Development of a deep learning model to distinguish the cause of optic disc atrophy using retinal fundus photography.

The differential diagnosis for optic atrophy can be challenging and requires expensive, time-consuming ancillary testing to determine the cause. While Leber's hereditary optic neuropathy (LHON) and optic neuritis (ON) are both clinically significant causes for optic atrophy, both relatively rare in the general population, contributing to limitations in obtaining large imaging datasets. This study therefore aims to develop a deep learning (DL) model based on small datasets that could distinguish the cause of optic disc atrophy using only fundus photography. We retrospectively reviewed fundus photographs of 120 normal eyes, 30 eyes (15 patients) with genetically-confirmed LHON, and 30 eyes (26 patients) with ON. Images were split into a training dataset and a test dataset and used for model training with ResNet-18. To visualize the critical regions in retinal photographs that are highly associated with disease prediction, Gradient-Weighted Class Activation Map (Grad-CAM) was used to generate image-level attention heat maps and to enhance the interpretability of the DL system. In the 3-class classification of normal, LHON, and ON, the area under the receiver operating characteristic curve (AUROC) was 1.0 for normal, 0.988 for LHON, and 0.990 for ON, clearly differentiating each class from the others with an overall total accuracy of 0.93. Specifically, when distinguishing between normal and disease cases, the precision, recall, and F1 scores were perfect at 1.0. Furthermore, in the differentiation of LHON from other conditions, ON from others, and between LHON and ON, we consistently observed precision, recall, and F1 scores of 0.8. The model performance was maintained until only 10% of the pixel values of the image, identified as important by Grad-CAM, were preserved and the rest were masked, followed by retraining and evaluation.

Comprehensive Assessment and Early Prediction of Gross Motor Performance in Toddlers With Graph Convolutional Networks-Based Deep Learning: Development and Validation Study.

BACKGROUND: Accurate and timely assessment of children's developmental status is crucial for early diagnosis and intervention. More accurate and automated developmental assessments are essential due to the lack of trained health care providers and imprecise parental reporting. In various areas of development, gross motor development in toddlers is known to be predictive of subsequent childhood developments. OBJECTIVE: The purpose of this study was to develop a model to assess gross motor behavior and integrate the results to determine the overall gross motor status of toddlers. This study also aimed to identify behaviors that are important in the assessment of overall gross motor skills and detect critical moments and important body parts for the assessment of each behavior. METHODS: We used behavioral videos of toddlers aged 18-35 months. To assess gross motor development, we selected 4 behaviors (climb up the stairs, go down the stairs, throw the ball, and stand on 1 foot) that have been validated with the Korean Developmental Screening Test for Infants and Children. In the child behavior videos, we estimated each child's position as a bounding box and extracted human keypoints within the box. In the first stage, the videos with the extracted human keypoints of each behavior were evaluated separately using a graph convolutional networks (GCN)-based algorithm. The probability values obtained for each label in the first-stage model were used as input for the second-stage model, the extreme gradient boosting (XGBoost) algorithm, to predict the overall gross motor status. For interpretability, we used gradient-weighted class activation mapping (Grad-CAM) to identify important moments and relevant body parts during the movements. The Shapley additive explanations method was used for the assessment of variable importance, to determine the movements that contributed the most to the overall developmental assessment. RESULTS: Behavioral videos of 4 gross motor skills were collected from 147 children, resulting in a total of 2395 videos. The stage-1 GCN model to evaluate each behavior had an area under the receiver operating characteristic curve (AUROC) of 0.79 to 0.90. Keypoint-mapping Grad-CAM visualization identified important moments in each behavior and differences in important body parts. The stage-2 XGBoost model to assess the overall gross motor status had an AUROC of 0.90. Among the 4 behaviors, "go down the stairs" contributed the most to the overall developmental assessment. CONCLUSIONS: Using movement videos of toddlers aged 18-35 months, we developed objective and automated models to evaluate each behavior and assess each child's overall gross motor performance. We identified the important behaviors for assessing gross motor performance and developed methods to recognize important moments and body parts while evaluating gross motor performance.

Synthetic Data Improve Survival Status Prediction Models in Early-Onset Colorectal Cancer.

