Diffusion-weighted Imaging Datasets Research Articles

Liver fibrosis automatic diagnosis utilizing dense-fusion attention contrastive learning network.

Liver fibrosis poses a significant public health challenge given its elevated incidence and associated mortality rates. Diffusion-Weighted Imaging (DWI) serves as a non-invasive diagnostic tool for supporting the identification of liver fibrosis. Deep learning, as a computer-aided diagnostic technology, can assist in recognizing the stage of liver fibrosis by extracting abstract features from DWI images. However, gathering samples is often challenging, posing a common dilemma in previous research. Moreover, previous studies frequently overlooked the cross-comparison information and latent connections among different DWI parameters. Thus, it is becoming a challenge to identify effective DWI parameters and dig potential features from multiple categories in a dataset with limited samples. A self-defined Multi-view Contrastive Learning Network is developed to automatically classify multi-parameter DWI images and explore synergies between different DWI parameters. A Dense-fusion Attention Contrastive Learning Network (DACLN) is designed and used to recognize DWI images. Concretely, a multi-view contrastive learning framework is constructed to train and extract features from raw multi-parameter DWI. Besides, a Dense-fusion module is designed to integrate feature and output predicted labels. We evaluated the performance of the proposed model on a set of real clinical data and analyzed the interpretability by Grad-CAM and annotation analysis, achieving average scores of 0.8825, 0.8702, 0.8933, 0.8727, and 0.8779 for accuracy, precision, recall, specificity and F-1 score. Of note, the experimental results revealed that IVIM-f, CTRW-β, and MONO-ADC exhibited significant recognition ability and complementarity. Our method achieves competitive accuracy in liver fibrosis diagnosis using the limited multi-parameter DWI dataset and finds three types of DWI parameters with high sensitivity for diagnosing liver fibrosis, which suggests potential directions for future research.

Effect of Phase Encoding Direction on Image Quality in Single-Shot EPI Diffusion-Weighted Imaging of the Breast.

In breast diffusion-weighted imaging (DWI), distortion and physiologic artifacts affect clinical interpretation. Image quality can be optimized by addressing the effect of phase encoding (PE) direction on these artifacts. To compare distortion artifacts in breast DWI acquired with different PE directions and polarities, and to discuss their clinical implications. Prospective. Eleven healthy volunteers (median age: 47 years old; range: 22-74 years old) and a breast phantom. Single-shot echo planar DWI and three-dimensional fast gradient echo sequences at 3 T. All DWI data were acquired with left-right, right-left, posterior-anterior, and anterior-posterior PE directions. In phantom data, displacement magnitude was evaluated by comparing the location of landmarks in anatomical and DWI images. Three breast radiologists (5, 17, and 23 years of experience) assessed the presence or absence of physiologic artifacts in volunteers' DWI datasets and indicated their PE-direction preference. Analysis of variance with post-hoc tests were used to assess differences in displacement magnitude across DWI datasets and observers. A binomial test and a chi-squared test were used to evaluateif each in vivo DWI dataset had an equal probability (25%) of being preferred by radiologists. Inter-reader agreement was evaluated using Gwet's AC1 agreement coefficient. A P-value <0.05 was considered statistically significant. In the phantom study, median displacement was the significantly largest in posterior-anterior data. While the displacement in the anterior-posterior and left-right data were equivalent (P = 0.545). In the in vivo data, there were no physiological artifacts observed in any dataset, regardless of PE direction. In the reader study, there was a significant preference for the posterior-anterior datasets which were selected 94% of the time. There was good agreement between readers (0.936). This study showed the impact of PE direction on distortion artifacts in breast DWI. In healthy volunteers, the posterior-to-anterior PE direction was preferred by readers. 2 TECHNICAL EFFICACY: Stage 1.

