Voxel level dense prediction of acute stroke territory in DWI using deep learning segmentation models and image enhancement strategies.

PURPOSE: To build a stroke territory classifier model in DWI by designing the problem as a multiclass segmentation task by defining each stroke territory as distinct segmentation targets and leveraging the guidance of voxel wise dense predictions. MATERIALS AND METHODS: Retrospective analysis of DWI images of 218 consecutive acute anterior or posterior ischemic stroke patients examined between January 2017 to April 2020 in a single center was carried out. Each stroke area was defined as distinct segmentation target with different class labels. U-Net based network was trained followed by majority voting of the voxel wise predictions of the model to transform them into patient level stroke territory classes. Effects of bias field correction and registration to a common space were explored. RESULTS: Of the 218 patients included in this study, 141 (65%) were anterior stroke, and 77 were posterior stroke (35%) whereas 117 (53%) were male and 101 (47%) were female. The model built with original images reached 0.77 accuracy, while the model built with N4 bias corrected images reached 0.80 and the model built with images which were N4 bias corrected and then registered into a common space reached 0.83 accuracy values. CONCLUSION: Voxel wise dense prediction coupled with bias field correction to eliminate artificial signal increase and registration to a common space help models for better performance than using original images. Knowing the properties of target domain while designing deep learning models is important for the overall success of these models.

DeepN4: Learning N4ITK Bias Field Correction for T1-weighted Images.

T1-weighted (T1w) MRI has low frequency intensity artifacts due to magnetic field inhomogeneities. Removal of these biases in T1w MRI images is a critical preprocessing step to ensure spatially consistent image interpretation. N4ITK bias field correction, the current state-of-the-art, is implemented in such a way that makes it difficult to port between different pipelines and workflows, thus making it hard to reimplement and reproduce results across local, cloud, and edge platforms. Moreover, N4ITK is opaque to optimization before and after its application, meaning that methodological development must work around the inhomogeneity correction step. Given the importance of bias fields correction in structural preprocessing and flexible implementation, we pursue a deep learning approximation / reinterpretation of the N4ITK bias fields correction to create a method which is portable, flexible, and fully differentiable. In this paper, we trained a deep learning network "DeepN4" on eight independent cohorts from 72 different scanners and age ranges with N4ITK-corrected T1w MRI and bias field for supervision in log space. We found that we can closely approximate N4ITK bias fields correction with naïve networks. We evaluate the peak signal to noise ratio (PSNR) in test dataset against the N4ITK corrected images. The median PSNR of corrected images between N4ITK and DeepN4 was 47.96 dB. In addition, we assess the DeepN4 model on eight additional external datasets and show the generalizability of the approach. This study establishes that incompatible N4ITK preprocessing steps can be closely approximated by naïve deep neural networks, facilitating more flexibility. All code and models are released at https://github.com/MASILab/DeepN4 .

DeepN4: Learning N4ITK Bias Field Correction for T1-weighted Images.

T1-weighted (T1w) MRI has low frequency intensity artifacts due to magnetic field inhomogeneities. Removal of these biases in T1w MRI images is a critical preprocessing step to ensure spatially consistent image interpretation. N4ITK bias field correction, the current state-of-the-art, is implemented in such a way that makes it difficult to port between different pipelines and workflows, thus making it hard to reimplement and reproduce results across local, cloud, and edge platforms. Moreover, N4ITK is opaque to optimization before and after its application, meaning that methodological development must work around the inhomogeneity correction step. Given the importance of bias fields correction in structural preprocessing and flexible implementation, we pursue a deep learning approximation / reinterpretation of the N4ITK bias fields correction to create a method which is portable, flexible, and fully differentiable. In this paper, we trained a deep learning network "DeepN4" on eight independent cohorts from 72 different scanners and age ranges with N4ITK-corrected T1w MRI and bias field for supervision in log space. We found that we can closely approximate N4ITK bias fields correction with naïve networks. We evaluate the peak signal to noise ratio (PSNR) in test dataset against the N4ITK corrected images. The median PSNR of corrected images between N4ITK and DeepN4 was 47.96 dB. In addition, we assess the DeepN4 model on eight additional external datasets and show the generalizability of the approach. This study establishes that incompatible N4ITK preprocessing steps can be closely approximated by naïve deep neural networks, facilitating more flexibility. All code and models are released at https://github.com/MASILab/DeepN4.

