Nonrigid registration of joint histograms for intensity standardization in magnetic resonance imaging

Quantitative MRI biomarker for classification of clinically significant prostate cancer: Calibration for reproducibility across echo times

PURPOSE

The purpose of the present study is to develop a calibration method to account for differences in echo times (TE) and facilitate the use of restriction spectrum imaging restriction score (RSIrs) as a quantitative biomarker for the detection of clinically significant prostate cancer (csPCa).

METHODS

This study included 197 consecutive patients who underwent MRI and biopsy examination; 97 were diagnosed with csPCa (grade group ≥ 2). RSI data were acquired three times during the same session: twice at minimum TE ~75 ms and once at TE = 90 ms (TEmin1, TEmin2, and TE90, respectively). A linear regression model was determined to match the C-maps of TE90 to the reference C-maps of TEmin1 within the interval ranging from 95th to 99th percentile of signal intensity within the prostate. RSIrs comparisons were made at the 98th percentile within each patient's prostate. We compared RSIrs from calibrated TE90 (RSIrsTE90corr) and uncorrected TE90 (RSIrsTE90) to RSIrs from reference TEmin1 (RSIrsTEmin1) and repeated TEmin2 (RSIrsTEmin2). Calibration performance was evaluated with sensitivity, specificity and area under the ROC curve (AUC).

RESULTS

Scaling factors for C1, C2, C3, and C4 were estimated as 1.68, 1.33, 1.02, and 1.13, respectively. In non-csPCa cases, the 98th percentile of RSIrsTEmin2 and RSIrsTEmin1 differed by 0.27 ± 0.86SI (mean ± standard deviation), whereas RSIrsTE90 differed from RSIrsTEmin1 by 1.82 ± 1.20SI. After calibration, this bias was reduced to -0.51 ± 1.21SI, representing a 72% reduction in absolute error. For patients with csPCa, the difference was 0.54 ± 1.98SI between RSIrsTEmin2 and RSIrsTEmin1 and 2.28 ± 2.06SI between RSIrsTE90 and RSIrsTEmin1. After calibration, the mean difference decreased to -1.03SI, a 55% reduction in absolute error. At the Youden index for patient-level classification of csPCa (8.94SI), RSIrsTEmin1 has a sensitivity of 66% and a specificity of 72%.

CONCLUSIONS

The proposed linear calibration method produces similar quantitative biomarker values for acquisitions with different TE, reducing TE-induced error by 72% and 55% for non-csPCa and csPCa, respectively.

Deep Learning for Delineation of the Spinal Canal in Whole-Body Diffusion-Weighted Imaging: Normalising Inter- and Intra-Patient Intensity Signal in Multi-Centre Datasets

BACKGROUND

Whole-Body Diffusion-Weighted Imaging (WBDWI) is an established technique for staging and evaluating treatment response in patients with multiple myeloma (MM) and advanced prostate cancer (APC). However, WBDWI scans show inter- and intra-patient intensity signal variability. This variability poses challenges in accurately quantifying bone disease, tracking changes over follow-up scans, and developing automated tools for bone lesion delineation. Here, we propose a novel automated pipeline for inter-station, inter-scan image signal standardisation on WBDWI that utilizes robust segmentation of the spinal canal through deep learning.

METHODS

We trained and validated a supervised 2D U-Net model to automatically delineate the spinal canal (both the spinal cord and surrounding cerebrospinal fluid, CSF) in an initial cohort of 40 patients who underwent WBDWI for treatment response evaluation (80 scans in total). Expert-validated contours were used as the target standard. The algorithm was further semi-quantitatively validated on four additional datasets (three internal, one external, 207 scans total) by comparing the distributions of average apparent diffusion coefficient (ADC) and volume of the spinal cord derived from a two-component Gaussian mixture model of segmented regions. Our pipeline subsequently standardises WBDWI signal intensity through two stages: (i) normalisation of signal between imaging stations within each patient through histogram equalisation of slices acquired on either side of the station gap, and (ii) inter-scan normalisation through histogram equalisation of the signal derived within segmented spinal canal regions. This approach was semi-quantitatively validated in all scans available to the study (N = 287).

RESULTS

The test dice score, precision, and recall of the spinal canal segmentation model were all above 0.87 when compared to manual delineation. The average ADC for the spinal cord (1.7 × 10-3 mm2/s) showed no significant difference from the manual contours. Furthermore, no significant differences were found between the average ADC values of the spinal cord across the additional four datasets. The signal-normalised, high-b-value images were visualised using a fixed contrast window level and demonstrated qualitatively better signal homogeneity across scans than scans that were not signal-normalised.

CONCLUSION

Our proposed intensity signal WBDWI normalisation pipeline successfully harmonises intensity values across multi-centre cohorts. The computational time required is less than 10 s, preserving contrast-to-noise and signal-to-noise ratios in axial diffusion-weighted images. Importantly, no changes to the clinical MRI protocol are expected, and there is no need for additional reference MRI data or follow-up scans.

Quantitative MRI biomarker for classification of clinically significant prostate cancer: calibration for reproducibility across echo times

Background

Restriction Spectrum Imaging restriction score (RSIrs) is a quantitative biomarker for detecting clinically significant prostate cancer (csPCa). However, the quantitative value of the RSIrs is affected by imaging parameters such as echo time (TE).