Breast lesion morphology assessment with high and standard b values in diffusion-weighted imaging at 3 Tesla

IntroductionWith increasing spatial resolution, diffusion-weighted imaging (DWI) may be suitable for morphologic lesion characterization in breast MRI – an area that has traditionally been occupied by dynamic contrast-enhanced imaging (DCE). This investigation compared DWI with b values of 800 and 1600 s/mm2 to DCE for lesion morphology assessment in high-resolution breast MRI at 3 Tesla. Material and methodsMultiparametric breast MRI was performed in 91 patients with 93 histopathologically proven lesions (31 benign, 62 malignant). Two radiologists independently evaluated three datasets per patient (DWIb800; DWIb1600; DCE) and assessed lesion visibility and BIRADS morphology criteria. Diagnostic accuracy was compared among readers and datasets using Cochran's Q test and pairwise post-hoc McNemar tests. Bland-Altman analyses were conducted for lesion size comparisons. ResultsDiscrimination of carcinomas was superior compared to benign findings in both DWIb800 and DWIb1600 (p < 0.001) with no b value-dependent difference. Similarly, assessability of mass lesions was better than of non-mass lesions, irrespective of b value (p < 0.001). Intra-reader reliability for the analysis of morphologic BIRADS criteria among DCE and DWI datasets was at least moderate (Fleiss κ≥0.557), while at least substantial inter-reader agreement was ascertained over all assessed categories (κ≥0.776). In pairwise Bland-Altman analyses, the measurement bias between DCE and DWIb800 was 0.7 mm, whereas the difference between DCE and DWIb1600 was 2.8 mm. DWIb1600 allowed for higher specificity than DCE (p = 0.007/0.062). ConclusionsDWI can be employed for reliable morphologic lesion characterization in high-resolution breast MRI. High b values increase diagnostic specificity, while lesion size assessment is more precise with standard 800 s/mm2 images.

Diffusion-weighted imaging on an MRI-linear accelerator to identify adversely prognostic tumour regions in glioblastoma during chemoradiation

Background and purposeSurvival in glioblastoma might be extended by escalating the radiotherapy dose to treatment-resistant tumour and adapting to tumour changes. Diffusion-weighted imaging (DWI) on MRI-linear accelerators (MR-Linacs) could be used to identify a dose escalation target, but its prognostic value must be demonstrated. The purpose of this study was to determine whether MR-Linac DWI can assess treatment response in glioblastoma and whether changes in DWI show greater prognostic value than changes in the contrast-enhancing gross tumour volume (GTV). Materials and methodsSeventy-five patients with glioblastoma were treated with chemoradiotherapy, of which 32 were treated on a 1.5 T MRI-linear accelerator (MR-Linac). Patients were imaged with simulation MRI scanners (MR-sim) at treatment planning and weeks 2, 4, and 10 after treatment start. Twenty-eight patients had additional MR-Linac DWI sequences. Cox modelling was used to evaluate the correlation of overall and progression-free survival (OS and PFS) with clinical variables and volumetric changes in the GTV and low-ADC regions (ADC < 1.25 µm2/ms within GTV). ResultsIn total, 479 MR-Linac DWI and 289 MR-sim DWI datasets were analyzed. MR-Linac low-ADC changes between weeks 2 and 5 inclusive were prognostic for OS (hazard ratio lower limits ≥ 1.2, p-values ≤ 0.02). MR-sim low-ADC changes showed greater correlation with OS and PFS than GTV changes (e.g., OS hazard ratio at week 2 was 3.4 (p <0.001) for low-ADC versus 2.0 (p = 0.022) for GTV). ConclusionMR-Linac DWI can measure low-ADC tumour volumes that correlate with OS and PFS better than contrast-enhancing GTV. Low-ADC regions could serve as dose escalation targets.

Accelerating High b-Value Diffusion-Weighted MRI Using a Convolutional Recurrent Neural Network (CRNN-DWI).

