Assessment of Background Parenchymal Enhancement at Dynamic Contrast-enhanced MRI in Predicting Breast Cancer Recurrence Risk.

Background Background parenchymal enhancement (BPE) at dynamic contrast-enhanced (DCE) MRI of cancer-free breasts increases the risk of developing breast cancer; implications of quantitative BPE in ipsilateral breasts with breast cancer are largely unexplored. Purpose To determine whether quantitative BPE measurements in one or both breasts could be used to predict recurrence risk in women with breast cancer, using the Oncotype DX recurrence score as the reference standard. Materials and Methods This HIPAA-compliant retrospective single-institution study included women diagnosed with breast cancer between January 2007 and January 2012 (development set) and between January 2012 and January 2017 (internal test set). Quantitative BPE was automatically computed using an in-house-developed computer algorithm in both breasts. Univariable logistic regression was used to examine the association of BPE with Oncotype DX recurrence score binarized into high-risk (recurrence score >25) and low- or intermediate-risk (recurrence score ≤25) categories. Models including BPE measures were assessed for their ability to distinguish patients with high risk versus those with low or intermediate risk and the actual recurrence outcome. Results The development set included 127 women (mean age, 58 years ± 10.2 [SD]; 33 with high risk and 94 with low or intermediate risk) with an actual local or distant recurrence rate of 15.7% (20 of 127) at a minimum 10 years of follow-up. The test set included 60 women (mean age, 57.8 years ± 11.6; 16 with high risk and 44 with low or intermediate risk). BPE measurements quantified in both breasts were associated with increased odds of a high-risk Oncotype DX recurrence score (odds ratio range, 1.27-1.66 [95% CI: 1.02, 2.56]; P < .001 to P = .04). Measures of BPE combined with tumor radiomics helped distinguish patients with a high-risk Oncotype DX recurrence score from those with a low- or intermediate-risk score, with an area under the receiver operating characteristic curve of 0.94 in the development set and 0.79 in the test set. For the combined models, the negative predictive values were 0.97 and 0.93 in predicting actual distant recurrence and local recurrence, respectively. Conclusion Ipsilateral and contralateral DCE MRI measures of BPE quantified in patients with breast cancer can help distinguish patients with high recurrence risk from those with low or intermediate recurrence risk, similar to Oncotype DX recurrence score. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Zhou and Rahbar in this issue.

Automated Preoperative and Postoperative Volume Estimates Risk of Retreatment in Chronic Subdural Hematoma: A Retrospective, Multicenter Study.

BACKGROUND AND OBJECTIVES: Several neurosurgical pathologies, ranging from glioblastoma to hemorrhagic stroke, use volume thresholds to guide treatment decisions. For chronic subdural hematoma (cSDH), with a risk of retreatment of 10%-30%, the relationship between preoperative and postoperative cSDH volume and retreatment is not well understood. We investigated the potential link between preoperative and postoperative cSDH volumes and retreatment. METHODS: We performed a retrospective chart review of patients operated for unilateral cSDH from 4 level 1 trauma centers, February 2009-August 2021. We used a 3-dimensional deep learning, automated segmentation pipeline to calculate preoperative and postoperative cSDH volumes. To identify volume thresholds, we constructed a receiver operating curve with preoperative and postoperative volumes to predict cSDH retreatment rates and selected the threshold with the highest Youden index. Then, we developed a light gradient boosting machine to predict the risk of cSDH recurrence. RESULTS: We identified 538 patients with unilateral cSDH, of whom 62 (12%) underwent surgical retreatment within 6 months of the index surgery. cSDH retreatment was associated with higher preoperative (122 vs 103 mL; P < .001) and postoperative (62 vs 35 mL; P < .001) volumes. Patients with >140 mL preoperative volume had nearly triple the risk of cSDH recurrence compared with those below 140 mL, while a postoperative volume >46 mL led to an increased risk for retreatment (22% vs 6%; P < .001). On multivariate modeling, our model had an area under the receiver operating curve of 0.76 (95% CI: 0.60-0.93) for predicting retreatment. The most important features were preoperative and postoperative volume, platelet count, and age. CONCLUSION: Larger preoperative and postoperative cSDH volumes increase the risk of retreatment. Volume thresholds may allow identification of patients at high risk of cSDH retreatment who would benefit from adjunct treatments. Machine learning algorithm can quickly provide accurate estimates of preoperative and postoperative volumes.

