Retrospective Dataset Research Articles

A comparative analysis of image harmonization techniques in mitigating differences in CT acquisition and reconstruction.

The study aims to systematically characterize the effect of CT parameter variations on images and lung radiomic and deep features, and to evaluate the ability of different image harmonization methods to mitigate the observed variations. A retrospective in-house sinogram dataset of 100 low-dose chest CT scans was reconstructed by varying radiation dose (100%, 25%, 10%) and reconstruction kernels (smooth, medium, sharp). A set of image processing, convolutional neural network (CNNs), and generative adversarial network-based (GANs) methods were trained to harmonize all image conditions to a reference condition (100% dose, medium kernel). Harmonized scans were evaluated for image similarity using peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and learned perceptual image patch similarity (LPIPS), and for the reproducibility of radiomic and deep features using concordance correlation coefficient (CCC). CNNs consistently yielded higher image similarity metrics amongst others; for Sharp/10%, which exhibited the poorest visual similarity, PSNR increased from a mean ± CI of 17.763 ± 0.492 to 31.925 ± 0.571, SSIM from 0.219 ± 0.009 to 0.754 ± 0.017, and LPIPS decreased from 0.490 ± 0.005 to 0.275 ± 0.016. Texture-based radiomic features exhibited a greater degree of variability across conditions, i.e. a CCC of 0.500 ± 0.332, compared to intensity-based features (0.972 ± 0.045). GANs achieved the highest CCC (0.969 ± 0.009 for radiomic and 0.841 ± 0.070 for deep features) amongst others. Convolutional neural networks are suitable if downstream applications necessitate visual interpretation of images, whereas generative adversarial networks are better alternatives for generating reproducible quantitative image features needed for machine learning applications. Understanding the efficacy of harmonization in addressing multi-parameter variability is crucial for optimizing diagnostic accuracy and a critical step toward building generalizable models suitable for clinical use.

Complexity of intrinsic brain activity in relapsing-remitting multiple sclerosis patients: patterns, association with structural damage, and clinical disability.

Functional plasticity has been demonstrated in multiple sclerosis (MS) studies. However, the intrinsic brain activity complexity alterations remain unclear. Here, using a coarse-graining time-series procedure algorithm, we obtained multiscale entropy (MSE) from a retrospective multi-centre dataset (208 relapsing-remitting MS patients and 228 healthy controls). By linear mixed model analysis, we demonstrated (1) increased entropy at scale 1 and decreased entropy at scale 6, indicating that regional brain activity shifted towards randomness in the stable MS subgroups (n = 159), and (2) decreased entropy across scales 1-6, trending towards regularity in the acute MS subgroups (n = 49). The main results of the correlation analysis included the following: (1) Decreased entropy was associated with lesion volume and brain volume specifically on longer time scales (scale 3-5), and (2) increased entropy of scale 3 was associated with clinical disability scores. These findings reflect the critical role of structural disruption in the brain activity complexity of BOLD signals in MS patients.

Comparable performance between automatic and manual laryngeal and hypopharyngeal GTV delineations validated with pathology.

Deep learning is a promising approach to increase reproducibility and time-efficiency of GTV delineation in head and neck cancer, but model evaluation primarily relies on manual GTV delineations as reference annotation, which are subjective and tend to overestimate tumor volume. This study aimed to validate a deep learning model for laryngeal and hypopharyngeal GTV segmentation with pathology and to compare its performance with clinicians' manual delineations. A retrospective dataset of 193 laryngeal and hypopharyngeal cancer patients was used to train a deep learning model with clinical GTV delineations as reference. For validation, a dataset comprising 18 patients who underwent imaging before total laryngectomy was used, with histopathology-based (n=16) tumor delineations as ground truth. The performance of the automatic segmentations was compared to that of clinicians' manual delineations, both quantitatively and qualitatively. Median sensitivity (0.90 and 0.91) and largest required CTV margin (6.4 and 6.6 mm) were comparable between automatic and manual GTV delineations. The positive predictive value yielded the only significant difference between automatic and manual GTV delineations, with medians of 0.52 and 0.61, respectively (P=.03). Clinical target volumes derived from automatic and manual GTVs exhibited similar sizes (median of 44.5 and 40.1 mL) and achieved a sensitivity of 1.00 in 13/16 and 14/16 tumors, respectively. Automatic segmentations were considered clinically acceptable in 67% of cases, compared to 63% of manual delineations. The proposed deep learning model for laryngeal and hypopharyngeal GTV segmentation achieved comparable results to clinicians' manual delineations, showing the potential for more consistency and efficiency in the radiotherapy workflow.