Purpose

The purpose of the present study is to develop a calibration method to account for differences in echo times and facilitate use of RSIrs as a quantitative biomarker for the detection of csPCa.

Methods

This study included 197 consecutive patients who underwent MRI and biopsy examination; 97 were diagnosed with csPCa (grade group ≥ 2). RSI data were acquired three times during the same session: twice at minimum TE∼75ms and once at TE=90ms (TEmin 1 , TEmin 2 , and TE90, respectively). A proposed calibration method, trained on patients without csPCa, estimated a linear scaling factor (f) for each of the four diffusion compartments (C) of the RSI signal model. A linear regression model was determined to match C-maps of TE90 to the reference C-maps of TEmin 1 within the interval ranging from 95 th to 99 th percentile of signal intensity within the prostate. RSIrs comparisons were made at 98 th percentile within each patient's prostate. We compared RSIrs from calibrated TE90 (RSIrs TE90corr ) and uncorrected TE90 (RSIrs TE90 ) to RSIrs from reference TEmin 1 (RSIrs TEmin1 ) and repeated TEmin 2 (RSIrs TEmin2 ). Calibration performance was evaluated with sensitivity, specificity, area under the ROC curve, positive predicted value, negative predicted value, and F1-score.

Results

Scaling factors for C 1 , C 2 , C 3 , and C 4 were estimated as 1.70, 1.38, 1.03, and 1.19, respectively. In non-csPCa cases, the 98 th percentile of RSIrs TEmin2 and RSIrs TEmin1 differed by 0.27±0.86SI (mean±standard deviation), whereas RSIrs TE90 differed from RSIrs TEmin1 by 1.81±1.20SI. After calibration, this bias was reduced to -0.41±1.20SI, representing a 78% reduction in absolute error. For patients with csPCa, the difference was 0.54±1.98SI between RSIrs TEmin2 and RSIrs TEmin1 and 2.28±2.06SI between RSIrs TE90 and RSIrs TEmin1 . After calibration, the mean difference decreased to -0.86SI, a 38% reduction in absolute error. At the Youden index for patient-level classification of csPCa (8.94SI), RSIrs TEmin1 has a sensitivity of 66% and a specificity of 72%. Prior to calibration, RSIrs TE90 at the same threshold tended to over-diagnose benign cases (sensitivity 44%, specificity 88%). Post-calibration, RSIrs TE90corr performs more similarly to the reference (sensitivity 71%, specificity 62%).

Conclusion

The proposed linear calibration method produces similar quantitative biomarker values for acquisitions with different TE, reducing TE-induced error by 78% and 38% for non-csPCa and csPCa, respectively.

A multi-sequences MRI deep framework study applied to glioma classfication

Glioma is one of the most important central nervous system tumors, ranked 15th in the most common cancer for men and women. Magnetic Resonance Imaging (MRI) represents a common tool for medical experts to the diagnosis of glioma. A set of multi-sequences from an MRI is selected according to the severity of the pathology. Our proposed approach aims moreto create a computer-aided system that is capable of helping morethe expert diagnose the brain gliomas. moreWe propose a supervised learning regime based on a convolutional neural network based framework and transfer learning techniques. Our research morefocuses on the performance of different pre-trained deep learning models with respect to different MRI sequences. We highlight the best combinations of such model-MRI sequence couple for our specific task of classifying healthy brain against brain with glioma. moreWe also propose to visually analyze the extracted deep features for studying the existing relation of the MRI sequences and models. This interpretability analysis gives some hints for medical expert to understand the diagnosis made by the models. Our study is based on the well-known BraTS datasets including multi-sequence images and expert diagnosis.

Whole-body magnetic resonance imaging for prostate cancer assessment: Current status and future directions

Over the past decade, updated definitions for the different stages of prostate cancer and risk for distant disease, along with the advent of new therapies, have remarkably changed the management of patients. The two expectations from imaging are accurate staging and appropriate assessment of disease response to therapies. Modern, next-generation imaging (NGI) modalities, including whole-body magnetic resonance imaging (WB-MRI) and nuclear medicine (most often prostate-specific membrane antigen [PSMA] positron emission tomography [PET]/computed tomography [CT]) bring added value to these imaging tasks. WB-MRI has proven its superiority over bone scintigraphy (BS) and CT for the detection of distant metastasis, also providing reliable evaluations of disease response to treatment. Comparison of the effectiveness of WB-MRI and molecular nuclear imaging techniques with regard to indications and the definition of their respective/complementary roles in clinical practice is ongoing. This paper illustrates the evolution of WB-MRI imaging protocols, defines the current state-of-the art, and highlights the latest developments and future challenges. The paper presents and discusses WB-MRI indications in the care pathway of men with prostate cancer in specific key situations: response assessment of metastatic disease, "all in one" cancer staging, and oligometastatic disease.

