The Study Of Image Quality And Repeatability Of Apparent Diffusion Coefficient Measurements On Liverdiffusion-Weighted Magnetic Resonance Imaging Based On Gating Techniques

ObjectiveThe purpose of this research was to evaluate the signal-to-noise ratio and reproducibility of apparent diffusion coefficient measurements on liver diffusion-weighted magnetic resonance imaging obtained with different techniques, which included respiratory-triggered, breath-hold, free-breath and electrocardiograph-triggered with different triggered delay time (Oms, 200ms,400ms and 600ms), and to discuss the clinical optimized scanning plan with gating technology in liver DWI scan.MethodsExperiment one:Data were collected from 36 young healthy volunteers from April to June 2014. Each volunteer underwent liver diffusion-weighted magnetic resonance imaging twice with respiratory-triggered, breath-hold, free-breath and electrocardiograph-triggered with different triggered delay time (Oms, 200ms,400ms and 600ms). The average ADC image was determined by linear regression analysis with a single exponential model as b was 500 and 0 s/mm2. Three-median-layer images were chosen. The left and right livers were divided into superior, medianand inferior parts, with a 40mm2 region of interest (ROI) placed in each part. The left and right ADC measurements were the means of nine parts ADC measurements.Experiment two:Data were collected from 10 young healthy volunteers from October to November 2015. Each volunteer underwent DW-MRI with respiratory-triggered, breath-hold, free-breath and electrocardiograph triggered with different triggered delay time (Oms,200ms,400ms and 600ms). We got 30 images in the same layer which was shown the biggest left and right lobes. Three 100mm2area ROIs were placed in the backgrounds, left and right lobes, respectively. The ROIs of left and right lobes were calculated signal intensity and the ROIs of backgrounds were calculated standard deviation. The signal-to-noise ratio was calculated. The mean and standard deviation of SNR were done a statistical analysis.Experiment three:Data were collectedat the same time of experiment two. In a similar acquisition time (3 minutes), each volunteer underwent DW-MRI twice with respiratory-triggered, free-breath and electrocardiograph triggered with different triggered delay time (Oms,200ms,400ms and 600ms) and different time of average (NEX). The average ADC image was determined by linear regression analysis with a single exponential model as b was 500 and 50 s/mm2. Three-layer images were chosen. Three 100mm2 area ROIs placed in the left and right lobes were calculated the ADC measurements. The left and right ADC measurements were the means of nine ADC measurements.All data above were statistically analyzed by SPSS19.0 and Medcalc software. We used paired t test to compare the ADCs and SNR of the left and right liver lobes. An analysis of variance was used on the ADCand SNR obtained with the different gating technology. The Bland-Altman method was used to evaluate the reproducibility of the ADC values.ResultsExperiment one:(1) There was no difference in heart rates and breath between two times of liver diffusion-weighted magnetic resonance imaging which the volunteers underwent (P>0.05).(2) The ADC measurements of left robe were significantly higher than the right lobe in DW-MRI with respiratory-triggered, breath-hold, free-breath and electrocardiograph triggered with different triggered delay time (Oms, 200ms,400ms and 600ms) (P<0.001)(3) The ADC measurements increased in RT and ECG-T (P<0.001) technologies. In different triggered time, the ADC measurements were similar(Left: P=0.819-1.000; Right:P=1.000)(4) In coronal liver scanning, repeatability of ADC measurements obtained by BH was better than the other technologies [Left:0.08(0.385); Right:0.02 (0.210)].Experiment two:(1) The SNR of left robes were higher than the right lobes [No statisticallydifference between the left and right lobes with BH(P=0.059)、 ECG-T with 400ms triggered delay (P=0.153) and ECG-T with 600ms triggered delay time (P=0.210). but the other sequences were significant difference P<0.001)].(2) The SNR increased in DWI with RT and ECG-T (P<0.001).The undulation of SNR with RT was significantly obvious than ECG-T. As the triggered delay time increased, the SNR would improve. The highest SNR was obtained by 600ms triggered delay time.Experiment three:(1) There was no difference in heart rates and breath between two times of liver diffusion-weighted magnetic resonance imaging which the volunteers underwent (P>0.05)(2) The ADC measurements of left robe were significantly higher than the right lobe in DW-MRI with respiratory-triggered, breath-hold, free-breath and electrocardiograph triggered with different triggered delay time (Oms, 200ms,400ms and 600ms) (P<0.001)(3) In the same triggered delay time, the ADC measurements with different NEX were no significant difference (P=0.038-0.982);but the repeatability of ADC measurements improved [In 600ms triggered delay time,1NEX Left:-0.02 (0.490), Right:0.01 (0.355); 4NEX Left:0.02 (0.265), Right:0.02(0.090)].(4) The repeatability of ADC measurements obtained by FB with 10 NEX was the best repeatability[Left:-0.02 (0.115); Right:0.01 (0.085)].Conclusion(1) In liver DW-MRI, the SNR and image quality could be improved by RT and ECG, but the undulation ECG-T was stable. The result indicated that the DWI with ECG in 600ms triggered delay time provided well SNR of liver images and clearstructure of anatomy.(2) The gating technologies influence the ADC measurements. But in a short acquisition time, FB with incremental NEX would improve the repeatability of ADC measurements.