PURPOSE: In artificial intelligence-based modeling, working with a limited number of patient groups is challenging. This retrospective study aimed to evaluate whether applying synthetic data generation methods to the clinical data of small patient groups can enhance the performance of prediction models. MATERIALS AND METHODS: A data set collected by the Cancer Registry Library Project from the Yonsei Cancer Center (YCC), Severance Hospital, between January 2008 and October 2020 was reviewed. Patients with colorectal cancer younger than 50 years who started their initial treatment at YCC were included. A Bayesian network-based synthesizing model was used to generate a synthetic data set, combined with the differential privacy (DP) method. RESULTS: A synthetic population of 5,005 was generated from a data set of 1,253 patients with 93 clinical features. The Hellinger distance and correlation difference metric were below 0.3 and 0.5, respectively, indicating no statistical difference. The overall survival by disease stage did not differ between the synthetic and original populations. Training with the synthetic data and validating with the original data showed the highest performances of 0.850, 0.836, and 0.790 for the Decision Tree, Random Forest, and XGBoost models, respectively. Comparison of synthetic data sets with different epsilon parameters from the original data sets showed improved performance >0.1%. For extremely small data sets, models using synthetic data outperformed those using only original data sets. The reidentification risk measures demonstrated that the epsilons between 0.1 and 100 fell below the baseline, indicating a preserved privacy state. CONCLUSION: The synthetic data generation approach enhances predictive modeling performance by maintaining statistical and clinical integrity, and simultaneously reduces privacy risks through the application of DP techniques.

Screening of Moyamoya Disease From Retinal Photographs: Development and Validation of Deep Learning Algorithms.

BACKGROUND: Moyamoya disease (MMD) is a rare and complex pathological condition characterized by an abnormal collateral circulation network in the basal brain. The diagnosis of MMD and its progression is unpredictable and influenced by many factors. MMD can affect the blood vessels supplying the eyes, resulting in a range of ocular symptoms. In this study, we developed a deep learning model using real-world data to assist a diagnosis and determine the stage of the disease using retinal photographs. METHODS: This retrospective observational study conducted from August 2006 to March 2022 included 498 retinal photographs from 78 patients with MMD and 3835 photographs from 1649 healthy participants. Photographs were preprocessed, and an ResNeXt50 model was developed. Model performance was measured using receiver operating curves and their area under the receiver operating characteristic curve, accuracy, sensitivity, and F1-score. Heatmaps and progressive erasing plus progressive restoration were performed to validate the faithfulness. RESULTS: Overall, 322 retinal photographs from 67 patients with MMD and 3752 retinal photographs from 1616 healthy participants were used to develop a screening and stage prediction model for MMD. The average age of the patients with MMD was 44.1 years, and the average follow-up time was 115 months. Stage 3 photographs were the most prevalent, followed by stages 4, 5, 2, 1, and 6 and healthy. The MMD screening model had an average area under the receiver operating characteristic curve of 94.6%, with 89.8% sensitivity and 90.4% specificity at the best cutoff point. MMD stage prediction models had an area under the receiver operating characteristic curve of 78% or higher, with stage 3 performing the best at 93.6%. Heatmap identified the vascular region of the fundus as important for prediction, and progressive erasing plus progressive restoration result shows an area under the receiver operating characteristic curve of 70% only with 50% of the important regions. CONCLUSIONS: This study demonstrated that retinal photographs could be used as potential biomarkers for screening and staging of MMD and the disease stage could be classified by a deep learning algorithm.

Development of Deep Ensembles to Screen for Autism and Symptom Severity Using Retinal Photographs.

Importance: Screening for autism spectrum disorder (ASD) is constrained by limited resources, particularly trained professionals to conduct evaluations. Individuals with ASD have structural retinal changes that potentially reflect brain alterations, including visual pathway abnormalities through embryonic and anatomic connections. Whether deep learning algorithms can aid in objective screening for ASD and symptom severity using retinal photographs is unknown. Objective: To develop deep ensemble models to differentiate between retinal photographs of individuals with ASD vs typical development (TD) and between individuals with severe ASD vs mild to moderate ASD. Design, Setting, and Participants: This diagnostic study was conducted at a single tertiary-care hospital (Severance Hospital, Yonsei University College of Medicine) in Seoul, Republic of Korea. Retinal photographs of individuals with ASD were prospectively collected between April and October 2022, and those of age- and sex-matched individuals with TD were retrospectively collected between December 2007 and February 2023. Deep ensembles of 5 models were built with 10-fold cross-validation using the pretrained ResNeXt-50 (32×4d) network. Score-weighted visual explanations for convolutional neural networks, with a progressive erasing technique, were used for model visualization and quantitative validation. Data analysis was performed between December 2022 and October 2023. Exposures: Autism Diagnostic Observation Schedule-Second Edition calibrated severity scores (cutoff of 8) and Social Responsiveness Scale-Second Edition T scores (cutoff of 76) were used to assess symptom severity. Main Outcomes and Measures: The main outcomes were participant-level area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity. The 95% CI was estimated through the bootstrapping method with 1000 resamples. Results: This study included 1890 eyes of 958 participants. The ASD and TD groups each included 479 participants (945 eyes), had a mean (SD) age of 7.8 (3.2) years, and comprised mostly boys (392 [81.8%]). For ASD screening, the models had a mean AUROC, sensitivity, and specificity of 1.00 (95% CI, 1.00-1.00) on the test set. These models retained a mean AUROC of 1.00 using only 10% of the image containing the optic disc. For symptom severity screening, the models had a mean AUROC of 0.74 (95% CI, 0.67-0.80), sensitivity of 0.58 (95% CI, 0.49-0.66), and specificity of 0.74 (95% CI, 0.67-0.82) on the test set. Conclusions and Relevance: These findings suggest that retinal photographs may be a viable objective screening tool for ASD and possibly for symptom severity. Retinal photograph use may speed the ASD screening process, which may help improve accessibility to specialized child psychiatry assessments currently strained by limited resources.

