Metadata-conditioned generative models to synthesize anatomically-plausible 3D brain MRIs

Recent advances in generative models have paved the way for enhanced generation of natural and medical images, including synthetic brain MRIs. However, the mainstay of current AI research focuses on optimizing synthetic MRIs with respect to visual quality (such as signal-to-noise ratio) while lacking insights into their relevance to neuroscience. To generate high-quality T1-weighted MRIs relevant for neuroscience discovery, we present a two-stage Diffusion Probabilistic Model (called BrainSynth) to synthesize high-resolution MRIs conditionally-dependent on metadata (such as age and sex). We then propose a novel procedure to assess the quality of BrainSynth according to how well its synthetic MRIs capture macrostructural properties of brain regions and how accurately they encode the effects of age and sex. Results indicate that more than half of the brain regions in our synthetic MRIs are anatomically plausible, i.e., the effect size between real and synthetic MRIs is small relative to biological factors such as age and sex. Moreover, the anatomical plausibility varies across cortical regions according to their geometric complexity. As is, the MRIs generated by BrainSynth significantly improve the training of a predictive model to identify accelerated aging effects in an independent study. These results indicate that our model accurately capture the brain's anatomical information and thus could enrich the data of underrepresented samples in a study. The code of BrainSynth will be released as part of the MONAI project at https://github.com/Project-MONAI/GenerativeModels.

PCP-GC-LM: single-sequence-based protein contact prediction using dual graph convolutional neural network and convolutional neural network

BACKGROUND

Recently, the process of evolution information and the deep learning network has promoted the improvement of protein contact prediction methods. Nevertheless, still remain some bottleneck: (1) One of the bottlenecks is the prediction of orphans and other fewer evolution information proteins. (2) The other bottleneck is the method of predicting single-sequence-based proteins mainly focuses on selecting protein sequence features and tuning the neural network architecture, However, while the deeper neural networks improve prediction accuracy, there is still the problem of increasing the computational burden. Compared with other neural networks in the field of protein prediction, the graph neural network has the following advantages: due to the advantage of revealing the topology structure via graph neural network and being able to take advantage of the hierarchical structure and local connectivity of graph neural networks has certain advantages in capturing the features of different levels of abstraction in protein molecules. When using protein sequence and structure information for joint training, the dependencies between the two kinds of information can be better captured. And it can process protein molecular structures of different lengths and shapes, while traditional neural networks need to convert proteins into fixed-size vectors or matrices for processing.

RESULTS

Here, we propose a single-sequence-based protein contact map predictor PCP-GC-LM, with dual-level graph neural networks and convolution networks. Our method performs better with other single-sequence-based predictors in different independent tests. In addition, to verify the validity of our method against complex protein structures, we will also compare it with other methods in two homodimers protein test sets (DeepHomo test dataset and CASP-CAPRI target dataset). Furthermore, we also perform ablation experiments to demonstrate the necessity of a dual graph network. In all, our framework presents new modules to accurately predict inter-chain contact maps in protein and it's also useful to analyze interactions in other types of protein complexes.

A Clinical and Imaging Fused Deep Learning Model Matches Expert Clinician Prediction of 90-Day Stroke Outcomes

BACKGROUND AND PURPOSE

Predicting long-term clinical outcome in acute ischemic stroke is beneficial for prognosis, clinical trial design, resource management, and patient expectations. This study used a deep learning-based predictive model (DLPD) to predict 90-day mRS outcomes and compared its predictions with those made by physicians.

MATERIALS AND METHODS

A previously developed DLPD that incorporated DWI and clinical data from the acute period was used to predict 90-day mRS outcomes in 80 consecutive patients with acute ischemic stroke from a single-center registry. We assessed the predictions of the model alongside those of 5 physicians (2 stroke neurologists and 3 neuroradiologists provided with the same imaging and clinical information). The primary analysis was the agreement between the ordinal mRS predictions of the model or physician and the ground truth using the Gwet Agreement Coefficient. We also evaluated the ability to identify unfavorable outcomes (mRS >2) using the area under the curve, sensitivity, and specificity. Noninferiority analyses were undertaken using limits of 0.1 for the Gwet Agreement Coefficient and 0.05 for the area under the curve analysis. The accuracy of prediction was also assessed using the mean absolute error for prediction, percentage of predictions ±1 categories away from the ground truth (±1 accuracy [ACC]), and percentage of exact predictions (ACC).

RESULTS

To predict the specific mRS score, the DLPD yielded a Gwet Agreement Coefficient score of 0.79 (95% CI, 0.71-0.86), surpassing the physicians' score of 0.76 (95% CI, 0.67-0.84), and was noninferior to the readers (P < .001). For identifying unfavorable outcome, the model achieved an area under the curve of 0.81 (95% CI, 0.72-0.89), again noninferior to the readers' area under the curve of 0.79 (95% CI, 0.69-0.87) (P < .005). The mean absolute error, ±1ACC, and ACC were 0.89, 81%, and 36% for the DLPD.

CONCLUSIONS

A deep learning method using acute clinical and imaging data for long-term functional outcome prediction in patients with acute ischemic stroke, the DLPD, was noninferior to that of clinical readers.

