Deep reinforcement learning identifies personalized intermittent androgen deprivation therapy for prostate cancer.

The evolution of drug resistance leads to treatment failure and tumor progression. Intermittent androgen deprivation therapy (IADT) helps responsive cancer cells compete with resistant cancer cells in intratumoral competition. However, conventional IADT is population-based, ignoring the heterogeneity of patients and cancer. Additionally, existing IADT relies on pre-determined thresholds of prostate-specific antigen to pause and resume treatment, which is not optimized for individual patients. To address these challenges, we framed a data-driven method in two steps. First, we developed a time-varied, mixed-effect and generative Lotka-Volterra (tM-GLV) model to account for the heterogeneity of the evolution mechanism and the pharmacokinetics of two ADT drugs Cyproterone acetate and Leuprolide acetate for individual patients. Then, we proposed a reinforcement-learning-enabled individualized IADT framework, namely, I$^{2}$ADT, to learn the patient-specific tumor dynamics and derive the optimal drug administration policy. Experiments with clinical trial data demonstrated that the proposed I$^{2}$ADT can significantly prolong the time to progression of prostate cancer patients with reduced cumulative drug dosage. We further validated the efficacy of the proposed methods with a recent pilot clinical trial data. Moreover, the adaptability of I$^{2}$ADT makes it a promising tool for other cancers with the availability of clinical data, where treatment regimens might need to be individualized based on patient characteristics and disease dynamics. Our research elucidates the application of deep reinforcement learning to identify personalized adaptive cancer therapy.

Lanthanide Inorganic Nanoparticles Enhance Semiconducting Polymer Nanoparticles Afterglow Luminescence for In Vivo Afterglow/Magnetic Resonance Imaging.

Dual/multimodal imaging strategies are increasingly recognized for their potential to provide comprehensive diagnostic insights in cancer imaging by harnessing complementary data. This study presents an innovative probe that capitalizes on the synergistic benefits of afterglow luminescence and magnetic resonance imaging (MRI), effectively eliminating autofluorescence interference and delivering a superior signal-to-noise ratio. Additionally, it facilitates deep tissue penetration and enables noninvasive imaging. Despite the advantages, only a limited number of probes have demonstrated the capability to simultaneously enhance afterglow luminescence and achieve high-resolution MRI and afterglow imaging. Herein, we introduce a cutting-edge imaging platform based on semiconducting polymer nanoparticles (PFODBT) integrated with NaYF4@NaGdF4 (Y@Gd@PFO-SPNs), which can directly amplify afterglow luminescence and generate MRI and afterglow signals in tumor tissues. The proposed mechanism involves lanthanide nanoparticles producing singlet oxygen (1O2) upon white light irradiation, which subsequently oxidizes PFODBT, thereby intensifying afterglow luminescence. This innovative platform paves the way for the development of high signal-to-background ratio imaging modalities, promising noninvasive diagnostics for cancer.

Deep learning identifies explainable reasoning paths of mechanism of action for drug repurposing from multilayer biological network.

The discovery and repurposing of drugs require a deep understanding of the mechanism of drug action (MODA). Existing computational methods mainly model MODA with the protein-protein interaction (PPI) network. However, the molecular interactions of drugs in the human body are far beyond PPIs. Additionally, the lack of interpretability of these models hinders their practicability. We propose an interpretable deep learning-based path-reasoning framework (iDPath) for drug discovery and repurposing by capturing MODA on by far the most comprehensive multilayer biological network consisting of the complex high-dimensional molecular interactions between genes, proteins and chemicals. Experiments show that iDPath outperforms state-of-the-art machine learning methods on a general drug repurposing task. Further investigations demonstrate that iDPath can identify explicit critical paths that are consistent with clinical evidence. To demonstrate the practical value of iDPath, we apply it to the identification of potential drugs for treating prostate cancer and hypertension. Results show that iDPath can discover new FDA-approved drugs. This research provides a novel interpretable artificial intelligence perspective on drug discovery.