Unsupervised MRI Homogenization: Application to Pediatric Anterior Visual Pathway Segmentation

Deep learning strategies have become ubiquitous optimization tools for medical image analysis. With the appropriate amount of data, these approaches outperform classic methodologies in a variety of image processing tasks. However, rare diseases and pediatric imaging often lack extensive data. Specially, MRI are uncommon because they require sedation in young children. Moreover, the lack of standardization in MRI protocols introduces a strong variability between different datasets. In this paper, we present a general deep learning architecture for MRI homogenization that also provides the segmentation map of an anatomical region of interest. Homogenization is achieved using an unsupervised architecture based on variational autoencoder with cycle generative adversarial networks, which learns a common space (i.e. a representation of the optimal imaging protocol) using an unpaired image-to-image translation network. The segmentation is simultaneously generated by a supervised learning strategy. We evaluated our method segmenting the challenging anterior visual pathway using three brain T1-weighted MRI datasets (variable protocols and vendors). Our method significantly outperformed a non-homogenized multi-protocol U-Net.

A Universal Intensity Standardization Method Based on a Many-to-One Weak-Paired Cycle Generative Adversarial Network for Magnetic Resonance Images

In magnetic resonance imaging (MRI), different imaging settings lead to various intensity distributions for a specific imaging object, which brings huge diversity to data-driven medical applications. To standardize the intensity distribution of magnetic resonance (MR) images from multiple centers and multiple machines using one model, a cycle generative adversarial network (CycleGAN)-based framework is proposed. It utilizes a unified forward generative adversarial network (GAN) path and multiple independent backward GAN paths to transform images in different groups into a single reference one. To preserve image details and prevent resolution loss, two jump connections are applied in the CycleGAN generators. A weak-pair strategy is designed to fully utilize the prior knowledge of the organ structure and promote the performance of the GANs. The experiments were conducted on a T2-FLAIR image database with 8192 slices from 489 patients. The database was obtained from four hospitals and five MRI scanners and was divided into nine groups with different imaging parameters. Compared with the representative algorithms, the peak signal-to-noise ratio, the histogram correlation, and the structural similarity were increased by 3.7%, 5.1%, and 0.1% on average, respectively; the gradient magnitude similarity deviation, the mean square error, and the average disparity were reduced by 19.0%, 15.7%, and 9.9% on average, respectively. Experiments also showed the robustness of the proposed model with a different training set configuration and effectiveness of the proposed framework over the original CycleGAN. Therefore, the MR images with different imaging settings could be efficiently standardized by the proposed method, which would benefit various data-driven applications.

Automatic classification of tissues on pelvic MRI based on relaxation times and support vector machine

Tissue segmentation and classification in MRI is a challenging task due to a lack of signal intensity standardization. MRI signal is dependent on the acquisition protocol, the coil profile, the scanner type, etc. While we can compute quantitative physical tissue properties independent of the hardware and the sequence parameters, it is still difficult to leverage these physical properties to segment and classify pelvic tissues. The proposed method integrates quantitative MRI values (T1 and T2 relaxation times and pure synthetic weighted images) and machine learning (Support Vector Machine (SVM)) to segment and classify tissues in the pelvic region, i.e.: fat, muscle, prostate, bone marrow, bladder, and air. Twenty-two men with a mean age of 30±14 years were included in this prospective study. The images were acquired with a 3 Tesla MRI scanner. An inversion recovery-prepared turbo spin echo sequence was used to obtain T1-weighted images at different inversion times with a TR of 14000 ms. A 32-echo spin echo sequence was used to obtain the T2-weighted images at different echo times with a TR of 5000 ms. T1 and T2 relaxation times, synthetic T1- and T2-weighted images and anatomical probabilistic maps were calculated and used as input features of a SVM for segmenting and classifying tissues within the pelvic region. The mean SVM classification accuracy across subjects was calculated for the different tissues: prostate (94.2%), fat (96.9%), muscle (95.8%), bone marrow (91%) and bladder (82.1%) indicating an excellent classification performance. However, the segmentation and classification for air (within the rectum) may not always be successful (mean SVM accuracy 47.5%) due to the lack of air data in the training and testing sets. Our findings suggest that SVM can reliably segment and classify tissues in the pelvic region.

Transfer learning by feature-space transformation: A method for Hippocampus segmentation across scanners

Many successful approaches in MR brain segmentation use supervised voxel classification, which requires manually labeled training images that are representative of the test images to segment. However, the performance of such methods often deteriorates if training and test images are acquired with different scanners or scanning parameters, since this leads to differences in feature representations between training and test data.

In this paper we propose a feature-space transformation (FST) to overcome such differences in feature representations. The proposed FST is derived from unlabeled images of a subject that was scanned with both the source and the target scan protocol. After an affine registration, these images give a mapping between source and target voxels in the feature space. This mapping is then used to map all training samples to the feature representation of the test samples.

We evaluated the benefit of the proposed FST on hippocampus segmentation. Experiments were performed on two datasets: one with relatively small differences between training and test images and one with large differences. In both cases, the FST significantly improved the performance compared to using only image normalization. Additionally, we showed that our FST can be used to improve the performance of a state-of-the-art patch-based-atlas-fusion technique in case of large differences between scanners.

•

We present a feature-space transformation for image segmentation across scanners.

•

This FST is trained on unlabeled images of subjects scanned with multiple scanners.

•

These are used to transform training samples to values observed in target samples.

•

The FST makes SVM hippocampus segmentation across scanners significantly better.

•

Our FST can also increase performance of patch-based fusion methods.