A segmentation-based method improving the performance of N4 bias field correction on T2weighted MR imaging data of the prostate.

Magnetic Resonance (MR) images suffer from spatial inhomogeneity, known as bias field corruption. The N4ITK filter is a state-of-the-art method used for correcting the bias field to optimize MR-based quantification. In this study, a novel approach is presented to quantitatively evaluate the performance of N4 bias field correction for pelvic prostate imaging. An exploratory analysis, regarding the different values of convergence threshold, shrink factor, fitting level, number of iterations and use of mask, is performed to quantify the performance of N4 filter in pelvic MR images. The performance of a total of 240 different N4 configurations is examined using the Full Width at Half Maximum (FWHM) of the segmented periprostatic fat distribution as evaluation metric. Phantom T2weighted images were used to assess the performance of N4 for a uniform test tissue mimicking material, excluding factors such as patient related susceptibility and anatomy heterogeneity. Moreover, 89 and 204 T2weighted patient images from two public datasets acquired by scanners with a combined surface and endorectal coil at 1.5 T and a surface coil at 3 T, respectively, were utilized and corrected with a variable set of N4 parameters. Furthermore, two external public datasets were used to validate the performance of the N4 filter in T2weighted patient images acquired by various scanning conditions with different magnetic field strengths and coils. The results show that the set of N4 parameters, converging to optimal representations of fat in the image, were: convergence threshold 0.001, shrink factor 2, fitting level 6, number of iterations 100 and the use of default mask for prostate images acquired by a combined surface and endorectal coil at both 1.5 T and 3 T. The corresponding optimal N4 configuration for MR prostate images acquired by a surface coil at 1.5 T or 3 T was: convergence threshold 0.001, shrink factor 2, fitting level 5, number of iterations 25 and the use of default mask. Hence, periprostatic fat segmentation can be used to define the optimal settings for achieving T2weighted prostate images free from bias field corruption to provide robust input for further analysis.

B1 and magnetization decay correction for hyperpolarized 129 Xe lung imaging using sequential 2D spiral acquisitions.

PURPOSE: To mitigate signal variations caused by inhomogeneous RF and magnetization decay in hyperpolarized 129 Xe ventilation images using flip-angle maps generated from sequential 2D spiral ventilation images acquired in a breath-hold. Images and correction maps were compared with those obtained using conventional, 2D gradient-recalled echo. THEORY AND METHODS: Analytical expressions to predict signal intensity and uncertainty in flip-angle measurements were derived from the Bloch equations and validated by simulations and phantom experiments. Imaging in 129 Xe phantoms and human subjects (1 healthy, 1 cystic fibrosis) was performed using 2D gradient-recalled echo and spiral. For both sequences, consecutive images were acquired with the same slice position during a breath-hold (Cartesian scan time = 15 s; spiral scan time = 5 s). The ratio of these images was used to calculate flip-angle maps and correct intensity inhomogeneities in ventilation images. RESULTS: Mean measured flip angle showed excellent agreement with the applied flip angle in simulations (R2 = 0.99) for both sequences. Mean measured flip angle agreed well with the globally applied flip angle (∼15% difference) in 129 Xe phantoms and in vivo imaging using both sequences. Corrected images displayed reduced coil-dependent signal nonuniformity relative to uncorrected images. CONCLUSIONS: Flip-angle maps were obtained using sequentially acquired, 2D spiral, 129 Xe ventilation images. Signal intensity variations caused by RF-coil inhomogeneity can be corrected by acquiring sequential single-breath ventilation images in less than 5-s scan time. Thus, this method can be used to remove undesirable heterogeneity while preserving physiological effects on the signal distribution.

Effect of magnetic resonance imaging pre-processing on the performance of model-based prostate tumor probability mapping.