The Impact of Expert Opinions in COVID-19 News Coverage on Vaccination During the Fifth Outbreak in Hong Kong: A Time Series Analysis.

The role of experts in news coverage has become increasingly prominent, but the evidence regarding the effectiveness of expert opinions in affecting public behavior remains mixed. This study seeks to examine the influence of expert opinions covered in the news on the public's response to public health crises. By adopting a macro-level framing perspective, we investigated how framing consistency, a macro-level concept indicating the agreement between expert opinions in news coverage and government policies or among peer experts, evolves over time and its temporal causal relationship with public behavior. Specifically, this study collected all press news coverage in Hong Kong over four months during the fifth outbreak, including 1,416 articles with 650 expert opinions, as well as the vaccination data that paralleled with this period. We constructed time series of expert opinions and vaccination behavior, and then conducted Vector Autoregressive (VAR) models with Granger causality analysis to examine how framing consistency of expert opinions in news coverage influenced vaccination. The results indicate that the consistent framing between expert opinions and government policies increased COVID-19 vaccination during the fifth outbreak in Hong Kong, while conflicting opinions responding to government policies had no significant effect on vaccination. Opinions among medical experts on COVID-19 issues also did not significantly impact vaccination. The implications for designing communication strategies and enhancing public behavioral support during public health crises are discussed.

Multi-omic analysis suggests tumor suppressor genes evolved specific promoter features to optimize cancer resistance.

Tumor suppressor genes (TSGs) exhibit distinct evolutionary features. We speculated that TSG promoters could have evolved specific features that facilitate their tumor-suppressing functions. We found that the promoter CpG dinucleotide frequencies of TSGs are significantly higher than that of non-cancer genes across vertebrate genomes, and positively correlated with gene expression across tissue types. The promoter CpG dinucleotide frequencies of all genes gradually increase with gene age, for which young TSGs have been subject to a stronger evolutionary pressure. Transcription-related features, namely chromatin accessibility, methylation and ZNF263-, SP1-, E2F4- and SP2-binding elements, are associated with gene expression. Moreover, higher promoter CpG dinucleotide frequencies and chromatin accessibility are positively associated with the ability of TSGs to resist downregulation during tumorigenesis. These results were successfully validated with independent datasets. In conclusion, TSGs evolved specific promoter features that optimized cancer resistance through achieving high expression in normal tissues and resistance to downregulation during tumorigenesis.

Lower risk of gout in sodium glucose cotransporter 2 (SGLT2) inhibitors versus dipeptidyl peptidase-4 (DPP4) inhibitors in type-2 diabetes.

OBJECTIVES: The effects of sodium-glucose cotransporter 2 inhibitors (SGLT2I) vs dipeptidyl peptidase-4 inhibitors (DPP4I) on the risk of new-onset gout remains unknown. This study aims to compare the effects of SGLT2I against DPP4I on gout risks. METHODS: This was a retrospective population-based cohort study of patients with type-2 diabetes mellitus treated with SGLT2I or DPP4I between 1 January 2015 and 31 December 2020 in Hong Kong. The study outcomes are new-onset gout and all-cause mortality. Propensity score matching (1:1 ratio) between SGLT2I and DPP4I was performed. Univariable and multivariable Cox regression models were conducted. Competing risks models and multiple approaches based on the propensity score were applied. RESULTS: This study included 43 201 patients [median age: 63.23 years old (Interquartile range, IQR): 55.21-71.95, 53.74% males; SGLT2I group: n = 16 144; DPP4I group: n = 27 057] with a median follow-up of 5.59 years (IQR: 5.27-5.81 years) since initial drug exposure. The incidence rate of developing gout [Incidence rate (IR): 2.5; 95% CI: 2.2, 2.9] among SGLT2I users was significantly lower than DPP4I users (IR: 5.2; 95% CI: 4.8, 5.8). SGLT2I was associated with 51% lower risks of gout (HR: 0.49; 95% CI: 0.42, 0.58; P-value < 0.0001) and 51% lower risks of all-cause mortality (HR: 0.49; 95% CI: 0.42, 0.58; P-value < 0.0001) after adjusting for significant demographics, past comorbidities, medications and laboratory results. The results remained consistent on competing risk and other propensity score approaches. CONCLUSIONS: SGLT2I use was associated with lower risks of new gout diagnosis compared with DPP4I use.

