Deep learning-accelerated T2-weighted imaging versus conventional T2-weighted imaging in the female pelvic cavity: image quality and diagnostic performance.

BACKGROUND: The deep learning (DL)-based reconstruction algorithm reduces noise in magnetic resonance imaging (MRI), thereby enabling faster MRI acquisition. PURPOSE: To compare the image quality and diagnostic performance of conventional turbo spin-echo (TSE) T2-weighted (T2W) imaging with DL-accelerated sagittal T2W imaging in the female pelvic cavity. METHODS: This study evaluated 149 consecutive female pelvic MRI examinations, including conventional T2W imaging with TSE (acquisition time = 2:59) and DL-accelerated T2W imaging with breath hold (DL-BH) (1:05 [0:14 × 3 breath-holds]) in the sagittal plane. In 294 randomly ordered sagittal T2W images, two radiologists independently assessed image quality (sharpness, subjective noise, artifacts, and overall image quality), made a diagnosis for uterine leiomyomas, and scored diagnostic confidence. For the uterus and piriformis muscle, quantitative imaging analysis was also performed. Wilcoxon signed rank tests were used to compare the two sets of T2W images. RESULTS: In the qualitative analysis, DL-BH showed similar or significantly higher scores for all features than conventional T2W imaging (P <0.05). In the quantitative analysis, the noise in the uterus was lower in DL-BH, but the noise in the muscle was lower in conventional T2W imaging. In the uterus and muscle, the signal-to-noise ratio was significantly lower in DL-BH than in conventional T2W imaging (P <0.001). The diagnostic performance of the two sets of T2W images was not different for uterine leiomyoma. CONCLUSIONS: DL-accelerated sagittal T2W imaging obtained with three breath-holds demonstrated superior or comparable image quality to conventional T2W imaging with no significant difference in diagnostic performance for uterine leiomyomas.

Deep Learning Reconstruction for Accelerated Spine MRI: Prospective Analysis of Interchangeability.

Background Deep learning (DL)-based MRI reconstructions can reduce examination times for turbo spin-echo (TSE) acquisitions. Studies that prospectively employ DL-based reconstructions of rapidly acquired, undersampled spine MRI are needed. Purpose To investigate the diagnostic interchangeability of an unrolled DL-reconstructed TSE (hereafter, TSEDL) T1- and T2-weighted acquisition method with standard TSE and to test their impact on acquisition time, image quality, and diagnostic confidence. Materials and Methods This prospective single-center study included participants with various spinal abnormalities who gave written consent from November 2020 to July 2021. Each participant underwent two MRI examinations: standard fully sampled T1- and T2-weighted TSE acquisitions (reference standard) and prospectively undersampled TSEDL acquisitions with threefold and fourfold acceleration. Image evaluation was performed by five readers. Interchangeability analysis and an image quality-based analysis were used to compare the TSE and TSEDL images. Acquisition time and diagnostic confidence were also compared. Interchangeability was tested using the individual equivalence index regarding various degenerative and nondegenerative entities, which were analyzed on each vertebra and defined as discordant clinical judgments of less than 5%. Interreader and intrareader agreement and concordance (κ and Kendall τ and W statistics) were computed and Wilcoxon and McNemar tests were used. Results Overall, 50 participants were evaluated (mean age, 46 years ± 18 [SD]; 26 men). The TSEDL method enabled up to a 70% reduction in total acquisition time (100 seconds for TSEDL vs 328 seconds for TSE, P < .001). All individual equivalence indexes were less than 4%. TSEDL acquisition was rated as having superior image noise by all readers (P < .001). No evidence of a difference was found between standard TSE and TSEDL regarding frequency of major findings, overall image quality, or diagnostic confidence. Conclusion The deep learning (DL)-reconstructed turbo spin-echo (TSE) method was found to be interchangeable with standard TSE for detecting various abnormalities of the spine at MRI. DL-reconstructed TSE acquisition provided excellent image quality, with a 70% reduction in examination time. German Clinical Trials Register no. DRKS00023278 © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Hallinan in this issue.