Competing risks multi-state model for time-to-event data analysis of HIV/AIDS: a retrospective cohort national datasets, Ethiopia

IntroductionTuberculosis (TB) remains the most common opportunistic infection and leading cause of death among individuals living with HIV/AIDS in Ethiopia. Its significant impact on morbidity and mortality underscores the crucial link between these two diseases. While the advent of antiretroviral therapy (ART) has led to a dramatic decline in mortality rates among HIV/AIDS patients, TB continues to pose a substantial threat. This study aims to estimate the probability of death due to TB among HIV/AIDS patients on ART, considering the presence of various competing risks, including diarrhea, other infections, and unknown/unspecified causes. Also we have assessed the effects of prognostic factors on HIV/AIDS cause specific deaths, compared with the death from other competing risks, and exploring leading cause of death among HIV/AIDS patients on Antiretroviral Therapy.MethodsData from a retrospective research examining the effectiveness of antiretroviral therapy (ART) in Ethiopia were used in this investigation. The data came from medical records of patients who were part of the national ART program. A total of 39,590 records were gathered between October 2019 and March 2020 from all regions of Ethiopia as well as the administration cities of Addis Ababa and Dire Dawa. The study facilities were grouped using a multi-stage sample technique and simple random selection was used to select health facility and a person record from medical records. In the presence of the competing causes of death, Cause specific hazard, subdistribution hazard model and flexible parametric proportional hazard model have been used to assess the effect of covariates on the risk of death, with the cmprisk package in R4.3.2 software.ResultsOut of the total 1212 deaths, 542(44.7%) died competing with other opportunistic infection (TE-Esophageal Candidiasis, TO-oral, CT-CNS Toxoplasmosis, CM-Crypotococcal Meningitis…), 421 (34.7%) died due to tuberculosis and the remaining death were unknown/Not specified infection 222(18.3%) and diarrhea 27(2.2%). Rates of mortality caused by tuberculosis, competing with other opportunistic infection, diarrhea and unknown/Not specified were 3.5, 4.5, 0.2 and 1.8 per 1000 person-months, respectively. Having a higher CD4 count at diagnosis, responding to combination antiretroviral treatment (cART) six months after start, and having prophylactic treatment for pneumocystis pneumonia (PCP) decreased the risk of tuberculosis, other opportunistic infections, and unidentified and diarrheal causes of death. However, older age, late HIV.AIDS diagnosis, and the last HIV/AIDS WHO clinical stages increased the hazard of tuberculosis and other opportunistic disease mortality. Additionally, male gender, older age and last HIV clinical stages increased the mortality HIV/AIDS patients.ConclusionThe findings of this study demonstrated that TB, an opportunistic infection, was the primary cause of death in HIV/AIDS patients, despite the presence of several competing risks, such as diarrhea, other infections, and an undetermined or unclear cause. It's important to use effective techniques to quickly detect those who have HIV or AIDS and provide them with care and treatment to increase their chances of surviving.

Breast radiotherapy planning: A decision-making framework using deep learning.

