Multitask Learning Based Three-Dimensional Striatal Segmentation of MRI: fMRI and PET Objective Assessments.

BACKGROUND: Recent studies have established a clear topographical and functional organization of projections to and from complex subdivisions of the striatum. Manual segmentation of these functional subdivisions is labor-intensive and time-consuming, and automated methods are not as reliable as manual segmentation. PURPOSE: To utilize multitask learning (MTL) as a method to segment subregions of the striatum consisting of pre-commissural putamen (prePU), pre-commissural caudate (preCA), post-commissural putamen (postPU), post-commissural caudate (postCA), and ventral striatum (VST). STUDY TYPE: Retrospective. POPULATION: Eighty-seven total data sets from patients with schizophrenia and matched controls. FIELD STRENGTH/SEQUENCE: 1.5 T and 3.0 T, T1 -weighted (SPGR SENSE, 3D BRAVO). ASSESSMENT: MTL-generated segmentations were compared to the Imperial College London Clinical Imaging Center (CIC) atlas. Dice similarity coefficient (DSC) was used to compare the automated methods to manual segmentations. Positron emission tomography (PET) imaging: 60 minutes of emission data were acquired using [11 C]raclopride. Data were reconstructed by filtered back projection (FBP) with computed tomography (CT) used for attenuation correction. Binding potential values, BPND , and region of interest (ROI) time series and whole-brain connectivity using functional magnetic resonance imaging (fMRI) images were compared between manual and both automated segmentations. STATISTICAL TESTS: Pearson correlation and paired t-test. RESULTS: MTL-generated segmentations showed excellent spatial agreement with manual (DSC ≥0.72 across all striatal subregions). BPND values from MTL-generated segmentations were shown to correlate well with manual segmentations with R2 ≥ 0.91 in all caudate and putamen subregions, and R2  = 0.69 in VST. Mean Pearson correlation coefficients of the fMRI data between MTL-generated and manual segmentations were also high in time series (≥0.86) and whole-brain connectivity (≥0.89) across all subregions. DATA CONCLUSION: Across both PET and fMRI task-based assessments, results from MTL-generated segmentations more closely corresponded to results from manually drawn ROIs than CIC-generated segmentations did. Therefore, the proposed MTL approach is a fast and reliable method for three-dimensional striatal subregion segmentation with results comparable to manually segmented ROIs. LEVEL OF EVIDENCE: 2 TECHNICAL EFFICACY STAGE: 1.

Task-based assessment of a convolutional neural network for segmenting breast lesions for radiomic analysis.

PURPOSE: Radiomics allows for powerful data-mining and feature extraction techniques to guide clinical decision making. Image segmentation is a necessary step in such pipelines and different techniques can significantly affect results. We demonstrate that a convolutional neural network (CNN) segmentation method performs comparably to expert manual segmentations in an established radiomics pipeline. METHODS: Using the manual regions of interest (ROIs) of an expert radiologist (R1), a CNN was trained to segment breast lesions from dynamic contrast-enhanced MRI (DCE-MRI). Following network training, we segmented lesions for the testing set of a previously established radiomics pipeline for predicting lymph node metastases using DCE-MRI of breast cancer. Prediction accuracy of CNN segmentations relative to manual segmentations by R1 from the original study, a resident (R2), and another expert radiologist (R3) were determined. We then retrained the CNN and radiomics model using R3's manual segmentations to determine the effects of different expert observers on end-to-end prediction. RESULTS: Using R1's ROIs, the CNN achieved a mean Dice coefficient of 0.71 ± 0.16 in the testing set. When input to our previously published radiomics pipeline, these CNN segmentations achieved comparable prediction performance to R1's manual ROIs, and superior performance to those of the other radiologists. Similar results were seen when training the CNN and radiomics model using R3's ROIs. CONCLUSION: A CNN architecture is able to provide DCE-MRI breast lesion segmentations which are suitable for input to our radiomics model. Moreover, the previously established radiomics model and CNN can be accurately trained end-to-end using ground truth data provided by distinct experts.

Mapping hepatic blood oxygenation by quantitative BOLD (qBOLD) MRI.