Characteristics and risk factors of suicide among people who attempted self-harm in South Korea: A longitudinal National Cohort Study in South Korea.

Although self-harm is known as a significant risk factor for suicide, there are insufficient studies on the characteristics of people who self-harmed and the factors affecting suicide using a national dataset in Asia. This study aimed to identify demographic, clinical, and socioeconomic factors of individuals who attempted self-harm concerning suicide mortality. By analyzing the Korean National Health Insurance Service data from 2002 to 2020, we compared the people who attempted self-harm to the general population and explored factors affecting suicide by using the Cox proportional hazards model. Older age, female sex, lower socioeconomic status, and psychiatric conditions were associated with higher self-harm attempts. Suicide was more prevalent among males with mild disabilities, using fatal self-harm methods, and higher Charlson Comorbidity Index (CCI) scores. Socioeconomic factors that were significantly related to self-harm attempt were relatively less significant in the suicide survival analysis, while male gender, older age, fatal self-harm methods, high CCI scores, psychiatric diagnosis, and drinking habits were significantly associated with lower suicide survival rates. These results showed that demographic, clinical and socioeconomic factors affecting self-harm differ from those affecting actual suicidal death after self-harm. These insights may assist in developing targeted prevention strategies for specific populations.

Three-dimensional label-free morphology of CD8 + T cells as a sepsis biomarker.

Sepsis is a dysregulated immune response to infection that leads to organ dysfunction and is associated with a high incidence and mortality rate. The lack of reliable biomarkers for diagnosing and prognosis of sepsis is a major challenge in its management. We aimed to investigate the potential of three-dimensional label-free CD8 + T cell morphology as a biomarker for sepsis. This study included three-time points in the sepsis recovery cohort (N = 8) and healthy controls (N = 20). Morphological features and spatial distribution within cells were compared among the patients' statuses. We developed a deep learning model to predict the diagnosis and prognosis of sepsis using the internal cell morphology. Correlation between the morphological features and clinical indices were analysed. Cell morphological features and spatial distribution differed significantly between patients with sepsis and healthy controls and between the survival and non-survival groups. The model for predicting the diagnosis and prognosis of sepsis showed an area under the receiver operating characteristic curve of nearly 100% with only a few cells, and a strong correlation between the morphological features and clinical indices was observed. Our study highlights the potential of three-dimensional label-free CD8 + T cell morphology as a promising biomarker for sepsis. This approach is rapid, requires a minimum amount of blood samples, and has the potential to provide valuable information for the early diagnosis and prognosis of sepsis.

Prognostic value of baseline and early treatment response of neutrophil-lymphocyte ratio, C-reactive protein, and lactate dehydrogenase in non-small cell lung cancer patients undergoing immunotherapy.

Background: Not all non-small cell lung cancer (NSCLC) patients will benefit from immune checkpoint therapy and use of these medications carry serious autoimmune adverse effects. Therefore, biomarkers are needed to better identify patients who will benefit from its use. Here, the correlation of overall survival (OS) with baseline and early treatment period serum biomarker responses was evaluated in patients with NSCLC undergoing immunotherapy. Methods: Patients diagnosed with NSCLC undergoing immunotherapy (n=597) at a tertiary academic medical center in South Korea were identified between January 2010 and November 2021. The neutrophil-lymphocyte ratio (NLR), C-reactive protein (CRP), and lactate dehydrogenase (LDH) levels in the survival and non-survival groups were examined at baseline and early treatment periods. Additionally, aberrant laboratory parameters at each period were used to stratify survival curves and examine their correlation with one-year OS. Results: In the non-survival group, the NLR, CRP, and LDH levels at the early treatment period were higher than those at the baseline (P<0.001). The survival curves stratified based on aberrant laboratory findings in each period varied (log-rank test P<0.001). Multivariate Cox regression analysis revealed that having prescribed more than 3rd line of chemotherapy [hazard ratio (HR) =3.19, 95% confidence interval (CI): 1.04-9.82; P=0.043] and early treatment period CRP (HR =3.88; 95% CI: 1.55-9.72; P=0.004) and LDH (HR =4.04; 95% CI: 2.01-8.12; P<0.001) levels were significant predictors of one-year OS. Conclusions: Early treatment period CRP and LDH levels were significant predictors of OS in patients with NSCLC undergoing immunotherapy.

The Impact of a Mobile Support Group on Distress and Physical Activity in Breast Cancer Survivors: Randomized, Parallel-Group, Open-Label, Controlled Trial.