Reconstructing the social network of HIV key populations from locally observed information.

Traditional surveys only provide local observations about the topological structure of isolated individuals. This study aims to develop a novel data-driven approach to reconstructing the social network of men who have sex with men (MSM) communities from locally observed information by surveys. A large social network consisting of 1075 users and their public relationships was obtained manually from BlueD.com. We followed the same survey-taking procedure to sample locally observed information and adapted an Exponential Random Graph Model (ERGM) to model the full structure of the BlueD social network (number of local nodes N = 1075, observed average degree k = 6.46). The parameters were learned and then used to reconstruct the MSM social networks by two real-world survey datasets in Hong Kong (N = 600, k = 5.61) and Guangzhou (N = 757, k = 5). Our method performed well on reconstructing the BlueD social network, with a high accuracy (90.3%). In conclusion, this study demonstrates the feasibility of using parameters learning methods to reconstruct the social networks of HIV key populations. The method has the potential to inform data-driven intervention programs that need global social network structures.

Predicting oral cancer survival-Development and validation of an Asia-Pacific nomogram.

BACKGROUND: Nomograms are graphical calculating devices that predict response to treatment during cancer management. Oral squamous cell carcinoma (OSCC) is a lethal and deforming disease of rising incidence and global significance. The aim of this study was to develop a nomogram to predict individualized OSCC survival using a population-based dataset obtained from Queensland, Australia and externally validated using a cohort of OSCC patients treated in Hong Kong. METHODS: Clinico-pathological data for newly diagnosed OSCC patients, including age, sex, tumour site and grading, were accessed retrospectively from the Queensland Cancer Registry (QCR) in Australia and the Clinical Data Analysis and Reporting System (CDARS) in Hong Kong. Multivariate Cox proportional hazard regression was used to construct overall survival (OS) and cancer-specific survival (CSS) prediction models. Nomograms were internally validated using 10-fold cross validation, and externally validated against the Hong Kong dataset. RESULTS: Data from 9885 OSCC patients in Queensland and 465 patients from Hong Kong were analysed. All clinico-pathological variables significantly influenced survival outcomes. Nomogram calibration curves demonstrated excellent agreement between predicted and actual probability for Queensland patients. External validation in the Hong Kong population demonstrated slightly poorer nomogram performance, but predictive power remained strong. CONCLUSION: Based upon readily available data documenting patient demographic and clinico-pathological variables, predictive nomograms offer pragmatic aid to clinicians in individualized treatment planning and prognosis assessment in contemporary OSCC management.

Comparison of Sodium-Glucose Cotransporter-2 Inhibitor and Dipeptidyl Peptidase-4 Inhibitor on the Risks of New-Onset Atrial Fibrillation, Stroke and Mortality in Diabetic Patients: A Propensity Score-Matched Study in Hong Kong.