PURPOSE: Abnormalities in hepatic oxygen delivery and oxygen consumption may serve as a significant indicator of hepatic cellular dysfunction and may predict treatment response. However, conventional and oxygen-enhanced hepatic BOLD MRI can only provide semiquantitative assessment of hepatic oxygenation. METHODS: A hepatic quantitative BOLD (qBOLD) model was proposed for noninvasive mapping of hepatic venous blood oxygen saturation (Yv ) and deoxygenated blood volume (DBV) in human subjects. The validity and the estimation bias of the proposed model were evaluated by Monte Carlo simulations. Eight healthy subjects were scanned after written consent with institutional review board approval. RESULTS: Monte Carlo simulations demonstrated that the proposed single-compartment hepatic qBOLD model leads to significant deviation of the predicted T2* decay profile from the simulated signal due to high hepatic blood volume fraction. Small relative estimation bias for hepatic Yv and significant overestimation for hepatic DBV were observed, which can be corrected by applying the calibration curves established from simulations. After correction, the mean hepatic Yv in human subjects was 56.8 ± 6.8%, and the mean hepatic DBV was 0.190 ± 0.035, consistent with measurements from other invasive approaches. Except in regions with significant vascular contamination, the maps for hepatic Yv and DBV were relatively homogenous. CONCLUSIONS: With estimation bias correction, the hepatic qBOLD approach enables noninvasive mapping of hepatic blood volume and oxygenation in human subjects. The established protocol may be used to quantitatively assess hepatic tissue hypoxia in multiple liver diseases.

Preoperative prediction of sentinel lymph node metastasis in breast cancer by radiomic signatures from dynamic contrast-enhanced MRI.

BACKGROUND: Sentinel lymph node (SLN) status is an important prognostic factor for patients with breast cancer, which is currently determined in clinical practice by invasive SLN biopsy. PURPOSE: To noninvasively predict SLN metastasis in breast cancer using dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) intra- and peritumoral radiomics features combined with or without clinicopathologic characteristics of the primary tumor. STUDY TYPE: Retrospective. POPULATION: A total of 163 breast cancer patients (55 positive SLN and 108 negative SLN). FIELD STRENGTH/SEQUENCE: 1.5T, T1 -weighted DCE-MRI. ASSESSMENT: A total of 590 radiomic features were extracted for each patient from both intratumoral and peritumoral regions of interest. To avoid overfitting, the dataset was randomly separated into a training set (∼67%) and a validation set (∼33%). The prediction models were built with the training set using logistic regression on the most significant radiomic features in the training set combined with or without clinicopathologic characteristics. The prediction performance was further evaluated in the independent validation set. STATISTICAL TESTS: Mann-Whitney U-test, Spearman correlation, least absolute shrinkage selection operator (LASSO) regression, logistic regression, and receiver operating characteristic (ROC) analysis were performed. RESULTS: Combining radiomic features with clinicopathologic characteristics, six features were automatically selected in the training set to establish the prediction model of SLN metastasis. In the independent validation set, the area under ROC curve (AUC) was 0.869 (NPV = 0.886). Using radiomic features alone in the same procedure, 4 features were selected and the validation set AUC was 0.806 (NPV = 0.824). DATA CONCLUSION: This is the first attempt to demonstrate the feasibility of using DCE-MRI radiomics to predict SLN metastasis in breast cancer. Clinicopathologic characteristics improved the prediction performance. This study provides noninvasive methods to evaluate SLN status for guiding further treatment of breast cancer patients, and can potentially benefit those with negative SLN, by eliminating unnecessary invasive lymph node removal and the associated complications, which is a step further towards precision medicine. LEVEL OF EVIDENCE: 1 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2019;49:131-140.

Bone microarchitectural parameters can detect oxytocin induced changes prior to bone density on mitigating bone deterioration in rabbit osteoporosis model using micro-CT.

BACKGROUND: This study is aimed to determine the efficacy of X-Ray Microtomography (micro-CT) in predicting oxytocin (OT) treatment response in rabbit osteoporosis(OP) model. METHODS: Sixty-five rabbits were randomly divided into three groups: control group, ovariectomy (OVX) -vehicle and OVX-oxytocin group. The controls underwent sham surgery. OVX-vehicle and OVX-oxytocin groups were subjected to bilateral OVX. The rabbits in OVX-oxytocin group were injected with oxytocin. In the 0th, 4th, 8th, 10th and 12th weeks post OVX operation, bone mineral density (BMD) and bone micro-architectural parameters were measured in three groups. RESULTS: Bone mineral density (BMD), bone volume fraction (BV/TV), Trabecular Number (Tb.N), and Trabecular Thickness (Tb.Th) decreased, while Trabecular Spacing (Tb.Sp) and Structure Model Index (SMI) increased overtime in all the three groups. In OVX-oxytocin group, the bone deterioration tendency is slowing down compared with that of the OVX-vehicle group. The BMD of the OVX-oxytocin group was significantly lower than those in the OVX-vehicle group at 12th week (P = 0.017). BV/TV and Tb.Sp in OVX-oxytocin group changed significantly from 8th week (P = 0.043) and 12th week (P = 0.014), which is earlier than that of BMD and other bone micro-architectural parameters. CONCLUSION: BV/TV and Tb.Sp changed prior to BMD and other bone micro-architectural parameters with oxytocin intervention, which indicate that they are more sensitive markers for predicting early osteoporosis and treatment monitoring when using micro-CT to evaluate osteoporosis rabbit model.