BACKGROUND: While mobile health apps have demonstrated their potential in revolutionizing health behavior changes, the impact of a mobile community built on these apps on the level of physical activity and mental well-being in cancer survivors remains unexplored. OBJECTIVE: In this randomized controlled trial, we examine the effects of participation in a mobile health community specifically designed for breast cancer survivors on their physical activity levels and mental distress. METHODS: We performed a single-center, randomized, parallel-group, open-label, controlled trial. This trial enrolled women between 20 and 60 years of age with stage 0 to III breast cancer, an Eastern Cooperative Oncology Group performance status of 0, and the capability of using their own smartphone apps. From January 7, 2019, to April 17, 2020, a total of 2,616 patients were consecutively screened for eligibility after breast cancer surgery. Overall, 202 patients were enrolled in this trial, and 186 patients were randomly assigned (1:1) to either the intervention group (engagement in a mobile peer support community using an app for tracking steps; n=93) or the control group (using the app for step tracking only; n=93) with a block size of 10 without stratification. The mobile app provides a visual interface of daily step counts, while the community function also provides rankings among its members and regular notifications encouraging physical activity. The primary end point was the rate of moderate to severe distress for the 24-week study period, measured through an app-based survey using the Distress Thermometer. The secondary end point was the total weekly steps during the 24-week period. RESULTS: After excluding dropouts, 85 patients in the intervention group and 90 patients in the control group were included in the analysis. Multivariate analyses showed that patients in the intervention group had a significantly lower degree of moderate to severe distress (B=-0.558; odds ratio 0.572; P<.001) and a higher number of total weekly step counts (B=0.125; rate ratio 1.132; P<.001) during the 24-week period. CONCLUSIONS: Engagement in a mobile app-based patient community was effective in reducing mental distress and increasing physical activity in breast cancer survivors. TRIAL REGISTRATION: ClinicalTrials.gov NCT03783481; https://classic.clinicaltrials.gov/ct2/show/NCT03783481.

Development and Verification of Time-Series Deep Learning for Drug-Induced Liver Injury Detection in Patients Taking Angiotensin II Receptor Blockers: A Multicenter Distributed Research Network Approach.

OBJECTIVES: The objective of this study was to develop and validate a multicenter-based, multi-model, time-series deep learning model for predicting drug-induced liver injury (DILI) in patients taking angiotensin receptor blockers (ARBs). The study leveraged a national-level multicenter approach, utilizing electronic health records (EHRs) from six hospitals in Korea. METHODS: A retrospective cohort analysis was conducted using EHRs from six hospitals in Korea, comprising a total of 10,852 patients whose data were converted to the Common Data Model. The study assessed the incidence rate of DILI among patients taking ARBs and compared it to a control group. Temporal patterns of important variables were analyzed using an interpretable timeseries model. RESULTS: The overall incidence rate of DILI among patients taking ARBs was found to be 1.09%. The incidence rates varied for each specific ARB drug and institution, with valsartan having the highest rate (1.24%) and olmesartan having the lowest rate (0.83%). The DILI prediction models showed varying performance, measured by the average area under the receiver operating characteristic curve, with telmisartan (0.93), losartan (0.92), and irbesartan (0.90) exhibiting higher classification performance. The aggregated attention scores from the models highlighted the importance of variables such as hematocrit, albumin, prothrombin time, and lymphocytes in predicting DILI. CONCLUSIONS: Implementing a multicenter-based timeseries classification model provided evidence that could be valuable to clinicians regarding temporal patterns associated with DILI in ARB users. This information supports informed decisions regarding appropriate drug use and treatment strategies.

Protocol for the development of joint attention-based subclassification of autism spectrum disorder and validation using multi-modal data.

BACKGROUND: Heterogeneity in clinical manifestation and underlying neuro-biological mechanisms are major obstacles to providing personalized interventions for individuals with autism spectrum disorder (ASD). Despite various efforts to unify disparate data modalities and machine learning techniques for subclassification, replicable ASD clusters remain elusive. Our study aims to introduce a novel method, utilizing the objective behavioral biomarker of gaze patterns during joint attention, to subclassify ASD. We will assess whether behavior-based subgrouping yields clinically, genetically, and neurologically distinct ASD groups. METHODS: We propose a study involving 60 individuals with ASD recruited from a specialized psychiatric clinic to perform joint attention tasks. Through the examination of gaze patterns in social contexts, we will conduct a semi-supervised clustering analysis, yielding two primary clusters: good gaze response group and poor gaze response group. Subsequent comparison will occur across these clusters, scrutinizing neuroanatomical structure and connectivity using structural as well as functional brain imaging studies, genetic predisposition through single nucleotide polymorphism data, and assorted socio-demographic and clinical information. CONCLUSIONS: The aim of the study is to investigate the discriminative properties and the validity of the joint attention-based subclassification of ASD using multi-modality data. TRIAL REGISTRATION: Clinical trial, KCT0008530, Registered 16 June 2023, https://cris.nih.go.kr/cris/index/index.do .