Human Not in the Loop: Objective Sample Difficulty Measures for Curriculum Learning.

Curriculum learning is a learning method that trains models in a meaningful order from easier to harder samples. A key here is to devise automatic and objective difficulty measures of samples. In the medical domain, previous work applied domain knowledge from human experts to qualitatively assess classification difficulty of medical images to guide curriculum learning, which requires extra annotation efforts, relies on subjective human experience, and may introduce bias. In this work, we propose a new automated curriculum learning technique using the variance of gradients (VoG) to compute an objective difficulty measure of samples and evaluated its effects on elbow fracture classification from X-ray images. Specifically, we used VoG as a metric to rank each sample in terms of the classification difficulty, where high VoG scores indicate more difficult cases for classification, to guide the curriculum training process We compared the proposed technique to a baseline (without curriculum learning), a previous method that used human annotations on classification difficulty, and anti-curriculum learning. Our experiment results showed comparable and higher performance for the binary and multi-class bone fracture classification tasks.

Comparison of machine learning models to predict long-term outcomes after severe traumatic brain injury.

OBJECTIVE: An estimated 1.5 million people die every year worldwide from traumatic brain injury (TBI). Physicians are relatively poor at predicting long-term outcomes early in patients with severe TBI. Machine learning (ML) has shown promise at improving prediction models across a variety of neurological diseases. The authors sought to explore the following: 1) how various ML models performed compared to standard logistic regression techniques, and 2) if properly calibrated ML models could accurately predict outcomes up to 2 years posttrauma. METHODS: A secondary analysis of a prospectively collected database of patients with severe TBI treated at a single level 1 trauma center between November 2002 and December 2018 was performed. Neurological outcomes were assessed at 3, 6, 12, and 24 months postinjury with the Glasgow Outcome Scale. The authors used ML models including support vector machine, neural network, decision tree, and naïve Bayes models to predict outcome across all 4 time points by using clinical information available on admission, and they compared performance to a logistic regression model. The authors attempted to predict unfavorable versus favorable outcomes (Glasgow Outcome Scale scores of 1-3 vs 4-5), as well as mortality. Models' performance was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC) with 95% confidence interval and balanced accuracy. RESULTS: Of the 599 patients in the database, the authors included 501, 537, 469, and 395 at 3, 6, 12, and 24 months posttrauma. Across all time points, the AUCs ranged from 0.71 to 0.85 for mortality and from 0.62 to 0.82 for unfavorable outcomes with various modeling strategies. Decision tree models performed worse than all other modeling approaches for multiple time points regarding both unfavorable outcomes and mortality. There were no statistically significant differences between any other models. After proper calibration, the models had little variation (0.02-0.05) across various time points. CONCLUSIONS: The ML models tested herein performed with equivalent success compared with logistic regression techniques for prognostication in TBI. The TBI prognostication models could predict outcomes beyond 6 months, out to 2 years postinjury.

Interpretable machine learning model for imaging-based outcome prediction after cardiac arrest.

INTRODUCTION: Early identification of brain injury patterns in computerized tomography (CT) imaging is crucial for post-cardiac arrest prognostication. Lack of interpretability of machine learning prediction reduces trustworthiness by clinicians and prevents translation to clinical practice. We aimed to identify CT imaging patterns associated with prognosis with interpretable machine learning. METHODS: In this IRB-approved retrospective study, we included consecutive comatose adult patients hospitalized at a single academic medical center after resuscitation from in- and out-of-hospital cardiac arrest between August 2011 and August 2019 who underwent unenhanced CT imaging of the brain within 24 hours of their arrest. We decomposed the CT images into subspaces to identify interpretable and informative patterns of injury, and developed machine learning models to predict patient outcomes (i.e., survival and awakening status) using the identified imaging patterns. Practicing physicians visually examined the imaging patterns to assess clinical relevance. We evaluated machine learning models using 80%-20% random data split and reported AUC values to measure the model performance. RESULTS: We included 1284 subjects of whom 35% awakened from coma and 34% survived hospital discharge. Our expert physicians were able to visualize decomposed image patterns and identify those believed to be clinically relevant on multiple brain locations. For machine learning models, the AUC was 0.710 ± 0.012 for predicting survival and 0.702 ± 0.053 for predicting awakening, respectively. DISCUSSION: We developed an interpretable method to identify patterns of early post-cardiac arrest brain injury on CT imaging and showed these imaging patterns are predictive of patient outcomes (i.e., survival and awakening status).