Feasibility of an accelerated 2D-multi-contrast knee MRI protocol using deep-learning image reconstruction: a prospective intraindividual comparison with a standard MRI protocol.

OBJECTIVES: The aim of this study was to evaluate the image quality and diagnostic performance of a deep-learning (DL)-accelerated two-dimensional (2D) turbo spin echo (TSE) MRI of the knee at 1.5 and 3 T in clinical routine in comparison to standard MRI. MATERIAL AND METHODS: Sixty participants, who underwent knee MRI at 1.5 and 3 T between October/2020 and March/2021 with a protocol using standard 2D-TSE (TSES) and DL-accelerated 2D-TSE sequences (TSEDL), were enrolled in this prospective institutional review board-approved study. Three radiologists assessed the sequences regarding structural abnormalities and evaluated the images concerning overall image quality, artifacts, noise, sharpness, subjective signal-to-noise ratio, and diagnostic confidence using a Likert scale (1-5, 5 = best). RESULTS: Overall image quality for TSEDL was rated to be excellent (median 5, IQR 4-5), significantly higher compared to TSES (median 5, IQR 4 - 5, p < 0.05), showing significantly lower extents of noise and improved sharpness (p < 0.001). Inter- and intra-reader agreement was almost perfect (κ = 0.92-1.00) for the detection of internal derangement and substantial to almost perfect (κ = 0.58-0.98) for the assessment of cartilage defects. No difference was found concerning the detection of bone marrow edema and fractures. The diagnostic confidence of TSEDL was rated to be comparable to that of TSES (median 5, IQR 5-5, p > 0.05). Time of acquisition could be reduced to 6:11 min using TSEDL compared to 11:56 min for a protocol using TSES. CONCLUSION: TSEDL of the knee is clinically feasible, showing excellent image quality and equivalent diagnostic performance compared to TSES, reducing the acquisition time about 50%. KEY POINTS: • Deep-learning reconstructed TSE imaging is able to almost halve the acquisition time of a three-plane knee MRI with proton density and T1-weighted images, from 11:56 min to 6:11 min at 3 T. • Deep-learning reconstructed TSE imaging of the knee provided significant improvement of noise levels (p < 0.001), providing higher image quality (p < 0.05) compared to conventional TSE imaging. • Deep-learning reconstructed TSE imaging of the knee had similar diagnostic performance for internal derangement of the knee compared to standard TSE.

Deep learning-accelerated T2-weighted imaging of the prostate: Impact of further acceleration with lower spatial resolution on image quality.

PURPOSE: To compare image quality in prostate MRI among standard T2-weighted imaging (T2-std), accelerated T2-weighted imaging (T2WI) with high resolution (T2-HR) and more accelerated T2WI with lower resolution (T2-LR) using both conventional reconstruction (C) and deep learning reconstruction (DL). MATERIALS AND METHODS: In 46 consecutive patients, T2-std, T2-HR and T2-LR were acquired in 3:32 min, 1:06 min and 0.52 min, respectively. Both reconstruction techniques (C and DL) were applied to T2-HR and T2-LR. Five sets of images (T2-std, T2-HRC, T2-LRC, T2-HRDL, and T2-LRDL) for each patient were independently evaluated by two radiologists. Quantitative analysis including the signal-to-noise ratio (SNR) and contrast ratio (CR) and qualitative analysis with a 5-point scale for the sharpness of structures, ghosting or other artifacts, noise and overall image quality were performed. RESULTS: The SNR was not different in either the peripheral zone (PZ) or transition zone (TZ) between T2-LRDL and T2-std with the median value of 21.7 versus 22.6 in PZ and 16.5 versus 17.3 in TZ, respectively. The CR between the prostate gland and muscle was significantly lower on T2-HRC and T2-LRC than on T2-std. Most of the evaluated factors showed significantly lower scores on T2-HRC and T2-LRC than on T2-std. Although noise and overall image quality on T2-HRDL and other artifacts on T2-LRDL were rated significantly lower than on T2-std (median value 4.0 versus 4.5, P < 0.001; 4.5 versus 5.0, P = 0.001; 4.5 versus 5.0, P = 0.006, respectively), other factors did not differ between T2-std and T2-HRDL or T2-LRDL. CONCLUSION: DL is useful to improve image quality in accelerated T2WI of the prostate gland. Using DL, accelerated T2WI with lower spatial resolution than T2-std can be achieved with similar image quality in much shorter scan time (75.5% reduction in the acquisition time).