Preoperative prediction of lymph node metastasis using deep learning-based features.

Lymph node involvement increases the risk of breast cancer recurrence. An accurate non-invasive assessment of nodal involvement is valuable in cancer staging, surgical risk, and cost savings. Radiomics has been proposed to pre-operatively predict sentinel lymph node (SLN) status; however, radiomic models are known to be sensitive to acquisition parameters. The purpose of this study was to develop a prediction model for preoperative prediction of SLN metastasis using deep learning-based (DLB) features and compare its predictive performance to state-of-the-art radiomics. Specifically, this study aimed to compare the generalizability of radiomics vs DLB features in an independent test set with dissimilar resolution. Dynamic contrast-enhancement images from 198 patients (67 positive SLNs) were used in this study. Of these subjects, 163 had an in-plane resolution of 0.7 × 0.7 mm2, which were randomly divided into a training set (approximately 67%) and a validation set (approximately 33%). The remaining 35 subjects with a different in-plane resolution (0.78 × 0.78 mm2) were treated as independent testing set for generalizability. Two methods were employed: (1) conventional radiomics (CR), and (2) DLB features which replaced hand-curated features with pre-trained VGG-16 features. The threshold determined using the training set was applied to the independent validation and testing dataset. Same feature reduction, feature selection, model creation procedures were used for both approaches. In the validation set (same resolution as training), the DLB model outperformed the CR model (accuracy 83% vs 80%). Furthermore, in the independent testing set of the dissimilar resolution, the DLB model performed markedly better than the CR model (accuracy 77% vs 71%). The predictive performance of the DLB model outperformed the CR model for this task. More interestingly, these improvements were seen particularly in the independent testing set of dissimilar resolution. This could indicate that DLB features can ultimately result in a more generalizable model.

Optimizing the Peritumoral Region Size in Radiomics Analysis for Sentinel Lymph Node Status Prediction in Breast Cancer.

RATIONALE AND OBJECTIVES: Peritumoral features have been suggested to be useful in improving the prediction performance of radiomic models. The aim of this study is to systematically investigate the prediction performance improvement for sentinel lymph node (SLN) status in breast cancer from peritumoral features in radiomic analysis by exploring the effect of peritumoral region sizes. MATERIALS AND METHODS: This retrospective study was performed using dynamic contrast-enhanced MRI scans of 162 breast cancer patients. The effect of peritumoral features was evaluated in a radiomics pipeline for predicting SLN metastasis in breast cancer. Peritumoral regions were generated by dilating the tumor regions-of-interest (ROIs) manually annotated by two expert radiologists, with thicknesses of 2 mm, 4 mm, 6 mm, and 8 mm. The prediction models were established in the training set (∼67% of cases) using the radiomics pipeline with and without peritumoral features derived from different peritumoral thicknesses. The prediction performance was tested in an independent validation set (the remaining ∼33%). RESULTS: For this specific application, the accuracy in the validation set when using the two radiologists' ROIs could be both improved from 0.704 to 0.796 by incorporating peritumoral features. The choice of the peritumoral size could affect the level of improvement. CONCLUSION: This study systematically investigates the effect of peritumoral region sizes in radiomic analysis for prediction performance improvement. The choice of the peritumoral size is dependent on the ROI drawing and would affect the final prediction performance of radiomic models, suggesting that peritumoral features should be optimized in future radiomics studies.

Fully automated radiomic screening pipeline for osteoporosis and abnormal bone density with a deep learning-based segmentation using a short lumbar mDixon sequence.