Registration strategies for multi-modal whole-body MRI mosaicing

PURPOSE

To test and compare different registration approaches for performing whole-body diffusion-weighted (wbDWI) image station mosaicing, and its alignment to corresponding anatomical T₁ whole-body image.

METHODS

Four different registration strategies aiming at mosaicing of diffusion-weighted image stations, and their alignment to the corresponding whole-body anatomical image, were proposed and evaluated. These included two-step approaches, where diffusion-weighted stations are first combined in a pairwise (Strategy 1) or groupwise (Strategy 2) manner and later non-rigidly aligned to the anatomical image; a direct pairwise mapping of DWI stations onto the anatomical image (Strategy 3); and simultaneous mosaicing of DWI and alignment to the anatomical image (Strategy 4). Additionally, different images driving the registration were investigated. Experiments were performed for 20 whole-body images of patients with bone metastases.

RESULTS

Strategies 1 and 2 showed significant improvement in mosaicing accuracy with respect to the non-registered images (P < 0.006). Strategy 2 based on ADC images increased the alignment accuracy between DWI stations and the T₁ whole-body image (P = 0.0009).

CONCLUSIONS

A two-step registration strategy, relying on groupwise mosaicing of the ADC stations and subsequent registration to T₁ , provided the best compromise between whole-body DWI image quality and multi-modal alignment. Magn Reson Med 79:1684-1695, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Intensity standardisation of 7T MR images for intensity-based segmentation of the human hypothalamus

The high spatial resolution of 7T MRI enables us to identify subtle volume changes in brain structures, providing potential biomarkers of mental disorders. Most volumetric approaches require that similar intensity values represent similar tissue types across different persons. By applying colour-coding to T1-weighted MP2RAGE images, we found that the high measurement accuracy achieved by high-resolution imaging may be compromised by inter-individual variations in the image intensity. To address this issue, we analysed the performance of five intensity standardisation techniques in high-resolution T1-weighted MP2RAGE images. Twenty images with extreme intensities in the GM and WM were standardised to a representative reference image. We performed a multi-level evaluation with a focus on the hypothalamic region—analysing the intensity histograms as well as the actual MR images, and requiring that the correlation between the whole-brain tissue volumes and subject age be preserved during standardisation. The results were compared with T1 maps. Linear standardisation using subcortical ROIs of GM and WM provided good results for all evaluation criteria: it improved the histogram alignment within the ROIs and the average image intensity within the ROIs and the whole-brain GM and WM areas. This method reduced the inter-individual intensity variation of the hypothalamic boundary by more than half, outperforming all other methods, and kept the original correlation between the GM volume and subject age intact. Mixed results were obtained for the other four methods, which sometimes came at the expense of unwarranted changes in the age-related pattern of the GM volume. The mapping of the T1 relaxation time with the MP2RAGE sequence is advertised as being especially robust to bias field inhomogeneity. We found little evidence that substantiated the T1 map’s theoretical superiority over the T1-weighted images regarding the inter-individual image intensity homogeneity.

Inter-station intensity standardization for whole-body MR data

PURPOSE

To develop and validate a method for performing inter-station intensity standardization in multispectral whole-body MR data.

METHODS

Different approaches for mapping the intensity of each acquired image stack into the reference intensity space were developed and validated. The registration strategies included: "direct" registration to the reference station (Strategy 1), "progressive" registration to the neighboring stations without (Strategy 2), and with (Strategy 3) using information from the overlap regions of the neighboring stations. For Strategy 3, two regularized modifications were proposed and validated. All methods were tested on two multispectral whole-body MR data sets: a multiple myeloma patients data set (48 subjects) and a whole-body MR angiography data set (33 subjects).

RESULTS

For both data sets, all strategies showed significant improvement of intensity homogeneity with respect to vast majority of the validation measures (P < 0.005). Strategy 1 exhibited the best performance, closely followed by Strategy 2. Strategy 3 and its modifications were performing worse, in majority of the cases significantly (P < 0.05).

CONCLUSIONS

We propose several strategies for performing inter-station intensity standardization in multispectral whole-body MR data. All the strategies were successfully applied to two types of whole-body MR data, and the "direct" registration strategy was concluded to perform the best. Magn Reson Med 77:422-433, 2017. © 2016 The Authors Magnetic Resonance in Medicine published by Wiley Periodicals, Inc. on behalf of International Society for Magnetic Resonance in Medicine.

Robust Intensity Standardization in Brain Magnetic Resonance Images

The paper is focused on a tiSsue-Based Standardization Technique (SBST) of magnetic resonance (MR) brain images. Magnetic Resonance Imaging intensities have no fixed tissue-specific numeric meaning, even within the same MRI protocol, for the same body region, or even for images of the same patient obtained on the same scanner in different moments. This affects postprocessing tasks such as automatic segmentation or unsupervised/supervised classification methods, which strictly depend on the observed image intensities, compromising the accuracy and efficiency of many image analyses algorithms. A large number of MR images from public databases, belonging to healthy people and to patients with different degrees of neurodegenerative pathology, were employed together with synthetic MRIs. Combining both histogram and tissue-specific intensity information, a correspondence is obtained for each tissue across images. The novelty consists of computing three standardizing transformations for the three main brain tissues, for each tissue class separately. In order to create a continuous intensity mapping, spline smoothing of the overall slightly discontinuous piecewise-linear intensity transformation is performed. The robustness of the technique is assessed in a post hoc manner, by verifying that automatic segmentation of images before and after standardization gives a high overlapping (Dice index >0.9) for each tissue class, even across images coming from different sources. Furthermore, SBST efficacy is tested by evaluating if and how much it increases intertissue discrimination and by assessing gaussianity of tissue gray-level distributions before and after standardization. Some quantitative comparisons to already existing different approaches available in the literature are performed.

