Cloud-based large-scale curation of medical imaging data using AI segmentation.

Rapid advances in medical imaging Artificial Intelligence (AI) offer unprecedented opportunities for automatic analysis and extraction of data from large imaging collections. Computational demands of such modern AI tools may be difficult to satisfy with the capabilities available on premises. Cloud computing offers the promise of economical access and extreme scalability. Few studies examine the price/performance tradeoffs of using the cloud, in particular for medical image analysis tasks. We investigate the use of cloud-provisioned compute resources for AI-based curation of the National Lung Screening Trial (NLST) Computed Tomography (CT) images available from the National Cancer Institute (NCI) Imaging Data Commons (IDC). We evaluated NCI Cancer Research Data Commons (CRDC) Cloud Resources - Terra (FireCloud) and Seven Bridges-Cancer Genomics Cloud (SB-CGC) platforms - to perform automatic image segmentation with TotalSegmentator and pyradiomics feature extraction for a large cohort containing >126,000 CT volumes from >26,000 patients. Utilizing >21,000 Virtual Machines (VMs) over the course of the computation we completed analysis in under 9 hours, as compared to the estimated 522 days that would be needed on a single workstation. The total cost of utilizing the cloud for this analysis was $1,011.05. Our contributions include: 1) an evaluation of the numerous tradeoffs towards optimizing the use of cloud resources for large-scale image analysis; 2) CloudSegmentator, an open source reproducible implementation of the developed workflows, which can be reused and extended; 3) practical recommendations for utilizing the cloud for large-scale medical image computing tasks. We also share the results of the analysis: the total of 9,565,554 segmentations of the anatomic structures and the accompanying radiomics features in IDC as of release v18.

Enrichment of lung cancer computed tomography collections with AI-derived annotations.

Public imaging datasets are critical for the development and evaluation of automated tools in cancer imaging. Unfortunately, many do not include annotations or image-derived features, complicating downstream analysis. Artificial intelligence-based annotation tools have been shown to achieve acceptable performance and can be used to automatically annotate large datasets. As part of the effort to enrich public data available within NCI Imaging Data Commons (IDC), here we introduce AI-generated annotations for two collections containing computed tomography images of the chest, NSCLC-Radiomics, and a subset of the National Lung Screening Trial. Using publicly available AI algorithms, we derived volumetric annotations of thoracic organs-at-risk, their corresponding radiomics features, and slice-level annotations of anatomical landmarks and regions. The resulting annotations are publicly available within IDC, where the DICOM format is used to harmonize the data and achieve FAIR (Findable, Accessible, Interoperable, Reusable) data principles. The annotations are accompanied by cloud-enabled notebooks demonstrating their use. This study reinforces the need for large, publicly accessible curated datasets and demonstrates how AI can aid in cancer imaging.

National Cancer Institute Imaging Data Commons: Toward Transparency, Reproducibility, and Scalability in Imaging Artificial Intelligence.

The remarkable advances of artificial intelligence (AI) technology are revolutionizing established approaches to the acquisition, interpretation, and analysis of biomedical imaging data. Development, validation, and continuous refinement of AI tools requires easy access to large high-quality annotated datasets, which are both representative and diverse. The National Cancer Institute (NCI) Imaging Data Commons (IDC) hosts large and diverse publicly available cancer image data collections. By harmonizing all data based on industry standards and colocalizing it with analysis and exploration resources, the IDC aims to facilitate the development, validation, and clinical translation of AI tools and address the well-documented challenges of establishing reproducible and transparent AI processing pipelines. Balanced use of established commercial products with open-source solutions, interconnected by standard interfaces, provides value and performance, while preserving sufficient agility to address the evolving needs of the research community. Emphasis on the development of tools, use cases to demonstrate the utility of uniform data representation, and cloud-based analysis aim to ease adoption and help define best practices. Integration with other data in the broader NCI Cancer Research Data Commons infrastructure opens opportunities for multiomics studies incorporating imaging data to further empower the research community to accelerate breakthroughs in cancer detection, diagnosis, and treatment. Published under a CC BY 4.0 license.

Deep Learning Based Parameterization of Diffeomorphic Image Registration for Cardiac Image Segmentation.

