Improving drug-target affinity prediction via feature fusion and knowledge distillation.

Rapid and accurate prediction of drug-target affinity can accelerate and improve the drug discovery process. Recent studies show that deep learning models may have the potential to provide fast and accurate drug-target affinity prediction. However, the existing deep learning models still have their own disadvantages that make it difficult to complete the task satisfactorily. Complex-based models rely heavily on the time-consuming docking process, and complex-free models lacks interpretability. In this study, we introduced a novel knowledge-distillation insights drug-target affinity prediction model with feature fusion inputs to make fast, accurate and explainable predictions. We benchmarked the model on public affinity prediction and virtual screening dataset. The results show that it outperformed previous state-of-the-art models and achieved comparable performance to previous complex-based models. Finally, we study the interpretability of this model through visualization and find it can provide meaningful explanations for pairwise interaction. We believe this model can further improve the drug-target affinity prediction for its higher accuracy and reliable interpretability.

Artificial intelligence assists identification and pathologic classification of glomerular lesions in patients with diabetic nephropathy.

BACKGROUND: Glomerular lesions are the main injuries of diabetic nephropathy (DN) and are used as a crucial index for pathologic classification. Manual quantification of these morphologic features currently used is semi-quantitative and time-consuming. Automatically quantifying glomerular morphologic features is urgently needed. METHODS: A series of convolutional neural networks (CNN) were designed to identify and classify glomerular morphologic features in DN patients. Associations of these digital features with pathologic classification and prognosis were further analyzed. RESULTS: Our CNN-based model achieved a 0.928 F1-score for global glomerulosclerosis and 0.953 F1-score for Kimmelstiel-Wilson lesion, further obtained a dice of 0.870 for the mesangial area and F1-score beyond 0.839 for three glomerular intrinsic cells. As the pathologic classes increased, mesangial cell numbers and mesangial area increased, and podocyte numbers decreased (p for all < 0.001), while endothelial cell numbers remained stable (p = 0.431). Glomeruli with Kimmelstiel-Wilson lesion showed more severe podocyte deletion compared to those without (p < 0.001). Furthermore, CNN-based classifications showed moderate agreement with pathologists-based classification, the kappa value between the CNN model 3 and pathologists reached 0.624 (ranging from 0.529 to 0.688, p < 0.001). Notably, CNN-based classifications obtained equivalent performance to pathologists-based classifications on predicting baseline and long-term renal function. CONCLUSION: Our CNN-based model is promising in assisting the identification and pathologic classification of glomerular lesions in DN patients.

Ocular biomarkers of cognitive decline based on deep-learning retinal vessel segmentation.

BACKGROUND: The current literature shows a strong relationship between retinal neuronal and vascular alterations in dementia. The purpose of the study was to use NFN+ deep learning models to analyze retinal vessel characteristics for cognitive impairment (CI) recognition. METHODS: We included 908 participants from a community-based cohort followed for over 15 years (the prospective KaiLuan Study) who underwent brain magnetic resonance imaging (MRI) and fundus photography between 2021 and 2022. The cohort consisted of both cognitively healthy individuals (N = 417) and those with cognitive impairment (N = 491). We employed the NFN+ deep learning framework for retinal vessel segmentation and measurement. Associations between Retinal microvascular parameters (RMPs: central retinal arteriolar / venular equivalents, arteriole to venular ratio, fractal dimension) and CI were assessed by Pearson correlation. P < 0.05 was considered statistically significant. The correlation between the CI and RMPs were explored, then the correlation coefficients between CI and RMPs were analyzed. Random Forest nonlinear classification model was used to predict whether one having cognitive decline or not. The assessment criterion was the AUC value derived from the working characteristic curve. RESULTS: The fractal dimension (FD) and global vein width were significantly correlated with the CI (P < 0.05). Age (0.193), BMI (0.154), global vein width (0.106), retinal vessel FD (0.099), and CRAE (0.098) were the variables in this model that were ranked in order of feature importance. The AUC values of the model were 0.799. CONCLUSIONS: Establishment of a predictive model based on the extraction of vascular features from fundus images has a high recognizability and predictive power for cognitive function and can be used as a screening method for CI.

Comparison of Changes in Retinal Vascular Parameters and Density in Patients with Moyamoya Disease: A Retrospective Study.