Multi-scale MRI spectrum detects differences in myelin integrity between MS lesion types

Background:

Lesions with different extents of myelin pathology are found at autopsy in multiple sclerosis (MS), but the differences are not discernible in magnetic resonance imaging (MRI).

Objective:

To determine whether analysis of the local spectrum in MRI is sensitive to lesion differences in myelin integrity.

Methods:

We imaged fresh brain slices from 21 MS patients using 1.5T scanners. White matter lesions were identified in T2-weighted MRI, matched to corresponding specimens, and then classified into five categories in histology: pre-active (intact myelin); active, chronic active, chronic inactive (complete demyelination); and remyelinated lesions. Voxel-based frequency spectrum was calculated using T2-weighted MRI to characterize lesion structure (image texture).

Results:

MRI texture heterogeneity resulting from all spectral scales was greater in completely demyelinated lesions than in myelin-preserved lesions ( p = 0.02) and normal-appearing white matter ( p < 0.01). Moreover, the spectral distribution pattern over low-frequency scales differentiated demyelinated lesions from remyelinated and pre-active lesions ( p < 0.01), where different lesion types also showed distinct texture scales.

Conclusion:

Using multi-scale spectral analysis, it may be possible for standard MRI to evaluate myelin integrity in MS lesions. This can be critical for monitoring disease activity and assessing remyelination therapies for MS patients.

Fully Automatic Localization and Segmentation of 3D Vertebral Bodies from CT/MR Images via a Learning-Based Method

In this paper, we address the problems of fully automatic localization and segmentation of 3D vertebral bodies from CT/MR images. We propose a learning-based, unified random forest regression and classification framework to tackle these two problems. More specifically, in the first stage, the localization of 3D vertebral bodies is solved with random forest regression where we aggregate the votes from a set of randomly sampled image patches to get a probability map of the center of a target vertebral body in a given image. The resultant probability map is then further regularized by Hidden Markov Model (HMM) to eliminate potential ambiguity caused by the neighboring vertebral bodies. The output from the first stage allows us to define a region of interest (ROI) for the segmentation step, where we use random forest classification to estimate the likelihood of a voxel in the ROI being foreground or background. The estimated likelihood is combined with the prior probability, which is learned from a set of training data, to get the posterior probability of the voxel. The segmentation of the target vertebral body is then done by a binary thresholding of the estimated probability. We evaluated the present approach on two openly available datasets: 1) 3D T2-weighted spine MR images from 23 patients and 2) 3D spine CT images from 10 patients. Taking manual segmentation as the ground truth (each MR image contains at least 7 vertebral bodies from T11 to L5 and each CT image contains 5 vertebral bodies from L1 to L5), we evaluated the present approach with leave-one-out experiments. Specifically, for the T2-weighted MR images, we achieved for localization a mean error of 1.6 mm, and for segmentation a mean Dice metric of 88.7% and a mean surface distance of 1.5 mm, respectively. For the CT images we achieved for localization a mean error of 1.9 mm, and for segmentation a mean Dice metric of 91.0% and a mean surface distance of 0.9 mm, respectively.

Localization and Segmentation of 3D Intervertebral Discs in MR Images by Data Driven Estimation

This paper addresses the problem of fully-automatic localization and segmentation of 3D intervertebral discs (IVDs) from MR images. Our method contains two steps, where we first localize the center of each IVD, and then segment IVDs by classifying image pixels around each disc center as foreground (disc) or background. The disc localization is done by estimating the image displacements from a set of randomly sampled 3D image patches to the disc center. The image displacements are estimated by jointly optimizing the training and test displacement values in a data-driven way, where we take into consideration both the training data and the geometric constraint on the test image. After the disc centers are localized, we segment the discs by classifying image pixels around disc centers as background or foreground. The classification is done in a similar data-driven approach as we used for localization, but in this segmentation case we are aiming to estimate the foreground/background probability of each pixel instead of the image displacements. In addition, an extra neighborhood smooth constraint is introduced to enforce the local smoothness of the label field. Our method is validated on 3D T2-weighted turbo spin echo MR images of 35 patients from two different studies. Experiments show that compared to state of the art, our method achieves better or comparable results. Specifically, we achieve for localization a mean error of 1.6-2.0 mm, and for segmentation a mean Dice metric of 85%-88% and a mean surface distance of 1.3-1.4 mm.

Histogram-based normalization technique on human brain magnetic resonance images from different acquisitions

BACKGROUND

Intensity normalization is an important preprocessing step in brain magnetic resonance image (MRI) analysis. During MR image acquisition, different scanners or parameters would be used for scanning different subjects or the same subject at a different time, which may result in large intensity variations. This intensity variation will greatly undermine the performance of subsequent MRI processing and population analysis, such as image registration, segmentation, and tissue volume measurement.

