| PurposeTo optimize IVIM-DWI scan parameters to ensure the success of scan.To explore the repeatability of measurement of parameters of IVIM-DWI.To compare the performances of apparent diffusion coefficient(ADC)and IVIM parameters,including the pure diffusion coefficient(D),perfusion coefficient(D*),and perfusion fraction(f)between active tumor region and normal parenchyma before treatment.To study the characterization of DWI functional parameters on active tumor region of metastatic tumor before treatment in group of positive treatment response and in group of negative treatment response.The feasibility and diagnostic value of predicting response of treatment response were researched by DWI parameters onhepatic metastatic tumor.To investigate reliability and feasibility of CTP using low-dose protocol on hepatic metastatic tumor.To assess the radiation dose on low-dose CTP and tri-phase enhancing scans.To research the affect of different image reconstruction(iterative reconstruction algorithm on idose 4 level and filter back-projection algorithm)on measurement of parameters of CTP.To compare the performances of CTP parameters,including the blood flow(BF),blood volume(BV),mean transit time(MTT),permeability surface(PS)and hepatic arterial fraction(HAF)between active tumor region and normal parenchyma before treatment.To study the characterization of CTP functional parameters on active tumor region of metastatic tumor before treatment in group of positive treatment response and in group of negative treatment response.The feasibility and diagnostic value of predicting response of treatment response were researched by CTP quantitative parameters on hepatic metastatic tumor.To investigate the correlations of DWI parameters and CTP parameters on active tumor region of metastatic tumor.To compare the diagnostic abilities of each quantitative functional parameters before treatment in predicting treatment response using receiver operating curve method and to find out the cut-point.Then to explore the sensibilities and specificities of predicting treatment response on hepatic metastatic tumor using multi-parameter indexes.Materials and MethodsTotal 24 patients with confirmed liver metastases underwent liver IVIM-DWI and(or)CTP within two weeks before chemotherapy from May 2014 to August 2015.Except 7 patients was loss of follow up,17 patents had assessments of target tumor size by CT or MRI at after one cycles of chemotherapies.Inclusion criteria:a.liver metastases proven by histopathology or comprehensive diagnosis of laboratory indexes,imaging and clinical data.b.initial treatment.c.agreed to participate and complete the study.Exclusion criteria:contraindications to MRI or CT scan.Primary tumors were composed of 11 cases of lung cancer,7 cases of nasopharyngeal carcinoma(NPC),3 cases of gastrointestinal cancer,1 case of paranasal sinuses cancer,1 case of promoting fiber hyperplasia of small round cell tumor,1 case of melanoma of parotid gland.According the different of treatment response,all patients were divided into two groups,namely group of positive treatment response and group of negative treatment response.51 tumors of 20 patients undergoing 25 DWI scans were enrolled to study the IVIM-DW quantitative parameters(positive group n=18,negative group n=19).56 tumors of 24 patients undergoing 30 CTP scans were enrolled to study the CTP quantitative parameters(positive group n=20,negative group n=19).This study was approved by the institutional review board of our hospital.All the examinations in this study were carried out as part of clinical practice.The possible adverse effects were explained to the patients by radiologist,and written informed consents were obtained before the MRI or CT scans.The patients had right to terminate the study whenever he(she)wanted.All MRI were performed on a 1.5 T MR imaging system(Philips Healthcare,Best,Netherlands)using a 16-channel torso-coil.Patients were in the supine position for image acquisition.The routine MRI protocol was used for the clinical need,including breath-hold axial T1-weighted fast spoiled gradient-echo in-and out-of-phase,respiratory-triggered