Objective. Multi-parametric magnetic resonance imaging (mpMRI) has become an important tool for the detection of prostate cancer in the past two decades. Despite the high sensitivity of MRI for tissue characterization, it often suffers from a lack of specificity. Several well-established pre-processing tools are publicly available for improving image quality and removing both intra- and inter-patient variability in order to increase the diagnostic accuracy of MRI. To date, most of these pre-processing tools have largely been assessed individually. In this study we present a systematic evaluation of a multi-step mpMRI pre-processing pipeline to automate tumor localization within the prostate using a previously trained model.Approach. The study was conducted on 31 treatment-naïve prostate cancer patients with a PI-RADS-v2 compliant mpMRI examination. Multiple methods were compared for each pre-processing step: (1) bias field correction, (2) normalization, and (3) deformable multi-modal registration. Optimal parameter values were estimated for each step on the basis of relevant individual metrics. Tumor localization was then carried out via a model-based approach that takes both mpMRI and prior clinical knowledge features as input. A sequential optimization approach was adopted for determining the optimal parameters and techniques in each step of the pipeline.Main results. The application of bias field correction alone increased the accuracy of tumor localization (area under the curve (AUC) = 0.77;p-value = 0.004) over unprocessed data (AUC = 0.74). Adding normalization to the pre-processing pipeline further improved diagnostic accuracy of the model to an AUC of 0.85 (p-value = 0.000 12). Multi-modal registration of apparent diffusion coefficient images to T2-weighted images improved the alignment of tumor locations in all but one patient, resulting in a slight decrease in accuracy (AUC = 0.84;p-value = 0.30).Significance. Overall, our findings suggest that the combined effect of multiple pre-processing steps with optimal values has the ability to improve the quantitative classification of prostate cancer using mpMRI. Clinical trials: NCT03378856 and NCT03367702.

Jointly estimating bias field and reconstructing uniform MRI image by deep learning.

Bias field is one of the main artifacts that degrade the quality of magnetic resonance images. It introduces intensity inhomogeneity and affects image analysis such as segmentation. In this work, we proposed a deep learning approach to jointly estimate bias field and reconstruct uniform image. By modeling the quality degradation process as the product of a spatially varying field and a uniform image, the network was trained on 800 images with true bias fields from 12 healthy subjects. A network structure of bias field estimation and uniform image reconstruction was designed to compensate for the intensity loss. To further evaluate the benefit of bias field correction, a quantitative analysis was made on image segmentation. Experimental results show that the proposed BFCNet improves the image uniformity by 8.3% and 10.1%, the segmentation accuracy by 4.1% and 6.8% on white and grey matter in T2-weighted brain images. Moreover, BFCNet outperforms the state-of-the-art traditional methods and deep learning methods on estimating bias field and preserving image structure, and BFCNet is robust to different levels of bias field and noise.

Bias field correction in hyperpolarized 129 Xe ventilation MRI using templates derived by RF-depolarization mapping.

PURPOSE: To correct for RF inhomogeneity for in vivo 129 Xe ventilation MRI using flip-angle mapping enabled by randomized 3D radial acquisitions. To extend this RF-depolarization mapping approach to create a flip-angle map template applicable to arbitrary acquisition strategies, and to compare these approaches to conventional bias field correction. METHODS: RF-depolarization mapping was evaluated first in digital simulations and then in 51 subjects who had undergone radial 129 Xe ventilation MRI in the supine position at 3T (views = 3600; samples/view = 128; TR/TE = 4.5/0.45 ms; flip angle = 1.5; FOV = 40 cm). The images were corrected using newly developed RF-depolarization and templated-based methods and the resulting quantitative ventilation metrics (mean, coefficient of variation, and gradient) were compared to those resulting from N4ITK correction. RESULTS: RF-depolarization and template-based mapping methods yielded a pattern of RF-inhomogeneity consistent with the expected variation based on coil architecture. The resulting corrected images were visually similar, but meaningfully distinct from those generated using standard N4ITK correction. The N4ITK algorithm eliminated the physiologically expected anterior-posterior gradient (-0.04 ± 1.56%/cm, P < 0.001). These 2 newly introduced methods of RF-depolarization and template correction retained the physiologically expected anterior-posterior ventilation gradient in healthy subjects (2.77 ± 2.09%/cm and 2.01 ± 2.73%/cm, respectively). CONCLUSIONS: Randomized 3D 129 Xe MRI ventilation acquisitions can inherently be corrected for bias field, and this technique can be extended to create flip angle templates capable of correcting images from a given coil regardless of acquisition strategy. These methods may be more favorable than the de facto standard N4ITK because they can remove undesirable heterogeneity caused by RF effects while retaining results from known physiology.