METHODS

In this work, we proposed a new histogram normalization method to reduce the intensity variation between MRIs obtained from different acquisitions. In our experiment, we scanned each subject twice on two different scanners using different imaging parameters. With noise estimation, the image with lower noise level was determined and treated as the high-quality reference image. Then the histogram of the low-quality image was normalized to the histogram of the high-quality image. The normalization algorithm includes two main steps: (1) intensity scaling (IS), where, for the high-quality reference image, the intensities of the image are first rescaled to a range between the low intensity region (LIR) value and the high intensity region (HIR) value; and (2) histogram normalization (HN),where the histogram of low-quality image as input image is stretched to match the histogram of the reference image, so that the intensity range in the normalized image will also lie between LIR and HIR.

RESULTS

We performed three sets of experiments to evaluate the proposed method, i.e., image registration, segmentation, and tissue volume measurement, and compared this with the existing intensity normalization method. It is then possible to validate that our histogram normalization framework can achieve better results in all the experiments. It is also demonstrated that the brain template with normalization preprocessing is of higher quality than the template with no normalization processing.

CONCLUSIONS

We have proposed a histogram-based MRI intensity normalization method. The method can normalize scans which were acquired on different MRI units. We have validated that the method can greatly improve the image analysis performance. Furthermore, it is demonstrated that with the help of our normalization method, we can create a higher quality Chinese brain template.

Joint intensity inhomogeneity correction for whole-body MR data

Whole-body MR receives increasing interest as potential alternative to many conventional diagnostic methods. Typical whole-body MR scans contain multiple data channels and are acquired in a multistation manner. Quantification of such data typically requires correction of two types of artefacts: different intensity scaling on each acquired image stack, and intensity inhomogeneity (bias) within each stack. In this work, we present an all-in-one method that is able to correct for both mentioned types of acquisition artefacts. The most important properties of our method are: 1) All the processing is performed jointly on all available data channels, which is necessary for preserving the relation between them, and 2) It allows easy incorporation of additional knowledge for estimation of the bias field. Performed validation on two types of whole-body MR data confirmed superior performance of our approach in comparison with state-of-the-art bias removal methods.

Cross gender-age trabecular texture analysis in cone beam CT

OBJECTIVES

To investigate whether multiple texture features in different regions of interest (ROIs) on cone beam CT (CBCT) are correlated with gender-age variation of trabecular patterns.

METHODS

CBCT volumes from 96 subjects were used. The data set was divided into four gender-age subgroups, including males younger than 40 years, males older than 40 years, females younger than 40 years and females older than 40 years. For each volume, cubes containing trabecular patterns at four ROIs in the jaws were manually cropped. 18 distinct texture features were calculated and their correlation with gender-age variations at different ROIs was studied through t-test statistical analysis.

RESULTS

For the 432 test pairs with different gender-age groups at different ROIs and texture features tested, 149 of them were shown to be statistically different at the 0.05 significance level and 60 of them at the 0.001 significance level. These features can therefore capture changes in trabecular patterns and have the potential to be used for trabecular analysis. Furthermore, fractal features were found to be better than intensity features in separating different gender-age groups. Trabecular patterns in the body of the mandible were more correlated with gender-age changes than other ROIs.

CONCLUSIONS

Multiple texture features on CBCT were found to be correlated with the cross gender-age variation of trabecular patterns. The results support the use of CBCT for advanced trabecular analysis, including osteoporosis screening tools in the jaws.

Characterization of multifocal T2*-weighted MRI hypointensities in the basal ganglia of elderly, community-dwelling subjects

Multifocal T2*-weighted (T2*w) hypointensities in the basal ganglia, which are believed to arise predominantly from mineralized small vessels and perivascular spaces, have been proposed as a biomarker for cerebral small vessel disease. This study provides baseline data on their appearance on conventional structural MRI for improving and automating current manual segmentation methods. Using a published thresholding method, multifocal T2*w hypointensities were manually segmented from whole brain T2*w volumes acquired from 98 community-dwelling subjects in their early 70s. Connected component analysis was used to derive the average T2*w hypointensity count and load per basal ganglia nucleus, as well as the morphology of their connected components, while nonlinear spatial probability mapping yielded their spatial distribution. T1-weighted (T1w), T2-weighted (T2w) and T2*w intensity distributions of basal ganglia T2*w hypointensities and their appearance on T1w and T2w MRI were investigated to gain further insights into the underlying tissue composition. In 75/98 subjects, on average, 3 T2*w hypointensities with a median total volume per intracranial volume of 50.3 ppm were located in and around the globus pallidus. Individual hypointensities appeared smooth and spherical with a median volume of 12 mm³ and median in-plane area of 4 mm². Spatial probability maps suggested an association between T2*w hypointensities and the point of entry of lenticulostriate arterioles into the brain parenchyma. T1w and T2w and especially the T2*w intensity distributions of these hypointensities, which were negatively skewed, were generally not normally distributed indicating an underlying inhomogeneous tissue structure. Globus pallidus T2*w hypointensities tended to appear hypo- and isointense on T1w and T2w MRI, whereas those from other structures appeared iso- and hypointense. This pattern could be explained by an increased mineralization of the globus pallidus. In conclusion, the characteristic spatial distribution and appearance of multifocal basal ganglia T2*w hypointensities in our elderly cohort on structural MRI appear to support the suggested association with mineralized proximal lenticulostriate arterioles and perivascular spaces.