coronal and axial T2-weighted fast spin-echo sequences combined with fat suppression of spectral attenuated inversion recovery(SPAIR),and delayed-enhanced T1-weighted high resolution isotropic volume examination(e-THRIVE)sequences.Axial IVIM DW imaging was acquired prior to contrast media injection.Respiratory triggered single-shot spin-echo echo planar imaging(SS-SE-EPI)sequence[repetition time(TR)/echo time(TE),1285/64ms;field of view(FOV),375 mm × 300 mm;matrix,124 x 100;EPI factor,53;number of signals averaged(NSA),2;slice thickness/gap,6/1 mm]was adopted in combination with fat suppression and parallel sampling technique.Diffusion sensitive gradient pulses were applied in three orthogonal directions,with 12 weighting factors of 0,10,20,30,40,50,75,100,150,300,500 and 800s/mm2.DW1 data were transferred to a post-processing workstation(Extended Workspace,Philips Healthcare)equipped with manufacturer-supplied software(PRIDE DWI Tool,v.1.5,Philips Healthcare)to extract the ADC and IV1M parameters,including the true diffusion coefficient(D),the pseudo-diffusion coefficient(D*),and the perfusion fraction(f).ADCs were automatically calculated by using all b values with a mono-exponential fit The IVIM parameters were calculated consecutively using a nonlinear bi-exponential fit based on the following equation:SI/SI0=(1-f)·exp(-bD)+f·exp(-bD*),where S10 is the mean signal intensity of the ROl for a b value of 0,and SI is the signal intensity for higher b values.Four parametric maps of ADC,D,D*,and f were created on a pixel-by-pixel basis for each patient.Two radiologists with 10 and 7 years of experience in liver MR,respectively)jointly reviewed the MR examinations.Image J software(National Institutes of Health,Bethesda,MD)was used to manually draw ROIs on each lesion at the level of maximum transverse diameter of lesions.Conventional T2-weighted MR images and DW images(b=150 s/mm2)were used as references to determine lesion areas on corresponding ADC maps and IVIM parametric maps.The area that visibly had the most restricted diffusion(MRDA)on the ADC map was also determined.Two elliptic ROIs were drawn respectively at visibly area of the lesion and normal parenchyma.The ROIs were then copied from the ADC map and placed on the other parametric maps for D,D*,and F.The ROIs measurements were repeated a month later by same radiologists for repeatability.All CT perfusion imaging were carried on 256 slice CT(PHILIPS)using given scan protocol:120kv,40mA tuber current,512×512 matrix size,non-jog axial maximal coverage,breath hold.Scan time per cycle was 1 second,the total scan time was 40 seconds.Patients were in the supine position to localize the central slice of the tumor for image acquisition.contrast agent protocol:Non-ionic iodinated contrast agent with an iodine concentration of 370 mg I/ml was injected intravenously as a bolus at 5 ml/sec using a binocular high-pressure syringe before 6 sec scanning start,and then 20 ml physiological saline was injected at the same rate.Patient preparations were done including oral 1000-2000ml water within 1hour and breath-hold exercising before scan.All the CTP images were transferred to a workstation(ADW 4.3)for images post-processing.CTP parameters were computed respectively using two series of images based on iterative algorithm on idose4 level and filter back-projection algorithm.The blood flow(BF,ml/min/100 ml),blood volume(BV,ml/100 ml),mean transit time(MTT,s),permeability of capillary vessel surface(PS,ml/min/100 ml)and hepatic arterial fraction(HAF)as the percentage of total blood flow of arterial origin were calculated using liver tumor model of perfusion3.Regions of interests(ROI)were placed at aorta as artery input and at portal vein as vein input in a similar manner as in our IVIM-DWI study.The CT values and the noise at active tumor region and normal parenchymal were measured.Signal noise ratio(SNR)and Contrast noise ratio(CNR)were calculated.The radiation doses presented by dose length product(DLP)on CTP scans and three phase enhancing scan were recorded.The rate of IVIM-DWI and CTP accomplished for functional analysis were