BACKGROUND: Although lumbar bone marrow fat fraction (BMFF) has been demonstrated to be predictive of osteoporosis, its utility is limited by the requirement of manual segmentation. Additionally, quantitative features beyond simple BMFF average remain to be explored. In this study, we developed a fully automated radiomic pipeline using deep learning-based segmentation to detect osteoporosis and abnormal bone density (ABD) using a <20 s modified Dixon (mDixon) sequence. METHODS: In total, 222 subjects underwent quantitative computed tomography (QCT) and lower back magnetic resonance imaging (MRI). Bone mineral density (BMD) were extracted from L1-L3 using QCT as the reference standard; 206 subjects (48.8±14.9 years old, 140 females) were included in the final analysis, and were divided temporally into the training/validation set (142/64 subjects). A deep-learning network was developed to perform automated segmentation. Radiomic models were built using the same training set to predict ABD and osteoporosis using the mDixon maps. The performance was evaluated using the temporal validation set comprised of 64 subjects, along with the automated segmentation. Additional 25 subjects (56.1±8.8 years, 14 females) from another site and a different scanner vendor was included as independent validation to evaluate the performance of the pipeline. RESULTS: The automated segmentation achieved an outstanding mean dice coefficient of 0.912±0.062 compared to manual in the temporal validation. Task-based evaluation was performed in the temporal validation set, for predicting ABD and osteoporosis, the area under the curve, sensitivity, specificity, and accuracy were 0.925/0.899, 0.923/0.667, 0.789/0.873, 0.844/0.844, respectively. These values were comparable to that of manual segmentation. External validation (cross-vendor) was also performed; the area under the curve, sensitivity, specificity, and accuracy were 0.688/0.913, 0.786/0.857, 0.545/0.944, 0.680/0.920 for ABD and osteoporosis prediction, respectively. CONCLUSIONS: Our work is the first attempt using radiomics to predict osteoporosis with BMFF map, and the deep-learning based segmentation will further facilitate the clinical utility of the pipeline as a screening tool for early detection of ABD.

White Matter Connectivity in Incident Mild Cognitive Impairment: A Diffusion Spectrum Imaging Study of World Trade Center Responders at Midlife.

BACKGROUND: Individuals who participated in response efforts at the World Trade Center (WTC) following 9/11/2001 are experiencing elevated incidence of mild cognitive impairment (MCI) at midlife. OBJECTIVE: We hypothesized that white matter connectivity measured using diffusion spectrum imaging (DSI) would be restructured in WTC responders with MCI versus cognitively unimpaired responders. METHODS: Twenty responders (mean age 56; 10 MCI/10 unimpaired) recruited from an epidemiological study were characterized using NIA-AA criteria alongside controls matched on demographics (age/sex/occupation/race/education). Axial DSI was acquired on a 3T Siemen's Biograph mMR scanner (12-channel head coil) using a multi-band diffusion sequence. Connectometry examined whole-brain tract-level differences in white matter integrity. Fractional anisotropy (FA), mean diffusivity (MD), and quantified anisotropy were extracted for region of interest (ROI) analyses using the Desikan-Killiany atlas. RESULTS: Connectometry identified both increased and decreased connectivity within regions of the brains of responders with MCI identified in the corticothalamic pathway and cortico-striatal pathway that survived adjustment for multiple comparisons. MCI was also associated with higher FA values in five ROIs including in the rostral anterior cingulate; lower MD values in four ROIs including the left rostral anterior cingulate; and higher MD values in the right inferior circular insula. Analyses by cognitive domain revealed nominal associations in domains of response speed, verbal learning, verbal retention, and visuospatial learning. CONCLUSIONS: WTC responders with MCI at midlife showed early signs of neurodegeneration characterized by both increased and decreased white matter diffusivity in regions commonly affected by early-onset Alzheimer's disease.

Full-count PET recovery from low-count image using a dilated convolutional neural network.

PURPOSE: Positron emission tomography (PET) is an essential technique in many clinical applications that allows for quantitative imaging at the molecular level. This study aims to develop a denoising method using a novel dilated convolutional neural network (CNN) to recover full-count images from low-count images. METHODS: We adopted similar hierarchical structures as the conventional U-Net and incorporated dilated kernels in each convolution to allow the network to observe larger, more robust features within the image without the requirement of downsampling and upsampling internal representations. Our dNet was trained alongside a U-Net for comparison. Both models were evaluated using a leave-one-out cross-validation procedure on a dataset of 35 subjects (~3500 slabs), which were obtained from an ongoing 18 F-Fluorodeoxyglucose (FDG) study. Low-count PET data (10% count) were generated by randomly selecting one-tenth of all events in the associated listmode file. Analysis was done on the static image from the last 10 minutes of emission data. Both low-count PET and full-count PET were reconstructed using ordered subset expectation maximization (OSEM). Objective image quality metrics, including mean absolute percent error (MAPE), peak signal-to-noise ratio (PSNR), and structural similarity index metric (SSIM), were used to analyze the deep learning methods. Both deep learning methods were further compared to a traditional Gaussian filtering method. Further, region of interest (ROI) quantitative analysis was also used to compare U-Net and dNet architectures. RESULTS: Both the U-Net and our proposed network were successfully trained to synthesize full-count PET images from the generated low-count PET images. Compared to low-count PET and Gaussian filtering, both deep learning methods improved MAPE, PSNR, and SSIM. Our dNet also systematically outperformed U-Net on all three metrics (MAPE: 4.99 ± 0.68 vs 5.31 ± 0.76, P < 0.01; PSNR: 31.55 ± 1.31 dB vs 31.05 ± 1.39, P < 0.01; SSIM: 0.9513 ± 0.0154 vs 0.9447 ± 0.0178, P < 0.01). ROI quantification showed greater quantitative improvements using dNet over U-Net. CONCLUSION: This study proposed a novel approach of using dilated convolutions for recovering full-count PET images from low-count PET images.