Predicting FDG-PET Images From Multi-Contrast MRI Using Deep Learning in Patients With Brain Neoplasms

BACKGROUND

18 F-fluorodeoxyglucose (FDG) positron emission tomography (PET) is valuable for determining presence of viable tumor, but is limited by geographical restrictions, radiation exposure, and high cost.

PURPOSE

To generate diagnostic-quality PET equivalent imaging for patients with brain neoplasms by deep learning with multi-contrast MRI.

STUDY TYPE

Retrospective.

SUBJECTS

Patients (59 studies from 51 subjects; age 56 ± 13 years; 29 males) who underwent 18 F-FDG PET and MRI for determining recurrent brain tumor.

FIELD STRENGTH/SEQUENCE

3T; 3D GRE T1, 3D GRE T1c, 3D FSE T2-FLAIR, and 3D FSE ASL, 18 F-FDG PET imaging.

ASSESSMENT

Convolutional neural networks were trained using four MRIs as inputs and acquired FDG PET images as output. The agreement between the acquired and synthesized PET was evaluated by quality metrics and Bland-Altman plots for standardized uptake value ratio. Three physicians scored image quality on a 5-point scale, with score ≥3 as high-quality. They assessed the lesions on a 5-point scale, which was binarized to analyze diagnostic consistency of the synthesized PET compared to the acquired PET.

STATISTICAL TESTS

The agreement in ratings between the acquired and synthesized PET were tested with Gwet's AC and exact Bowker test of symmetry. Agreement of the readers was assessed by Gwet's AC. P = 0.05 was used as the cutoff for statistical significance.

RESULTS

The synthesized PET visually resembled the acquired PET and showed significant improvement in quality metrics (+21.7% on PSNR, +22.2% on SSIM, -31.8% on RSME) compared with ASL. A total of 49.7% of the synthesized PET were considered as high-quality compared to 73.4% of the acquired PET which was statistically significant, but with distinct variability between readers. For the positive/negative lesion assessment, the synthesized PET had an accuracy of 87% but had a tendency to overcall.

CONCLUSION

The proposed deep learning model has the potential of synthesizing diagnostic quality FDG PET images without the use of radiotracers.

EVIDENCE LEVEL

3 TECHNICAL EFFICACY: Stage 2.

[Construction and preliminary validation of a risk prediction model for the recurrence of diabetic foot ulcer in diabetic patients]

Objective: To develop a risk prediction model for the recurrence of diabetic foot ulcer (DFU) in diabetic patients and primarily validate its predictive value. Methods: Meta-analysis combined with retrospective cohort study was conducted. The Chinese and English papers on risk factors related to DFU recurrence publicly published in China Biology Medicine disc, China National Knowledge Infrastructure, Wanfang Database, VIP Database, and PubMed, Embase, Cochrane Library, and Web of Science, and the search time was from the establishment date of each database until March 31st, 2022. The papers were screened and evaluated, the data were extracted, a meta-analysis was performed using RevMan 5.4.1 statistical software to screen risk factors for DFU recurrence, and Egger's linear regression was used to assess the publication bias of the study results. Risk factors for DFU recurrence mentioned in ≥3 studies and with statistically significant differences in the meta-analysis were selected as the independent variables to develop a logistic regression model for risk prediction of DFU recurrence. The medical records of 101 patients with DFU who met the inclusion criteria and were admitted to Affiliated Hospital of Guizhou Medical University from January 2019 to June 2022 were collected. There were 69 males and 32 females, aged (63±14) years. The receiver operating characteristic (ROC) curve of the predictive performance of the above constructed predictive model for DFU recurrence was drawn, and the area under the ROC curve, maximum Youden index, and sensitivity and specificity at the point were calculated. Dataset including data of 8 risk factors for DFU recurrence and the DFU recurrence rates of 10 000 cases was simulated using RStudio software and a scatter plot was drawn to determine two probabilities for risk division of DFU recurrence. Using the β coefficients corresponding to 8 DFU recurrence risk factors ×10 and taking the integer as the score of coefficient weight of each risk factor, the total score was obtained by summing up, and the cutoff scores for risk level division were calculated based on the total score × two probabilities for risk division of DFU recurrence. Results: Finally, 20 papers were included, including 3 case-control studies and 17 cohort studies, with a total of 4 238 cases and DFU recurrence rate of 22.7% to 71.2%. Meta-analysis showed that glycosylated hemoglobin >7.5% and with plantar ulcer, diabetic peripheral neuropathy, diabetic peripheral vascular disease, smoking, osteomyelitis, history of amputation/toe amputation, and multidrug-resistant bacterial infection were risk factors for the recurrence of DFU (with odds ratios of 3.27, 3.66, 4.05, 3.94, 1.98, 7.17, 11.96, 3.61, 95% confidence intervals of 2.79-3.84, 2.06-6.50, 2.50-6.58, 2.65-5.84, 1.65-2.38, 2.29-22.47, 4.60-31.14, 3.13-4.17, respectively, P<0.05). There were no statistically significant differences in publication biases of diabetic peripheral neuropathy, diabetic peripheral vascular disease, glycosylated hemoglobin >7.5%, plantar ulcer, smoking, multidrug-resistant bacterial infection, or osteomyelitis (P>0.05), but there was a statistically significant difference in the publication bias of amputation/toe amputation (t=-30.39, P<0.05). The area under the ROC curve of the predictive model was 0.81 (with 95% confidence interval of 0.71-0.91) and the maximum Youden index was 0.59, at which the sensitivity was 72% and the specificity was 86%. Ultimately, 29.0% and 44.8% were identified respectively as the cutoff for dividing the probability of low risk and medium risk, and medium risk and high risk for DFU recurrence, while the corresponding total scores of low, medium, and high risks of DFU recurrence were <37, 37-57, and 58-118, respectively. Conclusions: Eight risk factors for DFU recurrence are screened through meta-analysis and the risk prediction model for DFU recurrence is developed, which has moderate predictive accuracy and can provide guidance for healthcare workers to take interventions for patient with DFU recurrence risk.

