RESUMEN
OBJECTIVES: Quantitative synthetic magnetic resonance imaging (MRI) enables synthesis of various contrast-weighted images as well as simultaneous quantification of T1 and T2 relaxation times and proton density. However, to date, it has been challenging to generate magnetic resonance angiography (MRA) images with synthetic MRI. The purpose of this study was to develop a deep learning algorithm to generate MRA images based on 3D synthetic MRI raw data. MATERIALS AND METHODS: Eleven healthy volunteers and 4 patients with intracranial aneurysms were included in this study. All participants underwent a time-of-flight (TOF) MRA sequence and a 3D-QALAS synthetic MRI sequence. The 3D-QALAS sequence acquires 5 raw images, which were used as the input for a deep learning network. The input was converted to its corresponding MRA images by a combination of a single-convolution and a U-net model with a 5-fold cross-validation, which were then compared with a simple linear combination model. Image quality was evaluated by calculating the peak signal-to-noise ratio (PSNR), structural similarity index measurements (SSIMs), and high frequency error norm (HFEN). These calculations were performed for deep learning MRA (DL-MRA) and linear combination MRA (linear-MR), relative to TOF-MRA, and compared with each other using a nonparametric Wilcoxon signed-rank test. Overall image quality and branch visualization, each scored on a 5-point Likert scale, were blindly and independently rated by 2 board-certified radiologists. RESULTS: Deep learning MRA was successfully obtained in all subjects. The mean PSNR, SSIM, and HFEN of the DL-MRA were significantly higher, higher, and lower, respectively, than those of the linear-MRA (PSNR, 35.3 ± 0.5 vs 34.0 ± 0.5, P < 0.001; SSIM, 0.93 ± 0.02 vs 0.82 ± 0.02, P < 0.001; HFEN, 0.61 ± 0.08 vs 0.86 ± 0.05, P < 0.001). The overall image quality of the DL-MRA was comparable to that of TOF-MRA (4.2 ± 0.7 vs 4.4 ± 0.7, P = 0.99), and both types of images were superior to that of linear-MRA (1.5 ± 0.6, for both P < 0.001). No significant differences were identified between DL-MRA and TOF-MRA in the branch visibility of intracranial arteries, except for ophthalmic artery (1.2 ± 0.5 vs 2.3 ± 1.2, P < 0.001). CONCLUSIONS: Magnetic resonance angiography generated by deep learning from 3D synthetic MRI data visualized major intracranial arteries as effectively as TOF-MRA, with inherently aligned quantitative maps and multiple contrast-weighted images. Our proposed algorithm may be useful as a screening tool for intracranial aneurysms without requiring additional scanning time.
Asunto(s)
Interpretación de Imagen Asistida por Computador/métodos , Imagenología Tridimensional/métodos , Aneurisma Intracraneal/diagnóstico por imagen , Angiografía por Resonancia Magnética/métodos , Adulto , Algoritmos , Aprendizaje Profundo , Femenino , Humanos , Masculino , Relación Señal-Ruido , Adulto JovenRESUMEN
PURPOSE: Gadolinium-based contrast agents (GBCA) provide valuable information for assessing and differentiating lesions in the body. However, contrast enhancement evaluation on conventional MRI is qualitative because the signal intensity uses an arbitrary scale. An approach that allows more quantitative assessment of tissue enhancement that can be integrated into clinical use is desirable. This study aimed to provide a method that can estimate GBCA concentration in a clinically applicable scan-time. METHODS: Gadolinium-based contrast agent concentrations were quantified in phantoms containing water and nine different concentrations of Gadoteridol (Gd-HP-DO3A), ranging from 0.02 to 1.00 mmol/L, using quantitative synthetic MRI. Simple linear regression analysis between the estimated GBCA concentration and the reference values were performed to assess the accuracy. We performed region of interest analysis on each phantom, and recorded the mean and standard deviation. We evaluated the precision of the GBCA map by calculating the coefficient of variation (CV) for each concentration. The GBCA concentration quantification method was applied for 10 patients with metastatic brain tumors to demonstrate the feasibility of this method. RESULTS: For the phantom study, estimated GBCA concentrations were in a strong linear relationship (R2 = 0.998) with reference values, with a slope and intercept on simple linear regression analysis of 0.98 and 0.02, respectively. On precision assessment, the CV was <5%, with the exception of concentrations under 0.07 mmol/L. In the range of 0.07-0.99 mmol/L, the mean CV was 1.5 ± 1.2%. For application to brain metastases, the maximum estimated GBCA concentration in the metastases was 0.73 mmol/L, which was under the upper limit evaluated in the phantom study (i.e. -0.99 mmol/L). CONCLUSION: The concentration of Gd-HP-DO3A in the range of 0.07-0.99 mmol/L can be measured in a clinically applicable scan time using quantitative synthetic MRI, even though this study's results are only preliminary due to several limitations.
Asunto(s)
Neoplasias Encefálicas/diagnóstico por imagen , Medios de Contraste , Gadolinio , Imagen por Resonancia Magnética/métodos , Neoplasias Encefálicas/secundario , Estudios de Factibilidad , Humanos , Fantasmas de ImagenRESUMEN
The purpose of this study was to show the efficacy of dynamic field correction (DFC), a technique provided by the scanner software, in comparison to the FMRIB Software Library (FSL) post-processing "eddy" tool. DFC requires minimal additional scan time for the correction of eddy-current and motion-induced geometrical image distortions in diffusion-weighted echo-planar images. The fractional anisotropy derived from images corrected with DFC were comparable to images corrected with the "eddy" tool and significantly higher than images without correction, which demonstrates the utility of DFC.