To develop a novel convolutional recurrent neural network (CRNN-DWI) and apply it to reconstruct a highly undersampled (up to six-fold) multi-b-value, multi-direction diffusion-weighted imaging (DWI) dataset. A deep neural network that combines a convolutional neural network (CNN) and recurrent neural network (RNN) was first developed by using a set of diffusion images as input. The network was then used to reconstruct a DWI dataset consisting of 14 b-values, each with three diffusion directions. For comparison, the dataset was also reconstructed with zero-padding and 3D-CNN. The experiments were performed with undersampling rates (R) of 4 and 6. Standard image quality metrics (SSIM and PSNR) were employed to provide quantitative assessments of the reconstructed image quality. Additionally, an advanced non-Gaussian diffusion model was employed to fit the reconstructed images from the different approaches, thereby generating a set of diffusion parameter maps. These diffusion parameter maps from the different approaches were then compared using SSIM as a metric. Both the reconstructed diffusion images and diffusion parameter maps from CRNN-DWI were better than those from zero-padding or 3D-CNN. Specifically, the average SSIM and PSNR of CRNN-DWI were 0.750 ± 0.016 and 28.32 ± 0.69 (R = 4), and 0.675 ± 0.023 and 24.16 ± 0.77 (R = 6), respectively, both of which were substantially higher than those of zero-padding or 3D-CNN reconstructions. The diffusion parameter maps from CRNN-DWI also yielded higher SSIM values for R = 4 (>0.8) and for R = 6 (>0.7) than the other two approaches (for R = 4, <0.7, and for R = 6, <0.65). CRNN-DWI is a viable approach for reconstructing highly undersampled DWI data, providing opportunities to reduce the data acquisition burden.

The value of whole-volume apparent diffusion coefficient histogram analysis in preoperatively distinguishing intracranial solitary fibrous tumor and transitional meningioma.

To investigate the value of whole-volume apparent diffusion coefficient (ADC) histogram analysis in preoperatively distinguishing intracranial solitary fibrous tumors (SFT) from transitional meningiomas (TM), thereby assisting the establishment of the treatment protocol. Preoperative diffusion-weighted imaging datasets of 24 patients with SFT and 28 patients with TM were used to extract whole-volume ADC histogram parameters, including variance, skewness, kurtosis, and mean, as well as 1st (AP1), 10th (AP10), 50th (AP50), 90th (AP90), and 99th (AP99) percentiles of ADC using MaZda software. The independent t-test or Mann-Whitney U test was used to compare the differences between ADC histogram parameters of SFT and TM. Receiver operating characteristic (ROC) curves were generated to evaluate the performance of significant ADC histogram parameters. Spearman's correlation coefficients were calculated to evaluate correlations between these parameters and the Ki-67 expression levels. SFT exhibited significantly higher variance, and lower AP1 and AP10 (all P < 0.05) than TM. The best diagnostic performance was obtained by variance, with an area under the ROC curve of 0.848 (0.722-0.933). However, there was no significant difference in skewness, kurtosis, mean, or other percentiles of ADC between the two groups (all P > 0.05). Significant correlations were also observed between the Ki-67 proliferation index and variance (r = 0.519), AP1 (r = -0.425), and AP10 (r = -0.372) (all P < 0.05). Whole-volume ADC histogram analysis is a feasible tool for non-invasive preoperative discrimination between intracranial SFT and TM, with variance being the most promising prospective parameter.

A new open, high-resolution, multishell, diffusion-weighted imaging dataset of the living squirrel monkey

Although very well adapted to brain study, Magnetic Resonance Imaging (MRI) remains limited by the facilities and capabilities required to acquire data, especially for non-human primates. Addressing the data gaps resulting from these limitations requires making data more accessible and open. In contempt of the regular use of Saimiri sciureus in neuroscience research, in vivo diffusion has yet to be openly available for this species. Here we built and made openly available a unique new resource consisting of a high-resolution, multishell diffusion-weighted dataset in the anesthetized Saimiri sciureus. The data were acquired on 11 individuals with an 11.7 T MRI scanner (isotropic resolution of 400 µm3). This paper presents an overview of our dataset and illustrates some of its possible use through example analyses. To assess the quality of our data, we analyzed long-range connections (whole-brain tractography), microstructure (Neurite Orientation Dispersion and Density Imaging), and axon diameter in the corpus callosum (ActiveAx). Constituting an essential new resource for primate evolution studies, all data are openly available.