Feasibility and Implementation of a Deep Learning MR Reconstruction for TSE Sequences in Musculoskeletal Imaging.

Magnetic Resonance Imaging (MRI) of the musculoskeletal system is one of the most common examinations in clinical routine. The application of Deep Learning (DL) reconstruction for MRI is increasingly gaining attention due to its potential to improve the image quality and reduce the acquisition time simultaneously. However, the technology has not yet been implemented in clinical routine for turbo spin echo (TSE) sequences in musculoskeletal imaging. The aim of this study was therefore to assess the technical feasibility and evaluate the image quality. Sixty examinations of knee, hip, ankle, shoulder, hand, and lumbar spine in healthy volunteers at 3 T were included in this prospective, internal-review-board-approved study. Conventional (TSES) and DL-based TSE sequences (TSEDL) were compared regarding image quality, anatomical structures, and diagnostic confidence. Overall image quality was rated to be excellent, with a significant improvement in edge sharpness and reduced noise compared to TSES (p < 0.001). No difference was found concerning the extent of artifacts, the delineation of anatomical structures, and the diagnostic confidence comparing TSES and TSEDL (p > 0.05). Therefore, DL image reconstruction for TSE sequences in MSK imaging is feasible, enabling a remarkable time saving (up to 75%), whilst maintaining excellent image quality and diagnostic confidence.

Accelerated T2-Weighted TSE Imaging of the Prostate Using Deep Learning Image Reconstruction: A Prospective Comparison with Standard T2-Weighted TSE Imaging.

Multiparametric MRI (mpMRI) of the prostate has become the standard of care in prostate cancer evaluation. Recently, deep learning image reconstruction (DLR) methods have been introduced with promising results regarding scan acceleration. Therefore, the aim of this study was to investigate the impact of deep learning image reconstruction (DLR) in a shortened acquisition process of T2-weighted TSE imaging, regarding the image quality and diagnostic confidence, as well as PI-RADS and T2 scoring, as compared to standard T2 TSE imaging. Sixty patients undergoing 3T mpMRI for the evaluation of prostate cancer were prospectively enrolled in this institutional review board-approved study between October 2020 and March 2021. After the acquisition of standard T2 TSE imaging (T2S), the novel T2 TSE sequence with DLR (T2DLR) was applied in three planes. Overall, the acquisition time for T2S resulted in 10:21 min versus 3:50 min for T2DLR. The image evaluation was performed by two radiologists independently using a Likert scale ranging from 1-4 (4 best) applying the following criteria: noise levels, artifacts, overall image quality, diagnostic confidence, and lesion conspicuity. Additionally, T2 and PI-RADS scoring were performed. The mean patient age was 69 ± 9 years (range, 49-85 years). The noise levels and the extent of the artifacts were evaluated to be significantly improved in T2DLR versus T2S by both readers (p < 0.05). Overall image quality was also evaluated to be superior in T2DLR versus T2S in all three acquisition planes (p = 0.005-<0.001). Both readers evaluated the item lesion conspicuity to be superior in T2DLR with a median of 4 versus a median of 3 in T2S (p = 0.001 and <0.001, respectively). T2-weighted TSE imaging of the prostate in three planes with an acquisition time reduction of more than 60% including DLR is feasible with a significant improvement of image quality.