accounted.The repeatability of measurement of parameters of IVIM-DWI and CTP were tested using Pearson correlation analysis(correlation coefficient(r)between 0-0.3 is a weak positive correlation,r between 03-0.5 for low positive correlation,r between 0.5-0.8 for moderate positive correlation,r between 0.8-1 for highly positive correlation).The differences of noise,SNR and CNR between two series of CTP images were compared with paired samples t test.The difference of radiation doses were compared with paired samples t test.The maximum value,minimum value,quartile value,mean value of parameters(ADC,D,F,D*for DWI and BF,BV,MTT,PS,HAF for CTP)were compared in active tumor region and in normal hepatic parenchyma using paired samples t test.The statistical comparison of the mean values of all functional parameters at active region of metastasis foci used 2 independent samples t test between the group of positive treatment response and the group of negative treatment response.Relationships among DWI and CTP parameters were researched by Pearson correlation analysis.Receiver operating curve(ROC)was used to screen the valuable DWI or CTP parameters in predicting treatment response before treatment and then to find out the cut-points of every parameters according to Youden's index(to determine the parameter optimal value,namely sensitivity+specificity-1,the index value of the maximum value is the best field values).The area under the ROC curve(AUC)values is range 1-0.5.The closer to 1 AUC is,indicating better diagnosis effect:AUC in 0.5-0.7 with low accuracy,AUC in 0.7-0.9 with moderate accuracy,AUC more than 0.9 with higher accuracy.AUC=0.5,diagnostic method completely does not work,there is no diagnostic value.AUC<0.5 does not conform to the actual situation.Statistical analyses were performed using SPSS(v.20.0;SPSS,Chicago,IL).A P value of less than 0.05 was considered indicative of a statistically significant difference.Results100%of IVIM-DW images satisfied diagnostic quality and measured parameters taking 7-12 minutes acquisition time.The repeatability of measurement of DWI parameters had different degrees.D was high level(r=0.888,p<0.001),ADC was moderate level(r=0.543,p=0.005),D*did not show the repeatability of measurement(r=0.134,p=0.522)in normal hepatic parenchyma.While the repeatability of measurements of ADC,D,F and D*were all moderate level(r=0.797 p<0.001,r=0.621,p=0.0015;r=0.604,p=0.001 and r=0.545,p-0.005,respectively)in active tumor region.The mean values and quartile values of ADC,D and F were all significant lower in active tumor region than those values in normal parenchyma(mean:ADC:92.6130.3 vs.119.1 ±23.6×10-1 mm2/s p<0.001,D:83.8±30.5 vs.104.4±23.0×10-1 mm2/s p<0.001,F:8.2±6.5 vs.19.2±8.0 p<0.001),D*values was higher in active tumor region than those values in normal parenchyma,the difference was significant(14437.4±7181.1 vs.17007.3±3633.0 ×10-1 1mm2/s p=0.027),while 1/4 quartile value and median value were all lower in active tumor region than those values in normal parenchyma.Considering 1/4 quartile value was out of 95%confidence interval,variation of D*was large.Significant lower values of ADC and D in active tumor region were measured in group of positive treatment response compared with those values in group of negative treatment response(ADC:81.5±23.9 vs.102.9±33.6 x10-1 mm2/s p=0.033,D:66.2±29.7 vs.100.0±28.2×10-1 mm2/s p=0.001).While the D*and F values had no significant difference between two groups(D*:14673.5±6891.4 vs.14444.0±7320.7×10-1 mm2/s p=0.922,F:9.9±6.6 vs.7.2±4.9 p=0.151).The mean values of ADC,D and F were all significant lower in active tumor region than those values in normal parenchyma in group of positive treatment response(ADC:81.5±23.9 vs.125.2±15.7×10-1 mm2/s p<0.001,D:66.2±29.7 vs.117.18±17.6×10-1mm2/s p<0.001,F:9.9±6.6 vs.20.3±9.1 p<0.001),showing similar trend which total data manifested.In group of negative treatment response,the differences of ADC and D were not significant between in active tumor region and in normal parenchyma(p>0.05).At given same scan protocol,noise of image derived of iterative algorithm on idose4 level was