RU-Net: skull stripping in rat brain MR images after ischemic stroke with rat U-Net

BackgroundExperimental ischemic stroke models play a fundamental role in interpreting the mechanism of cerebral ischemia and appraising the development of pathological extent. An accurate and automatic skull stripping tool for rat brain image volumes with magnetic resonance imaging (MRI) are crucial in experimental stroke analysis. Due to the deficiency of reliable rat brain segmentation methods and motivated by the demand for preclinical studies, this paper develops a new skull stripping algorithm to extract the rat brain region in MR images after stroke, which is named Rat U-Net (RU-Net).MethodsBased on a U-shape like deep learning architecture, the proposed framework integrates batch normalization with the residual network to achieve efficient end-to-end segmentation. A pooling index transmission mechanism between the encoder and decoder is exploited to reinforce the spatial correlation. Two different modalities of diffusion-weighted imaging (DWI) and T2-weighted MRI (T2WI) corresponding to two in-house datasets with each consisting of 55 subjects were employed to evaluate the performance of the proposed RU-Net.ResultsExtensive experiments indicated great segmentation accuracy across diversified rat brain MR images. It was suggested that our rat skull stripping network outperformed several state-of-the-art methods and achieved the highest average Dice scores of 98.04% (p < 0.001) and 97.67% (p < 0.001) in the DWI and T2WI image datasets, respectively.ConclusionThe proposed RU-Net is believed to be potential for advancing preclinical stroke investigation and providing an efficient tool for pathological rat brain image extraction, where accurate segmentation of the rat brain region is fundamental.

Autocalibrating segmented diffusion-weighted acquisitions.

Segmented echo-planar imaging enables high-resolution diffusion-weighted imaging (DWI). However, phase differences between segments can lead to severe artifacts. This work investigates an algorithm to enable reconstruction of interleaved segmented acquisitions without the need of additional calibration or navigator measurements. A parallel imaging algorithm is presented that jointly reconstructs all segments of one DWI frame maintaining their phase information. Therefore, the algorithm allows for an iterative improvement of the phase estimates included in the joint reconstruction. Given a limited number of interleaves, the initial-phase estimates can be calculated by a traditional parallel-imaging reconstruction, using the unweighted scan of the DWI measurement as a reference. Reconstruction of phantom data and g-factor simulations show substantial improvement (up to 93% reduction in root mean square error) compared with a generalized auto-calibrating partially parallel-acquisition reconstruction. In vivo experiments show robust reconstruction outcomes in critical imaging situations, including small numbers of receiver channels or low signal-to-noise ratio. An algorithm for the robust reconstruction of segmented DWI data is presented. The method requires neither navigator nor calibration measurements; therefore, it can be applied to existing DWI data sets.

Precise Identification of Prostate Cancer from DWI Using Transfer Learning.

Background and Objective: The use of computer-aided detection (CAD) systems can help radiologists make objective decisions and reduce the dependence on invasive techniques. In this study, a CAD system that detects and identifies prostate cancer from diffusion-weighted imaging (DWI) is developed. Methods: The proposed system first uses non-negative matrix factorization (NMF) to integrate three different types of features for the accurate segmentation of prostate regions. Then, discriminatory features in the form of apparent diffusion coefficient (ADC) volumes are estimated from the segmented regions. The ADC maps that constitute these volumes are labeled by a radiologist to identify the ADC maps with malignant or benign tumors. Finally, transfer learning is used to fine-tune two different previously-trained convolutional neural network (CNN) models (AlexNet and VGGNet) for detecting and identifying prostate cancer. Results: Multiple experiments were conducted to evaluate the accuracy of different CNN models using DWI datasets acquired at nine distinct b-values that included both high and low b-values. The average accuracy of AlexNet at the nine b-values was with average sensitivity and specificity of and . These results improved with the use of the deeper CNN model (VGGNet). The average accuracy of VGGNet was with sensitivity and specificity of and . Conclusions: The results of the conducted experiments emphasize the feasibility and accuracy of the developed system and the improvement of this accuracy using the deeper CNN.

Image quality and diagnostic value of diffusion-weighted breast magnetic resonance imaging: Comparison of acquired and computed images.