INTRODUCTION: This study aimed to compare retinal vascular parameters and density in patients with moyamoya disease using the optical coherence tomography angiography. METHODS: This clinical trial totally enrolls 78 eyes from 39 participants, and all these patients with moyamoya disease (N = 13) are set as experimental group and participants with health who matched with age and gender are considered as the control group (N = 26). Then all these participants receive optical coherence tomography angiography detection. Participants' general data are collected and analyzed. Skeleton density (SD) value, vessel density (VD) value, fractal dimension (FD) value, vessel diameter index (VDI) value, foveal avascular zone (FAZ) value are analyzed. RESULTS: A total of 39 participants are included in this study. The SD value in the experimental group was significantly lower than that in control group (0.175 [0.166, 0.181] vs. 0.184 [0.175, 0.188], p = 0.017). Similarly, the VD value in the experimental group was significantly lower than that in the control group (0.333 [0.320, 0.350] vs. 0.354 [0.337, 0.364], p = 0.024). Additionally, the FD value in the experimental group was significantly lower than that in the control group (2.088 [2.083, 2.094] vs. 2.096 [2.090, 2.101], p = 0.022). As for the VDI and FAZ, VDI and FAZ values in the experimental group were lower than those in the control group, there was no significant difference in VDI and FAZ values between the two groups. CONCLUSIONS: Our study, using non-invasive and rapid OCTA imaging, confirmed decreased retinal vascular parameters and density in patients with moyamoya disease.

Advancing the application of the analytical renal pathology system in allograft IgA nephropathy patients.

BACKGROUND: The analytical renal pathology system (ARPS) based on convolutional neural networks has been used successfully in native IgA nephropathy (IgAN) patients. Considering the similarity of pathologic features, we aim to evaluate the performance of the ARPS in allograft IgAN patients and broaden its implementation. METHODS: Biopsy-proven allograft IgAN patients from two different centers were enrolled for internal and external validation. We implemented the ARPS to identify glomerular lesions and intrinsic glomerular cells, and then evaluated its performance. Consistency between the ARPS and pathologists was assessed using intraclass correlation coefficients. The association of digital pathological features with clinical and pathological data was measured. Kaplan-Meier survival curve and cox proportional hazards model were applied to investigate prognosis prediction. RESULTS: A total of 56 biopsy-proven allograft IgAN patients from the internal center and 17 biopsy-proven allograft IgAN patients from the external center were enrolled in this study. The ARPS was successfully applied to identify the glomerular lesions (F1-score, 0.696-0.959) and quantify intrinsic glomerular cells (F1-score, 0.888-0.968) in allograft IgAN patients rapidly and precisely. Furthermore, the mesangial hypercellularity score was positively correlated with all mesangial metrics provided by ARPS [Spearman's correlation coefficient (r), 0.439-0.472, and all p values < 0.001]. Besides, a higher allograft survival was noticed among patients in the high-level groups of the maximum and ratio of endothelial cells, as well as the maximum and density of podocytes. CONCLUSION: We propose that the ARPS could be implemented in future clinical practice with outstanding capability.

An effective self-supervised framework for learning expressive molecular global representations to drug discovery.

How to produce expressive molecular representations is a fundamental challenge in artificial intelligence-driven drug discovery. Graph neural network (GNN) has emerged as a powerful technique for modeling molecular data. However, previous supervised approaches usually suffer from the scarcity of labeled data and poor generalization capability. Here, we propose a novel molecular pre-training graph-based deep learning framework, named MPG, that learns molecular representations from large-scale unlabeled molecules. In MPG, we proposed a powerful GNN for modelling molecular graph named MolGNet, and designed an effective self-supervised strategy for pre-training the model at both the node and graph-level. After pre-training on 11 million unlabeled molecules, we revealed that MolGNet can capture valuable chemical insights to produce interpretable representation. The pre-trained MolGNet can be fine-tuned with just one additional output layer to create state-of-the-art models for a wide range of drug discovery tasks, including molecular properties prediction, drug-drug interaction and drug-target interaction, on 14 benchmark datasets. The pre-trained MolGNet in MPG has the potential to become an advanced molecular encoder in the drug discovery pipeline.

Novel in-house knowledge-based automated planning system for lung cancer treated with intensity-modulated radiotherapy.