Effective breast cancer treatment planning requires balancing tumor control while minimizing radiation exposure to healthy tissues. Choosing between intensity-modulated radiation therapy (IMRT) and three-dimensional conformal radiation therapy (3D-CRT) remains pivotal, influenced by patient anatomy and dosimetric constraints. This study aims to develop a decision-making framework utilizing deep learning to predict dose distributions, aiding in the selection of optimal treatmenttechniques. A 2D U-Net convolutional neural network (CNN) model was used to predict dose distribution maps and dose-volume histogram (DVH) metrics for breast cancer patients undergoing IMRT and 3D-CRT. The model was trained and fine-tuned using retrospective datasets from two medical centers, accounting for variations in CT systems, dosimetric protocols, and clinical practices, over 346 patients. An additional 30 consecutive patients were selected for external validation, where both 3D-CRT and IMRT plans were manually created. To show the potential of the approach, an independent medical physicist evaluated both dosimetric plans and selected the most appropriate one based on applicable clinical criteria. Confusion matrices were used to compare the decisions of the independent observer with the historical decision and the proposed decision-makingframework. Evaluation metrics, including dice similarity coefficients (DSC) and DVH analyses, demonstrated high concordance between predicted and clinical dose distribution for both IMRT and 3D-CRT techniques, especially for organs at risk (OARs). The decision-making framework demonstrated high accuracy (90 ), recall (95.7 ), and precision (91.7 ) when compared to independent clinical evaluations, while the historical decision-making had lower accuracy (50 ), recall (47.8 ), and precision (78.6 ). The proposed decision-making model accurately predicts dose distributions for both 3D-CRT and IMRT, ensuring reliable OAR dose estimation. This decision-making framework significantly outperforms historical decision-making, demonstrating higher accuracy, recall, and precision.

Open-source machine learning pipeline automatically flags instances of acute respiratory distress syndrome from electronic health records.

Physicians could greatly benefit from automated diagnosis and prognosis tools to help address information overload and decision fatigue. Intensive care physicians stand to benefit greatly from such tools as they are at particularly high risk for those factors. Acute Respiratory Distress Syndrome (ARDS) is a life-threatening condition affecting >10% of critical care patients and has a mortality rate over 40%. However, recognition rates for ARDS have been shown to be low (30-70%) in clinical settings. In this work, we present a reproducible computational pipeline that automatically adjudicates ARDS on retrospective datasets of mechanically ventilated adult patients. This pipeline automates the steps outlined by the Berlin Definition through implementation of natural language processing tools and classification algorithms. First, we used labeled chest imaging reports from two different hospitals over three different time periods to train an XGBoost model to detect bilateral infiltrates, and a subset of attending physician notes from one hospital labeled for the most common ARDS risk factor (pneumonia) to train another XGBoost model to detect a pneumonia diagnosis. Both models achieve high performance when tested on out-of-bag samples-an area under the receiver operating characteristic curve (AUROC) of 0.88 for adjudicating chest imaging reports, and an AUROC of 0.86 for detecting pneumonia on attending physician notes. Next, we integrate the models and validate the entire pipeline on a fourth cohort from a third hospital (MIMIC-III) and find a sensitivity of 93.5% - an extraordinary improvement over the 22.6% ARDS recognition rate reported for these encounters - along with a false positive rate of 18.8%. We conclude that our reproducible, automated diagnostic pipeline exhibits promising ARDS retrospective adjudication performance, thus providing a valuable resource for physicians aiming to enhance ARDS diagnosis and treatment strategies. We surmise that real-time integration of the pipeline with EHR systems has the potential to aid clinical practice by facilitating the recognition of ARDS cases at scale.

Automated Evaluation for Pericardial Effusion and Cardiac Tamponade with Echocardiographic Artificial Intelligence.

Timely and accurate detection of pericardial effusion and assessment cardiac tamponade remain challenging and highly operator dependent. Artificial intelligence has advanced many echocardiographic assessments, and we aimed to develop and validate a deep learning model to automate the assessment of pericardial effusion severity and cardiac tamponade from echocardiogram videos. We developed a deep learning model (EchoNet-Pericardium) using temporal-spatial convolutional neural networks to automate pericardial effusion severity grading and tamponade detection from echocardiography videos. The model was trained using a retrospective dataset of 1,427,660 videos from 85,380 echocardiograms at Cedars-Sinai Medical Center (CSMC) to predict PE severity and cardiac tamponade across individual echocardiographic views and an ensemble approach combining predictions from five standard views. External validation was performed on 33,310 videos from 1,806 echocardiograms from Stanford Healthcare (SHC). In the held out CSMC test set, EchoNet-Pericardium achieved an AUC of 0.900 (95% CI: 0.884- 0.916) for detecting moderate or larger pericardial effusion, 0.942 (95% CI: 0.917-0.964) for large pericardial effusion, and 0.955 (95% CI: 0.939-0.968) for cardiac tamponade. In the SHC external validation cohort, the model achieved AUCs of 0.869 (95% CI: 0.794-0.933) for moderate or larger pericardial effusion, 0.959 (95% CI: 0.945-0.972) for large pericardial effusion, and 0.966 (95% CI: 0.906-0.995) for cardiac tamponade. Subgroup analysis demonstrated consistent performance across ages, sexes, left ventricular ejection fraction, and atrial fibrillation statuses. Our deep learning-based framework accurately grades pericardial effusion severity and detects cardiac tamponade from echocardiograms, demonstrating consistent performance and generalizability across different cohorts. This automated tool has the potential to enhance clinical decision-making by reducing operator dependence and expediting diagnosis.