PurposeTo compare the image quality of acquired diffusion-weighted imaging (DWI) and computed DWI and evaluate the lesion detectability and likelihood of malignancy in these datasets.Materials and methodsThis prospective study was approved by our institutional review board. A total of 29 women (mean age, 43.5 years) underwent DWI between August 2018 and April 2019 for 32 breast cancers and 16 benign breast lesions. Three radiologists independently reviewed the acquired DWI with b-values of 1000 and 2000 s/mm2 (A-b1000 and A-b2000) and the computed DWI with a b-value of 2000 s/mm2 (C-b2000). Image quality was scored and compared between the three DWI datasets. Lesion detectability was recorded, and the lesion’s likelihood for malignancy was scored using a five-point scale.ResultsThe A-b1000 images were superior to the A-b2000 and C-b2000 images in chest distinction, fat suppression, and overall image quality. The A-b2000 and C-b2000 images showed comparable scores for all image quality parameters. C-b2000 showed the highest values for lesion detection among all readers, although there was no statistical difference in sensitivity, specificity, positive predictive value, negative predictive value, and accuracy between the DWI datasets. The malignancy scores of the DWI images were not significantly different among the three readers.ConclusionsA-b1000 DWI is suitable for breast lesion evaluations, considering its better image quality and comparable diagnostic values compared to that of A-b2000 and C-b2000 images. The additional use of computed high b-value DWI may have the potential to increase the detectability of breast masses.

Automatic Assessment of ASPECTS Using Diffusion-Weighted Imaging in Acute Ischemic Stroke Using Recurrent Residual Convolutional Neural Network.

The early detection and rapid quantification of acute ischemic lesions play pivotal roles in stroke management. We developed a deep learning algorithm for the automatic binary classification of the Alberta Stroke Program Early Computed Tomographic Score (ASPECTS) using diffusion-weighted imaging (DWI) in acute stroke patients. Three hundred and ninety DWI datasets with acute anterior circulation stroke were included. A classifier algorithm utilizing a recurrent residual convolutional neural network (RRCNN) was developed for classification between low (1–6) and high (7–10) DWI-ASPECTS groups. The model performance was compared with a pre-trained VGG16, Inception V3, and a 3D convolutional neural network (3DCNN). The proposed RRCNN model demonstrated higher performance than the pre-trained models and 3DCNN with an accuracy of 87.3%, AUC of 0.941, and F1-score of 0.888 for classification between the low and high DWI-ASPECTS groups. These results suggest that the deep learning algorithm developed in this study can provide a rapid assessment of DWI-ASPECTS and may serve as an ancillary tool that can assist physicians in making urgent clinical decisions.

In Reply: Early Diffusion Magnetic Resonance Imaging Changes in Normal-Appearing Brain in Pediatric Moyamoya Disease.

BACKGROUND Moyamoya disease often leads to ischemic strokes visible on diffusion-weighted imaging (DWI) and T2-weighted magnetic resonance imaging (MRI) with subsequent cognitive impairment. In adults with moyamoya, apparent diffusion coefficient (ADC) is correlated with regions of steal phenomenon and executive dysfunction prior to white matter changes. OBJECTIVE To investigate quantitative global diffusion changes in pediatric moyamoya patients prior to explicit structural ischemic damage. METHODS We retrospectively reviewed children (<20 yr old) with moyamoya disease and syndrome who underwent bypass surgery at our institution. We identified 29 children with normal structural preoperative MRI and without findings of cortical infarction or chronic white matter ischemic changes. DWI datasets were used to calculate ADC maps for each subject as well as for 60 age-matched healthy controls. Using an atlas-based approach, the cerebral white matter, cerebral cortex, thalamus, caudate, putamen, pallidum, hippocampus, amygdala, nucleus accumbens, and brainstem were segmented in each DWI dataset and used to calculate regional volumes and ADC values. RESULTS Multivariate analysis of covariance using the regional ADC and volume values as dependent variables and age and gender as covariates revealed a significant difference between the groups (P < .001). Post hoc analysis demonstrated significantly elevated ADC values for children with moyamoya in the cerebral cortex, white matter, caudate, putamen, and nucleus accumbens. No significant volume differences were found. CONCLUSION Prior to having bypass surgery, and in the absence of imaging evidence of ischemic stroke, children with moyamoya exhibit cerebral diffusion changes. These findings could reflect microstructural changes stemming from exhaustion of cerebrovascular reserve.