PURPOSE: The goal of this study was to propose a knowledge-based planning system which could automatically design plans for lung cancer patients treated with intensity-modulated radiotherapy (IMRT). METHODS AND MATERIALS: From May 2018 to June 2020, 612 IMRT treatment plans of lung cancer patients were retrospectively selected to construct a planning database. Knowledge-based planning (KBP) architecture named αDiar was proposed in this study. It consisted of two parts separated by a firewall. One was the in-hospital workstation, and the other was the search engine in the cloud. Based on our previous study, A­Net in the in-hospital workstation was used to generate predicted virtual dose images. A search engine including a three-dimensional convolutional neural network (3D CNN) was constructed to derive the feature vectors of dose images. By comparing the similarity of the features between virtual dose images and the clinical dose images in the database, the most similar feature was found. The optimization parameters (OPs) of the treatment plan corresponding to the most similar feature were assigned to the new plan, and the design of a new treatment plan was automatically completed. After αDiar was developed, we performed two studies. The first retrospective study was conducted to validate whether this architecture was qualified for clinical practice and involved 96 patients. The second comparative study was performed to investigate whether αDiar could assist dosimetrists in improving the quality of planning for the patients. Two dosimetrists were involved and designed plans for only one trial with and without αDiar; 26 patients were involved in this study. RESULTS: The first study showed that about 54% (52/96) of the automatically generated plans would achieve the dosimetric constraints of the Radiation Therapy Oncology Group (RTOG) and about 93% (89/96) of the automatically generated plans would achieve the dosimetric constraints of the National Comprehensive Cancer Network (NCCN). The second study showed that the quality of treatment planning designed by junior dosimetrists was improved with the help of αDiar. CONCLUSIONS: Our results showed that αDiar was an effective tool to improve planning quality. Over half of the patients' plans could be designed automatically. For the remaining patients, although the automatically designed plans did not fully meet the clinical requirements, their quality was also better than that of manual plans.

Identification of atrial fibrillation phenotypes at low risk of stroke in patients with CHA2DS2-VASc ≥2: Insight from the China-AF study.

OBJECTIVE: Patients with atrial fibrillation (AF) are highly heterogeneous, and current risk stratification scores are only modestly good at predicting an individual's stroke risk. We aim to identify distinct AF clinical phenotypes with cluster analysis to optimize stroke prevention practices. METHODS: From the prospective Chinese Atrial Fibrillation Registry cohort study, we included 4337 AF patients with CHA2 DS2 -VASc≥2 for males and 3 for females who were not treated with oral anticoagulation. We randomly split the patients into derivation and validation sets by a ratio of 7:3. In the derivation set, we used outcome-driven patient clustering with metric learning to group patients into clusters with different risk levels of ischemic stroke and systemic embolism, and identify clusters of patients with low risks. Then we tested the results in the validation set, using the clustering rules generated from the derivation set. Finally, the survival decision tree was applied as a sensitivity analysis to confirm the results. RESULTS: Up to the follow-up of 1 year, 140 thromboembolic events (ischemic stroke or systemic embolism) occurred. After supervised metric learning from six variables involved in CHA2 DS2 -VASc scheme, we identified a cluster of patients (255/3035, 8.4%) at an annual thromboembolism risk of 0.8% in the derivation set. None of the patients in the low-risk cluster had prior thromboembolism, heart failure, diabetes, or age older than 70 years. After applying the regularities from metric learning on the validation set, we also identified a cluster of patients (137/1302, 10.5%) with an incident thromboembolism rate of 0.7%. Sensitivity analysis based on the survival decision tree approach selected a subgroup of patients with the same phenotypes as the metric-learning algorithm. CONCLUSIONS: Cluster analysis identified a distinct clinical phenotype at low risk of stroke among high-risk [CHA2 DS2 -VASc≥2 (3 for females)] patients with AF. The use of the novel analytic approach has the potential to prevent a subset of AF patients from unnecessary anticoagulation and avoid the associated risk of major bleeding.

Widely targeted quantitative lipidomics and prognostic model reveal plasma lipid predictors for nasopharyngeal carcinoma.

BACKGROUND: Dysregulation of lipid metabolism is closely associated with cancer progression. The study aimed to establish a prognostic model to predict distant metastasis-free survival (DMFS) in patients with nasopharyngeal carcinoma (NPC), based on lipidomics. METHODS: The plasma lipid profiles of 179 patients with locoregionally advanced NPC (LANPC) were measured and quantified using widely targeted quantitative lipidomics. Then, patients were randomly split into the training (125 patients, 69.8%) and validation (54 patients, 30.2%) sets. To identify distant metastasis-associated lipids, univariate Cox regression was applied to the training set (P < 0.05). A deep survival method called DeepSurv was employed to develop a proposed model based on significant lipid species (P < 0.01) and clinical biomarkers to predict DMFS. Concordance index and receiver operating curve analyses were performed to assess model effectiveness. The study also explored the potential role of lipid alterations in the prognosis of NPC. RESULTS: Forty lipids were recognized as distant metastasis-associated (P < 0.05) by univariate Cox regression. The concordance indices of the proposed model were 0.764 (95% confidence interval (CI), 0.682-0.846) and 0.760 (95% CI, 0.649-0.871) in the training and validation sets, respectively. High-risk patients had poorer 5-year DMFS compared with low-risk patients (Hazard ratio, 26.18; 95% CI, 3.52-194.80; P < 0.0001). Moreover, the six lipids were significantly correlated with immunity- and inflammation-associated biomarkers and were mainly enriched in metabolic pathways. CONCLUSIONS: Widely targeted quantitative lipidomics reveals plasma lipid predictors for LANPC, the prognostic model based on that demonstrated superior performance in predicting metastasis in LANPC patients.