The importance of identifying functional Val158Met polymorphism in catechol-O- Methyltransferase when assessing MRI-based volumetric measurements in major depressive disorder.

Many studies have shown volumetric differences in the hippocampus between COMT gene polymorphisms and other studies have shown differences between depressed patients and controls; yet, few studies have been completed to identify the volumetric differences when taking both factors into consideration. Using voxel-based morphology (VBM) we investigated, in major depressive disorder (MDD) patients and healthy controls, the relationship between COMT gene polymorphism and volumetric abnormalities. Data from 60 MDD patients and 25 healthy controls were included in this study. Volumetric measurements and genotyping of COMTval158met polymorphism were conducted to determine its impact on gray matter volume (GMV) in the hippocampus and amygdala using a Met dominant model (Val/Val vs Met/Val & Met/Met). In the analysis, a significant difference in the right hippocampus (p = 0.015), right amygdala (p = 0.003) and entire amygdala (p = 0.019) was found between the interaction of diagnosis and genotype after MRI scanner, age and sex correction. Healthy controls (HC) with the Met dominant genotype exhibited a larger right hippocampal, right amygdalar and entire amydgalar volume than MDD patients with the Met dominant genotype. Conversely, HC with the Val/Val genotype displayed a lower right hippocampal, right amygdalar and entire amygdalar volume than their MDD counterparts. This study shows that COMT polymorphism and depression may have a confounding effect on neuroimaging studies.

Fatty infiltration of paraspinal muscles is associated with bone mineral density of the lumbar spine.

A total of 88 subjects were enrolled to investigate the relationship between paraspinal muscle fatty infiltration and lumbar bone mineral density (BMD) using chemical shift encoding-based water-fat MRI and quantitative computed tomography (QCT), respectively. A moderate inverse correlation between paraspinal muscle proton density fat fraction and lumbar QCT-BMD was found with age, sex, and BMI controlled. PURPOSE: To investigate the relationship between paraspinal muscle fatty infiltration and lumbar bone mineral density (BMD). METHODS: A total of 88 subjects were enrolled in this study (52 females, 36 males; age, 46.6 ± 14.2 years old; BMI, 23.2 ± 3.49 kg/m2). Proton density fat fractions (PDFF) of paraspinal muscles (erector spinae, multifidus, and psoas) were measured at L2/3, L3/4, and L4/5 levels using chemical shift encoding-based water-fat MRI. Quantitative computed tomography (QCT) was used to assess BMD of L1, L2, and L3. The differences in paraspinal muscle PDFF among subjects with normal bone density, osteopenia, and osteoporosis were tested using one-way ANOVA. The relationship between paraspinal muscle PDFF and QCT-BMD was analyzed using linear regression with age, sex, and BMI variables. RESULTS: PDFF of the erector spinae, multifidus, and psoas of subjects with normal bone density were all significantly less than those with osteopenia and those with osteoporosis (all p < 0.001). There was an inverse correlation between paraspinal muscle PDFF and BMD after controlling for age, sex, and BMI (standardized beta coefficient, - 0.21~- 0.29; all p < 0.05). CONCLUSIONS: Paraspinal muscle fatty infiltration increased while lumbar BMD decreased after adjusting for age, sex, and BMI. Paraspinal muscles and vertebrae are interacting tissues. Paraspinal muscle fatty infiltration may be a marker of low lumbar BMD. Chemical shift imaging is an efficient and fast quantitative method and can be easily added to the clinical protocol to measure paraspinal muscle PDFF when the patient underwent the routine lumbar MRI with low-back pain.