Quality Comparison of 3 Tesla multiparametric MRI of the prostate using a flexible surface receiver coil versus conventional surface coil plus endorectal coil setup

PurposeTo subjectively and quantitatively compare the quality of 3 Tesla magnetic resonance imaging of the prostate acquired with a novel flexible surface coil (FSC) and with a conventional endorectal coil (ERC).MethodsSix radiologists independently reviewed 200 pairs of axial, high-resolution T2-weighted and diffusion-weighted image data sets, each containing one examination acquired with the FSC and one with the ERC, respectively. Readers selected their preferred examination from each pair and assessed every single examination using six quality criteria on 4-point scales. Signal-to-noise ratios were measured and compared.ResultsTwo readers preferred FSC acquisition (36.5–45%) over ERC acquisition (13.5–15%) for both sequences combined, and four readers preferred ERC acquisition (41–46%). Analysis of pooled responses for both sequences from all readers shows no significant preference for FSC or ERC. Analysis of the individual sequences revealed a pooled preference for the FSC in T2WI (38.7% vs 17.8%) and for the ERC in DWI (50.9% vs 19.6%). Patients’ weight was the only weak predictor of a preference for the ERC acquisition (p = 0.04). SNR and CNR were significantly higher in the ERC acquisitions (p<0.001) except CNR differentiating tumor lesions from benign prostate (p=0.1).ConclusionAlthough readers have strong individual preferences, comparable subjective image quality can be obtained for prostate MRI with an ERC and the novel FSC. ERC imaging might be particularly valuable for sequences with inherently lower SNR as DWI and larger patients whereas the FSC is generally preferred in T2WI. FSC imaging generates a lower SNR than with an ERC.

The impact of position-orientation adaptive smoothing in diffusion weighted imaging-From diffusion metrics to fiber tractography.

In contrast to commonly used approaches to improve data quality in diffusion weighted imaging, position-orientation adaptive smoothing (POAS) provides an edge-preserving post-processing approach. This study aims to investigate its potential and effects on image quality, diffusion metrics, and fiber tractography of the corticospinal tract in relation to non-post-processed and averaged data. 22 healthy volunteers were included in this study. For each volunteer five clinically applicable diffusion weighted imaging data sets were acquired and post-processed by standard averaging and POAS. POAS post-processing led to significantly higher signal-to-noise-ratios (p < 0.001), lower fractional anisotropy across the whole brain (p < 0.05) and reduced intra-subject variability of diffusion weighted imaging signal intensity and fractional anisotropy (p < 0.001, p = 0.006). Fiber tractography of the corticospinal tract resulted in significantly (p = 0.027, p = 0.014) larger tract volumes while fiber density was the lowest. Similarity across tractography results was highest for POAS post-processed data (p < 0.001). POAS post-processing enhances image quality, decreases the intra-subject variability of signal intensity and fractional anisotropy, increases fiber tract volume of the corticospinal tract, and leads to higher reproducibility of tractography results. Thus, POAS post-processing supports a reliable and more accurate fiber tractography of the corticospinal tract, being mandatory for the clinical use.

Early Diffusion Magnetic Resonance Imaging Changes in Normal-Appearing Brain in Pediatric Moyamoya Disease.

Moyamoya disease often leads to ischemic strokes visible on diffusion-weighted imaging (DWI) and T2-weighted magnetic resonance imaging (MRI) with subsequent cognitive impairment. In adults with moyamoya, apparent diffusion coefficient (ADC) is correlated with regions of steal phenomenon and executive dysfunction prior to white matter changes. To investigate quantitative global diffusion changes in pediatric moyamoya patients prior to explicit structural ischemic damage. We retrospectively reviewed children (<20 yr old) with moyamoya disease and syndrome who underwent bypass surgery at our institution. We identified 29 children with normal structural preoperative MRI and without findings of cortical infarction or chronic white matter ischemic changes. DWI datasets were used to calculate ADC maps for each subject as well as for 60 age-matched healthy controls. Using an atlas-based approach, the cerebral white matter, cerebral cortex, thalamus, caudate, putamen, pallidum, hippocampus, amygdala, nucleus accumbens, and brainstem were segmented in each DWI dataset and used to calculate regional volumes and ADC values. Multivariate analysis of covariance using the regional ADC and volume values as dependent variables and age and gender as covariates revealed a significant difference between the groups (P<.001). Post hoc analysis demonstrated significantly elevated ADC values for children with moyamoya in the cerebral cortex, white matter, caudate, putamen, and nucleus accumbens. No significant volume differences were found. Prior to having bypass surgery, and in the absence of imaging evidence of ischemic stroke, children with moyamoya exhibit cerebral diffusion changes. These findings could reflect microstructural changes stemming from exhaustion of cerebrovascular reserve.