LSOR: Longitudinally-Consistent Self-Organized Representation Learning

Interpretability is a key issue when applying deep learning models to longitudinal brain MRIs. One way to address this issue is by visualizing the high-dimensional latent spaces generated by deep learning via self-organizing maps (SOM). SOM separates the latent space into clusters and then maps the cluster centers to a discrete (typically 2D) grid preserving the high-dimensional relationship between clusters. However, learning SOM in a high-dimensional latent space tends to be unstable, especially in a self-supervision setting. Furthermore, the learned SOM grid does not necessarily capture clinically interesting information, such as brain age. To resolve these issues, we propose the first self-supervised SOM approach that derives a high-dimensional, interpretable representation stratified by brain age solely based on longitudinal brain MRIs (i.e., without demographic or cognitive information). Called Longitudinally-consistent Self-Organized Representation learning (LSOR), the method is stable during training as it relies on soft clustering (vs. the hard cluster assignments used by existing SOM). Furthermore, our approach generates a latent space stratified according to brain age by aligning trajectories inferred from longitudinal MRIs to the reference vector associated with the corresponding SOM cluster. When applied to longitudinal MRIs of the Alzheimer's Disease Neuroimaging Initiative (ADNI, N = 632 ), LSOR generates an interpretable latent space and achieves comparable or higher accuracy than the state-of-the-art representations with respect to the downstream tasks of classification (static vs. progressive mild cognitive impairment) and regression (determining ADAS-Cog score of all subjects). The code is available at https://github.com/ouyangjiahong/longitudinal-som-single-modality.

Generative Adversarial Network-Enhanced Ultra-Low-Dose [18F]-PI-2620 τ PET/MRI in Aging and Neurodegenerative Populations

BACKGROUND AND PURPOSE

With the utility of hybrid τ PET/MR imaging in the screening, diagnosis, and follow-up of individuals with neurodegenerative diseases, we investigated whether deep learning techniques can be used in enhancing ultra-low-dose [18F]-PI-2620 τ PET/MR images to produce diagnostic-quality images.

MATERIALS AND METHODS

Forty-four healthy aging participants and patients with neurodegenerative diseases were recruited for this study, and [18F]-PI-2620 τ PET/MR data were simultaneously acquired. A generative adversarial network was trained to enhance ultra-low-dose τ images, which were reconstructed from a random sampling of 1/20 (approximately 5% of original count level) of the original full-dose data. MR images were also used as additional input channels. Region-based analyses as well as a reader study were conducted to assess the image quality of the enhanced images compared with their full-dose counterparts.

RESULTS

The enhanced ultra-low-dose τ images showed apparent noise reduction compared with the ultra-low-dose images. The regional standard uptake value ratios showed that while, in general, there is an underestimation for both image types, especially in regions with higher uptake, when focusing on the healthy-but-amyloid-positive population (with relatively lower τ uptake), this bias was reduced in the enhanced ultra-low-dose images. The radiotracer uptake patterns in the enhanced images were read accurately compared with their full-dose counterparts.

CONCLUSIONS

The clinical readings of deep learning-enhanced ultra-low-dose τ PET images were consistent with those performed with full-dose imaging, suggesting the possibility of reducing the dose and enabling more frequent examinations for dementia monitoring.

Functional Outcome Prediction in Acute Ischemic Stroke Using a Fused Imaging and Clinical Deep Learning Model

BACKGROUND

Predicting long-term clinical outcome based on the early acute ischemic stroke information is valuable for prognostication, resource management, clinical trials, and patient expectations. Current methods require subjective decisions about which imaging features to assess and may require time-consuming postprocessing. This study's goal was to predict ordinal 90-day modified Rankin Scale (mRS) score in acute ischemic stroke patients by fusing a Deep Learning model of diffusion-weighted imaging images and clinical information from the acute period.

METHODS

A total of 640 acute ischemic stroke patients who underwent magnetic resonance imaging within 1 to 7 days poststroke and had 90-day mRS follow-up data were randomly divided into 70% (n=448) for model training, 15% (n=96) for validation, and 15% (n=96) for internal testing. Additionally, external testing on a cohort from Lausanne University Hospital (n=280) was performed to further evaluate model generalization. Accuracy for ordinal mRS, accuracy within ±1 mRS category, mean absolute prediction error, and determination of unfavorable outcome (mRS score >2) were evaluated for clinical only, imaging only, and 2 fused clinical-imaging models.