Results of the 2020 fastMRI Challenge for Machine Learning MR Image Reconstruction.

Accelerating MRI scans is one of the principal outstanding problems in the MRI research community. Towards this goal, we hosted the second fastMRI competition targeted towards reconstructing MR images with subsampled k-space data. We provided participants with data from 7,299 clinical brain scans (de-identified via a HIPAA-compliant procedure by NYU Langone Health), holding back the fully-sampled data from 894 of these scans for challenge evaluation purposes. In contrast to the 2019 challenge, we focused our radiologist evaluations on pathological assessment in brain images. We also debuted a new Transfer track that required participants to submit models evaluated on MRI scanners from outside the training set. We received 19 submissions from eight different groups. Results showed one team scoring best in both SSIM scores and qualitative radiologist evaluations. We also performed analysis on alternative metrics to mitigate the effects of background noise and collected feedback from the participants to inform future challenges. Lastly, we identify common failure modes across the submissions, highlighting areas of need for future research in the MRI reconstruction community.

Deep learning-accelerated T2-weighted imaging of the prostate: Reduction of acquisition time and improvement of image quality.

PURPOSE: To introduce a novel deep learning (DL) T2-weighted TSE imaging (T2DL) sequence in prostate MRI and investigate its impact on examination time, image quality, diagnostic confidence, and PI-RADS classification compared to standard T2-weighted TSE imaging (T2S). METHOD: Thirty patients who underwent multiparametric MRI (mpMRI) of the prostate due to suspicion of prostatic cancer were included in this retrospective study. Standard sequences were acquired consisting of T1- and T2-weighted imaging and diffusion-weighted imaging as well as the novel T2DL. Axial acquisition time of T2S was 4:37 min compared to 1:38 min of T2DL. Two radiologists independently evaluated all imaging datasets in a blinded reading regarding image quality, lesion detectability, and diagnostic confidence using a Likert-scale ranging from 1 to 4 with 4 being the best. T2 score as well as PI-RADS score were obtained for the most malignant lesion. RESULTS: Mean patient age was 65 ±â€¯11 years. Noise levels and overall image quality were rated significantly superior by both readers with a median of 4 in T2DL compared to a median of 3 in T2S (all p < 0.001). Lesion detectability was also rated higher in T2DL by both readers with a median of 4 versus a median of 3 in T2S (p = 0.005 and <0.001, respectively). There was no difference regarding PI-RADS scoring between T2DL and T2S affecting patient management. CONCLUSIONS: Deep learning axial T2w TSE imaging of the prostate is feasible with reduction of examination time of 65 % compared to standard imaging and improvement of image quality and lesion detectability.

A Novel Method for High-Dimensional Anatomical Mapping of Extra-Axial Cerebrospinal Fluid: Application to the Infant Brain.

Cerebrospinal fluid (CSF) plays an essential role in early postnatal brain development. Extra-axial CSF (EA-CSF) volume, which is characterized by CSF in the subarachnoid space surrounding the brain, is a promising marker in the early detection of young children at risk for neurodevelopmental disorders. Previous studies have focused on global EA-CSF volume across the entire dorsal extent of the brain, and not regionally-specific EA-CSF measurements, because no tools were previously available for extracting local EA-CSF measures suitable for localized cortical surface analysis. In this paper, we propose a novel framework for the localized, cortical surface-based analysis of EA-CSF. The proposed processing framework combines probabilistic brain tissue segmentation, cortical surface reconstruction, and streamline-based local EA-CSF quantification. The quantitative analysis of local EA-CSF was applied to a dataset of typically developing infants with longitudinal MRI scans from 6 to 24 months of age. There was a high degree of consistency in the spatial patterns of local EA-CSF across age using the proposed methods. Statistical analysis of local EA-CSF revealed several novel findings: several regions of the cerebral cortex showed reductions in EA-CSF from 6 to 24 months of age, and specific regions showed higher local EA-CSF in males compared to females. These age-, sex-, and anatomically-specific patterns of local EA-CSF would not have been observed if only a global EA-CSF measure were utilized. The proposed methods are integrated into a freely available, open-source, cross-platform, user-friendly software tool, allowing neuroimaging labs to quantify local extra-axial CSF in their neuroimaging studies to investigate its role in typical and atypical brain development.