Proactive cervical cancer risk assessment using data-driven analytics

This study introduces a sophisticated predictive model integrating clinical and lifestyle data addressing the critical public health challenge of cervical cancer, particularly in regions lacking routine screenings. Leveraging data driven analytics, the proposed model undergoes comprehensive preprocessing, including exploratory data analysis, missing value imputation, and feature extraction. Feature selection is carried out using the XGBoost classifier to ensure model efficacy. Data normalization and class balance via oversampling techniques are applied, with model validation conducted through stratified cross-validation. The optimized feature vector is then employed to train a LightGBM model. Utilizing a retrospective dataset of 858 patients from the Hospital Universitario de Caracas, Venezuela, comprising demographic, lifestyle, and medical history data, the LightGBM model achieves an impressive accuracy of 98%, outperforming similar existing approaches. The study outcome demonstrates the effectiveness of the proposed data modelling framework and feature selection, along with the choice of LightGBM as a suitable classifier. The proposed predictive framework can efficiently aid healthcare professionals in prioritizing high-risk patients for further evaluation and intervention.

SMART-PET: a Self-SiMilARiTy-aware generative adversarial framework for reconstructing low-count [18F]-FDG-PET brain imaging.

In Positron Emission Tomography (PET) imaging, the use of tracers increases radioactive exposure for longitudinal evaluations and in radiosensitive populations such as pediatrics. However, reducing injected PET activity potentially leads to an unfavorable compromise between radiation exposure and image quality, causing lower signal-to-noise ratios and degraded images. Deep learning-based denoising approaches can be employed to recover low count PET image signals: nonetheless, most of these methods rely on structural or anatomic guidance from magnetic resonance imaging (MRI) and fails to effectively preserve global spatial features in denoised PET images, without impacting signal-to-noise ratios. In this study, we developed a novel PET only deep learning framework, the Self-SiMilARiTy-Aware Generative Adversarial Framework (SMART), which leverages Generative Adversarial Networks (GANs) and a self-similarity-aware attention mechanism for denoising [18F]-fluorodeoxyglucose (18F-FDG) PET images. This study employs a combination of prospective and retrospective datasets in its design. In total, 114 subjects were included in the study, comprising 34 patients who underwent 18F-Fluorodeoxyglucose PET (FDG) PET imaging for drug-resistant epilepsy, 10 patients for frontotemporal dementia indications, and 70 healthy volunteers. To effectively denoise PET images without anatomical details from MRI, a self-similarity attention mechanism (SSAB) was devised. which learned the distinctive structural and pathological features. These SSAB-enhanced features were subsequently applied to the SMART GAN algorithm and trained to denoise the low-count PET images using the standard dose PET image acquired from each individual participant as reference. The trained GAN algorithm was evaluated using image quality measures including structural similarity index measure (SSIM), peak signal-to-noise ratio (PSNR), normalized root mean square (NRMSE), Fréchet inception distance (FID), signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR). In comparison to the standard-dose, SMART-PET had on average a SSIM of 0.984 ± 0.007, PSNR of 38.126 ± 2.631 dB, NRMSE of 0.091 ± 0.028, FID of 0.455 ± 0.065, SNR of 0.002 ± 0.001, and CNR of 0.011 ± 0.011. Regions of interest measurements obtained with datasets decimated down to 10% of the original counts, showed a deviation of less than 1.4% when compared to the ground-truth values. In general, SMART-PET shows promise in reducing noise in PET images and can synthesize diagnostic quality images with a 90% reduction in standard of care injected activity. These results make it a potential candidate for clinical applications in radiosensitive populations and for longitudinal neurological studies.