DEVELOPMENT AND VALIDATION OF AN EXPLAINABLE ARTIFICIAL INTELLIGENCE FRAMEWORK FOR MACULAR DISEASE DIAGNOSIS BASED ON OPTICAL COHERENCE TOMOGRAPHY IMAGES.

PURPOSE: To develop and validate an artificial intelligence framework for identifying multiple retinal lesions at image level and performing an explainable macular disease diagnosis at eye level in optical coherence tomography images. METHODS: A total of 26,815 optical coherence tomography images were collected from 865 eyes, and 9 retinal lesions and 3 macular diseases were labeled by ophthalmologists, including diabetic macular edema and dry/wet age-related macular degeneration. We applied deep learning to classify retinal lesions at image level and random forests to achieve an explainable disease diagnosis at eye level. The performance of the integrated two-stage framework was evaluated and compared with human experts. RESULTS: On testing data set of 2,480 optical coherence tomography images from 80 eyes, the deep learning model achieved an average area under curve of 0.978 (95% confidence interval, 0.971-0.983) for lesion classification. In addition, random forests performed accurate disease diagnosis with a 0% error rate, which achieved the same accuracy as one of the human experts and was better than the other three experts. It also revealed that the detection of specific lesions in the center of macular region had more contribution to macular disease diagnosis. CONCLUSION: The integrated method achieved high accuracy and interpretability in retinal lesion classification and macular disease diagnosis in optical coherence tomography images and could have the potential to facilitate the clinical diagnosis.

Development and quantitative assessment of deep learning-based image enhancement for optical coherence tomography.

PURPOSE: To develop a deep learning-based framework to improve the image quality of optical coherence tomography (OCT) and evaluate its image enhancement effect with the traditional image averaging method from a clinical perspective. METHODS: 359 normal eyes and 456 eyes with various retinal conditions were included. A deep learning framework with high-resolution representation was developed to achieve image quality enhancement for OCT images. The quantitative comparisons, including expert subjective scores from ophthalmologists and three objective metrics of image quality (structural similarity index measure (SSIM), peak signal-to-noise ratio (PSNR) and contrast-to-noise ratio (CNR)), were performed between deep learning method and traditional image averaging. RESULTS: With the increase of frame count from 1 to 20, our deep learning method always obtained higher SSIM and PSNR values than the image averaging method while importing the same number of frames. When we selected 5 frames as inputs, the local objective assessment with CNR illustrated that the deep learning method had more obvious tissue contrast enhancement than averaging method. The subjective scores of image quality were all highest in our deep learning method, both for normal retinal structure and various retinal lesions. All the objective and subjective indicators had significant statistical differences (P < 0.05). CONCLUSION: Compared to traditional image averaging methods, our proposed deep learning enhancement framework can achieve a reasonable trade-off between image quality and scanning times, reducing the number of repeated scans.

Early prediction of moderate-to-severe condition of inhalation-induced acute respiratory distress syndrome via interpretable machine learning.