RESULTS

The fused models demonstrated superior performance in predicting ordinal mRS score and unfavorable outcome in both internal and external test cohorts when compared with the clinical and imaging models. For the internal test cohort, the top fused model had the highest area under the curve of 0.92 for unfavorable outcome prediction and the lowest mean absolute error (0.96 [95% CI, 0.77-1.16]), with the highest proportion of mRS score predictions within ±1 category (79% [95% CI, 71%-88%]). On the external Lausanne University Hospital cohort, the best fused model had an area under the curve of 0.90 for unfavorable outcome prediction and outperformed other models with an mean absolute error of 0.90 (95% CI, 0.79-1.01), and the highest percentage of mRS score predictions within ±1 category (83% [95% CI, 78%-87%]).

CONCLUSIONS

A Deep Learning-based imaging model fused with clinical variables can be used to predict 90-day stroke outcome with reduced subjectivity and user burden.

[Summary of the best evidence on exercise for the prevention and treatment of diabetic foot]

Objective: To summarize the best evidence on exercise for the prevention and treatment of diabetic foot. Methods: A bibliometric approach was used. Systematic searches were carried out to retrieve all the publicly published evidences till July 2022 on exercise for the prevention and treatment of diabetic foot, including guidelines, evidence summary, recommended practices, expert consensus, systematic review, and original research, from foreign language databases including BMJ Best Practice, UpToDate, Joanna Briggs Institute Evidence-Based Practice Database, Cochrane Library, Embase, PubMed, Guideline International Network, National Guideline Clearinghouse, Chinese databases including China National Knowledge Infrastructure, Wanfang Database, VIP Database, China Biology Medicine disc, China Clinical Guidelines Library, and the official websites of relevant academic organizations including National Institute for Health and Care Excellence of the United Kingdom, Registered Nurses' Association of Ontario of Canada, the International Working Group on the Diabetic Foot, International Diabetes Federation, American College of Sports Medicine, American Diabetes Association, and Chinese Diabetes Society. The literature was screened and evaluated for the quality, from which the evidences were extracted and evaluated to summarize the best evidences. Results: Nine guidelines, three expert consensuses, one evidence summary (with two systematic reviews being traced), two systematic reviews, 6 randomized controlled trials were retrieved and included, with good quality of literature. Totally 33 pieces of best evidences on exercise for the prevention and treatment of diabetic foot were summarized from the aspects of appropriate exercise prevention of diabetic foot, exercise therapy of diabetic foot, precautions for exercise, health education, and establishment of a multidisciplinary limb salvage team. Conclusions: Totally 33 pieces of best evidences on exercise for the prevention and treatment of diabetic foot were summarized from 5 aspects, providing decision-making basis for clinical guidance on exercise practice for patients with diabetic foot.

Predicting Hypoperfusion Lesion and Target Mismatch in Stroke from Diffusion-weighted MRI Using Deep Learning

Background Perfusion imaging is important to identify a target mismatch in stroke but requires contrast agents and postprocessing software. Purpose To use a deep learning model to predict the hypoperfusion lesion in stroke and identify patients with a target mismatch profile from diffusion-weighted imaging (DWI) and clinical information alone, using perfusion MRI as the reference standard. Materials and Methods Imaging data sets of patients with acute ischemic stroke with baseline perfusion MRI and DWI were retrospectively reviewed from multicenter data available from 2008 to 2019 (Imaging Collaterals in Acute Stroke, Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution 2, and University of California, Los Angeles stroke registry). For perfusion MRI, rapid processing of perfusion and diffusion software automatically segmented the hypoperfusion lesion (time to maximum, ≥6 seconds) and ischemic core (apparent diffusion coefficient [ADC], ≤620 × 10-6 mm2/sec). A three-dimensional U-Net deep learning model was trained using baseline DWI, ADC, National Institutes of Health Stroke Scale score, and stroke symptom sidedness as inputs, with the union of hypoperfusion and ischemic core segmentation serving as the ground truth. Model performance was evaluated using the Dice score coefficient (DSC). Target mismatch classification based on the model was compared with that of the clinical-DWI mismatch approach defined by the DAWN trial by using the McNemar test. Results Overall, 413 patients (mean age, 67 years ± 15 [SD]; 207 men) were included for model development and primary analysis using fivefold cross-validation (247, 83, and 83 patients in the training, validation, and test sets, respectively, for each fold). The model predicted the hypoperfusion lesion with a median DSC of 0.61 (IQR, 0.45-0.71). The model identified patients with target mismatch with a sensitivity of 90% (254 of 283; 95% CI: 86, 93) and specificity of 77% (100 of 130; 95% CI: 69, 83) compared with the clinical-DWI mismatch sensitivity of 50% (140 of 281; 95% CI: 44, 56) and specificity of 89% (116 of 130; 95% CI: 83, 94) (P < .001 for all). Conclusion A three-dimensional U-Net deep learning model predicted the hypoperfusion lesion from diffusion-weighted imaging (DWI) and clinical information and identified patients with a target mismatch profile with higher sensitivity than the clinical-DWI mismatch approach. ClinicalTrials.gov registration nos. NCT02225730, NCT01349946, NCT02586415 © RSNA, 2022 Supplemental material is available for this article. See also the editorial by Kallmes and Rabinstein in this issue.