Technical feasibility of automated blur detection in digital mammography using convolutional neural network

BackgroundThe presence of a blurred area, depending on its localization, in a mammogram can limit diagnostic accuracy. The goal of this study was to develop a model for automatic detection of blur in diagnostically relevant locations in digital mammography.MethodsA retrospective dataset consisting of 152 examinations acquired with mammography machines from three different vendors was utilized. The blurred areas were contoured by expert breast radiologists. Normalized Wiener spectra (nWS) were extracted in a sliding window manner from each mammogram. These spectra served as input for a convolutional neural network (CNN) generating the probability of the spectra originating from a blurred region. The resulting blur probability mask, upon thresholding, facilitated the classification of a mammogram as either blurred or sharp. Ground truth for the test set was defined by the consensus of two radiologists.ResultsA significant correlation between the view (p < 0.001), as well as between the laterality and the presence of blur (p = 0.004) was identified. The developed model AUROC of 0.808 (95% confidence interval 0.794–0.821) aligned with the consensus in 78% (67–83%) of mammograms classified as blurred. For mammograms classified by consensus as sharp, the model achieved agreement in 75% (67–83%) of them.ConclusionA model for blur detection was developed and assessed. The results indicate that a robust approach to blur detection, based on feature extraction in frequency space, tailored to radiologist expertise regarding clinical relevance, could eliminate the subjectivity associated with the visual assessment.Relevance statementThis blur detection model, if implemented in clinical practice, could provide instantaneous feedback to technicians, allowing for prompt mammogram retakes and ensuring that only high-quality mammograms are sent for screening and diagnostic tasks.Key PointsBlurring in mammography limits radiologist interpretation and diagnostic accuracy.This objective blur detection tool ensures image quality, and reduces retakes and unnecessary exposures.Wiener spectrum analysis and CNN enabled automated blur detection in mammography.Graphical

Analysis of Prognostic Factors for Mortality in Patients With Gastrointestinal Bleeding: Application of Machine Learning Tools

Introduction: Treatment of upper gastrointestinal bleeding (UGIB) is a complex challenge due to the wide range of causes and factors affecting hospitalization outcomes.Objective: To study the impact of various factors on 30-day hospital outcomes using machine learning (ML) tools.Materials and methods: We compiled a retrospective data set that includes clinical, laboratory, and imaging data of 101 patients. The database was divided into 2 groups by UGIB etiology: ulcer and variceal bleedings. Both etiological groups were processed using ML tools in 2 steps: imputation by the MICE (multiple imputation by chained equations) model and factor importance analysis using the Random Forest model.Results: Analysis revealed that the most prognostically valuable parameters in both groups were well-known mortality predictors and emerging predictive factors, such as creatinine, blood pressure, activated partial thromboplastin time, level of consciousness, urea, lactate, comorbidity status, procalcitonin, ferritin, and total protein.Conclusions: The application of advanced tools confirmed the significance of popular and validated mortality predictors and contributed to the development of predictors, both explored and unexplored ones.

Reliability optimization through soft digital twins

This paper presents a framework for developing a “soft digital twin” (SDT) for reliability optimization through modeling and simulation. The focus is on demonstrating the effectiveness of the SDT in modeling maintenance task processes within mission-critical facilities, which is particularly essential for reliable operations in critical infrastructure such as nuclear facilities. The need for efficient completion of maintenance tasks (MTs) to ensure high reliability and minimal downtime underscores the challenge faced by managers, who desire a predictive tool akin to a “crystal ball” for optimal resource configuration in the face of uncertainties from equipment failures, staffing issues, and supply chain disruptions. The proposed SDT functions as this predictive tool, leveraging the simulation’s versatility to provide insights into resource configurations and staff planning. This work extends the prior research by incorporating additional model validation, sensitivity testing, and analysis of the impact of various resource changes on the system. It employs a combination of data-driven frameworks and stochastic modeling methods to construct an adaptive SDT capable of accommodating changes in system behavior. The work provides a framework for the construction of SDT, testing and validation of its results, and its application to a case study of a mission-critical facility focused on replication and then optimizing of two key performance indicators based on human resource mixtures. Initial findings suggest that the SDT yields reliable results comparable to retrospective test datasets, offering the potential to minimize MTs’ time in the system while maximizing throughput within specified time frames through scenario experimentation and optimization.