BACKGROUND: Several studies have investigated the correlation between physiological parameters and the risk of acute respiratory distress syndrome (ARDS), in addition, etiology-associated heterogeneity in ARDS has become an emerging topic quite recently; however, the intersection between the two, which is early prediction of target conditions in etiology-specific ARDS, has not been well-studied. We aimed to develop and validate a machine-learning model for the early prediction of moderate-to-severe condition of inhalation-induced ARDS. METHODS: Clinical expertise was applied with data-driven analysis. Using data from electronic intensive care units (retrospective derivation cohort) and the three most accessible vital signs (i.e. heart rate, temperature, and respiratory rate) together with feature engineering, we applied a random forest approach during the time window of 90 h that ended 6 h prior to the onset of moderate-to-severe respiratory failure (the ratio of partial pressure of arterial oxygen to fraction of inspired oxygen ≤ 200 mmHg). RESULTS: The trained random forest classifier was validated using two independent validation cohorts, with an area under the curve of 0.9127 (95% confidence interval 0.8713-0.9542) and 0.9026 (95% confidence interval 0.8075-1), respectively. A Stable and Interpretable RUle Set (SIRUS) was used to extract rules from the RF to provide guidelines for clinicians. We identified several predictive factors, including resp_96h_6h_min < 9, resp_96h_6h_mean ≥ 16.1, HR_96h_6h_mean ≥ 102, and temp_96h_6h_max > 100, that could be used for predicting inhalation-induced ARDS (moderate-to-severe condition) 6 h prior to onset in critical care units. ('xxx_96h_6h_min/mean/max': the minimum/mean/maximum values of the xxx vital sign collected during a 90 h time window beginning 96 h prior to the onset of ARDS and ending 6 h prior to the onset from every recorded blood gas test). CONCLUSIONS: This newly established random forest­based interpretable model shows good predictive ability for moderate-to-severe inhalation-induced ARDS and may assist clinicians in decision-making, as well as facilitate the enrolment of patients in prevention programmes to improve their outcomes.

Machine Learning for Prediction and Risk Stratification of Lupus Nephritis Renal Flare.

BACKGROUND: Renal flare of lupus nephritis (LN) is strongly associated with poor kidney outcomes, and predicting renal flare and stratifying its risk are important for clinical decision-making and individualized management to reduce LN flare. METHODS: We randomly divided 1,694 patients with biopsy-proven LN, who had achieved remission after treatment, into a derivation cohort (n = 1,186) and an internal validation cohort (n = 508), at a ratio of 7:3. The risk of renal flare 5 years after remission was predicted using an eXtreme Gradient Boosting (XGBoost) method model, developed from 59 variables, including demographic, clinical, immunological, pathological, and therapeutic characteristics. A simplified risk score prediction model (SRSPM) was developed from important variables selected by XGBoost model using stepwise Cox regression for practical convenience. RESULTS: The 5-year relapse rates were 39.5% and 38.2% in the derivation and internal validation cohorts, respectively. Both the XGBoost model and the SRSPM had good predictive performance, with a C-index of 0.819 (95% confidence interval [CI]: 0.774-0.857) and 0.746 (95% CI: 0.697-0.795), respectively, in the validation cohort. The SRSPM comprised 6 variables, including partial remission and endocapillary hypercellularity at baseline, age, serum Alb, anti-dsDNA, and serum complement C3 at the point of remission. Using Kaplan-Meier analysis, the SRSPM identified significant risk stratification for renal flares (p < 0.001). CONCLUSIONS: Renal flare of LN can be readily predicted using the XGBoost model and the SRSPM, and the SRSPM can also stratify flare risk. Both models are useful for clinical decision-making and individualized management in LN.

Deep learning shows good reliability for automatic segmentation and volume measurement of brain hemorrhage, intraventricular extension, and peripheral edema.

OBJECTIVES: To evaluate for the first time the performance of a deep learning method based on no-new-Net for fully automated segmentation and volumetric measurements of intracerebral hemorrhage (ICH), intraventricular extension of intracerebral hemorrhage (IVH), and perihematomal edema (PHE) in primary ICH on CT. METHODS: Three hundred and eighty primary ICH patients who underwent CT at hospital arrival were divided into a training cohort (n = 300) and a validation cohort (n = 80). An independent cohort with 80 patients was used for testing. Ground truth (segmentation masks) was manually generated by radiologists. Model performance on lesion segmentation and volumetric measurement of ICH, IVH, and PHE were evaluated by comparing the model results with the segmentations performed by radiologists. RESULTS: In the test cohort, the Dice scores of lesion segmentation were 0.92, 0.79, and 0.71 for ICH, IVH, and PHE, respectively. The sensitivities were 0.93 for ICH, 0.88 for IVH, and 0.81 for PHE. The positive predictive values were 0.92, 0.76, and 0.69 for ICH, IVH, and PHE, respectively. Excellent concordance (concordance correlation coefficients [CCCs] ≥ 0.98) of ICH and IVH and good concordance of PHE (CCCs ≥ 0.92) were demonstrated between manually and automatically measured volumes. The model took approximately 15 s to provide automatic segmentation and volume analysis for each patient. CONCLUSION: Our model demonstrates good reliability for automatic segmentation and volume measurement of ICH, IVH, and PHE in primary ICH, which can be useful to reduce the effort and time of doctors to calculate volumes of ICH, IVH, and PHE. KEY POINTS: • Deep learning algorithms can provide automatic and reliable assessment of intracerebral hemorrhage, intraventricular hemorrhage, and perihematomal edema on CT. • Non-contrast CT-based deep learning method can be helpful to provide efficient and accurate measurements of ICH, IVH, and PHE in primary ICH patients, thereby reducing the effort and time of doctors to segment and calculate volumes of ICH, IVH, and PHE in primary ICH patients.