Serological analysis of allergic components of house dust mite provides more insight in epidemiological characteristics and clinical symptom development in North China.

Background: House dust mite (HDM) is the most common airborne source causing complex allergy symptoms. There are geographic differences in the allergen molecule sensitization profiles. Serological testing with allergen components may provide more clues for diagnosis and clinical management. Objective: This study aims to investigate the sensitization profile of eight HDM allergen components in a large number of patients enrolled in the clinic and to analyze the relation of gender, age, and clinical symptoms in North China. Methods: The 548 serum samples of HDM-allergic patients (ImmunoCAP® d1 or d2 IgE ≥0.35) were collected in Beijing City and divided in four different age groups and three allergic symptoms. The specific IgE of HDM allergenic components, Der p 1/Der f 1, Der p 2/Der f 2, Der p 7, Der p 10, Der p 21, and Der p 23, was measured using the micro-arrayed allergen test kit developed by Hangzhou Zheda Dixun Biological Gene Engineering Co., Ltd. The new system was validated by comparing to single-component Der p 1, Der p 2, and Der p 23 tests by ImmunoCAP in 39 sera. The epidemiological study of these IgE profiles and the relation to age and clinical phenotypes were analyzed. Results: A greater proportion of male patients was in the younger age groups, while more female patients were in the adult groups. Both the sIgE levels and the positive rates (approximately 60%) against Der p 1/Der f 1 and Der p 2/Der f 2 were higher than for the Der p 7, Der p 10, and Der p 21 components (below 25%). The Der f 1 and Der p 2 positive rates were higher in 2-12-year-old children. The Der p 2 and Der f 2 IgE levels and positive rates were higher in the allergic rhinitis group. The positive rates of Der p 10 increased significantly with age. Der p 21 is relevant in allergic dermatitis symptom, while Der p 23 contributes to asthma development. Conclusion: HDM groups 1 and 2 were the major sensitizing allergens, with group 2 being the most important component relevant to respiratory symptoms in North China. The Der p 10 sensitization tends to increase with age. Der p 21 and Der p 23 might be associated with the development of allergic skin disease and asthma, respectively. Multiple allergen sensitizations increased the risk of allergic asthma.

Structures and Anti-Allergic Activities of Natural Products from Marine Organisms.

In recent years, allergic diseases have occurred frequently, affecting more than 20% of the global population. The current first-line treatment of anti-allergic drugs mainly includes topical corticosteroids, as well as adjuvant treatment of antihistamine drugs, which have adverse side effects and drug resistance after long-term use. Therefore, it is essential to find alternative anti-allergic agents from natural products. High pressure, low temperature, and low/lack of light lead to highly functionalized and diverse functional natural products in the marine environment. This review summarizes the information on anti-allergic secondary metabolites with a variety of chemical structures such as polyphenols, alkaloids, terpenoids, steroids, and peptides, obtained mainly from fungi, bacteria, macroalgae, sponges, mollusks, and fish. Molecular docking simulation is applied by MOE to further reveal the potential mechanism for some representative marine anti-allergic natural products to target the H1 receptor. This review may not only provide insight into information about the structures and anti-allergic activities of natural products from marine organisms but also provides a valuable reference for marine natural products with immunomodulatory activities.

Treatment of allergy: Overview of synthetic anti-allergy small molecules in medicinal chemistry.