Cardiac segmentation from magnetic resonance imaging (MRI) is one of the essential tasks in analyzing the anatomy and function of the heart for the assessment and diagnosis of cardiac diseases. However, cardiac MRI generates hundreds of images per scan, and manual annotation of them is challenging and time-consuming, and therefore processing these images automatically is of interest. This study proposes a novel end-to-end supervised cardiac MRI segmentation framework based on a diffeomorphic deformable registration that can segment cardiac chambers from 2D and 3D images or volumes. To represent actual cardiac deformation, the method parameterizes the transformation using radial and rotational components computed via deep learning, with a set of paired images and segmentation masks used for training. The formulation guarantees transformations that are invertible and prevents mesh folding, which is essential for preserving the topology of the segmentation results. A physically plausible transformation is achieved by employing diffeomorphism in computing the transformations and activation functions that constrain the range of the radial and rotational components. The method was evaluated over three different data sets and showed significant improvements compared to exacting learning and non-learning based methods in terms of the Dice score and Hausdorff distance metrics.

Benefit of stereoscopic volume rendering for the identification of pediatric pulmonary vein stenosis from CT angiography.

The use of three-dimensional (3D) technologies in medical practice is increasing; however, its use is largely untested. One 3D technology, stereoscopic volume-rendered 3D display, can improve depth perception. Pulmonary vein stenosis (PVS) is a rare cardiovascular pathology, often diagnosed by computed tomography (CT), where volume rendering may be useful. Depth cues may be lost when volume rendered CT is displayed on regular screens instead of 3D displays. The objective of this study was to determine whether the 3D stereoscopic display of volume-rendered CT improved perception compared to standard monoscopic display, as measured by PVS diagnosis. CT angiograms (CTAs) from 18 pediatric patients aged 3 weeks to 2 years were volume rendered and displayed with and without stereoscopic display. Patients had 0 to 4 pulmonary vein stenoses. Participants viewed the CTAs in 2 groups with half on monoscopic and half on stereoscopic display and the converse a minimum of 2 weeks later, and their diagnoses were recorded. A total of 24 study participants, comprised of experienced staff cardiologists, cardiovascular surgeons and radiologists, and their trainees viewed the CTAs and assessed the presence and location of PVS. Cases were classified as simple (2 or fewer lesions) or complex (3 or more lesions). Overall, there were fewer type 2 errors in diagnosis for stereoscopic display than standard display, an insignificant difference (p = 0.095). There was a significant decrease in type 2 errors for complex multiple lesion cases (≥3) vs simpler cases (p = 0.027) and improvement in localization of pulmonary veins (p = 0.011). Subjectively, 70% of participants stated that stereoscopy was helpful in the identification of PVS. The stereoscopic display did not result in significantly decreased errors in PVS diagnosis but was helpful for more complex cases.

Promoting the use of the PI-QUAL score for prostate MRI quality: results from the ESOR Nicholas Gourtsoyiannis teaching fellowship.

OBJECTIVES: The Prostate Imaging Quality (PI-QUAL) score is a new metric to evaluate the diagnostic quality of multiparametric magnetic resonance imaging (MRI) of the prostate. This study assesses the impact of an intervention, namely a prostate MRI quality training lecture, on the participant's ability to apply PI-QUAL. METHODS: Sixteen participants (radiologists, urologists, physicists, and computer scientists) of varying experience in reviewing diagnostic prostate MRI all assessed the image quality of ten examinations from different vendors and machines. Then, they attended a dedicated lecture followed by a hands-on workshop on MRI quality assessment using the PI-QUAL score. Five scans assessed by the participants were evaluated in the workshop using the PI-QUAL score for teaching purposes. After the course, the same participants evaluated the image quality of a new set of ten scans applying the PI-QUAL score. Results were assessed using receiver operating characteristic analysis. The reference standard was the PI-QUAL score assessed by one of the developers of PI-QUAL. RESULTS: There was a significant improvement in average area under the curve for the evaluation of image quality from baseline (0.59 [95 % confidence intervals: 0.50-0.66]) to post-teaching (0.96 [0.92-0.98]), an improvement of 0.37 [0.21-0.41] (p < 0.001). CONCLUSIONS: A teaching course (dedicated lecture + hands-on workshop) on PI-QUAL significantly improved the application of this scoring system to assess the quality of prostate MRI examinations. KEY POINTS: • A significant improvement in the application of PI-QUAL for the assessment of prostate MR image quality was observed after an educational intervention. • Appropriate training on image quality can be delivered to those involved in the acquisition and interpretation of prostate MRI. • Further investigation will be needed to understand the impact on improving the acquisition of high-quality diagnostic prostate MR examinations.