Self-supervised learning of neighborhood embedding for longitudinal MRI

In recent years, several deep learning models recommend first to represent Magnetic Resonance Imaging (MRI) as latent features before performing a downstream task of interest (such as classification or regression). The performance of the downstream task generally improves when these latent representations are explicitly associated with factors of interest. For example, we derived such a representation for capturing brain aging by applying self-supervised learning to longitudinal MRIs and then used the resulting encoding to automatically identify diseases accelerating the aging of the brain. We now propose a refinement of this representation by replacing the linear modeling of brain aging with one that is consistent in local neighborhoods in the latent space. Called Longitudinal Neighborhood Embedding (LNE), we derive an encoding so that neighborhoods are age-consistent (i.e., brain MRIs of different subjects with similar brain ages are in close proximity of each other) and progression-consistent, i.e., the latent space is defined by a smooth trajectory field where each trajectory captures changes in brain ages between a pair of MRIs extracted from a longitudinal sequence. To make the problem computationally tractable, we further propose a strategy for mini-batch sampling so that the resulting local neighborhoods accurately approximate the ones that would be defined based on the whole cohort. We evaluate LNE on three different downstream tasks: (1) to predict chronological age from T1-w MRI of 274 healthy subjects participating in a study at SRI International; (2) to distinguish Normal Control (NC) from Alzheimer's Disease (AD) and stable Mild Cognitive Impairment (sMCI) from progressive Mild Cognitive Impairment (pMCI) based on T1-w MRI of 632 participants of the Alzheimer's Disease Neuroimaging Initiative (ADNI); and (3) to distinguish no-to-low from moderate-to-heavy alcohol drinkers based on fractional anisotropy derived from diffusion tensor MRIs of 764 adolescents recruited by the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA). Across the three data sets, the visualization of the smooth trajectory vector fields and superior accuracy on downstream tasks demonstrate the strength of the proposed method over existing self-supervised methods in extracting information related to brain aging, which could help study the impact of substance use and neurodegenerative disorders. The code is available at https://github.com/ouyangjiahong/longitudinal-neighbourhood-embedding.

Disentangling Normal Aging From Severity of Disease via Weak Supervision on Longitudinal MRI

The continuous progression of neurological diseases are often categorized into conditions according to their severity. To relate the severity to changes in brain morphometry, there is a growing interest in replacing these categories with a continuous severity scale that longitudinal MRIs are mapped onto via deep learning algorithms. However, existing methods based on supervised learning require large numbers of samples and those that do not, such as self-supervised models, fail to clearly separate the disease effect from normal aging. Here, we propose to explicitly disentangle those two factors via weak-supervision. In other words, training is based on longitudinal MRIs being labelled either normal or diseased so that the training data can be augmented with samples from disease categories that are not of primary interest to the analysis. We do so by encouraging trajectories of controls to be fully encoded by the direction associated with brain aging. Furthermore, an orthogonal direction linked to disease severity captures the residual component from normal aging in the diseased cohort. Hence, the proposed method quantifies disease severity and its progression speed in individuals without knowing their condition. We apply the proposed method on data from the Alzheimer's Disease Neuroimaging Initiative (ADNI, N =632 ). We then show that the model properly disentangled normal aging from the severity of cognitive impairment by plotting the resulting disentangled factors of each subject and generating simulated MRIs for a given chronological age and condition. Moreover, our representation obtains higher balanced accuracy when used for two downstream classification tasks compared to other pre-training approaches. The code for our weak-supervised approach is available at https://github.com/ouyangjiahong/longitudinal-direction-disentangle.

Automated detection of arterial landmarks and vascular occlusions in patients with acute stroke receiving digital subtraction angiography using deep learning

BACKGROUND

Digital subtraction angiography (DSA) is the gold-standard method of assessing arterial blood flow and blockages prior to endovascular thrombectomy.

OBJECTIVE

To detect anatomical features and arterial occlusions with DSA using artificial intelligence techniques.

METHODS

We included 82 patients with acute ischemic stroke who underwent DSA imaging and whose carotid terminus was visible in at least one run. Two neurointerventionalists labeled the carotid location (when visible) and vascular occlusions on 382 total individual DSA runs. For detecting the carotid terminus, positive and negative image patches (either containing or not containing the internal carotid artery terminus) were extracted in a 1:1 ratio. Two convolutional neural network architectures (ResNet-50 pretrained on ImageNet and ResNet-50 trained from scratch) were evaluated. Area under the curve (AUC) of the receiver operating characteristic and pixel distance from the ground truth were calculated. The same training and analysis methods were used for detecting arterial occlusions.

RESULTS

The ResNet-50 trained from scratch most accurately detected the carotid terminus (AUC 0.998 (95% CI 0.997 to 0.999), p<0.00001) and arterial occlusions (AUC 0.973 (95% CI 0.971 to 0.975), p<0.0001). Average pixel distances from ground truth for carotid terminus and occlusion localization were 63±45 and 98±84, corresponding to approximately 1.26±0.90 cm and 1.96±1.68 cm for a standard angiographic field-of-view.

CONCLUSION

These results may serve as an unbiased standard for clinical stroke trials, as optimal standardization would be useful for core laboratories in endovascular thrombectomy studies, and also expedite decision-making during DSA-based procedures.