•

A rater segmented focal hypointensities on T2*w brain MRI from 98 elderly subjects.

•

On average 3 focal hypointensities were found in the basal ganglia of 75 subjects.

•

Their spatial distribution suggests an association with lenticulostriate arterioles.

•

Signal intensity distributions suggest an underlying inhomogeneous tissue structure.

Efficient and robust model-to-image alignment using 3D scale-invariant features

This paper presents feature-based alignment (FBA), a general method for efficient and robust model-to-image alignment. Volumetric images, e.g. CT scans of the human body, are modeled probabilistically as a collage of 3D scale-invariant image features within a normalized reference space. Features are incorporated as a latent random variable and marginalized out in computing a maximum a posteriori alignment solution. The model is learned from features extracted in pre-aligned training images, then fit to features extracted from a new image to identify a globally optimal locally linear alignment solution. Novel techniques are presented for determining local feature orientation and efficiently encoding feature intensity in 3D. Experiments involving difficult magnetic resonance (MR) images of the human brain demonstrate FBA achieves alignment accuracy similar to widely-used registration methods, while requiring a fraction of the memory and computation resources and offering a more robust, globally optimal solution. Experiments on CT human body scans demonstrate FBA as an effective system for automatic human body alignment where other alignment methods break down.

PATCH BASED INTENSITY NORMALIZATION OF BRAIN MR IMAGES

Magnetic resonance (MR) imaging (MRI) is widely used to study the structure of human brains. Unlike computed tomography (CT), MR image intensities do not have a tissue specific interpretation. Thus images of the same subject obtained with either the same imaging sequence on different scanners or with differing parameters have widely varying intensity scales. This inconsistency introduces errors in segmentation, and other image processing tasks, thus necessitating image intensity standardization. Compared to previous intensity normalization methods using histogram transformations-which try to find a global one-to-one intensity mapping based on histograms-we propose a patch based generative model for intensity normalization between images acquired under different scanners or different pulse sequence parameters. Our method outperforms histogram based methods when normalizing phantoms simulated with various parameters. Additionally, experiments on real data, acquired under a variety of scanners and acquisition parameters, have more consistent segmentations after our normalization.

Interactive segmentation of plexiform neurofibroma tissue: method and preliminary performance evaluation

Plexiform neurofibromas (PNs) are a major manifestation of neurofibromatosis-1 (NF1), a common genetic disease involving the nervous system. Treatment decisions are mostly based on a gross assessment of changes in tumor using MRI. Accurate volumetric measurements are rarely performed in this kind of tumors mainly due to its great dispersion, size, and multiple locations. This paper presents a semi-automatic method for segmentation of PN from STIR MRI scans. The method starts with a user-based delineation of the tumor area in a single slice and automatically segments the PN lesions in the entire image based on the tumor connectivity. Experimental results on seven datasets, with lesion volumes in the range of 75-690 ml, yielded a mean absolute volume error of 10 % (after manual adjustment) as compared to manual segmentation by an expert radiologist. The mean computation and interaction time was 13 versus 63 min for manual annotation.

Automated approaches for analysis of multimodal MRI acquisitions in a study of cognitive aging

In this work we describe an integrated and automated workflow for a comprehensive and robust analysis of multimodal MR images from a cohort of more than hundred subjects. Image examinations are done three years apart and consist of 3D high-resolution anatomical images, low resolution tensor-valued DTI recordings and 4D resting state fMRI time series. The integrated analysis of the data requires robust tools for segmentation, registration and fiber tracking, which we combine in an automated manner. Our automated workflow is strongly desired due to the large number of subjects. Especially, we introduce the use of histogram segmentation to processed fMRI data to obtain functionally important seed and target regions for fiber tracking between them. This enables analysis of individually important resting state networks. We also discuss various approaches for the assessment of white matter integrity parameters along tracts, and in particular we introduce the use of functional data analysis (FDA) for this task.

A discriminative model-constrained EM approach to 3D MRI brain tissue classification and intensity non-uniformity correction