A New Semi-automated Algorithm for Volumetric Segmentation of the Left Ventricle in Temporal 3D Echocardiography Sequences.

PURPOSE: Echocardiography is commonly used as a non-invasive imaging tool in clinical practice for the assessment of cardiac function. However, delineation of the left ventricle is challenging due to the inherent properties of ultrasound imaging, such as the presence of speckle noise and the low signal-to-noise ratio. METHODS: We propose a semi-automated segmentation algorithm for the delineation of the left ventricle in temporal 3D echocardiography sequences. The method requires minimal user interaction and relies on a diffeomorphic registration approach. Advantages of the method include no dependence on prior geometrical information, training data, or registration from an atlas. RESULTS: The method was evaluated using three-dimensional ultrasound scan sequences from 18 patients from the Mazankowski Alberta Heart Institute, Edmonton, Canada, and compared to manual delineations provided by an expert cardiologist and four other registration algorithms. The segmentation approach yielded the following results over the cardiac cycle: a mean absolute difference of 1.01 (0.21) mm, a Hausdorff distance of 4.41 (1.43) mm, and a Dice overlap score of 0.93 (0.02). CONCLUSION: The method performed well compared to the four other registration algorithms.

Automated detection of pneumonia in lung ultrasound using deep video classification for COVID-19.

There is a crucial need for quick testing and diagnosis of patients during the COVID-19 pandemic. Lung ultrasound is an imaging modality that is cost-effective, widely accessible, and can be used to diagnose acute respiratory distress syndrome in patients with COVID-19. It can be used to find important characteristics in the images, including A-lines, B-lines, consolidation, and pleural effusion, which all inform the clinician in monitoring and diagnosing the disease. With the use of portable ultrasound transducers, lung ultrasound images can be easily acquired, however, the images are often of poor quality. They often require an expert clinician interpretation, which may be time-consuming and is highly subjective. We propose a method for fast and reliable interpretation of lung ultrasound images by use of deep learning, based on the Kinetics-I3D network. Our learned model can classify an entire lung ultrasound scan obtained at point-of-care, without requiring the use of preprocessing or a frame-by-frame analysis. We compare our video classifier against ground truth classification annotations provided by a set of expert radiologists and clinicians, which include A-lines, B-lines, consolidation, and pleural effusion. Our classification method achieves an accuracy of 90% and an average precision score of 95% with the use of 5-fold cross-validation. The results indicate the potential use of automated analysis of portable lung ultrasound images to assist clinicians in screening and diagnosing patients.

Cardiac MRI-Derived Myocardial Deformation Parameters Correlate with Pulmonary Valve Replacement Indications in Repaired Tetralogy of Fallot.

Right ventricular (RV) volumetric cardiac magnetic resonance (CMR) criteria serve as indicators for pulmonary valve replacement (PVR) in repaired tetralogy of Fallot (rTOF). Myocardial deformation and tricuspid valve displacement parameters may be more sensitive measures of RV dysfunction. This study's aim was to describe rTOF RV deformation and tricuspid displacement patterns using novel CMR semi-automated software and determine associations with standard CMR measures. Retrospective study of 78 pediatric rTOF patients was compared to 44 normal controls. Global RV longitudinal and circumferential strain and strain rate (SR) and tricuspid valve (TV) displacement were measured. Correlation analysis between strain, SR, TV displacement, and volumes was performed between and within subgroups. The sensitivity and specificity of strain parameters in predicting CMR criteria for PVR was determined. Deformation variables were reduced in rTOF compared to controls. Decreased RV strain and TV shortening were associated with increased RV volumes and decreased RVEF. Longitudinal and circumferential parameters were predictive of RVESVi (> 80 ml/m2) and RVEF (< 47%), with circumferential strain (> - 15.88%) and SR (> - 0.62) being most sensitive. Longitudinal strain was unchanged between rTOF subgroups, while circumferential strain trended abnormal in those meeting PVR criteria compared to controls. RV deformation and TV displacement are abnormal in rTOF, and RV circumferential strain variation may reflect an adaptive response to chronic volume or pressure load. This coupled with associations of ventricular deformation with traditional PVR indications suggest importance of this analysis in the evolution of rTOF RV assessment.