Identification of glomerular lesions and intrinsic glomerular cell types in kidney diseases via deep learning.

Identification of glomerular lesions and structures is a key point for pathological diagnosis, treatment instructions, and prognosis evaluation in kidney diseases. These time-consuming tasks require a more accurate and reproducible quantitative analysis method. We established derivation and validation cohorts composed of 400 Chinese patients with immunoglobulin A nephropathy (IgAN) retrospectively. Deep convolutional neural networks and biomedical image processing algorithms were implemented to locate glomeruli, identify glomerular lesions (global and segmental glomerular sclerosis, crescent, and none of the above), identify and quantify different intrinsic glomerular cells, and assess a network-based mesangial hypercellularity score in periodic acid-Schiff (PAS)-stained slides. Our framework achieved 93.1% average precision and 94.9% average recall for location of glomeruli, and a total Cohen's kappa of 0.912 [95% confidence interval (CI), 0.892-0.932] for glomerular lesion classification. The evaluation of global, segmental glomerular sclerosis, and crescents achieved Cohen's kappa values of 1.0, 0.776, 0.861, and 95% CI of (1.0, 1.0), (0.727, 0.825), (0.824, 0.898), respectively. The well-designed neural network can identify three kinds of intrinsic glomerular cells with 92.2% accuracy, surpassing the about 5-11% average accuracy of junior pathologists. Statistical interpretation shows that there was a significant difference (P value < 0.0001) between this analytic renal pathology system (ARPS) and four junior pathologists for identifying mesangial and endothelial cells, while that for podocytes was similar, with P value = 0.0602. In addition, this study indicated that the ratio of mesangial cells, endothelial cells, and podocytes within glomeruli from IgAN was 0.41:0.36:0.23, and the performance of mesangial score assessment reached a Cohen's kappa of 0.42 and 95% CI (0.18, 0.69). The proposed computer-aided diagnosis system has feasibility for quantitative analysis and auxiliary recognition of glomerular pathological features. © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland.

An integrated framework for modelling quantitative effects of entry restrictions and travel quarantine on importation risk of COVID-19.

OBJECTIVE: As the potential spread of COVID-19 sparked by imported cases from overseas will pose continuous challenges, it is essential to estimate the effects of control measures on reducing the importation risk of COVID-19. Our objective is to provide a framework of methodology for quantifying the combined effects of entry restrictions and travel quarantine on managing the importation risk of COVID-19 and other pandemics by leveraging different sets of parameters. METHODS: Three major categories of control measures on controlling importation risk were parameterized and modelled by the framework: 1) entry restrictions, 2) travel quarantine, and 3) domestic containment measures. Integrating the parameterized intensity of control measures, a modified SEIR model was developed to simulate the case importation and local epidemic under different scenarios of global epidemic dynamics. A web-based tool was also provided to enable interactive visualization of epidemic simulation. RESULTS: The simulated number of case importation and local spread modelled by the proposed framework of methods fitted well to the historical epidemic curve of China and Singapore. Based on the simulation results, the total numbers of infected cases when reducing 30% of visitor arrivals would be 88·4 (IQR 87·5-89·6) and 58·8 (IQR 58·3-59·5) times more than those when reducing 99% of visitor arrivals in mainland China and Singapore respectively, assuming actual time-varying Rt and travel quarantine policy. If the number of global daily new infections reached 100,000, 85%-91% of inbound travels should be reduced to keep the daily new infected number below 100 for a country with a similar travel volume as Singapore (daily 52,000 tourist arrivals in 2019). Whereas if the number was lower than 10,000, the daily new infected case would be less than 100 even with no entry restrictions. DISCUSSIONS: We proposed a framework that first estimated the intensity of travel restrictions and local containment measures for countries since the first overseas imported case. Our approach then quantified the combined effects of entry restrictions and travel quarantine using a modified SEIR model to simulate the potential epidemic spread under hypothetical intensities of these control measures. We also developed a web-based system that enables interactive simulation, which could serve as a valuable tool for health system administrators to assess policy effects on managing the importation risk. By leveraging different sets of parameters, it could adapt to any specific country and specific type of epidemic. CONCLUSIONS: This framework has provided a valuable tool to parameterize the intensity of control measures, simulate both the case importation and local epidemic, and quantify the combined effects of entry restrictions and travel quarantine on managing the importation risk.