OBJECTIVE: To compare the effects of sodium-glucose cotransporter 2 inhibitors (SGLT2Is) and dipeptidyl peptidase-4 inhibitors (DPP4Is) on adverse outcomes in diabetic patients in Hong Kong. METHODS: This was a retrospective population-based cohort study of type 2 diabetes mellitus patients (n = 72,746) treated with SGLT2I or DPP4I between January 1, 2015, and December 31, 2020, in Hong Kong. Patients with exposure to both DPP4I and SGLT2I therapy, without complete demographics or mortality data, or who had prior atrial fibrillation (AF) were excluded. The study outcomes were new-onset AF, stroke/transient ischemic attack, cardiovascular mortality and all-cause mortality. Propensity score matching (1:1 ratio) between SGLT2I and DPP4I users was performed. RESULTS: The unmatched study cohort included 21,713 SGLT2I users and 39,510 DPP4I users (total: n = 61,233 patients; 55.37% males, median age: 62.7 years [interquartile range (IQR): 54.6-71.9 years]). Over a median follow-up of 2030 (IQR: 1912-2117) days, 2496 patients (incidence rate [IR]: 4.07%) developed new-onset AF, 2179 patients (IR: 3.55%) developed stroke/transient ischemic attack, 1963 (IR: 3.20%) died from cardiovascular causes and 6607 patients (IR: 10.79%) suffered from all-cause mortality. After propensity score matching (SGLT2I: n = 21,713; DPP4I: n = 21,713), SGLT2I users showed lower incidence of new-onset AF (1.96% vs. 2.78%, standardized mean difference [SMD] = 0.05), stroke (1.80% vs. 3.52%, SMD = 0.11), cardiovascular mortality (0.47% vs. 1.56%, SMD = 0.11) and all-cause mortality (2.59% vs. 7.47%, SMD = 0.22) compared to DPP4I users. Cox regression found that SGLT2I users showed lower risk of new-onset AF (hazard ratio [HR]: 0.68, 95% confidence interval [CI]: [0.56, 0.83], P = 0.0001), stroke (HR: 0.64, 95% CI: [0.53, 0.79], P < 0.0001), cardiovascular mortality (HR: 0.39, 95% CI: [0.27, 0.56], P < 0.0001) and all-cause mortality (HR: 0.44, 95% CI: [0.37, 0.51], P < 0.0001) after adjusting for significant demographics, past comorbidities, medications and laboratory tests. CONCLUSIONS: Based on real-world data of type 2 diabetic patients in Hong Kong, SGLT2I use was associated with lower risk of incident AF, stroke/transient ischemic attack, and cardiovascular and all-cause mortality outcomes compared to DPP4I use.

Five-year long-term comparison of robotic and laparoscopic gastrectomy for gastric cancer: a large single-center cohort study.

BACKGROUND: Robotic gastrectomy (RG) has been reported to be technically feasible and safe for patients with gastric cancer. However, 5-year long-term survival and recurrence outcomes for advanced gastric cancer have rarely been reported. This study aimed to compare the long-term oncologic outcomes between RG and laparoscopic gastrectomy (LG) for gastric cancer. METHODS: The general clinicopathological data of 1905 consecutive patients who underwent RG and LG were retrospectively collected at the Chinese People's Liberation Army General Hospital between November 2011 and October 2017. Propensity score matching (PSM) was used to match groups. The primary endpoints were 5-year disease-free survival (DFS) and overall survival (OS). RESULTS: After PSM, a well-balanced cohort of 283 patients in the RG group and 701 patients in the LG group were included in the analysis. The 5-year cumulative DFS rates were 67.28% in the robotic group and 70.41% in the laparoscopic group. The 5-year OS rate was 69.01% in the robotic group and 69.58% in the laparoscopic group. No significant differences in Kaplan-Meier survival curves for DFS (HR = 1.08, 95% CI 0.83-1.39, Log-rank P = 0.557) and OS (HR = 1.02, 95% CI 0.78-1.34, Log-rank P = 0.850) were observed between the 2 groups. In the subgroup analyses for potential confounding variables, there were no significant differences in 5-year DFS and 5-year OS survival between the 2 groups (P > 0.05), except for patients with pathological stage III and pathological stage N3 (P < 0.05). CONCLUSION: For patients with early gastric cancer, robotic and laparoscopic approaches have similar long-term survival. For patients with advanced gastric cancer, further studies need to be conducted to assess the long-term survival outcomes of RG.

Ultrasmall PtMn nanoparticles as sensitive manganese release modulator for specificity cancer theranostics.