The Development and Validation of a Glaucoma Health Score for Glaucoma Screening Based on Clinical Parameters and Optical Coherence Tomography Metrics.

Background/Objectives: This study aims to develop and validate a Glaucoma Health Score (GHS) that incorporates multiple individual glaucoma risk factors to enhance glaucoma detection in screening environments. Methods: The GHS was developed using a retrospective dataset from two clinical sites, including both eyes of glaucoma patients and controls. The model incorporated age, central corneal thickness, intraocular pressure, pattern standard deviation from a visual field threshold 24-2 test, and two parameters from an optical coherence tomography (OCT) test: the average circumpapillary retinal nerve fiber layer thickness and the minimum thickness of the six sectors of the macular ganglion cell plus the inner plexiform layer. The GHS was then validated in two independent datasets: one from primary care sites using Maestro OCT data (test dataset 1) and another from an academic center using DRI OCT Triton (test dataset 2). Results: Both eyes of 51 glaucoma patients and 67 controls were included in the development dataset. Setting the GHS cutoff at 75 points out of 100, test dataset 1, which comprised 41 subjects with glaucoma and 41 healthy controls, achieved an area under the receiver operating characteristic curve (AUROC) of 0.98, with a sensitivity of 71% and specificity of 98%; test dataset 2, which included 53 patients with glaucoma and 53 healthy controls, resulted in an AUROC of 0.95, with a sensitivity of 75% and specificity of 96%. A decision curve analysis across all datasets demonstrated a higher net benefit for the GHS model compared to individual OCT parameters. Conclusions: The GHS offers a feasible, standardized approach for early detection of glaucoma, providing strong specificity and acceptable sensitivity, with clear decision-making benefits in screening settings.

Multi-Institutional Evaluation and Training of Breast Density Classification AI Algorithm Using ACR Connect and AI-LAB

ObjectiveTo demonstrate and test the capabilities of the ACR Connect and AI-LAB software platform by implementing multi-institutional artificial intelligence (AI) training and validation for breast density classification. MethodsIn this proof-of-concept study, six US-based hospitals installed Connect and AI-LAB. A breast density algorithm was trained and tested on retrospective mammograms. We recorded time to receive institutional review board approval, to install software locally, and to complete the testing and training. We calculated the performance of the breast density algorithm at each participating hospital and compared it to the performance of a holdout multi-institutional clinical trial testing dataset and a retrospective multi-institutional dataset. We calculated the performance of the locally fine-tuned models on the holdout test datasets. ResultsThe median time to receive institutional review board approval was 66 days, and the median time to successfully install Connect and AI-LAB locally was 157 days. The median time to complete breast density algorithm testing and training was 216 days. The breast density algorithm performed worse at each hospital than on the holdout test dataset, suggesting poor generalizability of the base model. The fine-tuned models had mixed performance locally and performed poorly on the test dataset. DiscussionIn this study, we demonstrate the successful installation and implementation of Connect and AI-LAB software platforms at six facilities using a breast density algorithm. Our results suggest poor generalizability of an algorithm trained on a single dataset and algorithms fine-tuned at individual institutions, emphasizing the hypothetical importance of multi-institutional testing and training.