Optical coherence tomography-based short-term effect prediction of anti-vascular endothelial growth factor treatment in neovascular age-related macular degeneration using sensitive structure guided network.

PURPOSE: To predict short-term anti-vascular endothelial growth factor (anti-VEGF) treatment responder/non-responder for neovascular age-related macular degeneration (nAMD) patients based on optical coherence tomography (OCT) images. METHODS: A total of 4944 OCT scans from 206 patients with nAMD were involved to develop and evaluate a responder/non-responder prediction method for the short-term effect of anti-VEGF therapy. A deep learning architecture named sensitive structure guided network (SSG-Net) was proposed to make the prediction leveraging a sensitive structure guidance module trained from pre- and post-treatment images. To verify its clinical efficiency, other 2 deep learning methods and 4 experienced ophthalmologists were involved to evaluate the performance of the developed model. RESULTS: For the testing dataset, SSG-Net could predict the response by an accuracy of 84.6% and an area under the receiver curve (AUC) of 0.83, with a sensitivity of 0.692 and specificity of 1. In contrast, the 2 compared deep learning methods achieved an accuracy of 65.4% with a sensitivity of 0.461 and specificity of 0.846, and an accuracy of 73.1% with a sensitivity of 0.692 and specificity of 0.846, respectively. The predicted accuracy for 4 experienced ophthalmologists was 53.8 to 76.9%, with sensitivity of 0.538 to 0.923 and specificity of 0.385 to 0.846, respectively. CONCLUSION: Our proposed SSG-Net shows effective prediction on the short-term efficacy of anti-VEGF treatment for nAMD patients. This technique could potentially help clinicians explain the necessity of anti-VEGF treatment to the potential responder and avoid unnecessary treatment for the non-responder.

Differential impact of non-pharmaceutical public health interventions on COVID-19 epidemics in the United States.

BACKGROUND: The widespread pandemic of novel coronavirus disease 2019 (COVID-19) poses an unprecedented global health crisis. In the United States (US), different state governments have adopted various combinations of non-pharmaceutical public health interventions (NPIs), such as non-essential business closures and gathering bans, to mitigate the epidemic from February to April, 2020. Quantitative assessment on the effectiveness of NPIs is greatly needed to assist in guiding individualized decision making for adjustment of interventions in the US and around the world. However, the impacts of these approaches remain uncertain. METHODS: Based on the reported cases, the effective reproduction number (Rt) of COVID-19 epidemic for 50 states in the US was estimated. Measurements on the effectiveness of nine different NPIs were conducted by assessing risk ratios (RRs) between Rt and NPIs through a generalized linear model (GLM). RESULTS: Different NPIs were found to have led to different levels of reduction in Rt. Stay-at-home contributed approximately 51% (95% CI 46-57%), wearing (face) masks 29% (15-42%), gathering ban (more than 10 people) 19% (14-24%), non-essential business closure 16% (10-21%), declaration of emergency 13% (8-17%), interstate travel restriction 11% (5-16%), school closure 10% (7-14%), initial business closure 10% (6-14%), and gathering ban (more than 50 people) 7% (2-11%). CONCLUSIONS: This retrospective assessment of NPIs on Rt has shown that NPIs played critical roles on epidemic control in the US in the past several months. The quantitative results could guide individualized decision making for future adjustment of NPIs in the US and other countries for COVID-19 and other similar infectious diseases.

Characteristics of Online Health Care Services From China's Largest Online Medical Platform: Cross-sectional Survey Study.