Added value of contrast-enhanced and diffusion-weighted MRI data sets for characterization of perianal fistulas; single center experience

Aim: Our aim in this study is to determine the effectiveness of fat-suppressed T2-weighted (FS-T2W), FS-T2W+Diffusion-weighted image (DWI) and FS-T2W+Contrast-enhanced (CE) magnetic resonance imaging (MRI) data sets in the diagnosis of perianal fistula (PAF).Material and Methods: Retrospective analysis of 207 patients with a history of PAF operation between January 2018 and March 2020 was performed. The inclusion criteria were patients with PAF who underwent MRI prior to surgery and had a detailed intraoperative report. MRI data sets were evaluated by a radiologist, assessed fistula type, internal/external orifice position and presence of abscess/secondary tracts. The Cohen’s kappa (κ) statistic was used to define the level of between from intraoperative findings and MRI data sets agreement. The sensitivity and specificity were compared using the McNemar’s test.Results: All MRI data sets showed the presence of PAF correctly (n = 67, 100%). When the type of PAF per St James and AGA classification, presence of perianal abscesses/secondary tract agreement between intraoperative findings and MRI data sets is evaluated; moderate agreement for isolated FS-T2W data set, almost perfect agreement for combined FS-T2W+DWI and FS-T2W+CE data sets. In the evaluation of agreement for internal/external orifice position; substantial agreement for isolated FS-T2W data set, almost perfect agreement for combined FS-T2W+DWI and FS-T2W+CE data sets. Combined FS-T2W+DWI and FS-T2W+CE data sets have equal sensitivity and specificity each other (p = 0.544) and both data sets higher sensitivity and specificity than isolated FS-T2W data set (p 0.001, p 0.001, respectively).Conclusion: Adding DWI or CE data sets to the FS-T2W data set in the diagnosis of PAF has been shown to detect both diagnosis and complications with high accuracy. Combined FS-T2W+DWI and FS-T2W+CE data sets have equal sensitivity and specificity, and the use of the FS-T2W+DWI data set can prevent unnecessary use of contrast agent.

Segmentation of osteosarcoma tumor using diffusion weighted MRI: a comparative study using nine segmentation algorithms

Osteosarcoma is a primary malignant bone tumor in children and adolescents with significant morbidity and poor prognosis. Diffusion weighted imaging (DWI) plays a crucial role in diagnosis and prognosis of this malignant disease by capturing cellular changes in tumor tissue early in the course of treatment without any contrast injection. Segmentation of tumor in DWI is challenging due to low signal-to-noise ratio, partial-volume effects, intensity inhomogeneities and irregular shape of osteosarcoma. The purpose of this study was to segment osteosarcoma solely utilizing DWI and identify effective and robust technique(s) for tumor segmentation. DWI dataset of fifty-five (N = 55; male:female = 41:14; Age = 17.8 ± 7.4 years) patients with osteosarcoma was acquired before treatment. Total nine automated and semi-automated segmentation algorithms based on (1) Otsu thresholding (OT), (2) Otsu threshold-based region growing (OT-RG), (3) Active contour (AC), (4) Simple linear iterative clustering Superpixels (SLIC-S), (5) Fuzzy c-means clustering (FCM), (6) Graph cut (GC), (7) Logistic regression (LR) (8) Linear support vector machines (L-SVM) and (9) Deep feed-forward neural network (DNN) were implemented. Segmentation accuracy was estimated by Dice coefficient (DC), Jaccard Index (JI), precision (P) and recall (R) using manually demarcated ground-truth tumor mask by a radiologist. Evaluated apparent diffusion coefficient (ADC) in segmented tumor mask and ground-truth tumor mask was compared using paired t test for statistical significance (p < 0.05) and Pearson correlation coefficient (PCC). Automated SLIC-S and FCM showed quantitatively and qualitatively superior segmentation with DC: ~ 79–82%; JI: ~ 67–71%; P: ~ 81–83%; R: ~ 80–86% and PCC = 0.89, 0.88 among all methods. Among semi-automated methods, AC was quantitatively more accurate (DC: ~ 77%; JI: ~ 65%; P: ~ 72%; R: ~ 88%; PCC = 0.85) than OT-RG and GC (DC: ~ 74–75%; JI: ~ 60–61%; P: ~ 67–72%; R: ~ 84–89%; PCC = 0.78, 0.73). Among machine learning algorithms, DNN showed the highest accuracy (DC: ~ 73%; JI: ~ 62%; P: ~ 77%; R: ~ 86%; PCC = 0.79) than LR and L-SVM (DC: ~ 70–71%; JI: ~ 58–63%; P: ~ 73%; R: ~ 74–85%; PCC = 0.69, 0.71). Execution times were instantaneous for SLIC-S, FCM and machine learning methods, while OT-RG, AC and GC took comparable ~ 1–6 s/slice image. Automated SLIC-S, FCM and semi-automated AC methods produced promising tumor segmentation results using DWI of the osteosarcoma dataset.