significant lower compared with that of filter back-projection algorithm(tumor:12.58±2.21(HU)vs.19.4±5.14(HU)p<0.001,normal parenchyma:10.91 ±2.78(HU)vs.17.49±5.34(HU)p<0.001).Signal-noise ratio(SNR)and Contrast noise ratio(CNR)of the former were significant higher than the latter(SNR:4.45±2.75 vs.3.0±1.96 p<0.001,CNR:3.67±1.95 vs.2.47±1.39 p<0.001).When given 120kv,40mA non-jog axial maximal coverage scan protocol,the radiation dose of CTP was significant lower than tri-phase enhance spiral scan coverage of liver given auto tuber current modulation at 120kv(dose length product(DLP):863.8± 1.8 vs.970.6±294.6 mGy.cm p=0.028).Based on two different image reconstruction algorithms,the time-density curves of regions of interesting(ROI)at aorta or at portal vein were similar.The values of computed CTP parameters had high level or moderate consistency.BF,BV,MTT in active tumor region and HAF in normal parenchyma had good consistency(r=0.889,0.841,0.861 and 0.82,respectively)and PS,HAF in active tumor region and BF,BV,MTT,PS in normal parenchyma had moderate consistency(r=0.759,0.547,0.768,0.752,0.706 and 0.508 respectively).The repeatability of measurement of CTP parameters were satisfied,The repeatability of BF in active tumor region and BF,BV in normal parenchyma is excellent(r=0.852,0.9 and 0.81,respectively)had good consistency(r=0.889,0.841,0.861 and 0.82,respectively)and the moderate repeatability of BV,MTT,PS in active tumor region and MTT,PS,HAF in normal parenchyma were obtained(r=0.786,0.735,0.798,0.763,0.552 and 0.704,respectively).HAF in active tumor region had weak repeatability(r=0.443).The mean values of BV,PS and MTT were all significant lower in active tumor region than those values in normal parenchyma(BV:14.6±21.6 vs.19.8±26.7 ml/100g p=0.018,PS:20.8±18 vs.40.9±18.5 ml/min/100g,p<0.001,MTT:9.3±5.9 vs.12.5±18 s,p<0.001),HAF values was higher in active tumor region than those values in normal parenchyma,the difference was significant(0.48±0.21 vs.0.18±0.15,p<0.001).There were overlap for BF both in active tumor region and in normal parenchyma,the difference had no significance(143.2±107.2 vs.145.9±166.8 ml/min/100g,p=0.909).Significant lower values of BF,BV and PS in active tumor region were measured in group of positive treatment response compared with those values in group of negative treatment response(BF:77.95±51.7 vs.168.9± 109.1 ml/min/100g p=0.002,BV:6.9±4.6 vs.13.8±7.5 ml/100g p=0.001,PS:16.3±15.9 vs.29.2±21.6 ml/min/100g,p=0.042).While the MTT values was longer,but had no significant difference between two groups(11.3±7.94 vs.7.5±3.8 s,p=0.065).No significant difference had been found for HAF,too(0.41±0.22 vs.0.49±0.21,p=0.203).The difference of BV was significant between in active tumor region and in normal parenchyma within group of positive treatment response(6.9±4.6 vs.14.5±9.4 ml/100g,p=0.005),but no difference had been found within group of negative treatment response(13.8±7.5 vs.16.1±6.8 ml/lOOg,p=0.314).The differences of MTT,PS,HAF were significant between in active tumor region and in normal parenchyma within each group(MTT:11.317.94 vs.15.9±8.2 s p=0.024,PS:16.3±15.9 vs.38±21.4 ml/min/100g p=0.002,HAF:0.41±0.22 vs.0.19±0.17 p=0.001 in positive group.MTT:7.5±3.8 vs.11.0±5.0 s p=0.015,PS:29.2±21.6 vs.39.9±15.0 ml/min/100g p=0.046,HAF:0.49±0.21 vs.0.15±0.12 p<0.001 in negative group).While BF did not showed difference within each group(77.95±51.7 vs.113.7±143.3 ml/min/100g p=0.31 in positive group,168.9±109.1vs.141.6±109.3 ml/min/100g p=0.396 in negative group).In active tumor region,for within DWI parameters,D had significant positive correlations with ADC at a high level(r=0.827 p<0.001).F had significant negative correlations with D*at a low level(r=-0.329 p=0.018).No correlations had been found for other DWI parameters(p>0.05).For between DWI parameters and CTP parameters,ADC had significant positive correlations with BF at very low level(r=0.292 p=0.037)and with PS at low level(r=0.462 p=0.001).D had significant positive correlations with both BF and BV at low level(BF:r=0.312 p=0.026,BV:r=0.307 p=0.028)and with PS at moderate level(r=0.501 p<0.001).For