BACKGROUND: Internet hospitals in China are in great demand due to limited and unevenly distributed health care resources, lack of family doctors, increased burdens of chronic diseases, and rapid growth of the aged population. The COVID-19 epidemic catalyzed the expansion of online health care services. In recent years, internet hospitals have been rapidly developed. Ping An Good Doctor is the largest, national online medical entry point in China and is a widely used platform providing online health care services. OBJECTIVE: This study aims to give a comprehensive description of the characteristics of the online consultations and inquisitions in Ping An Good Doctor. The analyses tried to answer the following questions: (1) What are the characteristics of the consultations in Ping An Good Doctor in terms of department and disease profiles? (2) Who uses the online health services most frequently? and (3) How is the user experience of the online consultations of Ping An Good Doctor? METHODS: A total of 35.3 million consultations and inquisitions over the course of 1 year were analyzed with respect to the distributions of departments and diseases, user profiles, and consulting behaviors. RESULTS: The geographical distribution of the usage of Ping An Good Doctor showed that Shandong (18.4%), Yunnan (15.6%), Shaanxi (7.2%), and Guangdong (5.5%) were the provinces that used it the most; they accounted for 46.6% of the total consultations and inquisitions. In terms of department distribution, we found that gynecology and obstetrics (19.2%), dermatology (17.0%), and pediatrics (14.4%) were the top three departments in Ping An Good Doctor. The disease distribution analysis showed that, except for nondisease-specific consultations, acute upper respiratory infection (AURI) (4.1%), pregnancy (2.8%), and dermatitis (2.4%) were the most frequently consulted diseases. In terms of user profiles, females (60.4%) from 19 to 35 years of age were most likely to seek consultations online, in general. The user behavior analyses showed that the peak times of day for online consultations occurred at 10 AM, 3 PM, and 9 PM. Regarding user experience, 93.0% of users gave full marks following their consultations. For some disease-related health problems, such as AURI, dermatitis, and eczema, the feedback scores were above average. CONCLUSIONS: The prevalence of internet hospitals, such as Ping An Good Doctor, illustrated the great demand for online health care services that can go beyond geographical limitations. Our analyses showed that nondisease-specific issues and moderate health problems were much more frequently consulted about than severe clinical conditions. This indicated that internet hospitals played the role of the family doctor, which helped to relieve the stress placed on offline hospitals and facilitated people's lives. In addition, good user experiences, especially regarding disease-related inquisitions, suggested that online health services can help solve health problems. With support from the government and acceptance by the public, online health care services could develop at a fast pace and greatly benefit people's daily lives.

Effective Treatment Recommendations for Type 2 Diabetes Management Using Reinforcement Learning: Treatment Recommendation Model Development and Validation.

BACKGROUND: Type 2 diabetes mellitus (T2DM) and its related complications represent a growing economic burden for many countries and health systems. Diabetes complications can be prevented through better disease control, but there is a large gap between the recommended treatment and the treatment that patients actually receive. The treatment of T2DM can be challenging because of different comprehensive therapeutic targets and individual variability of the patients, leading to the need for precise, personalized treatment. OBJECTIVE: The aim of this study was to develop treatment recommendation models for T2DM based on deep reinforcement learning. A retrospective analysis was then performed to evaluate the reliability and effectiveness of the models. METHODS: The data used in our study were collected from the Singapore Health Services Diabetes Registry, encompassing 189,520 patients with T2DM, including 6,407,958 outpatient visits from 2013 to 2018. The treatment recommendation model was built based on 80% of the dataset and its effectiveness was evaluated with the remaining 20% of data. Three treatment recommendation models were developed for antiglycemic, antihypertensive, and lipid-lowering treatments by combining a knowledge-driven model and a data-driven model. The knowledge-driven model, based on clinical guidelines and expert experiences, was first applied to select the candidate medications. The data-driven model, based on deep reinforcement learning, was used to rank the candidates according to the expected clinical outcomes. To evaluate the models, short-term outcomes were compared between the model-concordant treatments and the model-nonconcordant treatments with confounder adjustment by stratification, propensity score weighting, and multivariate regression. For long-term outcomes, model-concordant rates were included as independent variables to evaluate if the combined antiglycemic, antihypertensive, and lipid-lowering treatments had a positive impact on reduction of long-term complication occurrence or death at the patient level via multivariate logistic regression. RESULTS: The test data consisted of 36,993 patients for evaluating the effectiveness of the three treatment recommendation models. In 43.3% of patient visits, the antiglycemic medications recommended by the model were concordant with the actual prescriptions of the physicians. The concordant rates for antihypertensive medications and lipid-lowering medications were 51.3% and 58.9%, respectively. The evaluation results also showed that model-concordant treatments were associated with better glycemic control (odds ratio [OR] 1.73, 95% CI 1.69-1.76), blood pressure control (OR 1.26, 95% CI, 1.23-1.29), and blood lipids control (OR 1.28, 95% CI 1.22-1.35). We also found that patients with more model-concordant treatments were associated with a lower risk of diabetes complications (including 3 macrovascular and 2 microvascular complications) and death, suggesting that the models have the potential of achieving better outcomes in the long term. CONCLUSIONS: Comprehensive management by combining knowledge-driven and data-driven models has good potential to help physicians improve the clinical outcomes of patients with T2DM; achieving good control on blood glucose, blood pressure, and blood lipids; and reducing the risk of diabetes complications in the long term.