Multisite reproducibility and test-retest reliability of the T1w/T2w-ratio: A comparison of processing methods.

BACKGROUND: The ratio of T1-weighted (T1w) and T2-weighted (T2w) magnetic resonance imaging (MRI) images is often used as a proxy measure of cortical myelin. However, the T1w/T2w-ratio is based on signal intensities that are inherently non-quantitative and known to be affected by extrinsic factors. To account for this a variety of processing methods have been proposed, but a systematic evaluation of their efficacy is lacking. Given the dependence of the T1w/T2w-ratio on scanner hardware and T1w and T2w protocols, it is important to ensure that processing pipelines perform well also across different sites. METHODS: We assessed a variety of processing methods for computing cortical T1w/T2w-ratio maps, including correction methods for nonlinear field inhomogeneities, local outliers, and partial volume effects as well as intensity normalisation. These were implemented in 33 processing pipelines which were applied to four test-retest datasets, with a total of 170 pairs of T1w and T2w images acquired on four different MRI scanners. We assessed processing pipelines across datasets in terms of their reproducibility of expected regional distributions of cortical myelin, lateral intensity biases, and test-retest reliability regionally and across the cortex. Regional distributions were compared both qualitatively with histology and quantitatively with two reference datasets, YA-BC and YA-B1+, from the Human Connectome Project. RESULTS: Reproducibility of raw T1w/T2w-ratio distributions was overall high with the exception of one dataset. For this dataset, Spearman rank correlations increased from 0.27 to 0.70 after N3 bias correction relative to the YA-BC reference and from -0.04 to 0.66 after N4ITK bias correction relative to the YA-B1+ reference. Partial volume and outlier corrections had only marginal effects on the reproducibility of T1w/T2w-ratio maps and test-retest reliability. Before intensity normalisation, we found large coefficients of variation (CVs) and low intraclass correlation coefficients (ICCs), with total whole-cortex CV of 10.13% and whole-cortex ICC of 0.58 for the raw T1w/T2w-ratio. Intensity normalisation with WhiteStripe, RAVEL, and Z-Score improved total whole-cortex CVs to 5.91%, 5.68%, and 5.19% respectively, whereas Z-Score and Least Squares improved whole-cortex ICCs to 0.96 and 0.97 respectively. CONCLUSIONS: In the presence of large intensity nonuniformities, bias field correction is necessary to achieve acceptable correspondence with known distributions of cortical myelin, but it can be detrimental in datasets with less intensity inhomogeneity. Intensity normalisation can improve test-retest reliability and inter-subject comparability. However, both bias field correction and intensity normalisation methods vary greatly in their efficacy and may affect the interpretation of results. The choice of T1w/T2w-ratio processing method must therefore be informed by both scanner and acquisition protocol as well as the given study objective. Our results highlight limitations of the T1w/T2w-ratio, but also suggest concrete ways to enhance its usefulness in future studies.

Predicting Infiltrative Hepatocellular Carcinoma Patient Outcome Post-TACE: MR Bias Field Correction Effect on 3D-quantitative Image Analysis.

Background and Aims: To investigate the impact of MR bias field correction on response determination and survival prediction using volumetric tumor enhancement analysis in patients with infiltrative hepatocellular carcinoma, after transcatheter arterial chemoembolization (TACE). Methods: This study included 101 patients treated with conventional or drug-eluting beads TACE between the years of 2001 and 2013. Semi-automated 3D quantification software was used to segment and calculate the enhancing tumor volume (ETV) of the liver with and without bias-field correction on multi-phasic contrast-enhanced MRI before and 1-month after initial TACE. ETV (expressed as cm3) at baseline imaging and the relative change in ETV (as % change, ETV%) before and after TACE were used to predict response and survival, respectively. Statistical survival analyses included Kaplan-Meier curve generation and Cox proportional hazards modeling. Q statistics were calculated and used to identify the best cut-off value for ETV to separate responders and non-responders (ETV cm3). The difference in survival was evaluated between responders and non-responders using Kaplan-Meier and Cox models. Results: MR bias field correction correlated with improved response calculation from baseline MR as well as survival after TACE; using a 415 cm3 cut-off for ETV at baseline (hazard ratio: 2.00, 95% confidence interval: 1.23-3.26, p=0.01) resulted in significantly improved response prediction (median survival in patients with baseline ETV <415 cm3: 19.66 months vs. ≥415 cm3: 9.21 months, p<0.001, log-rank test). A ≥41% relative decrease in ETV (hazard ratio: 0.58, 95%confidence interval: 0.37-0.93, p=0.02) was significant in predicting survival (ETV ≥41%: 19.20 months vs. ETV <41%: 8.71 months, p=0.008, log-rank test). Without MR bias field correction, response from baseline ETV could be predicted but survival after TACE could not. Conclusions: MR bias field correction improves both response assessment and accuracy of survival prediction using whole liver tumor enhancement analysis from baseline MR after initial TACE in patients with infiltrative hepatocellular carcinoma.