Metaverse-based social skills training programme for children with autism spectrum disorder to improve social interaction ability: an open-label, single-centre, randomised controlled pilot trial.

Background: Autism spectrum disorder (ASD) is characterised by abnormalities in social interactions and restricted and repetitive behaviors. Children with high-functioning ASD (HFASD), lack social communication skills, do not interact with others, and lack peer relationships. We aimed to develop, and evaluate the feasibility of, a metaverse-based programme to enhance the social skills of children with HFASD. Methods: This open-label, single-centre, pilot parallel randomised controlled trial (RCT) was conducted on boys aged 7-12 years with HFASD. Children were recruited from a treatment centre for children with HFASD in Korea or by self-referral through online community webpages for the parents of children with HFASD. Participants were randomly assigned (1:1) by a blinded researcher to receive either four weeks of a metaverse-based social skills training programme or a control group. Randomisation was stratified by age (children aged 7-9 and 10-12 years) using permuted blocks (block size 4). The metaverse-based social skills training programme was delivered via the metaverse platforms (Roblox) and Zoom. Children in the intervention group completed the metaverse-based social skills training programme at home for four weeks. The intervention consisted of four sessions, one session per week, for 60 min each. The control group did not receive any interventions. The primary outcome measure was the median change in the Social Responsiveness Scale-2 (SRS-2) scores from pre-to post-intervention. SRS-2 is an assessment tool used to confirm the effectiveness of social interactions. Higher scores indicate lower social functioning. The trial is registered with CRIS Registration Number; KCT0006859. Findings: Between February 14, 2022, and March 31, 2022, 20 participants were enrolled. Overall, 15 children (median [Interquartile range (IQR)] age, intervention group: 9.0 [8.0-10.0]; control group: 8.5 [8.0-10.0]) participated in the programme. The intervention group included nine participants (60%), and the control group included six participants (40%). The SRS-2 total scores for the intervention group decreased from baseline 96.0 (IQR: 74.0-112.0) to post-intervention 85.0 (IQR: 84.0-103.0). The group median difference in SRS-2 scores between the intervention and control groups was 11.5 (95% CI: 8.5-14.0), with a further reduction in the intervention group. Similar trends were seen for social cognition (group median difference, 95% CI: 2.0, 1.0-4.0), social communication (group median difference, 95% CI: 2.0, 1.0-4.0), and autistic mannerism (group median difference, 95% CI: 4.0, 1.0-5.0). There were no adverse events related to study participation. Interpretation: The findings of this feasibility study suggest that children with HFASD can potentially be familiarised, through metaverse-based programmes, with real-life social situations to improve sociality and reduce emotional and behavioural problems. Such interventions could be delivered at home and possibly be extended to target groups that have difficulty in interacting with peers offline. Funding: The Institute of Information & Communications Technology Planning & Evaluation grant, via the Ministry of Science and ICT of the South Korean Government.

Development and validation of an artificial intelligence model for the early classification of the aetiology of meningitis and encephalitis: a retrospective observational study.

Background: Early diagnosis and appropriate treatment are essential in meningitis and encephalitis management. We aimed to implement and verify an artificial intelligence (AI) model for early aetiological determination of patients with encephalitis and meningitis, and identify important variables in the classification process. Methods: In this retrospective observational study, patients older than 18 years old with meningitis or encephalitis at two centres in South Korea were enrolled for development (n = 283) and external validation (n = 220) of AI models, respectively. Their clinical variables within 24 h after admission were used for the multi-classification of four aetiologies including autoimmunity, bacteria, virus, and tuberculosis. The aetiology was determined based on the laboratory test results of cerebrospinal fluid conducted during hospitalization. Model performance was assessed using classification metrics, including the area under the receiver operating characteristic curve (AUROC), recall, precision, accuracy, and F1 score. Comparisons were performed between the AI model and three clinicians with varying neurology experience. Several techniques (eg, Shapley values, F score, permutation feature importance, and local interpretable model-agnostic explanations weights) were used for the explainability of the AI model. Findings: Between January 1, 2006, and June 30, 2021, 283 patients were enrolled in the training/test dataset. An ensemble model with extreme gradient boosting and TabNet showed the best performance among the eight AI models with various settings in the external validation dataset (n = 220); accuracy, 0.8909; precision, 0.8987; recall, 0.8909; F1 score, 0.8948; AUROC, 0.9163. The AI model outperformed all clinicians who achieved a maximum F1 score of 0.7582, by demonstrating a performance of F1 score greater than 0.9264. Interpretation: This is the first multiclass classification study for the early determination of the aetiology of meningitis and encephalitis based on the initial 24-h data using an AI model, which showed high performance metrics. Future studies can improve upon this model by securing and inputting time-series variables and setting various features about patients, and including a survival analysis for prognosis prediction. Funding: MD-PhD/Medical Scientist Training Program through the Korea Health Industry Development Institute, funded by the Ministry of Health & Welfare, Republic of Korea.