Validation of Diffeomorphic Registration on Cine Cardiac MR via Direct Frame-to-Frame Comparison with HARP Tracking on Tagged MR.

Cardiac magnetic resonance (MR) tissue tagging offers an excellent solution for tracking deformation and is considered the reference standard for the quantification of strain. However, due to the requirements for a dedicated acquisition sequence and post-processing software, tagged MR acquisitions are performed much less frequently in routine clinical practice than the anatomical cine MR sequence. Using tagged MR as the reference standard, this study proposes an approach to evaluate a diffeomorphic image registration algorithm applied on cine MR images to compute the cardiac deformation. In contrast to previous evaluation methods that compared the final results, such as strain, computed from cine and tagged MR sequences, the proposed method performs a direct frame-to-frame comparison in the evaluation. To overcome the problem of misalignment between the tagged and cine MR images, the proposed approach performs transformations to and from the two-dimensional image pixel coordinates and three-dimensional space using the meta-information encoded in the MR images. Linear temporal interpolation is performed using the frame acquisition time since the last R-wave peak value of the electrocardiogram signal recorded in the meta-information. Several statistic measures are computed and reported for the registration error using the Euclidean distances between the corresponding set of points obtained using cine and tagged MR images.

Validation of a diffeomorphic registration algorithm using true deformation computed from thin plate spline interpolation.

Despite the inter and intraobserver variabilities, manual contours are commonly used as surrogates for ground truth in the validation process for nonrigid medical image registration. In contrast, this study proposes the use of thin plate spline interpolation to create a true deformation field. A diffeomorphic registration method was compared to the true deformation field along with three other algorithms and was evaluated on simulated cardiac motion deformation over 10 subjects from the Automated Cardiac Diagnosis Challenge (ACDC) dataset. Two sequential registration approaches were undertaken: with respect to the first frame, and with respect to the previous frame. The Dice score was calculated between the simulated and warped contours for the two approaches: diffeomorphic registration method =0.991 and 0.997, RealTITracker (L2L2 method) = 0.971 and 0.977, RealTITracker (L2L1 method) = 0.975 and 0.978, and Elastix = 0.976 and 0.994. The results demonstrate the robust performance of the diffeomorphic registration method.Clinical relevance This establishes a validation of a registration method that can be used for segmentation of chambers of the heart.

A Novel 4D Semi-Automated Algorithm for Volumetric Segmentation in Echocardiography.

Segmentation of the left ventricle (LV) in temporal 3D echocardiography sequences poses a challenge. However, it is an essential component in generating quantitative clinical measurements for the diagnosis and treatment of various cardiac diseases. Identifying the endocardial borders of the left ventricle can be difficult due to the inherent properties of ultrasound. This study proposes a 4D segmentation algorithm that segments over temporal 3D volumes that has minimal user interaction and is based on a diffeomorphic registration approach. In contrast to several existing algorithms, the proposed method does not depend on training data or make any geometrical assumptions. The algorithm was evaluated on seven patients obtained from the Mazankowski Alberta Heart Institute, Edmonton, Canada in comparison to expert manual segmentation. The proposed approach yielded Dice scores of 0.94 (0.01), 0.91 (0.03) and 0.92 (0.02) at end diastole, at end systole and over the entire cardiac cycle, respectively. The corresponding Hausdorff distance values were 4.49 (1.01) mm, 4.94 (1.41) mm, and 5.05 (0.85) mm, respectively. These results demonstrate that the proposed 4D segmentation approach for the left ventricle is robust and can potentially be used in clinical practice.