Mechanism of ALDH2 improves the neuronal damage caused by hypoxia/reoxygenation

OBJECTIVE

To investigate the protective effect and mechanism of ALDH2 on PC12 cells and brain nerve tissue injury under hypoxia.

MATERIALS AND METHODS

The hypoxia model of PC12 cells with low ALDH2 expression was established and screened. The eukaryotic expression vector of wild type pEGFP-N1-ALDH2 and blank plasmid pEGFP-N1 were constructed and transfected into PC12 hypoxia cells respectively. After reoxygenation culture, the morphology, quantity, ALDH2 expression level and apoptosis rate of the two groups were observed, and the role of ALDH2 in cell hypoxia injury was analyzed. Eighty SD rats were randomly divided into model group (ischemia-reperfusion injury group), Alda-1 group (intraperitoneal injection of alda-1 12 hours before and after modeling), DMSO group (intraperitoneal injection of dimethyl sulfoxide) and sham operation group, with 20 rats in each group. The neurobehavioral score, apoptosis rate of nerve cells, the content and activity of ALDH2 in active cerebral cortex and hippocampal CA1 area were compared.

RESULTS

The number of PC12 cells in hypoxia group was lower than that in control group. The expression level of ALDH2 protein in PC12 cells after 4 hours of hypoxia was lower than that in normal culture group. The number of PC12 cells transfected with wild-type recombinant plasmid was significantly more than that of blank plasmid group. Compared with the hypoxia group, the pre apoptotic and post apoptotic cells in wild type transfection group decreased after hypoxia treatment. Compared with sham operation group, nerve injury and apoptosis were increased in group M and DMSO, while ALDH2 activity and expression did not change significantly. Compared with M group and DMSO group, the nerve injury and apoptosis in Alda-1 group were improved, ALDH2 activity was increased, and ALDH2 expression was not significantly changed in Alda-1 group.

CONCLUSIONS

Increasing the expression of ALDH2 or enhancing the activity of ALDH2 can improve the injury of neurons induced by hypoxia/reoxygenation.

Abstract 8: Hypoperfusion Lesion And Target Mismatch Prediction In Acute Ischemic Stroke From Baseline Mr Diffusion Imaging Using A 3d Convolutional Neural Network

Purpose: Perfusion imaging assesses target mismatch but requires contrast and processing software. Clinical/diffusion mismatch can miss cases that have target mismatch and could benefit from thrombectomy. We explored whether a neural network can predict hypoperfusion and identify target mismatch from diffusion-weighted imaging (DWI) and clinical information alone.

Methods: Acute ischemic stroke cases with baseline MR perfusion and DWI were included from two multi-center trials and one registry for model development and a separate randomized trial for external validation. MR perfusion images were processed by RAPID, which segments Tmax lesion (Tmax≥6s) and the ischemic core lesion (apparent diffusion coefficient [ADC]≤ 620). A 3D U-Net was trained using baseline DWI, ADC, NIH stroke scale, and side of stroke as input, and the union of Tmax and ischemic core segmentation as the ground truth. 5-fold cross-validation was performed for model development cohort. Model performance was evaluated by Dice score coefficient (DSC) and volume difference. Sensitivity and specificity of model target mismatch and clinical/diffusion mismatch criteria from the DAWN were compared, using the DEFUSE 3 target mismatch as reference.

Results: 413 patients were included for model development and 46 for external validation. In model development and external validation cohort, the model achieved median DSC of 0.61 (IQR 0.45, 0.71) and 0.62 (IQR 0.53, 0.72); and volume difference of 3 ml (IQR -37, 41) and 7 ml (IQR -24, 32), respectively. Compared to the clinical/diffusion mismatch approach, the model identified target mismatch with a sensitivity of 89.5% vs 49.3%, a specificity of 77.5% vs 89.2% in the model development cohort, and a sensitivity of 95.6% vs 41.3% in external validation cohort.

Conclusion: A 3D U-Net can predict hypoperfusion lesions from baseline DWI and clinical information, with more sensitive classification of target mismatch than clinical/diffusion mismatch.

Deep learning evaluation of biomarkers from echocardiogram videos

BACKGROUND

Laboratory testing is routinely used to assay blood biomarkers to provide information on physiologic state beyond what clinicians can evaluate from interpreting medical imaging. We hypothesized that deep learning interpretation of echocardiogram videos can provide additional value in understanding disease states and can evaluate common biomarkers results.

METHODS

We developed EchoNet-Labs, a video-based deep learning algorithm to detect evidence of anemia, elevated B-type natriuretic peptide (BNP), troponin I, and blood urea nitrogen (BUN), as well as values of ten additional lab tests directly from echocardiograms. We included patients (n = 39,460) aged 18 years or older with one or more apical-4-chamber echocardiogram videos (n = 70,066) from Stanford Healthcare for training and internal testing of EchoNet-Lab's performance in estimating the most proximal biomarker result. Without fine-tuning, the performance of EchoNet-Labs was further evaluated on an additional external test dataset (n = 1,301) from Cedars-Sinai Medical Center. We calculated the area under the curve (AUC) of the receiver operating characteristic curve for the internal and external test datasets.