We describe a fully automated method for tissue classification, which is the segmentation into cerebral gray matter (GM), cerebral white matter (WM), and cerebral spinal fluid (CSF), and intensity non-uniformity (INU) correction in brain magnetic resonance imaging (MRI) volumes. It combines supervised MRI modality-specific discriminative modeling and unsupervised statistical expectation maximization (EM) segmentation into an integrated Bayesian framework. While both the parametric observation models and the non-parametrically modeled INUs are estimated via EM during segmentation itself, a Markov random field (MRF) prior model regularizes segmentation and parameter estimation. Firstly, the regularization takes into account knowledge about spatial and appearance-related homogeneity of segments in terms of pairwise clique potentials of adjacent voxels. Secondly and more importantly, patient-specific knowledge about the global spatial distribution of brain tissue is incorporated into the segmentation process via unary clique potentials. They are based on a strong discriminative model provided by a probabilistic boosting tree (PBT) for classifying image voxels. It relies on the surrounding context and alignment-based features derived from a probabilistic anatomical atlas. The context considered is encoded by 3D Haar-like features of reduced INU sensitivity. Alignment is carried out fully automatically by means of an affine registration algorithm minimizing cross-correlation. Both types of features do not immediately use the observed intensities provided by the MRI modality but instead rely on specifically transformed features, which are less sensitive to MRI artifacts. Detailed quantitative evaluations on standard phantom scans and standard real-world data show the accuracy and robustness of the proposed method. They also demonstrate relative superiority in comparison to other state-of-the-art approaches to this kind of computational task: our method achieves average Dice coefficients of 0.93 ± 0.03 (WM) and 0.90 ± 0.05 (GM) on simulated mono-spectral and 0.94 ± 0.02 (WM) and 0.92 ± 0.04 (GM) on simulated multi-spectral data from the BrainWeb repository. The scores are 0.81 ± 0.09 (WM) and 0.82 ± 0.06 (GM) and 0.87 ± 0.05 (WM) and 0.83 ± 0.12 (GM) for the two collections of real-world data sets-consisting of 20 and 18 volumes, respectively-provided by the Internet Brain Segmentation Repository.

Hybrid cone-beam tomographic reconstruction: incorporation of prior anatomical models to compensate for missing data

We propose a method for improving the quality of cone-beam tomographic reconstruction done with a C-arm. C-arm scans frequently suffer from incomplete information due to image truncation, limited scan length, or other limitations. Our proposed "hybrid reconstruction" method injects information from a prior anatomical model, derived from a subject-specific computed tomography (CT) or from a statistical database (atlas), where the C-arm X-ray data is missing. This significantly reduces reconstruction artifacts with little loss of true information from the X-ray projections. The methods consist of constructing anatomical models, fast rendering of digitally reconstructed radiograph (DRR) projections of the models, rigid or deformable registration of the model and the X-ray images, and fusion of the DRR and X-ray projections, all prior to a conventional filtered back-projection algorithm. Our experiments, conducted with a mobile image intensifier C-arm, demonstrate visually and quantitatively the contribution of data fusion to image quality, which we assess through comparison to a "ground truth" CT. Importantly, we show that a significantly improved reconstruction can be obtained from a C-arm scan as short as 90° by complementing the observed projections with DRRs of two prior models, namely an atlas and a preoperative same-patient CT. The hybrid reconstruction principles are applicable to other types of C-arms as well.

Synthetic MRI signal standardization: application to multi-atlas analysis

From the image analysis perspective, a disadvantage of MRI is the lack of image intensity standardization. Differences in coil sensitivity, pulse sequence and acquisition parameters lead to very different mappings from tissue properties to image intensity levels. This presents challenges for image analysis techniques because the distribution of image intensities for different brain regions can change substantially from scan to scan. Though intensity correction can sometimes alleviate this problem, it fails in more difficult scenarios in which different types of tissue are mapped to similar gray levels in one scan but different intensities in another. Here, we propose using multi-spectral data to create synthetic MRI scans matched to the intensity distribution of a given dataset using a physical model of acquisition. If the multi-spectral data are manually annotated, the labels can be transfered to the synthetic scans to build a dataset-tailored gold standard. The approach was tested on a multi-atlas based hippocampus segmentation framework using a publicly available database, significantly improving the results obtained with other intensity correction methods.

An Automated Method for Segmenting White Matter Lesions through Multi-Level Morphometric Feature Classification with Application to Lupus

We demonstrate an automated, multi-level method to segment white matter brain lesions and apply it to lupus. The method makes use of local morphometric features based on multiple MR sequences, including T1-weighted, T2-weighted, and fluid attenuated inversion recovery. After preprocessing, including co-registration, brain extraction, bias correction, and intensity standardization, 49 features are calculated for each brain voxel based on local morphometry. At each level of segmentation a supervised classifier takes advantage of a different subset of the features to conservatively segment lesion voxels, passing on more difficult voxels to the next classifier. This multi-level approach allows for a fast lesion classification method with tunable trade-offs between sensitivity and specificity producing accuracy comparable to a human rater.

Fast and robust multi-atlas segmentation of brain magnetic resonance images

We introduce an optimised pipeline for multi-atlas brain MRI segmentation. Both accuracy and speed of segmentation are considered. We study different similarity measures used in non-rigid registration. We show that intensity differences for intensity normalised images can be used instead of standard normalised mutual information in registration without compromising the accuracy but leading to threefold decrease in the computation time. We study and validate also different methods for atlas selection. Finally, we propose two new approaches for combining multi-atlas segmentation and intensity modelling based on segmentation using expectation maximisation (EM) and optimisation via graph cuts. The segmentation pipeline is evaluated with two data cohorts: IBSR data (N=18, six subcortial structures: thalamus, caudate, putamen, pallidum, hippocampus, amygdala) and ADNI data (N=60, hippocampus). The average similarity index between automatically and manually generated volumes was 0.849 (IBSR, six subcortical structures) and 0.880 (ADNI, hippocampus). The correlation coefficient for hippocampal volumes was 0.95 with the ADNI data. The computation time using a standard multicore PC computer was about 3-4 min. Our results compare favourably with other recently published results.