Manganese-based nanomaterials (Mn-nanomaterials) hold immense potential in cancer diagnosis and therapies. However, most Mn-nanomaterials are limited by the low sensitivity and low efficiency toward mild weak acidity (pH 6.4-6.8) of the tumor microenvironment, resulting in unsatisfactory therapeutic effect and poor magnetic resonance imaging (MRI) performance. This study introduces pH-ultrasensitive PtMn nanoparticles as a novel platform for enhanced ferroptosis-based cancer theranostics. The PtMn nanoparticles were synthesized with different diameters from 5.3 to 2.7 nm with size-dominant catalytic activity and magnetic relaxation, and modified with an acidity-responsive polymer to create pH-sensitive agents. Importantly, R-PtMn-1 (3 nm core) presents "turn-on" oxidase-like activity, affording a significant enhancement ratio (pH 6.0/pH 7.4) in catalytic activity (6.7 folds), compared with R-PtMn-2 (4.2 nm core, 3.7 folds) or R-PtMn-3 (5.3 nm core, 2.1 folds), respectively. Moreover, R-PtMn-1 exhibits dual-mode contrast in high-field MRI. R-PtMn-1 possesses a good enhancement ratio (pH 6.4/pH 7.4) that is 3 or 3.2 folds for T1- or T2-MRI, respectively, which is higher than that of R-PtMn-2 (1.4 or 1.5 folds) or R-PtMn-3 (1.1 or 1.2 folds). Moreover, their pH-ultrasensitivity enabled activation specifically within the tumor microenvironment, avoiding off-target toxicity in normal tissues during delivery. In vitro studies demonstrated elevated intracellular reactive oxygen species production, lipid peroxidation, mitochondrial membrane potential changes, malondialdehyde content, and glutathione depletion, leading to enhanced ferroptosis in cancer cells. Meanwhile, normal cells remained unaffected by the nanoparticles. Overall, the pH-ultrasensitive PtMn nanoparticles offer a promising strategy for accurate cancer diagnosis and ferroptosis-based therapy.

Identification of Key Influencers for Secondary Distribution of HIV Self-Testing Kits Among Chinese Men Who Have Sex With Men: Development of an Ensemble Machine Learning Approach.

BACKGROUND: HIV self-testing (HIVST) has been rapidly scaled up and additional strategies further expand testing uptake. Secondary distribution involves people (defined as "indexes") applying for multiple kits and subsequently sharing them with people (defined as "alters") in their social networks. However, identifying key influencers is difficult. OBJECTIVE: This study aimed to develop an innovative ensemble machine learning approach to identify key influencers among Chinese men who have sex with men (MSM) for secondary distribution of HIVST kits. METHODS: We defined three types of key influencers: (1) key distributors who can distribute more kits, (2) key promoters who can contribute to finding first-time testing alters, and (3) key detectors who can help to find positive alters. Four machine learning models (logistic regression, support vector machine, decision tree, and random forest) were trained to identify key influencers. An ensemble learning algorithm was adopted to combine these 4 models. For comparison with our machine learning models, self-evaluated leadership scales were used as the human identification approach. Four metrics for performance evaluation, including accuracy, precision, recall, and F1-score, were used to evaluate the machine learning models and the human identification approach. Simulation experiments were carried out to validate our approach. RESULTS: We included 309 indexes (our sample size) who were eligible and applied for multiple test kits; they distributed these kits to 269 alters. We compared the performance of the machine learning classification and ensemble learning models with that of the human identification approach based on leadership self-evaluated scales in terms of the 2 nearest cutoffs. Our approach outperformed human identification (based on the cutoff of the self-reported scales), exceeding by an average accuracy of 11.0%, could distribute 18.2% (95% CI 9.9%-26.5%) more kits, and find 13.6% (95% CI 1.9%-25.3%) more first-time testing alters and 12.0% (95% CI -14.7% to 38.7%) more positive-testing alters. Our approach could also increase the simulated intervention's efficiency by 17.7% (95% CI -3.5% to 38.8%) compared to that of human identification. CONCLUSIONS: We built machine learning models to identify key influencers among Chinese MSM who were more likely to engage in secondary distribution of HIVST kits. TRIAL REGISTRATION: Chinese Clinical Trial Registry (ChiCTR) ChiCTR1900025433; https://www.chictr.org.cn/showproj.html?proj=42001.