Correction for Magnetic Field Inhomogeneities and Normalization of Voxel Values Are Needed to Better Reveal the Potential of MR Radiomic Features in Lung Cancer.

Purpose: To design and validate a preprocessing procedure dedicated to T2-weighted MR images of lung cancers so as to improve the ability of radiomic features to distinguish between adenocarcinoma and other histological types. Materials and Methods: A discovery set of 52 patients with advanced lung cancer who underwent T2-weighted MR imaging at 3 Tesla in a single center study from August 2017 to May 2019 was used. Findings were then validated using a validation set of 19 additional patients included from May to October 2019. Tumor type was obtained from the pathology report after trans-thoracic needle biopsy, metastatic lymph node or metastasis samples, or surgical excisions. MR images were preprocessed using N4ITK bias field correction and by normalizing voxel intensities with fat as a reference region. Segmentation and extraction of radiomic features were performed with LIFEx software on the raw images, on the N4ITK-corrected images and on the fully preprocessed images. Two analyses were conducted where radiomic features were extracted: (1) from the whole tumor volume (3D analysis); (2) from all slices encompassing the tumor (2D analysis). Receiver operating characteristic (ROC) analysis was used to identify features that could distinguish between adenocarcinoma and other histological types. Sham experiments were also designed to control the number of false positive findings. Results: There were 31 (12) adenocarcinomas and 21 (7) other histological types in the discovery (validation) set. In 2D, preprocessing increased the number of discriminant radiomic features from 8 without preprocessing to 22 with preprocessing. 2D analysis yielded more features able to identify adenocarcinoma than 3D analysis (12 discriminant radiomic features after preprocessing in 3D). Preprocessing did not increase false positive findings as no discriminant features were identified in any of the sham experiments. The greatest sensitivity of the 2D analysis applied to preprocessed data was confirmed in the validation set. Conclusion: Correction for magnetic field inhomogeneities and normalization of voxel values are essential to reveal the full potential of radiomic features to identify the tumor histological type from MR T2-weighted images, with classification performance similar to those reported in PET/CT and in multiphase CT in lung cancers.

Efficient segmentation and correction model for brain MR images with level set framework based on basis functions.

With the wide application of MR images to detect disease in human's brain deeply, the shortcomings of the technology are necessarily waiting to be solved. For example, MR images always show serious intensity inhomogeneity called the bias field, which may prevent to deduce exact analysis of images. To eliminate the distraction, many methods are proposed. Though experimental results already have stood for the advantages of those methods, there are still lots of problems that cannot be neglected, such as bad segmentation, wrong correction and over-correction which has not attracted much attention yet. Among all those methods, the multiplicative intrinsic component optimization (MICO) model influenced us more. Based on the MICO model and split Bregman method, in this paper, we put forward a new model to segment and correct bias field moderately and simultaneously for MR images. Then, we applied our model to a large quantity of MR images, and gained lots of expected results. For a better observation, we compared our model with the MICO model in both segmentation and bias correction results, it can be seen from the experimental results that our model has performed well for the challenging intensity inhomogeneity problems. Many good characteristics like accuracy, efficiency and robustness also have been exhibited in numerical results and comparisons with the MICO model.

Joint deformable liver registration and bias field correction for MR-guided HDR brachytherapy.