Automatic Measurement of Extra-Axial CSF from Infant MRI Data.

The quantification of cerebrospinal fluid (CSF) in the human brain has shown to play an important role in early postnatal brain development. Extra-axial fluid (EA-CSF), which is characterized by CSF in the subarachnoid space, is a promising marker for the early detection of children at risk for neurodevelopmental disorders, such as Autism Spectrum Disorder (ASD). Yet, non-ventricular CSF quantification, in particular extra-axial CSF quantification, is not supported in the major neuro-imaging software solutions, such as FreeSurfer. Most current structural image analysis packages mask out the extra-axial CSF space in one of the first pre-processing steps. A quantitative protocol was previously developed by our group to objectively measure the volume of total EA-CSF volume using a pipeline workflow implemented in a series of python scripts. While this solution worked for our specific lab, a graphical user interface-based tool is necessary to facilitate the computation of extra-axial CSF volume across a wide array of neuroimaging studies and research labs. This paper presents the development of a novel open-source, cross-platform, user-friendly software tool, called Auto-EACSF, for the automatic computation of such extra-axial CSF volume. Auto-EACSF allows neuroimaging labs to quantify extra-axial CSF in their neuroimaging studies in order to investigate its role in normal and atypical brain development.

Role of deep learning in infant brain MRI analysis.

Deep learning algorithms and in particular convolutional networks have shown tremendous success in medical image analysis applications, though relatively few methods have been applied to infant MRI data due numerous inherent challenges such as inhomogenous tissue appearance across the image, considerable image intensity variability across the first year of life, and a low signal to noise setting. This paper presents methods addressing these challenges in two selected applications, specifically infant brain tissue segmentation at the isointense stage and presymptomatic disease prediction in neurodevelopmental disorders. Corresponding methods are reviewed and compared, and open issues are identified, namely low data size restrictions, class imbalance problems, and lack of interpretation of the resulting deep learning solutions. We discuss how existing solutions can be adapted to approach these issues as well as how generative models seem to be a particularly strong contender to address them.

Model selection for spatiotemporal modeling of early childhood sub-cortical development.

Spatiotemporal shape models capture the dynamics of shape change over time and are an essential tool for monitoring and measuring anatomical growth or degeneration. In this paper we evaluate non-parametric shape regression on the challenging problem of modeling early childhood sub-cortical development starting from birth. Due to the flexibility of the model, it can be challenging to choose parameters which lead to a good model fit yet does not over fit. We systematically test a variety of parameter settings to evaluate model fit as well as the sensitivity of the method to specific parameters, and we explore the impact of missing data on model estimation.

Conformal initialization for shape analysis applications in SlicerSALT.

Shape analysis is an important method used in neuroimaging research community due to its potential to precisely locate morphological changes between healthy and pathological structures. A popular shape analysis framework in the neuroimaging community is based on the encoding surface locations as spherical harmonics for a representation called SPHARM-PDM.1 The SPHARM-PDM pipeline takes a set of brain segmentation of a single brain structure (for example, hippocampus) as input and converts them into a corresponding spherical harmonic description (SPHARM), which is then sampled into triangulated surface (SPHARM-PDM). At present, the SPHARM-PDM pipeline utilizes an area-preserving optimization of the spherical mapping based on an initial heat-equation based mapping of the surface mesh to the unit sphere. In the case of objects with complex shape, this initial spherical mapping suffers from a high degree of mapping distortion that cannot always be corrected by the following optimization procedure. Here we proposed the use of an alternative initialization based on a conformal flattening.2 This method adopts a bijective angle preserving conformal flattening scheme to replace the heat equation mapping scheme as initialization for use in the SPHARM-PDM pipeline. After quantitative measures of shape calculated from various complex structures, we concluded that in most cases, the new pipeline produced dramatically better results than the old pipeline. The main contribution of this paper is a command line tool based on the Slicer Execution Model, which merges the conformal flattening into the SPHARM-PDM pipeline for use in the SALT shape analysis toolbox.