Future trajectory of respiratory infections following the COVID-19 pandemic in Hong Kong.

The accumulation of susceptible populations for respiratory infectious diseases (RIDs) when COVID-19-targeted non-pharmaceutical interventions (NPIs) were in place might pose a greater risk of future RID outbreaks. We examined the timing and magnitude of RID resurgence after lifting COVID-19-targeted NPIs and assessed the burdens on the health system. We proposed the Threshold-based Control Method (TCM) to identify data-driven solutions to maintain the resilience of the health system by re-introducing NPIs when the number of severe infections reaches a threshold. There will be outbreaks of all RIDs with staggered peak times after lifting COVID-19-targeted NPIs. Such a large-scale resurgence of RID patients will impose a significant risk of overwhelming the health system. With a strict NPI strategy, a TCM-initiated threshold of 600 severe infections can ensure a sufficient supply of hospital beds for all hospitalized severely infected patients. The proposed TCM identifies effective dynamic NPIs, which facilitate future NPI relaxation policymaking.

Multiparametric MRI-Based Radiomic Signature for Preoperative Evaluation of Overall Survival in Intrahepatic Cholangiocarcinoma After Partial Hepatectomy.

BACKGROUND: The clinical outcomes of patients with intrahepatic cholangiocarcinoma (ICC) after partial hepatectomy remain suboptimal. Identifying patients with poor outcomes before surgery is urgently required. PURPOSE: To develop a multiparametric magnetic resonance imaging (MRI)-based radiomic signature to evaluate overall survival (OS) preoperatively and to investigate its incremental value for disease stratification. STUDY TYPE: Retrospective. SUBJECTS: One hundred and sixty-three patients with pathologically defined ICC, divided into training (N = 115) and validation sets (N = 48). SEQUENCE: Three-dimensional T1-weighted gradient-echo sequence with and without contrast agent, T2-weighted fast spin-echo sequence, and diffusion-weighted imaging with single-shot echo-planar sequence at 1.5 T or 3.0 T. ASSESSMENT: OS was defined as the time from the date of surgery to death or last contact. The radiomic signature was built based on the least absolute shrinkage and selection operator regression model. A clinicopathologic-radiographic (CPR) model and a combined model integrating radiomic signature with CPR factors were developed with multivariable Cox regression models. STATISTICAL TESTS: Harrell's concordance index (C-index) was used to compare the discrimination of different models. Net reclassification index (NRI) and integrated discrimination improvement (IDI) were used to quantify the improvement of prognostic accuracy after adding radiomic signature. RESULTS: The high-risk patients of death defined by the radiomic signature showed significantly lower OS compared with low-risk patients in validation set (3-year OS 17.1% vs. 56.4%, P < 0.001). Integrating radiomic signature into tumor, node, and metastasis (TNM) staging system significantly improved the prognostic accuracy compared with TNM stage alone (validation set C-index 0.745 vs. 0.649, P = 0.039, NRI improvement 39.9%-43.8%, IDI improvement 16.1%-19.4%). The radiomic signature showed no significant difference of C-index with postoperative CPR model (validation set, 0.698 vs. 0.674, P = 0.752). Incorporating the radiomic signature into CPR model significantly improved prognostic accuracy (NRI improvement 32.5%-34.3%, IDI improvement 8.1%-12.9%). DATA CONCLUSION: Multiparametric MRI-based radiomic signature is a potential biomarker for preoperative prognostic evaluation of ICC patients. LEVEL OF EVIDENCE: 4 TECHNICAL EFFICACY: Stage 4.