The prevalence of allergic diseases has been continuously increasing over the past few decades, affecting approximately 20-30% of the global population. Allergic reactions to infection of respiratory tract, digestive tract, and skin system involve multiple different targets. The main difficulty of anti-allergy research is how to develop drugs with good curative effect and less side effects by adopting new multi-targets and mechanisms according to the clinical characteristics of different allergic populations and different allergens. This review focuses on information concerning potential therapeutic targets as well as the synthetic anti-allergy small molecules with respect to their medicinal chemistry. The structure-activity relationship and the mechanism of compound-target interaction were highlighted with perspective to histamine-1/4 receptor antagonists, leukotriene biosynthesis, Th2 cytokines inhibitors, and calcium channel blockers. We hope that the study of chemical scaffold modification and optimization for different lead compounds summarized in this review not only lays the foundation for improvement of success rate and efficiency of virtual screening of antiallergic drugs, but also can provide valuable reference for the drug design of related promising research such as allergy, inflammation, and cancer.

PGFed: Personalize Each Client's Global Objective for Federated Learning.

Personalized federated learning has received an upsurge of attention due to the mediocre performance of conventional federated learning (FL) over heterogeneous data. Unlike conventional FL which trains a single global consensus model, personalized FL allows different models for different clients. However, existing personalized FL algorithms only implicitly transfer the collaborative knowledge across the federation by embedding the knowledge into the aggregated model or regularization. We observed that this implicit knowledge transfer fails to maximize the potential of each client's empirical risk toward other clients. Based on our observation, in this work, we propose Personalized Global Federated Learning (PGFed), a novel personalized FL framework that enables each client to personalize its own global objective by explicitly and adaptively aggregating the empirical risks of itself and other clients. To avoid massive (O(N2)) communication overhead and potential privacy leakage while achieving this, each client's risk is estimated through a first-order approximation for other clients' adaptive risk aggregation. On top of PGFed, we develop a momentum upgrade, dubbed PGFedMo, to more efficiently utilize clients' empirical risks. Our extensive experiments on four datasets under different federated settings show consistent improvements of PGFed over previous state-of-the-art methods. The code is publicly available at https://github.com/ljaiverson/pgfed.

SurvivalCNN: A deep learning-based method for gastric cancer survival prediction using radiological imaging data and clinicopathological variables.

Radiological images have shown promising effects in patient prognostication. Deep learning provides a powerful approach for in-depth analysis of imaging data and integration of multi-modal data for modeling. In this work, we propose SurvivalCNN, a deep learning structure for cancer patient survival prediction using CT imaging data and non-imaging clinical data. In SurvivalCNN, a supervised convolutional neural network is designed to extract volumetric image features, and radiomics features are also integrated to provide potentially different imaging information. Within SurvivalCNN, a novel multi-thread multi-layer perceptron module, namely, SurvivalMLP, is proposed to perform survival prediction from censored survival data. We evaluate the proposed SurvivalCNN framework on a large clinical dataset of 1061 gastric cancer patients for both overall survival (OS) and progression-free survival (PFS) prediction. We compare SurvivalCNN to three different modeling methods and examine the effects of various sets of data/features when used individually or in combination. With five-fold cross validation, our experimental results show that SurvivalCNN achieves averaged concordance index 0.849 and 0.783 for predicting OS and PFS, respectively, outperforming the compared state-of-the-art methods and the clinical model. After future validation, the proposed SurvivalCNN model may serve as a clinical tool to improve gastric cancer patient survival estimation and prognosis analysis.

Comparative genomic analysis of Citrobacter sp. XT1-2-2 reveals insights into the molecular mechanism of microbial immobilization of heavy metals.