A Novel Framework for the Local Extraction of Extra-Axial Cerebrospinal Fluid from MR Brain Images.

The quantification of cerebrospinal fluid (CSF) in the human brain has shown to play an important role in early postnatal brain developmental. Extr a-axial fluid (EA-CSF), which is characterized by the CSF in the subarachnoid space, is promising in the early detection of children at risk for neurodevelopmental disorders. Currently, though, there is no tool to extract local EA-CSF measurements in a way that is suitable for localized analysis. In this paper, we propose a novel framework for the localized, cortical surface based analysis of EA-CSF. In our proposed processing, we combine probabilistic brain tissue segmentation, cortical surface reconstruction as well as streamline based local EA-CSF quantification. For streamline computation, we employ the vector field generated by solving a Laplacian partial differential equation (PDE) between the cortical surface and the outer CSF hull. To achieve sub-voxel accuracy while minimizing numerical errors, fourth-order Runge-Kutta (RK4) integration was used to generate the streamlines. Finally, the local EA-CSF is computed by integrating the CSF probability along the generated streamlines. The proposed local EA-CSF extraction tool was used to study the early postnatal brain development in typically developing infants. The results show that the proposed localized EA-CSF extraction pipeline can produce statistically significant regions that are not observed in previous global approach.

NON-EUCLIDEAN, CONVOLUTIONAL LEARNING ON CORTICAL BRAIN SURFACES.

In recent years there have been many studies indicating that multiple cortical features, extracted at each surface vertex, are promising in the detection of various neurodevelopmental and neurodegenerative diseases. However, with limited datasets, it is challenging to train stable classifiers with such high-dimensional surface data. This necessitates a feature reduction that is commonly accomplished via regional volumetric morphometry from standard brain atlases. However, current regional summaries are not specific to the given age or pathology that is studied, which runs the risk of losing relevant information that can be critical in the classification process. To solve this issue, this paper proposes a novel data-driven approach by extending convolutional neural networks (CNN) for use on non-Euclidean manifolds such as cortical surfaces. The proposed network learns the most powerful features and brain regions from the extracted large dimensional feature space; thus creating a new feature space in which the dimensionality is reduced and feature distributions are better separated. We demonstrate the usability of the proposed surface-CNN framework in an example study classifying Alzheimers disease patients versus normal controls. The high performance in the cross-validation diagnostic results shows the potential of our proposed prediction system.

SlicerSALT: Shape AnaLysis Toolbox.

SlicerSALT is an open-source platform for disseminating state-of-the-art methods for performing statistical shape analysis. These methods are developed as 3D Slicer extensions to take advantage of its powerful underlying libraries. SlicerSALT itself is a heavily customized 3D Slicer package that is designed to be easy to use for shape analysis researchers. The packaged methods include powerful techniques for creating and visualizing shape representations as well as performing various types of analysis.

A Segmentation Editing Framework Based on Shape Change Statistics.