within CTP parameters,BV had significant positive correlations with BF at moderate level(r=0.749 p<0.001),with PS at low level(r=0.316 p=0.024)and with HAF at very low level(r=0.292 p=0.037).BV and MTT were not correlative(p>0.05).BF had significant negative correlations with MTT at a low level(r=-0.489 p<0.001),had significant positive correlations with PS at very low level(r=0.278 p=0.048),with HAF at a low level(r=0.404 p=0.003).No correlations had been found for other CTP parameters(p>0.05).The parameters BV,BF,PS,D had ability of predicting treatment response before treatment on hepatic metastatic tumor(p<0.05),especially BV,BF and D showed a medium degree diagnostic value.The areas under curve(AUC)of BV,BF,PS and D were 0.8?0.763?0.795 and 0.697,respectively.The parameters ADC,F,D*,HAF and MTT did not displayed the ability of predicting treatment response(p>0.05).The lower the BV and BF were,the better treatment response was*In contrast,the lower the D was,the more poor treatment response was.The cut-points of BV,BF,PS and D for predicting treatment response according Youden's index method were 8.15 ml/100g,169.45 ml/min/100g,19.7 ml/min/100g,71.44x10-1 mm2/s,respectively;and the corresponding sensibilities and specificities70%and 84.2%,40%and 100%,65%and 73.7%,94.7%and 61.1%,respectively.When BV and D at cut-points(BV=8.15ml/100g,D=71.44×10-1 mm2/s)were combined in parallei test,the sensibilities and specificities were 98.4%and 51.4%,while in series test,66.3%and 93.9%.When BV and BF at cut-points(BV=8.15ml/100g,BF=169.45ml/min/100g)were combined in parallel test,the sensibilities and specificities were 82%and 84.2%,while in series test,28%and 100%.When D and BF at cut-points(D=71.44×10-1 mm2/s,BF=169.45ml/min/100g)were combined in parallel test,the sensibilities and specificities were 96.8%and 61.1%,while in series test,37.9%and 100%.When BV and BF combined in parallel test were combined D in series test at cut-points(BV=8.15ml/100g,BF=169.45 ml/min/100g,D=71.44×10-1mm2/s),the sensibilities and specificities were 77.7%and 93.9%.ConclusionIVIM-DWI can acquire good image quality to measure DWI parameters using single-shot spin-echo echo planar imaging(SS-SE-EPI)series combining spectral pre-saturation inversion recovery(SPIR),sensitivity encoding(SENSE)and respiratory-triggered motion compensation techniques in an acceptable scan time.The repeatability of measurement of parameters of IVIM-DWI was medium degree.The values of ADC,D and F were lower in active tumor region than those values in normal parenchyma.Significant lower values of ADC and D in active tumor region were measured in group of positive treatment response compared with those values in group of negative treatment response.The parameter D had a medium degree ability of predicting treatment response before treatment on hepatic metastatic tumor.The reliability and feasibility of CTP using low-dose(low tube current)protocol were good on hepatic metastatic tumor.The iterative reconstruction algorithm has potency to decrease the radiation dose while keeping a diagnostic image quality and the different image reconstruction algorithms(iterative algorithm on idose4 level and filter back-projection algorithm)had no parenchymal effect on measurements of CTP parameters.The values of BV,MTT and PS were lower in active tumor region than those values in normal parenchyma;conversely,HAF was significantly higher.Significant lower values of BV,BF and PS in active tumor region were measured in group of positive treatment response compared with those values in group of negative treatment response.The parameters including BV,BF and PS,respectively,had a medium degree ability of predicting treatment response before treatment on hepatic metastatic tumor.ADC and D showed significant positive correlations with high degree.The significant correlations with low levels were observed between DWI parameters(ADC and D)and CTP parameters(BV,BF and PS).Combining multi-parameters was favorable to satisfy clinical needs for the sensibilities and specificities of predicting treatment response on hepatic metastatic tumor. |
PDF Full Text Request