Bleeding risk and mortality according to antithrombotic agents' exposure in cancer-related stroke patients: nationwide population-based cohort study in South Korea.

BACKGROUND: Ischemic stroke with active cancer is thought to have a unique mechanism compared to conventional stroke etiologies. There is no gold standard guideline for secondary prevention in patients with cancer-related stroke, hence, adequate type of antithrombotic agent for treatment is controversial. METHODS: Subjects who were enrolled in National Health Insurance System Customized Research data during the period between 2010 and 2015 were observed until 2019. Subject diagnosed with ischemic stroke within six months before and 12 months after a cancer diagnosis was defined as cancer-related stroke patient. To solve immeasurable time bias, the drug exposure evaluation was divided into daily units, and each person-day was classified as four groups: antiplatelet, anticoagulant, both types, and unexposed to antithrombotic drugs. To investigate bleeding risk and mortality, Cox proportional hazards regression model with time-dependent covariates were used. RESULTS: Two thousand two hundred eighty-five subjects with cancer-related stroke were followed and analyzed. A group with anticoagulation showed high estimated hazard ratios (HRs) of all bleeding events compared to a group with antiplatelet (major bleeding HR, 1.35; 95% confidence interval [CI], 1.20-1.52; p < 0.001). And the result was also similar in the combination group (major bleeding HR, 1.54; 95% CI, 1.13-2.09; p = 0.006). The combination group also showed increased mortality HR compared to antiplatelet group (HR, 1.72; 95% CI, 1.47-2.00; p < 0.001). CONCLUSIONS: Bleeding risk increased in the anticoagulant-exposed group compared to antiplatelet-exposed group in cancer-related stroke patients. Thus, this result should be considered when selecting a secondary prevention drug.

Development and Validation of a Joint Attention-Based Deep Learning System for Detection and Symptom Severity Assessment of Autism Spectrum Disorder.

Importance: Joint attention, composed of complex behaviors, is an early-emerging social function that is deficient in children with autism spectrum disorder (ASD). Currently, no methods are available for objectively quantifying joint attention. Objective: To train deep learning (DL) models to distinguish ASD from typical development (TD) and to differentiate ASD symptom severities using video data of joint attention behaviors. Design, Setting, and Participants: In this diagnostic study, joint attention tasks were administered to children with and without ASD, and video data were collected from multiple institutions from August 5, 2021, to July 18, 2022. Of 110 children, 95 (86.4%) completed study measures. Enrollment criteria were 24 to 72 months of age and ability to sit with no history of visual or auditory deficits. Exposures: Children were screened using the Childhood Autism Rating Scale. Forty-five children were diagnosed with ASD. Three types of joint attention were assessed using a specific protocol. Main Outcomes and Measures: Correctly distinguishing ASD from TD and different levels of ASD symptom severity using the DL model area under the receiver operating characteristic curve (AUROC), accuracy, precision, and recall. Results: The analytical population consisted of 45 children with ASD (mean [SD] age, 48.0 [13.4] months; 24 [53.3%] boys) vs 50 with TD (mean [SD] age, 47.9 [12.5] months; 27 [54.0%] boys). The DL ASD vs TD models showed good predictive performance for initiation of joint attention (IJA) (AUROC, 99.6% [95% CI, 99.4%-99.7%]; accuracy, 97.6% [95% CI, 97.1%-98.1%]; precision, 95.5% [95% CI, 94.4%-96.5%]; and recall, 99.2% [95% CI, 98.7%-99.6%]), low-level response to joint attention (RJA) (AUROC, 99.8% [95% CI, 99.6%-99.9%]; accuracy, 98.8% [95% CI, 98.4%-99.2%]; precision, 98.9% [95% CI, 98.3%-99.4%]; and recall, 99.1% [95% CI, 98.6%-99.5%]), and high-level RJA (AUROC, 99.5% [95% CI, 99.2%-99.8%]; accuracy, 98.4% [95% CI, 97.9%-98.9%]; precision, 98.8% [95% CI, 98.2%-99.4%]; and recall, 98.6% [95% CI, 97.9%-99.2%]). The DL-based ASD symptom severity models showed reasonable predictive performance for IJA (AUROC, 90.3% [95% CI, 88.8%-91.8%]; accuracy, 84.8% [95% CI, 82.3%-87.2%]; precision, 76.2% [95% CI, 72.9%-79.6%]; and recall, 84.8% [95% CI, 82.3%-87.2%]), low-level RJA (AUROC, 84.4% [95% CI, 82.0%-86.7%]; accuracy, 78.4% [95% CI, 75.0%-81.7%]; precision, 74.7% [95% CI, 70.4%-78.8%]; and recall, 78.4% [95% CI, 75.0%-81.7%]), and high-level RJA (AUROC, 84.2% [95% CI, 81.8%-86.6%]; accuracy, 81.0% [95% CI, 77.3%-84.4%]; precision, 68.6% [95% CI, 63.8%-73.6%]; and recall, 81.0% [95% CI, 77.3%-84.4%]). Conclusions and Relevance: In this diagnostic study, DL models for identifying ASD and differentiating levels of ASD symptom severity were developed and the premises for DL-based predictions were visualized. The findings suggest that this method may allow digital measurement of joint attention; however, follow-up studies are necessary for further validation.