Segmentation is a key task in medical image analysis because its accuracy significantly affects successive steps. Automatic segmentation methods often produce inadequate segmentations, which require the user to manually edit the produced segmentation slice by slice. Because editing is time-consuming, an editing tool that enables the user to produce accurate segmentations by only drawing a sparse set of contours would be needed. This paper describes such a framework as applied to a single object. Constrained by the additional information enabled by the manually segmented contours, the proposed framework utilizes object shape statistics to transform the failed automatic segmentation to a more accurate version. Instead of modeling the object shape, the proposed framework utilizes shape change statistics that were generated to capture the object deformation from the failed automatic segmentation to its corresponding correct segmentation. An optimization procedure was used to minimize an energy function that consists of two terms, an external contour match term and an internal shape change regularity term. The high accuracy of the proposed segmentation editing approach was confirmed by testing it on a simulated data set based on 10 in-vivo infant magnetic resonance brain data sets using four similarity metrics. Segmentation results indicated that our method can provide efficient and adequately accurate segmentations (Dice segmentation accuracy increase of 10%), with very sparse contours (only 10%), which is promising in greatly decreasing the work expected from the user.

Studying Autism Spectrum Disorder with Structural and Diffusion Magnetic Resonance Imaging: A Survey.

Magnetic resonance imaging (MRI) modalities have emerged as powerful means that facilitate non-invasive clinical diagnostics of various diseases and abnormalities since their inception in the 1980s. Multiple MRI modalities, such as different types of the sMRI and DTI, have been employed to investigate facets of ASD in order to better understand this complex syndrome. This paper reviews recent applications of structural magnetic resonance imaging (sMRI) and diffusion tensor imaging (DTI), to study autism spectrum disorder (ASD). Main reported findings are sometimes contradictory due to different age ranges, hardware protocols, population types, numbers of participants, and image analysis parameters. The primary anatomical structures, such as amygdalae, cerebrum, and cerebellum, associated with clinical-pathological correlates of ASD are highlighted through successive life stages, from infancy to adulthood. This survey demonstrates the absence of consistent pathology in the brains of autistic children and lack of research investigations in patients under 2 years of age in the literature. The known publications also emphasize advances in data acquisition and analysis, as well as significance of multimodal approaches that combine resting-state, task-evoked, and sMRI measures. Initial results obtained with the sMRI and DTI show good promise toward the early and non-invasive ASD diagnostics.

Infant Brain Extraction in T1-Weighted MR Images Using BET and Refinement Using LCDG and MGRF Models.

In this paper, we propose a novel framework for the automated extraction of the brain from T1-weighted MR images. The proposed approach is primarily based on the integration of a stochastic model [a two-level Markov-Gibbs random field (MGRF)] that serves to learn the visual appearance of the brain texture, and a geometric model (the brain isosurfaces) that preserves the brain geometry during the extraction process. The proposed framework consists of three main steps: 1) Following bias correction of the brain, a new three-dimensional (3-D) MGRF having a 26-pairwise interaction model is applied to enhance the homogeneity of MR images and preserve the 3-D edges between different brain tissues. 2) The nonbrain tissue found in the MR images is initially removed using the brain extraction tool (BET), and then the brain is parceled to nested isosurfaces using a fast marching level set method. 3) Finally, a classification step is applied in order to accurately remove the remaining parts of the skull without distorting the brain geometry. The classification of each voxel found on the isosurfaces is made based on the first- and second-order visual appearance features. The first-order visual appearance is estimated using a linear combination of discrete Gaussians (LCDG) to model the intensity distribution of the brain signals. The second-order visual appearance is constructed using an MGRF model with analytically estimated parameters. The fusion of the LCDG and MGRF, along with their analytical estimation, allows the approach to be fast and accurate for use in clinical applications. The proposed approach was tested on in vivo data using 300 infant 3-D MR brain scans, which were qualitatively validated by an MR expert. In addition, it was quantitatively validated using 30 datasets based on three metrics: the Dice coefficient, the 95% modified Hausdorff distance, and absolute brain volume difference. Results showed the capability of the proposed approach, outperforming four widely used BETs: BET, BET2, brain surface extractor, and infant brain extraction and analysis toolbox. Experiments conducted also proved that the proposed framework can be generalized to adult brain extraction as well.