Reproducibility of intravoxel incoherent motion of liver on a 3.0T scanner: free-breathing and respiratory-triggered sequences acquired with different numbers of excitations

Abstract

PurposeTo optimise the intravoxel incoherent motion (IVIM) imaging of the liver on a 3.0T scanner by assessing parameter reproducibility on free-breathing (FB) and respiratory-triggered (RT) sequences acquired with different numbers of signal averages (NSA).Material and methodsIn this prospective study 20 subjects (M/F: 10/10; age: 25-62 years, mean: 39 years) underwent IVIM magnetic resonance imaging (MRI) on a 3.0T scanner using an 18-channel phase-arrayed coil and four different echo-planar sequences, each with 10 b values: 0, 10, 30, 50, 75, 100, 150, 200, 500, and 900 s/mm2. Images were acquired with FB and RT with NSA = 1-4 (FBNSA1-4, RTNSA1-4) and with NSA = 3-6 (FBNSA3-6, RTNSA3-6). Subsequently, for the assessment of reproducibility of IVIM-derived parameters (f, D, D*), each subject was scanned again with an identical protocol during the same session. IVIM parameters were calculated. The distribution of IVIM-parameters for each DWI sequence were given as the median value with first and third quartile. Inter-scan reproducibility for each IVIM parameter was evaluated using coefficient of variance and Bland-Altman difference. Differences between FB sequence and RT sequence were tested using non-parametric Wilcoxon signed-rank test.ResultsMean coefficient of variance (%) for f, D, and D* ranged from 60 to 64, from 58 to 84, and from 82 to 99 for FBNSA1-4 sequence; from 50 to 69, from 41 to 97, and from 80 to 82 for RTNSA1-4 sequence; from 22 to 27, 15, and from 70 to 80 for FBNSA3-6 sequence; and from 21 to 32, from 12 to, and from 50 to 80 for RTNSA3-6 sequence, respectively.ConclusionsIncreasing the number of signal averages for IVIM acquisitions allows us to improve the reproducibility of IVIM-derived parameters. The sequence acquired during free-breathing with NSA = 3-6 was optimal in terms of reproducibility and acquisition time.

An optimized b‐value distribution for triexponential intravoxel incoherent motion (IVIM) in the liver

“…For the biexponential IVIM model, b‐value optimizations were already performed for different organs and purposes to increase IVIM parameter reproducibility 6,31‐34 . Besides the optimization of the b‐value sampling, lots of other optimizations for the biexponential IVIM model in the liver, for example, concerning breathing schemes or minimizing the acquisition time were performed 34‐39 . However, to our knowledge, no comparable optimization has been performed for triexponential IVIM yet.…”

“…6,[31][32][33][34] Besides the optimization of the b-value sampling, lots of other optimizations for the biexponential IVIM model in the liver, for example, concerning breathing schemes or minimizing the acquisition time were performed. [34][35][36][37][38][39] However, to our knowledge, no comparable optimization has been performed for triexponential IVIM yet. Within this study, we were able to find an optimized distribution of b-values for triexponential IVIM in the liver and demonstrated the advantage of optimized b-values compared to 3 representative b-value distributions used in literature by means of simulations and volunteer measurements.…”

Intravoxel incoherent motion magnetic resonance imaging: basic principles and clinical applications

“…However, the main problem with implementation of IVIM for routine MR imaging is lack of standardisation of both image acquisition and analysis, resulting in significant variance in calculated parameters among studies [ 75 , 76 ] ( Table 4 ). Thus, this technique still needs to be refined in terms of acquisition, analysis, and application.…”

Comparative Study of Amide Proton Transfer Imaging and Intravoxel Incoherent Motion Imaging for Predicting Histologic Grade of Hepatocellular Carcinoma

“…In addition, the IVIM and APT images were acquired with free breathing, resulting in decreased signal-tonoise ratio on parameter maps. However, the free-breathing protocol was recommended in several studies (30,31) because of its good reproducibility and shorter acquisition time compared with that of respiratory-triggered and breath-hold imaging. Third, the freehand ROI analysis could produce definite artificial errors, which might affect the accuracy of the values of those quantitative imaging parameters.…”