Investigation of the Relationship Between Psychiatry Visit and Suicide After Deliberate Self-harm: Longitudinal National Cohort Study.

BACKGROUND: Deliberate self-harm (DSH) along with old age, physical disability, and low socioeconomic status are well-known contributors to suicide-related deaths. In recent years, South Korea has the highest suicide death rate among all Organization for Economic Co-operation and Development countries. Owing to the difficulty of accessing data of individuals with DSH behavior who died by suicide, the factors associated with suicide death in these high-risk individuals have not been sufficiently explored. There have been conflicting findings with regard to the relationship between previous psychiatric visits and suicidal death. OBJECTIVE: We aimed to address the following 3 questions: Are there considerable differences in demographics, socioeconomic status, and clinical features in individuals who received psychiatric diagnosis (either before DSH or after DSH event) and those who did not? Does receiving a psychiatric diagnosis from the Department of Psychiatry, as opposed to other departments, affect survival? and Which factors related to DSH contribute to deaths by suicide? METHODS: We used the Korean National Health Insurance Service Database to design a cohort of 5640 individuals (3067/5640, 54.38% women) who visited the hospital for DSH (International Classification of Diseases codes X60-X84) between 2002 and 2020. We analyzed whether there were significant differences among subgroups of individuals with DSH behavior based on psychiatric diagnosis status (whether they had received a psychiatric diagnosis, either before or after the DSH event) and the department from which they had received the psychiatric diagnosis. Another main outcome of the study was death by suicide. Cox regression models yielded hazard ratios (HRs) for suicide risk. Patterns were plotted using Kaplan-Meier survival curves. RESULTS: There were significant differences in all factors including demographic, health-related, socioeconomic, and survival variables among the groups that were classified according to psychiatric diagnosis status (P<.001). The group that did not receive a psychiatric diagnosis had the lowest survival rate (867/1064, 81.48%). Analysis drawn using different departments from where the individual had received a psychiatric diagnosis showed statistically significant differences in all features of interest (P<.001). The group that had received psychiatric diagnoses from the Department of Psychiatry had the highest survival rate (888/951, 93.4%). These findings were confirmed using the Kaplan-Meier survival curves (P<.001). The severity of DSH (HR 4.31, 95% CI 3.55-5.26) was the most significant contributor to suicide death, followed by psychiatric diagnosis status (HR 1.84, 95% CI 1.47-2.30). CONCLUSIONS: Receiving psychiatric assessment from a health care professional, especially a psychiatrist, reduces suicide death in individuals who had deliberately harmed themselves before. The key characteristics of individuals with DSH behavior who die by suicide are male sex, middle age, comorbid physical disabilities, and higher socioeconomic status.

Investigation of the Trajectory of Muscle and Body Mass as a Prognostic Factor in Patients With Colorectal Cancer: Longitudinal Cohort Study.

BACKGROUND: Skeletal muscle and BMI are essential prognostic factors for survival in colorectal cancer (CRC). However, there is a lack of understanding due to scarce studies on the continuous aspects of these variables. OBJECTIVE: This study aimed to evaluate the prognostic impact of the initial status and trajectories of muscle and BMI on overall survival (OS) and assess whether these 4 profiles within 1 year can represent the profiles 6 years later. METHODS: We analyzed 4056 newly diagnosed patients with CRC between 2010 to 2020. The volume of the muscle with 5-mm thickness at the third lumbar spine level was measured using a pretrained deep learning algorithm. The skeletal muscle volume index (SMVI) was defined as the muscle volume divided by the square of the height. The correlation between BMI status at the first, third, and sixth years of diagnosis was analyzed and assessed similarly for muscle profiles. Prognostic significances of baseline BMI and SMVI and their 1-year trajectories for OS were evaluated by restricted cubic spline analysis and survival analysis. Patients were categorized based on these 4 dimensions, and prognostic risks were predicted and demonstrated using heat maps. RESULTS: Trajectories of SMVI were categorized as decreased (812/4056, 20%), steady (2014/4056, 49.7%), or increased (1230/4056, 30.3%). Similarly, BMI trajectories were categorized as decreased (792/4056, 19.5%), steady (2253/4056, 55.5%), or increased (1011/4056, 24.9%). BMI and SMVI values in the first year after diagnosis showed a statistically significant correlation with those in the third and sixth years (P<.001). Restricted cubic spline analysis showed a nonlinear relationship between baseline BMI and SMVI change ratio and OS; BMI, in particular, showed a U-shaped correlation. According to survival analysis, increased BMI (hazard ratio [HR] 0.83; P=.02), high baseline SMVI (HR 0.82; P=.04), and obesity stage 1 (HR 0.80; P=.02) showed a favorable impact, whereas decreased SMVI trajectory (HR 1.31; P=.001), decreased BMI (HR 1.23; P=.02), and initial underweight (HR 1.38; P=.02) or obesity stages 2-3 (HR 1.79; P=.01) were negative prognostic factors for OS. Considered simultaneously, BMI >30 kg/m2 with a low SMVI at the time of diagnosis resulted in the highest mortality risk. We observed improved survival in patients with increased muscle mass without BMI loss compared to those with steady muscle mass and BMI. CONCLUSIONS: Profiles within 1 year of both BMI and muscle were surrogate indicators for predicting the later profiles. Continuous trajectories of body and muscle mass are independent prognostic factors of patients with CRC. An automatic algorithm provides a unique opportunity to conduct longitudinal evaluations of body compositions. Further studies to understand the complicated natural courses of muscularity and adiposity are necessary for clinical application.