FINDINGS

On the held-out test set of Stanford patients not previously seen during model training, EchoNet-Labs achieved an AUC of 0.80 (0.79-0.81) in detecting anemia (low hemoglobin), 0.86 (0.85-0.88) in detecting elevated BNP, 0.75 (0.73-0.78) in detecting elevated troponin I, and 0.74 (0.72-0.76) in detecting elevated BUN. On the external test dataset from Cedars-Sinai, EchoNet-Labs achieved an AUC of 0.80 (0.77-0.82) in detecting anemia, of 0.82 (0.79-0.84) in detecting elevated BNP, of 0.75 (0.72-0.78) in detecting elevated troponin I, and of 0.69 (0.66-0.71) in detecting elevated BUN. We further demonstrate the utility of the model in detecting abnormalities in 10 additional lab tests. We investigate the features necessary for EchoNet-Labs to make successful detection and identify potential mechanisms for each biomarker using well-known and novel explainability techniques.

INTERPRETATION

These results show that deep learning applied to diagnostic imaging can provide additional clinical value and identify phenotypic information beyond current imaging interpretation methods.

FUNDING

J.W.H. and B.H. are supported by the NSF Graduate Research Fellowship. D.O. is supported by NIH K99 HL157421-01. J.Y.Z. is supported by NSF CAREER 1942926, NIH R21 MD012867-01, NIH P30AG059307 and by a Chan-Zuckerberg Biohub Fellowship.

Self-Supervised Longitudinal Neighbourhood Embedding

Longitudinal MRIs are often used to capture the gradual deterioration of brain structure and function caused by aging or neurological diseases. Analyzing this data via machine learning generally requires a large number of ground-truth labels, which are often missing or expensive to obtain. Reducing the need for labels, we propose a self-supervised strategy for representation learning named Longitudinal Neighborhood Embedding (LNE). Motivated by concepts in contrastive learning, LNE explicitly models the similarity between trajectory vectors across different subjects. We do so by building a graph in each training iteration defining neighborhoods in the latent space so that the progression direction of a subject follows the direction of its neighbors. This results in a smooth trajectory field that captures the global morphological change of the brain while maintaining the local continuity. We apply LNE to longitudinal T1w MRIs of two neuroimaging studies: a dataset composed of 274 healthy subjects, and Alzheimer's Disease Neuroimaging Initiative (ADNI, N = 632). The visualization of the smooth trajectory vector field and superior performance on downstream tasks demonstrate the strength of the proposed method over existing self-supervised methods in extracting information associated with normal aging and in revealing the impact of neurodegenerative disorders. The code is available at https://github.com/ouyangjiahong/longitudinal-neighbourhood-embedding.

Tissue at Risk and Ischemic Core Estimation Using Deep Learning in Acute Stroke

BACKGROUND AND PURPOSE

In acute stroke patients with large vessel occlusions, it would be helpful to be able to predict the difference in the size and location of the final infarct based on the outcome of reperfusion therapy. Our aim was to demonstrate the value of deep learning-based tissue at risk and ischemic core estimation. We trained deep learning models using a baseline MR image in 3 multicenter trials.

MATERIALS AND METHODS

Patients with acute ischemic stroke from 3 multicenter trials were identified and grouped into minimal (≤20%), partial (20%-80%), and major (≥80%) reperfusion status based on 4- to 24-hour follow-up MR imaging if available or into unknown status if not. Attention-gated convolutional neural networks were trained with admission imaging as input and the final infarct as ground truth. We explored 3 approaches: 1) separate: train 2 independent models with patients with minimal and major reperfusion; 2) pretraining: develop a single model using patients with partial and unknown reperfusion, then fine-tune it to create 2 separate models for minimal and major reperfusion; and 3) thresholding: use the current clinical method relying on apparent diffusion coefficient and time-to-maximum of the residue function maps. Models were evaluated using area under the curve, the Dice score coefficient, and lesion volume difference.

RESULTS

Two hundred thirty-seven patients were included (minimal, major, partial, and unknown reperfusion: n = 52, 80, 57, and 48, respectively). The pretraining approach achieved the highest median Dice score coefficient (tissue at risk = 0.60, interquartile range, 0.43-0.70; core = 0.57, interquartile range, 0.30-0.69). This was higher than the separate approach (tissue at risk = 0.55; interquartile range, 0.41-0.69; P = .01; core = 0.49; interquartile range, 0.35-0.66; P = .04) or thresholding (tissue at risk = 0.56; interquartile range, 0.42-0.65; P = .008; core = 0.46; interquartile range, 0.16-0.54; P < .001).

CONCLUSIONS

Deep learning models with fine-tuning lead to better performance for predicting tissue at risk and ischemic core, outperforming conventional thresholding methods.