BACKGROUND: In our previous study, Citrobacter sp. XT1-2-2 was isolated from high cadmium-contaminated soils, and demonstrated an excellent ability to decrease the bioavailability of cadmium in the soil and inhibit cadmium uptake in rice. In addition, the strain XT1-2-2 could significantly promote rice growth and increase rice biomass. Therefore, the strain XT1-2-2 shows great potential for remediation of cadmium -contaminated soils. However, the genome sequence of this organism has not been reported so far.  RESULTS: Here the basic characteristics and genetic diversity of the strain XT1-2-2 were described, together with the draft genome and comparative genomic results. The strain XT1-2-2 is 5040459 bp long with an average G + C content of 52.09%, and contains a total of 4801 genes. Putative genomic islands were predicted in the genome of Citrobacter sp. XT1-2-2. All genes of a complete set of sulfate reduction pathway and various putative heavy metal resistance genes in the genome were identified and analyzed. CONCLUSIONS: These analytical results provide insights into the genomic basis of microbial immobilization of heavy metals.

A multi-center study on glucometabolic response to bariatric surgery for different subtypes of obesity.

Objectives: To assess the benefit of a bariatric surgery in four artificial intelligence-identified metabolic (AIM) subtypes of obesity with respect to the improvement of glucometabolism and the remission of diabetes and hyperinsulinemia. Methods: This multicenter retrospective study prospectively collected data from five hospitals in China from 2010 to 2021. At baseline 1008 patients who underwent a bariatric surgery were enrolled (median age 31 years; median BMI 38.1kg/m2; 57.40% women) and grouped into the four AIM subtypes. Baseline and follow-up data (506 and 359 patients at 3- and 12-month post-surgery) were collected for longitudinal effect analysis. Results: Out of the four AIM subgroups, hypometabolic obesity (LMO) group was characterized by decompensated insulin secretion and high incidence of diabetes (99.2%) pre-surgery. After surgery, 62.1% of LMO patients with diabetes achieved remission, lower than the other three subgroups. Still, the bariatric surgery significantly reduced their blood glucose (median HbA1c decreased by 27.2%). The hypermetabolic obesity-hyperinsulinemia (HMO-I) group was characterized by severe insulin resistance and high incidence of hyperinsulinemia (87.8%) pre-surgery, which had been greatly alleviated post-surgery. For both metabolic healthy obesity (MHO) and hypermetabolic obesity-hyperuricemia (HMO-U) groups who showed a relatively healthy glucometabolism pre-surgery, rate of glucometabolic comorbidities improved moderately post-surgery. Conclusion: In terms of glucometabolism, the four AIM subtypes of patients benefited differently from a bariatric surgery, which significantly relieved hyperglycemia and hyperinsulinemia for the LMO and HMO-I patients, respectively. The AIM-based subtypes may help better inform clinical decisions on bariatric surgery and patient counseling pertaining to post-surgery outcomes.

SAH-NET: Structure-Aware Hierarchical Network for Clustered Microcalcification Classification in Digital Breast Tomosynthesis.

Benign and malignant classification of clustered microcalcifications (MCs) in digital breast tomosynthesis (DBT) is an essential task in computer-aided diagnosis. However, due to the anisotropic resolution of DBT, three-dimensional (3-D) convolutional neural network (CNN)-based methods cannot extract hierarchical features efficiently. Moreover, the sparse distribution of MC points in the cluster makes it difficult for the CNN to extract discriminative structural information for classification. To comprehensively address these challenges, we propose a novel structure-aware hierarchical network (SAH-Net) for benign and malignant classification of clustered MC in a DBT volume. Specifically, the two-dimensional (2-D) group convolution is used to extract intraslice features. The one-to-one correspondence between group convolutions and slices ensures the independence of hierarchical feature extraction. Then, a partial deformable Transformer-based 3-D structural feature learning module is proposed to capture the long-range dependency between MC points in the cluster. We evaluate the proposed method on an in-house dataset with 495 clustered MCs collected from 462 DBT images. Experimental results confirm the validity of our proposed modules. The results also show that the proposed SAH-Net outperforms several other representative methods on this topic, and achieves the best classification result, with an area under the receiver operation curve (AUC) of 86.87%. The implementation of the proposed model is available at https://github.com/sunhaotian130911/SAHNet.

Four clinical phenotypes of cow's milk protein allergy based on dairy product specific IgE antibody types in North China.

