Artificial neural network for suppression of banding artifacts in balanced steady-state free precession MRI

Abstract

The balanced steady-state free precession (bSSFP) MR sequence is frequently used in clinics, but is sensitive to off-resonance effects, which can cause banding artifacts. Often multiple bSSFP datasets are acquired at different phase cycling (PC) angles and then combined in a special way for banding artifact suppression. Many strategies of combining the datasets have been suggested for banding artifact suppression, but there are still limitations in their performance, especially when the number of phase-cycled bSSFP datasets is small. The purpose of this study is to develop a learning-based model to combine the multiple phase-cycled bSSFP datasets for better banding artifact suppression. Multilayer perceptron (MLP) is a feedforward artificial neural network consisting of three layers of input, hidden, and output layers. MLP models were trained by input bSSFP datasets acquired from human brain and knee at 3T, which were separately performed for two and four PC angles. Banding-free bSSFP images were generated by maximum-intensity projection (MIP) of 8 or 12 phase-cycled datasets and were used as targets for training the output layer. The trained MLP models were applied to another brain and knee datasets acquired with different scan parameters and also to multiple phase-cycled bSSFP functional MRI datasets acquired on rat brain at 9.4T, in comparison with the conventional MIP method. Simulations were also performed to validate the MLP approach. Both the simulations and human experiments demonstrated that MLP suppressed banding artifacts significantly, superior to MIP in both banding artifact suppression and SNR efficiency. MLP demonstrated superior performance over MIP for the 9.4T fMRI data as well, which was not used for training the models, while visually preserving the fMRI maps very well. Artificial neural network is a promising technique for combining multiple phase-cycled bSSFP datasets for banding artifact suppression.

Introduction

Balanced steady state free precession (bSSFP) sequence has been gaining importance in clinical practices due to its high speed, high signal-to-noise ratio (SNR), and minimal distortion. However, bSSFP often suffers from well-known banding artifacts, which are related to the field strength and repetition time (TR). Good shimming is also essential to reduce banding artifacts. Unfortunately, even with optimized scan parameters and good shimming, banding artifacts are frequently unavoidable.

In bSSFP, sum of gradients along each of the three directions (i.e., readout, phase encoding, slice selection) during one TR equals zero [1], [2], indicating no phase contributions of the gradients. The net phase of a proton spin after each TR is therefore affected only by (i) phase evolution per TR due to magnetic field inhomogeneity and (ii) increment in transmission RF phase per TR called “phase cycling” (PC) angle. Typically the former is minimized by using a minimum possible TR and the best shimming condition, while the latter is set to 180° for the residual transverse magnetization from one RF excitation to be used for the following RF excitation in a balanced way. When sum of the two phase terms corresponds to 0° (or integer multiple of 360°) rather than 180°, offset of the RF pulses relative to the precession angle is zero, leading to banding artifacts. Banding artifacts can spatially shift with PC angle, which can be used as a method to avoid banding artifacts in a specific region of interest [3], [4], [5], [6].

Various methods have been introduced to suppress banding artifacts in bSSFP [7], [8], [9], [10], [11], [12], [13], [14], [15], [16]. The most common way is to use multiple bSSFP acquisitions at various PC angles and combine them in a special way, including maximum-intensity projection (MIP) [7], complex sum [8], sum-of-square (SOS) [9], nonlinear averaging [11], sum of free induction decay and primary echo components from Fourier analysis [15], geometric solution [13], linearization with Gauss-Newton algorithm [12], and geometric-algebraic solution [14]. However, banding artifacts cannot be suppressed completely by these methods, especially when number of multiple PC bSSFP datasets is small.

In this study, we proposed a learning-based method to combine the multiple PC bSSFP datasets for better banding artifact suppression. Multilayer perceptron (MLP) is a feedforward artificial neural network consisting of three layers of input, hidden, and output layers. MLP models were trained by input bSSFP datasets acquired from human brain and knee at 3 T, which were separately performed for two and four PC angles. Banding-free bSSFP images were generated by maximum-intensity projection (MIP) of 8 or 12 phase-cycled datasets and were used as targets for training the output layer. The MLP models were applied to another brain and knee datasets acquired with different scan parameters and also to multiple phase-cycled bSSFP functional MRI datasets acquired on rat brain at 9.4 T, in comparison with the conventional MIP method, which has been widely used [11].