Magnetoencephalography-based identification of functional connectivity network disruption following mild traumatic brain injury.

Mild traumatic brain injury (mTBI) leads to long-term cognitive sequelae in a significant portion of patients. Disruption of normal neural communication across functional brain networks may explain the deficits in memory and attention observed after mTBI. In this study, we used magnetoencephalography (MEG) to examine functional connectivity during a resting state in a group of mTBI subjects (n = 9) compared with age-matched control subjects (n = 15). We adopted a data-driven, exploratory analysis in source space using phase locking value across different frequency bands. We observed a significant reduction in functional connectivity in band-specific networks in mTBI compared with control subjects. These networks spanned multiple cortical regions involved in the default mode network (DMN). The DMN is thought to subserve memory and attention during periods when an individual is not engaged in a specific task, and its disruption may lead to cognitive deficits after mTBI. We further applied graph theoretical analysis on the functional connectivity matrices. Our data suggest reduced local efficiency in different brain regions in mTBI patients. In conclusion, MEG can be a potential tool to investigate and detect network alterations in patients with mTBI. The value of MEG to reveal potential neurophysiological biomarkers for mTBI patients warrants further exploration.

An exploratory analysis linking neuropsychological testing to quantification of tractography using High Definition Fiber Tracking (HDFT) in military TBI.

To realize the potential value of tractography in traumatic brain injury (TBI), we must identify metrics that provide meaningful information about functional outcomes. The current study explores quantitative metrics describing the spatial properties of tractography from advanced diffusion imaging (High Definition Fiber Tracking, HDFT). In a small number of right-handed males from military TBI (N = 7) and civilian control (N = 6) samples, both tract homologue symmetry and tract spread (proportion of brain mask voxels contacted) differed for several tracts among civilian controls and extreme groups in the TBI sample (high scorers and low scorers) for verbal recall, serial reaction time, processing speed index, and trail-making. Notably, proportion of voxels contacted in the arcuate fasciculus distinguished high and low performers on the CVLT-II and PSI, potentially reflecting linguistic task demands, and GFA in the left corticospinal tract distinguished high and low performers in PSI and Trail Making Test Part A, potentially reflecting right hand motor response demands. The results suggest that, for advanced diffusion imaging, spatial properties of tractography may add analytic value to measures of tract anisotropy.

Quantifying white matter structural integrity with high-definition fiber tracking in traumatic brain injury.

There is an urgent, unmet demand for definitive biological diagnosis of traumatic brain injury (TBI) to pinpoint the location and extent of damage. We have developed High-Definition Fiber Tracking, a 3 T magnetic resonance imaging-based diffusion spectrum imaging and tractography analysis protocol, to quantify axonal injury in military and civilian TBI patients. A novel analytical methodology quantified white matter integrity in patients with TBI and healthy controls. Forty-one subjects (23 TBI, 18 controls) were scanned with the High-Definition Fiber Tracking diffusion spectrum imaging protocol. After reconstruction, segmentation was used to isolate bilateral hemisphere homologues of eight major tracts. Integrity of segmented tracts was estimated by calculating homologue correlation and tract coverage. Both groups showed high correlations for all tracts. TBI patients showed reduced homologue correlation and tract spread and increased outlier count (correlations>2.32 SD below control mean). On average, 6.5% of tracts in the TBI group were outliers with substantial variability among patients. Number and summed deviation of outlying tracts correlated with initial Glasgow Coma Scale score and 6-month Glasgow Outcome Scale-Extended score. The correlation metric used here can detect heterogeneous damage affecting a low proportion of tracts, presenting a potential mechanism for advancing TBI diagnosis.

Regulatory network analysis acceleration with reconfigurable hardware.

In medical research it is of great importance to be able to quickly obtain answers to inquiries about system response to different stimuli. Modeling the dynamics of biological regulatory networks is a promising approach to achieve this goal, but existing modeling approaches suffer from complexity issues and become inefficient with large networks. In order to improve the efficiency, we propose the implementation of models of regulatory networks in hardware, which allows for highly parallel simulation of these networks. We find that our FPGA implementation of an example model of peripheral naïve T cell differentiation provides five orders of magnitude speedup when compared to software simulation.