Longitudinal Pooling & Consistency Regularization to Model Disease Progression From MRIs

Many neurological diseases are characterized by gradual deterioration of brain structure andfunction. Large longitudinal MRI datasets have revealed such deterioration, in part, by applying machine and deep learning to predict diagnosis. A popular approach is to apply Convolutional Neural Networks (CNN) to extract informative features from each visit of the longitudinal MRI and then use those features to classify each visit via Recurrent Neural Networks (RNNs). Such modeling neglects the progressive nature of the disease, which may result in clinically implausible classifications across visits. To avoid this issue, we propose to combine features across visits by coupling feature extraction with a novel longitudinal pooling layer and enforce consistency of the classification across visits in line with disease progression. We evaluate the proposed method on the longitudinal structural MRIs from three neuroimaging datasets: Alzheimer's Disease Neuroimaging Initiative (ADNI, N=404), a dataset composed of 274 normal controls and 329 patients with Alcohol Use Disorder (AUD), and 255 youths from the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA). In allthree experiments our method is superior to other widely used approaches for longitudinal classification thus making a unique contribution towards more accurate tracking of the impact of conditions on the brain. The code is available at https://github.com/ouyangjiahong/longitudinal-pooling.

[Association of hyperuricemia-induced renal damage with sirtuin 1 and endothelial nitric oxide synthase in rats]

Objective: To investigate the association of hyperuricemia-induced renal damage with sirtuin 1 (SIRT1) and endothelial nitric oxide synthase (eNOS) in rats. Methods: Using the random number table method, 32 Sprague-Dawley rats were randomly divided into 4 groups: control group, model A group (the model was generated using oxonic acid potassium salt alone), model B group (hyperuricemia model was generated using oxonic acid potassium salt combined with uric acid) and resveratrol group, with 8 rats in each group. The experiment lasted 12 weeks. Serum uric acid and cystatin C levels were monitored regularly. In week 12, serum creatinine and urea nitrogen levels were measured, and the kidneys were extracted. The expression of SIRT1 and eNOS in renal tissues was measured and determined by immunohistochemistry, quantitative reverse-transcription polymerase chain reaction (RT-qPCR) and western blotting. Immunohistochemistry of alpha-smooth muscle actin combined with Masson staining was employed to evaluate the degree of renal fibrosis, and pathological changes were observed based on hematoxylin and eosin staining. Results: In week 12, the uric acid levels in both the model A and model B groups were higher than those in the control group [(316±43) μmol/L, (297±40) μmol/L vs (118±44) μmol/L, both P<0.05]. The levels of cystatin C in the model A, model B, and resveratrol groups were all higher than those in the control group [(156±20) ng/ml, (143±29) ng/ml, (128±26) ng/ml vs (62±18) ng/ml, all P<0.05]. Creatinine levels were higher in the model A and model B groups than those in the control group [(68.5±10.3) μmol/L, (64.5±13.9) μmol/L vs (43.2±10.6) μmol/L, both P<0.05]. The levels of uric acid, cystatin C and creatinine in the resveratrol group were lower than those in the model A group (all P<0.05). Immunohistochemistry, RT-qPCR, and Western blotting for renal SIRT1 and eNOS showed that the expression in the model A and model B groups was inhibited, while the expression in the resveratrol group was not significantly inhibited, compared with that in the control group. Microscopically, obvious abnormalities were not found in the renal tissue of the control group. Renal inflammatory cell aggregation and edema occurred, and interstitial fibrosis was obvious in both the model A and model B groups, while these lesions in the resveratrol group were significantly improved. Conclusions: Hyperuricemia may cause renal injury by inhibiting the expression of SIRT1 and eNOS.

Generalization of deep learning models for ultra-low-count amyloid PET/MRI using transfer learning

PURPOSE

We aimed to evaluate the performance of deep learning-based generalization of ultra-low-count amyloid PET/MRI enhancement when applied to studies acquired with different scanning hardware and protocols.

METHODS

Eighty simultaneous [18F]florbetaben PET/MRI studies were acquired, split equally between two sites (site 1: Signa PET/MRI, GE Healthcare, 39 participants, 67 ± 8 years, 23 females; site 2: mMR, Siemens Healthineers, 64 ± 11 years, 23 females) with different MRI protocols. Twenty minutes of list-mode PET data (90-110 min post-injection) were reconstructed as ground-truth. Ultra-low-count data obtained from undersampling by a factor of 100 (site 1) or the first minute of PET acquisition (site 2) were reconstructed for ultra-low-dose/ultra-short-time (1% dose and 5% time, respectively) PET images. A deep convolution neural network was pre-trained with site 1 data and either (A) directly applied or (B) trained further on site 2 data using transfer learning. Networks were also trained from scratch based on (C) site 2 data or (D) all data. Certified physicians determined amyloid uptake (+/-) status for accuracy and scored the image quality. The peak signal-to-noise ratio, structural similarity, and root-mean-squared error were calculated between images and their ground-truth counterparts. Mean regional standardized uptake value ratios (SUVR, reference region: cerebellar cortex) from 37 successful site 2 FreeSurfer segmentations were analyzed.

RESULTS

All network-synthesized images had reduced noise than their ultra-low-count reconstructions. Quantitatively, image metrics improved the most using method B, where SUVRs had the least variability from the ground-truth and the highest effect size to differentiate between positive and negative images. Method A images had lower accuracy and image quality than other methods; images synthesized from methods B-D scored similarly or better than the ground-truth images.

CONCLUSIONS

Deep learning can successfully produce diagnostic amyloid PET images from short frame reconstructions. Data bias should be considered when applying pre-trained deep ultra-low-count amyloid PET/MRI networks for generalization.