Development and validation of an early acute kidney injury risk prediction model for patients with sepsis in emergency departments

In this study, we aimed to develop and validate a nomogram to predicting the risk of sepsis-associated acute kidney injury (SA-AKI) in patients admitted to emergency departments (EDs). We randomly divided a retrospective dataset of 391 patients with sepsis into a 294-person training cohort and a 97-person validation cohort, and developed three predictive models using multivariate logistic regression analysis and clinical insight. No difference was observed between the three models using the DeLong test and Model 3 was selected as the risk prediction model based on the principle of least inclusion indicators. The use of vasopressor drugs, patient age, platelet count, procalcitonin, and D-dimer levels were included. The training and validation cohorts had a consistency index of 0.832 and 0.866, respectively, indicating high accuracy and stability in predicting SA-AKI risk. The area under the receiver operating characteristic curve was 0.832, showing excellent discrimination. The calibration curves for the training and validation cohorts showed excellent calibration. The decision curve and clinical impact curve analyses showed that the net clinical benefit of using the nomogram was greatest over a probability threshold of 0.05–0.90. In addition, the model showed moderate validity in predicting the 30-day survival and the incidence of major adverse renal events within 30 days. The nomogram developed for SA-AKI risk assessment in patients in EDs showed good discriminability and clinical utility. It can provide a theoretical basis for emergency physicians to prevent SA-AKI.

Prediction of atrial fibrillation recurrence after ablation: the impact of machine learning and recurrence during the blanking period

Abstract Introduction Pulmonary vein isolation is superior to antiarrhythmic drugs for rhythm control in patients with atrial fibrillation (AF). Nevertheless, recurrences within the first year are still significant. Several clinical risk scores were developed to predict AF recurrence; however, despite good results in development cohorts, the results in external validation are usually poor, limiting the general applicability of these scores. Purpose We aimed to develop machine learning (ML) models to predict AF recurrences within the first year after catheter ablation and compare their performance with conventional risk scores. Methods We used a retrospective dataset of a tertiary hospital center, including all consecutive patients submitted to the first AF ablation between 2017 and 2021. The dataset included 76 features: patient characteristics, medical history, baseline echocardiography, calcium score, procedure variables, and early recurrence (ER) during the 90-day blanking period. The outcome was defined as the occurrence of an electrocardiographically documented late recurrence (LR) of atrial tachycardia/AF &amp;gt; 30s between 3-12 months. Three different supervised ML models (penalized logistic regression, random forests, and XGBoost) were developed on the training cohort, and hyperparameters were tuned using 10-fold cross-validation. A testing dataset was held out to estimate the final performance (25%). The following risk scores were calculated: APPLE, CAAP-AF, and BASE-AF2 (the last includes ER). Areas under the curve (AUC) of receiver operating characteristic curves were compared using DeLong’s test. Results 679 patients were included: 62.7% males, median age of 59±16 years, and 78.2% with paroxysmal AF. The median time from diagnosis to ablation was 2±4 years. 74.2% underwent radiofrequency ablation and the remaining cryoablation. ER occurred in 12.5%, and most of these patients also experienced LR (68.7 vs 19.6% in those without ER, p&amp;lt;0.001). LR was observed in 25.6%. The XGBoost model showed the best performance with an AUC 0.774, 95% confidence interval (CI) 0.688–0.844) outperforming existing scores (picture 1): APPLE score AUC 0.607, 95% CI 0.562–0.653, p-value&amp;lt;0.001; CAAP-AF score AUC 0.622, 95% CI 0.576–0.668, p-value=0.002; BASE-AF2 AUC 0.628, 95% CI 0.581–0.675, p-value= 0.003. Variable importance analysis showed a significant drop in the performance of the models when ER was not considered, indicating its high importance in predicting LR (picture 2). Conclusions Recurrence during the blanking period was the most important predictor of LR in our population. The ML model was superior to conventional risk scores. ML models might be an essential tool to improve the prediction of outcomes and clinical decision-making for optimal follow-up.ROC curvesFeature Importance

Clinical trials for Lennox-Gastaut syndrome: Challenges and priorities.

Lennox-Gastaut syndrome (LGS) is a severe, childhood-onset epilepsy that is typically refractory to treatment. We surveyed the current landscape of LGS treatment, aiming to identify challenges to the development of efficacious therapies, and to articulate corresponding priorities toward clinical trials that improve outcomes. The LGS Special Interest Group of the Pediatric Epilepsy Research Consortium integrated evidence from the literature and expert opinion, into a narrative review. We provide an overview of approved and emerging medical, dietary, surgical and neuromodulation approaches for LGS. We note that quality of care could be improved by standardizing LGS treatment based on expert consensus and empirical data. Whereas LGS natural history is incompletely understood, prospective studies and use of large retrospective datasets to understand LGS across the lifespan would enable clinical trials that address these dynamics. Recent discoveries related to LGS pathophysiology should enable development of disease-modifying therapies, which are currently lacking. Finally, clinical trials have focused chiefly on seizures involving "drops," but should incorporate additional patient-centered outcomes, using emerging measures adapted to people with LGS. Clinicians and researchers should enact these priorities, with the goal of patient-centered clinical trials that are tailored to LGS pathophysiology and natural history.