PURPOSE: In interstitial high-dose rate brachytherapy, liver cancer is treated by internal radiation, requiring percutaneous placement of applicators within or close to the tumor. To maximize utility, the optimal applicator configuration is pre-planned on magnetic resonance images. The pre-planned configuration is then implemented via a magnetic resonance-guided intervention. Mapping the pre-planning information onto interventional data would reduce the radiologist's cognitive load during the intervention and could possibly minimize discrepancies between optimally pre-planned and actually placed applicators. METHODS: We propose a fast and robust two-step registration framework suitable for interventional settings: first, we utilize a multi-resolution rigid registration to correct for differences in patient positioning (rotation and translation). Second, we employ a novel iterative approach alternating between bias field correction and Markov random field deformable registration in a multi-resolution framework to compensate for non-rigid movements of the liver, the tumors and the organs at risk. In contrast to existing pre-correction methods, our multi-resolution scheme can recover bias field artifacts of different extents at marginal computational costs. RESULTS: We compared our approach to deformable registration via B-splines, demons and the SyN method on 22 registration tasks from eleven patients. Results showed that our approach is more accurate than the contenders for liver as well as for tumor tissues. We yield average liver volume overlaps of 94.0 ± 2.7% and average surface-to-surface distances of 2.02 ± 0.87 mm and 3.55 ± 2.19 mm for liver and tumor tissue, respectively. The reported distances are close to (or even below) the slice spacing (2.5 - 3.0 mm) of our data. Our approach is also the fastest, taking 35.8 ± 12.8 s per task. CONCLUSION: The presented approach is sufficiently accurate to map information available from brachytherapy pre-planning onto interventional data. It is also reasonably fast, providing a starting point for computer-aidance during intervention.

Resting State fMRI in the moving fetus: a robust framework for motion, bias field and spin history correction.

There is growing interest in exploring fetal functional brain development, particularly with Resting State fMRI. However, during a typical fMRI acquisition, the womb moves due to maternal respiration and the fetus may perform large-scale and unpredictable movements. Conventional fMRI processing pipelines, which assume that brain movements are infrequent or at least small, are not suitable. Previous published studies have tackled this problem by adopting conventional methods and discarding as much as 40% or more of the acquired data. In this work, we developed and tested a processing framework for fetal Resting State fMRI, capable of correcting gross motion. The method comprises bias field and spin history corrections in the scanner frame of reference, combined with slice to volume registration and scattered data interpolation to place all data into a consistent anatomical space. The aim is to recover an ordered set of samples suitable for further analysis using standard tools such as Group Independent Component Analysis (Group ICA). We have tested the approach using simulations and in vivo data acquired at 1.5 T. After full motion correction, Group ICA performed on a population of 8 fetuses extracted 20 networks, 6 of which were identified as matching those previously observed in preterm babies.

Multiplicative intrinsic component optimization (MICO) for MRI bias field estimation and tissue segmentation.

This paper proposes a new energy minimization method called multiplicative intrinsic component optimization (MICO) for joint bias field estimation and segmentation of magnetic resonance (MR) images. The proposed method takes full advantage of the decomposition of MR images into two multiplicative components, namely, the true image that characterizes a physical property of the tissues and the bias field that accounts for the intensity inhomogeneity, and their respective spatial properties. Bias field estimation and tissue segmentation are simultaneously achieved by an energy minimization process aimed to optimize the estimates of the two multiplicative components of an MR image. The bias field is iteratively optimized by using efficient matrix computations, which are verified to be numerically stable by matrix analysis. More importantly, the energy in our formulation is convex in each of its variables, which leads to the robustness of the proposed energy minimization algorithm. The MICO formulation can be naturally extended to 3D/4D tissue segmentation with spatial/sptatiotemporal regularization. Quantitative evaluations and comparisons with some popular softwares have demonstrated superior performance of MICO in terms of robustness and accuracy.

Image-guided regularization level set evolution for MR image segmentation and bias field correction.

Magnetic resonance (MR) image segmentation is a crucial step in surgical and treatment planning. In this paper, we propose a level-set-based segmentation method for MR images with intensity inhomogeneous problem. To tackle the initialization sensitivity problem, we propose a new image-guided regularization to restrict the level set function. The maximum a posteriori inference is adopted to unify segmentation and bias field correction within a single framework. Under this framework, both the contour prior and the bias field prior are fully used. As a result, the image intensity inhomogeneity can be well solved. Extensive experiments are provided to evaluate the proposed method, showing significant improvements in both segmentation and bias field correction accuracies as compared with other state-of-the-art approaches.