Risk of New-Onset Prostate Cancer for Metformin Versus Sulfonylurea Use in Type 2 Diabetes Mellitus: A Propensity Score-Matched Study.

BACKGROUND: The aim of this study was to compare the risks of new-onset prostate cancer between metformin and sulfonylurea users with type 2 diabetes mellitus (T2DM). METHODS: This population-based retrospective cohort study included male patients with T2DM presenting to public hospitals/clinics in Hong Kong between January 1, 2000, and December 31, 2009. We only included patients prescribed either, but not both, metformin or sulfonylurea. All patients were followed up until December 31, 2019. The primary outcome was new-onset prostate cancer and the secondary outcome was all-cause mortality. One-to-one propensity score matching was performed between metformin and sulfonylurea users based on demographics, comorbidities, antidiabetic and cardiovascular medications, fasting blood glucose level, and hemoglobin A1c level. Subgroup analyses based on age and use of androgen deprivation therapy were performed. RESULTS: The final study cohort consisted of 25,695 metformin users (mean [SD] age, 65.2 [11.8] years) and 25,695 matched sulfonylurea users (mean [SD] age, 65.3 [11.8] years) with a median follow-up duration of 119.6 months (interquartile range, 91.7-139.6 months) after 1:1 propensity score matching of 66,411 patients. Metformin users had lower risks of new-onset prostate cancer (hazard ratio, 0.80; 95% CI, 0.69-0.93; P=.0031) and all-cause mortality (hazard ratio, 0.89; 95% CI, 0.86-0.92; P<.0001) than sulfonylurea users. Metformin use was more protective against prostate cancer but less protective against all-cause mortality in patients aged <65 years (P for trend <.0001 for both) compared with patients aged ≥65 years. Metformin users had lower risk of all-cause mortality than sulfonylurea users, regardless of the use of androgen deprivation therapy (P for trend <.0001) among patients who developed prostate cancer. CONCLUSIONS: Metformin use was associated with significantly lower risks of new-onset prostate cancer and all-cause mortality than sulfonylurea use in male patients with T2DM.

Ureteral calculi lithotripsy for single ureteral calculi: can DNN-assisted model help preoperatively predict risk factors for sepsis?

OBJECTIVES: To explore the utility of radiomics and deep learning model in assessing the risk factors for sepsis after flexible ureteroscopy lithotripsy (FURL) or percutaneous nephrolithotomy (PCNL) in patients with ureteral calculi. METHODS: This retrospective analysis included 847 patients with treatment-naive proximal ureteral calculi who received FURL or PCNL. All participants were preoperatively conducted non-contrast computed tomography scans, and relevant clinical information was meanwhile collected. After propensity score matching, the radiomics model was established to predict the onset of sepsis. A deep learning model was also adapted to further improve the prediction accuracy. Performance of these trained models was verified in another independent external validation set including 40 cases of ureteral calculi patients. RESULTS: The overall incidence of sepsis after FURL or PCNL was 5.9%. The least absolute shrinkage and selection operator (LASSO) regression analysis revealed 26 predictive variables, with an overall AUC of 0.881 (95% CI, 0.813-0.931) and an AUC of 0.783 (95% CI, 0.766-0.801) in external validation cohort. Judicious adaption of a deep neural network (DNN) model to our dataset improved the AUC to 0.920 (95% CI, 0.906-0.933) in the internal validation. To eliminate the overfitting, external validation was carried out for DNN model (AUC = 0.874 (95% CI, 0.858-0.891)). CONCLUSIONS: The DNN was more effective than the LASSO model in revealing risk factors for sepsis after FURL or PCNL in single ureteral calculi patients, and females are more susceptible to sepsis than males. Deep learning models have the potential to act as gatekeepers to facilitate patient stratification. KEY POINTS: • Both the least absolute shrinkage and selection operator (LASSO) and deep neural network (DNN) models were shown to be effective in sepsis prediction. • The DNN model achieved superior prediction capability, with an AUC of 0.920 (95% CI, 0.906-0.933). • DNN-assisted model has potential to serve as a gatekeeper to facilitate patient stratification.