Background: Cow's milk protein allergy (CMPA) is a common allergy. Immunoglobulin E (IgE)-mediated cow's milk allergy is associated with a high mortality risk and poor prognosis. The study aims to investigate whether there are different clinically CMPA phenotypes in China and to explore the association between CMPA phenotypes and specific IgE (sIgE) antibodies against different dairy products. Methods: Serum sIgE against different animal milk and cow's milk products and different milk components was measured by an allergen array. Four CMPA classifications were identified by the presence of serum sIgE: boiled milk-positive, yogurt-positive, buttermilk-positive, and raw milk-positive. Results: Among the 234 participants included in the study, 9 were boiled milk sIgE-positive, 50 were yogurt sIgE-positive, 17 were buttermilk sIgE-positive, and 158 were only raw milk sIgE-positive. The boiled milk-positive group had the highest levels of raw milk sIgE and casein sIgE antibodies, followed sequentially by the yogurt-positive, buttermilk-positive, and raw milk-positive groups. The boiled milk group observed the highest levels of sIgE against raw milk, casein, α-lactalbumin, and ß-lactoglobulin. These levels differed significantly from those in the other three groups. Allergic symptoms were distributed differently among the four study groups. The percentages of allergic patients with gastrointestinal tract symptoms in the above mentioned four groups ranged from high to low, and the percentages of patients with skin symptoms in the four groups ranged from low to high, respectively. Conclusion: Based on dairy product sIgE antibody levels associated with different milk components and various clinical allergic symptom tendencies, we could distinguish four CMPA phenotypes.

A self-training teacher-student model with an automatic label grader for abdominal skeletal muscle segmentation.

Deep learning on a limited number of labels/annotations is a challenging task for medical imaging analysis. In this paper, we propose a novel self-training segmentation pipeline (Self-Seg in short) for segmenting skeletal muscle in CT images. Self-Seg starts with a small set of annotated images and then iteratively learns from unlabeled datasets to gradually improve the segmentation performance. Self-Seg follows a semi-supervised teacher-student learning scheme and there are two contributions: 1) we construct a self-attention UNet to improve segmentation over the classical UNet model, and 2) we implement an automatic label grader to implicitly incorporate medical knowledge for quality assurance of pseudo labels, from which good quality pseudo labels are identified to enhance learning of the segmentation model. We perform extensive experiments on three CT image datasets and show promising results on five evaluation settings, and we also compared our method to several baseline and related methods and achieved superior performance.

Outcome Prediction in Patients with Severe Traumatic Brain Injury Using Deep Learning from Head CT Scans.

Background After severe traumatic brain injury (sTBI), physicians use long-term prognostication to guide acute clinical care yet struggle to predict outcomes in comatose patients. Purpose To develop and evaluate a prognostic model combining deep learning of head CT scans and clinical information to predict long-term outcomes after sTBI. Materials and Methods This was a retrospective analysis of two prospectively collected databases. The model-building set included 537 patients (mean age, 40 years ± 17 [SD]; 422 men) from one institution from November 2002 to December 2018. Transfer learning and curriculum learning were applied to a convolutional neural network using admission head CT to predict mortality and unfavorable outcomes (Glasgow Outcomes Scale scores 1-3) at 6 months. This was combined with clinical input for a holistic fusion model. The models were evaluated using an independent internal test set and an external cohort of 220 patients with sTBI (mean age, 39 years ± 17; 166 men) from 18 institutions in the Transforming Research and Clinical Knowledge in Traumatic Brain Injury (TRACK-TBI) study from February 2014 to April 2018. The models were compared with the International Mission on Prognosis and Analysis of Clinical Trials in TBI (IMPACT) model and the predictions of three neurosurgeons. Area under the receiver operating characteristic curve (AUC) was used as the main model performance metric. Results The fusion model had higher AUCs than did the IMPACT model in the prediction of mortality (AUC, 0.92 [95% CI: 0.86, 0.97] vs 0.80 [95% CI: 0.71, 0.88]; P < .001) and unfavorable outcomes (AUC, 0.88 [95% CI: 0.82, 0.94] vs 0.82 [95% CI: 0.75, 0.90]; P = .04) on the internal data set. For external TRACK-TBI testing, there was no evidence of a significant difference in the performance of any models compared with the IMPACT model (AUC, 0.83; 95% CI: 0.77, 0.90) in the prediction of mortality. The Imaging model (AUC, 0.73; 95% CI: 0.66-0.81; P = .02) and the fusion model (AUC, 0.68; 95% CI: 0.60, 0.76; P = .02) underperformed as compared with the IMPACT model (AUC, 0.83; 95% CI: 0.77, 0.89) in the prediction of unfavorable outcomes. The fusion model outperformed the predictions of the neurosurgeons. Conclusion A deep learning model of head CT and clinical information can be used to predict 6-month outcomes after severe traumatic brain injury. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Haller in this issue.