Comorbidity Differences by Trajectory Groups as a Reference for Identifying Patients at Risk for Late Mortality in Childhood Cancer Survivors: Longitudinal National Cohort Study.

BACKGROUND: Childhood cancer has a high long-term morbidity and mortality rate. Five years after the initial cancer diagnosis, approximately two-thirds of childhood cancer survivors experience at least one late complication, with one-quarter experiencing severe, life-threatening complications. Chronic health conditions can impact survivors' life planning and daily activities, reducing their health-related quality of life. Comprehensive and longitudinal data are required for investigations of national claims data. OBJECTIVE: This study aimed to address clinical and health policy interventions and improved survival rates. A comprehensive categorization of the long-term morbidities associated with childhood cancer survivorship is required. We analyzed the trajectory groups associated with long-term mortality among childhood cancer survivors. METHODS: We collected data from a nationwide claims database of the entire Korean population. Between 2003 and 2007, patients diagnosed with and treated for cancer before the age of 20 years were included. With 8119 patients who survived >10 years, 3 trajectory groups were classified according to yearly changes in the number of diagnoses (the lowest in group 1 and the highest in group 3). RESULTS: The patterns of most comorbidities and survival rates differed significantly between the trajectory groups. Group 3 had a higher rate of mental and behavioral disorders, neoplasms, and blood organ diseases than the other two groups. Furthermore, there was a difference in the number of diagnoses by trajectory groups over the entire decade, and the disparity increased as the survival period increased. If a patient received more than four diagnoses, especially after the fourth year, the patient was likely to be assigned to group 3, which had the worst prognosis. Group 1 had the highest overall survival rate, and group 3 had the lowest (P<.001). Group 3 had the highest hazard ratio of 4.37 (95% CI 2.57-7.42; P<.001) in a multivariate analysis of late mortality. CONCLUSIONS: Our findings show that the pattern of comorbidities differed significantly among trajectory groups for late death, which could help physicians identify childhood cancer survivors at risk for late mortality. Patients with neoplasms, blood organ diseases, or mental and behavioral disorders should be identified as having an increased risk of late mortality. Furthermore, vigilance and prompt action are essential to mitigate the potential consequences of a child cancer survivor receiving four or more diagnoses within a year.

Stratifying non-small cell lung cancer patients using an inverse of the treatment decision rules: validation using electronic health records with application to an administrative database.

BACKGROUND: To validate a stratification method using an inverse of treatment decision rules that can classify non-small cell lung cancer (NSCLC) patients in real-world treatment records. METHODS: (1) To validate the index classifier against the TNM 7th edition, we analyzed electronic health records of NSCLC patients diagnosed from 2011 to 2015 in a tertiary referral hospital in Seoul, Korea. Predictive accuracy, stage-specific sensitivity, specificity, positive predictive value, negative predictive value, F1 score, and c-statistic were measured. (2) To apply the index classifier in an administrative database, we analyzed NSCLC patients in Korean National Health Insurance Database, 2002-2013. Differential survival rates among the classes were examined with the log-rank test, and class-specific survival rates were compared with the reference survival rates. RESULTS: (1) In the validation study (N = 1375), the overall accuracy was 93.8% (95% CI: 92.5-95.0%). Stage-specific c-statistic was the highest for stage I (0.97, 95% CI: 0.96-0.98) and the lowest for stage III (0.82, 95% CI: 0.77-0.87). (2) In the application study (N = 71,593), the index classifier showed a tendency for differentiating survival probabilities among classes. Compared to the reference TNM survival rates, the index classification under-estimated the survival probability for stages IA, IIIB, and IV, and over-estimated it for stages IIA and IIB. CONCLUSION: The inverse of the treatment decision rules has a potential to supplement a routinely collected database with information encoded in the treatment decision rules to classify NSCLC patients. It requires further validation and replication in multiple clinical settings.