Automated detection of motion artifacts in brain MR images using deep learning.

Quality assessment, including inspecting the images for artifacts, is a critical step during magnetic resonance imaging (MRI) data acquisition to ensure data quality and downstream analysis or interpretation success. This study demonstrates a deep learning (DL) model to detect rigid motion in T1-weighted brain images. We leveraged a 2D convolutional neural network (CNN) trained on motion-synthesized data for three-class classification and tested it on publicly available retrospective and prospective datasets. Grad-CAM heatmaps enabled the identification of failure modes and provided an interpretation of the model's results. The model achieved average precision and recall metrics of 85% and 80% on six motion-simulated retrospective datasets. Additionally, the model's classifications on the prospective dataset showed 93% agreement with the labeling of a radiologist a strong inverse correlation (-0.84) compared to average edge strength, an image quality metric indicative of motion. This model is aimed at inline automatic detection of motion artifacts, accelerating part of the time-consuming quality assessment (QA) process and augmenting expertise on-site, particularly relevant in low-resource settings where local MR knowledge is scarce.

Ecuador Towards Zero Leprosy: A Twenty-Three-Year Retrospective Epidemiologic and Spatiotemporal Analysis of Leprosy in Ecuador.

Ecuador has gone through a significant reduction in new cases from 2000 (106) to 2023 (12), suggesting a trend towards zero leprosy. An ecological spatiotemporal study design was used to describe the epidemiological distribution of the disease in the country during 2000-2023. Leprosy cases registered by the surveillance system of the Ecuadorian Ministry of Public Health were the data utilized for the study. From January 2000 to December 2023, 1539, incidence cases were diagnosed with leprosy in Ecuador. At the time of diagnosis, the median age was 54 years. Most of the cases were males (71.5%). The proportion of incidence cases in subjects over 50 years was 63% and 1.5% in children ≤ 15 years old. The yearly incidence rate ranged from 8.5/1,000,000 population in 2000 to 0.68/1,000,000 population in 2023, remaining within the low-endemic parameter. In total, 35 cantons reported newly detected leprosy cases in the year 2000. By the end of 2023, only eight cantons actively reported cases of leprosy. High-risk clusters for leprosy were detected in the tropical coastal region of Ecuador. The provinces with the highest number of cases during the study period were Guayas (44.8%) and Los Rios (15.7%), with zero cases being found in the Galapagos Islands. Our study is unique in that it documents a retrospective dataset over a two-decade timespan from a South American country that has effectively applied global guidelines for the control and elimination of leprosy.

Synthetic data in generalizable, learning-based neuroimaging

Abstract Synthetic data has emerged as an attractive option for developing machine learning methods in human neuroimaging, particularly in magnetic resonance imaging (MRI)— a modality where image contrast depends enormously on acquisition hardware and parameters. This retrospective paper reviews a family of recently proposed methods, based on synthetic data, for generalizable machine learning in brain MRI analysis. Central to this framework is the concept of domain randomization, which involves training neural networks on a vastly diverse array of synthetically generated images with random contrast properties. This technique has enabled robust, adaptable models that are capable of handling diverse MRI contrasts, resolutions, and pathologies, while working out-of-the-box, without retraining. We have successfully applied this method to tasks such as whole brain segmentation (SynthSeg), skull-stripping (SynthStrip), registration (SynthMorph, EasyReg), super-resolution and MR contrast transfer (SynthSR). Beyond these applications, the paper discusses other possible use cases and future work in our methodology. Neural networks trained with synthetic data enable the analysis of clinical MRI, including large retrospective datasets, while greatly alleviating (and sometimes eliminating) the need for substantial labeled datasets, and offer enormous potential as robust tools to address various research goals.