Deep learning of early brain imaging to predict post-arrest electroencephalography.

INTRODUCTION: Guidelines recommend use of computerized tomography (CT) and electroencephalography (EEG) in post-arrest prognostication. Strong associations between CT and EEG might obviate the need to acquire both modalities. We quantified these associations via deep learning. METHODS: We performed a single-center, retrospective study including comatose patients hospitalized after cardiac arrest. We extracted brain CT DICOMs, resized and registered each to a standard anatomical atlas, performed skull stripping and windowed images to optimize contrast of the gray-white junction. We classified initial EEG as generalized suppression, other highly pathological findings or benign activity. We extracted clinical information available on presentation from our prospective registry. We trained three machine learning (ML) models to predict EEG from clinical covariates. We used three state-of-the-art approaches to build multi-headed deep learning models using similar model architectures. Finally, we combined the best performing clinical and imaging models. We evaluated discrimination in test sets. RESULTS: We included 500 patients, of whom 218 (44%) had benign EEG findings, 135 (27%) showed generalized suppression and 147 (29%) had other highly pathological findings that were most commonly (93%) burst suppression with identical bursts. Clinical ML models had moderate discrimination (test set AUCs 0.73-0.80). Image-based deep learning performed worse (test set AUCs 0.51-0.69), particularly discriminating benign from highly pathological findings. Adding image-based deep learning to clinical models improved prediction of generalized suppression due to accurate detection of severe cerebral edema. DISCUSSION: CT and EEG provide complementary information about post-arrest brain injury. Our results do not support selective acquisition of only one of these modalities, except in the most severely injured patients.

Adapt to Adaptation: Learning Personalization for Cross-Silo Federated Learning.

Conventional federated learning (FL) trains one global model for a federation of clients with decentralized data, reducing the privacy risk of centralized training. However, the distribution shift across non-IID datasets, often poses a challenge to this one-model-fits-all solution. Personalized FL aims to mitigate this issue systematically. In this work, we propose APPLE, a personalized cross-silo FL framework that adaptively learns how much each client can benefit from other clients' models. We also introduce a method to flexibly control the focus of training APPLE between global and local objectives. We empirically evaluate our method's convergence and generalization behaviors, and perform extensive experiments on two benchmark datasets and two medical imaging datasets under two non-IID settings. The results show that the proposed personalized FL framework, APPLE, achieves state-of-the-art performance compared to several other personalized FL approaches in the literature. The code is publicly available at https://github.com/ljaiverson/pFL-APPLE.

FEDSLD: FEDERATED LEARNING WITH SHARED LABEL DISTRIBUTION FOR MEDICAL IMAGE CLASSIFICATION.

Federated learning (FL) enables collaboratively training a joint model for multiple medical centers, while keeping the data decentralized due to privacy concerns. However, federated optimizations often suffer from the heterogeneity of the data distribution across medical centers. In this work, we propose Federated Learning with Shared Label Distribution (FedSLD) for classification tasks, a method that adjusts the contribution of each data sample to the local objective during optimization via knowledge of clients' label distribution, mitigating the instability brought by data heterogeneity. We conduct extensive experiments on four publicly available image datasets with different types of non-IID data distributions. Our results show that FedSLD achieves better convergence performance than the compared leading FL optimization algorithms, increasing the test accuracy by up to 5.50 percentage points.