Replication and generalization in applied neuroimaging

There is much interest in translating neuroimaging findings into meaningful clinical diagnostics. The goal of scientific discoveries differs from clinical diagnostics. Scientific discoveries must replicate under a specific set of conditions; to translate to the clinic we must show that findings using purpose-built scientific instruments will be observable in clinical populations and instruments. Here we describe and evaluate data and computational methods designed to translate a scientific observation to a clinical setting. Using diffusion weighted imaging (DWI), Wahl et al. (2010) observed that across subjects the mean fractional anisotropy (FA) of homologous pairs of tracts is highly correlated. We hypothesize that this is a fundamental biological trait that should be present in most healthy participants, and deviations from this assessment may be a useful diagnostic metric. Using this metric as an illustration of our methods, we analyzed six pairs of homologous white matter tracts in nine different DWI datasets with 44 subjects each. Considering the original FA measurement as a baseline, we show that the new metric is between 2 and 4 times more precise when used in a clinical context. Our framework to translate research findings into clinical practice can be applied, in principle, to other neuroimaging results.

Automated deep learning method for whole‐breast segmentation in diffusion‐weighted breast MRI

Diffusion-weighted imaging (DWI) in MRI plays an increasingly important role in diagnostic applications and developing imaging biomarkers. Automated whole-breast segmentation is an important yet challenging step for quantitative breast imaging analysis. While methods have been developed on dynamic contrast-enhanced (DCE) MRI, automatic whole-breast segmentation in breast DWI MRI is still underdeveloped. To develop a deep/transfer learning-based segmentation approach for DWI MRI scans and conduct an extensive study assessment on four imaging datasets from both internal and external sources. Retrospective. In all, 98 patients (144 MRI scans; 11,035 slices) of four different breast MRI datasets from two different institutions. 1.5T scanners with DCE sequence (Dataset 1 and Dataset 2) and DWI sequence. A 3.0T scanner with one external DWI sequence. Deep learning models (UNet and SegNet) and transfer learning were used as segmentation approaches. The main DCE Dataset (4,251 2D slices from 39 patients) was used for pre-training and internal validation, and an unseen DCE Dataset (431 2D slices from 20 patients) was used as an independent test dataset for evaluating the pre-trained DCE models. The main DWI Dataset (6,343 2D slices from 75 MRI scans of 29 patients) was used for transfer learning and internal validation, and an unseen DWI Dataset (10 2D slices from 10 patients) was used for independent evaluation to the fine-tuned models for DWI segmentation. Manual segmentations by three radiologists (>10-year experience) were used to establish the ground truth for assessment. The segmentation performance was measured using the Dice Coefficient (DC) for the agreement between manual expert radiologist's segmentation and algorithm-generated segmentation. The mean value and standard deviation of the DCs were calculated to compare segmentation results from different deep learning models. For the segmentation on the DCE MRI, the average DC of the UNet was 0.92 (cross-validation on the main DCE dataset) and 0.87 (external evaluation on the unseen DCE dataset), both higher than the performance of the SegNet. When segmenting the DWI images by the fine-tuned models, the average DC of the UNet was 0.85 (cross-validation on the main DWI dataset) and 0.72 (external evaluation on the unseen DWI dataset), both outperforming the SegNet on the same datasets. The internal and independent tests show that the deep/transfer learning models can achieve promising segmentation effects validated on DWI data from different institutions and scanner types. Our proposed approach may provide an automated toolkit to help computer-aided quantitative analyses of breast DWI images. 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;51:635-643.