hemochromatosis.Fat fraction and iron fraction was calculated based on in/out phase sequences,the formula was as follow:Fat fraction=(SI_in-SI_out)/(2×SI_in)×100%.Hepatic steatosis was defined as 5%or above fat fraction in liver.Moreover,Iron fraction was defined as(SI_in-SI_out)<-10.According to inclusion criteria,69 patients were included.1 patient was excluded by incomplete Gd-EOB-DTPA MR study,2 patients were excluded by serious motion artifact,4 patients were excluded by hepatic steatosis,and 1 patient was excluded by hemochromatosis.Finally,61 patients were included for the study,47 males and 14females,age 18-82 year-old,mean age 54.43±11.23 year-old.There were 13hepatocellular carcinoma,2 cholangiocarcinoma,18 metastasis,21 hepatic cysts,and11 hemangioma.The largest lesion would be measured for the patients with multiple lesion in liver,and all lesion must be larger than 1 cm,the firse three lesions of largest size were allowed to be included for single patient.Hepatic cyst,hemangioma and metastasis were diagnosed based on clinical history and follow-up exam,while hepatocellular carcinoma and cholangiocarcinoma were diagnosed by biopsy or surgery.2.The Gd-EOB-DTPA MRI protocol:All patients in our study were performed with 10°,20°,30°FA for hepatobiliary phase imaging,which composed three groups of FA10°,FA20°,FA30°.Beside of FA,the other parameters among three groups were identical.3.The quantitative evaluation of MR images:All the measurement was performed by two radiologists with 3-and 4-year experience of abdominal imaging by PACS,they were blinded the clinical history and laboratory test result.The measurement criteria were as follows:（1）Four region of interest（ROI）were placed at right anterior and posterior liver lobes,left lateral and medial lobe at pre-and post-contrast administration,the average of measurement was used as final data for the calculation of relative enhancement ratio（RER）and signal-to-noise ratio of liver(SNR_liver).To match the position between different phase images,the ROIs by copy and paste were adopted.（2）Two ROIs were place at spleen and liver lesion respectively in order to calculate liver-to-spleen ratio（LSR）,liver-to-lesion ration（LLR）and signal-to-noise ratio of lesion(SNR_leison);（3）ROI was placed at portal vein to calculate the liver-to-portal vein ratio（LPR）;（4）ROIs were placed at common bile duct and pancreatic head to calculate bile duct-to-pancreas ratio（BPR）and bile duct-to-liver ratio（BLR）.All ROIs were placed avoiding vessel,bile duct,lesions,calcification or artifact,and the largest ROI was preferred with size in the range of 0.7–4.6 cm².The formulas were as follows:RER=(SI_pre-SI_post)/SI_preLSR=SI_liver/SI_spleenLPR=SI_liver/SI_{portal vein} LLR=SI_liver/SI_{liver lesion}BPR=SI_{bile duct}/SI_pancreas BLR=SI_{bile duct}/SI_liverSNR_liver=SI_liver/SD_liverSNR_lesion=SI_lesion/SD_lesion4.Statistical analysisAVONA was used to compare the RER,LSR,LPR,LLR,BPR,BLR among FA10°,FA20°,FA30°,LSD would be performed for subgroup comparisons.All statistical analysis was performed using commercially available software SPSS 14.0（SPSS Inc.,Chicago,IL,USA）,P<0.05 was considered as significant difference.Result:1.The influence of FA to the image contrast of liver,spleen and liver lesion:According to AVONA,there were significant difference among groups in RER,LSR,LPR and LLR.In RER,subgroup comparison showed that there was no significant difference between RER-10°and RER-30°,P>0.05.There were significant difference between RER-FA10°and RER-FA20°,RER-FA20°and RER-FA30°.The RER value in FA-20°was highest,and lowest in FA-30°.In LSR,LPR and LLR,the trend that the higher FA led to higher ratio was found.The subgroup comparison showed that there were significant difference between FA10°and FA20°,FA10°and FA30°,FA20°and FA30°in LSR,LPR and LLR,P<0.05.2.The influence of FA to the image contrast of bile duct:According to AVONA,there were significant difference in BPR and BLR（P<0.001）.The trend that higher FA leaded to higher ratio was found in BPR and BLR,and there were significant difference between FA10°and FA20°,FA10°and FA30°,FA20°and FA30°,P<0.05.3.The influence of FA to the image contrast of SNR:According to AVONA,there were significant difference in SNR_liveriver and SNR_leison（P<0.001）.The trend that higher FA leaded to lower ratio was found in SNR_liveriver and SNR_leison,and the subgroup comparison showed that there were significant difference between FA10°and FA20°,FA10°and FA30°,FA20°and FA30°,P<0.05.Conclusion:Increasing FA of Gd-EOB-DTPA-enhanced 3D-THRIVE imaging during hepatobiliary phase lead to increase the image contrast between hepatic parenchyma and splenic parenchyma,hepatic parenchyma and portal vein,hepatic parenchyma and liver lesion,bile duct and pancreatic parenchyma,bile duct and hepatic parenchyma.However,it is noted that higher FA could result in reduction of SNR.