Data science approaches to infectious disease surveillance.

Novel data science approaches are needed to confront large-scale infectious disease epidemics such as COVID-19, human immunodeficiency viruses, African swine flu and Ebola. Human beings are now equipped with richer data and more advanced data analytics methodologies, many of which have become available only in the last decade. The theme issue Data Science Approaches to Infectious Diseases Surveillance reports the latest interdisciplinary research on developing novel data science methodologies to capitalize on the rich 'big data' of human behaviours to confront infectious diseases, with a particular focus on combating the ongoing COVID-19 pandemic. Compared to conventional public health research, articles in this issue present innovative data science approaches that were not possible without the growing human behaviour data and the recent advances in information and communications technology. This issue has 12 research papers and one review paper from a strong lineup of contributors from multiple disciplines, including data science, computer science, computational social sciences, applied maths, statistics, physics and public health. This introductory article provides a brief overview of the issue and discusses the future of this emerging field. This article is part of the theme issue 'Data science approaches to infectious disease surveillance'.

Data science approaches to confronting the COVID-19 pandemic: a narrative review.

During the COVID-19 pandemic, more than ever, data science has become a powerful weapon in combating an infectious disease epidemic and arguably any future infectious disease epidemic. Computer scientists, data scientists, physicists and mathematicians have joined public health professionals and virologists to confront the largest pandemic in the century by capitalizing on the large-scale 'big data' generated and harnessed for combating the COVID-19 pandemic. In this paper, we review the newly born data science approaches to confronting COVID-19, including the estimation of epidemiological parameters, digital contact tracing, diagnosis, policy-making, resource allocation, risk assessment, mental health surveillance, social media analytics, drug repurposing and drug development. We compare the new approaches with conventional epidemiological studies, discuss lessons we learned from the COVID-19 pandemic, and highlight opportunities and challenges of data science approaches to confronting future infectious disease epidemics. This article is part of the theme issue 'Data science approaches to infectious disease surveillance'.

Optimal resource allocation in HIV self-testing secondary distribution among Chinese MSM: data-driven integer programming models.

Human immunodeficiency virus self-testing (HIVST) is an innovative and effective strategy important to the expansion of HIV testing coverage. Several innovative implementations of HIVST have been developed and piloted among some HIV high-risk populations like men who have sex with men (MSM) to meet the global testing target. One innovative strategy is the secondary distribution of HIVST, in which individuals (defined as indexes) were given multiple testing kits for both self-use (i.e.self-testing) and distribution to other people in their MSM social network (defined as alters). Studies about secondary HIVST distribution have mainly concentrated on developing new intervention approaches to further increase the effectiveness of this relatively new strategy from the perspective of traditional public health discipline. There are many points of HIVST secondary distribution in which mathematical modelling can play an important role. In this study, we considered secondary HIVST kits distribution in a resource-constrained situation and proposed two data-driven integer linear programming models to maximize the overall economic benefits of secondary HIVST kits distribution based on our present implementation data from Chinese MSM. The objective function took expansion of normal alters and detection of positive and newly-tested 'alters' into account. Based on solutions from solvers, we developed greedy algorithms to find final solutions for our linear programming models. Results showed that our proposed data-driven approach could improve the total health economic benefit of HIVST secondary distribution. This article is part of the theme issue 'Data science approaches to infectious disease surveillance'.

Summarizing methods for estimating population size for key populations: a global scoping review for human immunodeficiency virus research.

BACKGROUND: Estimating the population sizes of key populations(people who inject drugs, men who have sex with men, transgender persons, and commercial sex workers) is critical for understanding the overall Human Immunodeficiency Virus burden. This scoping review aims to synthesize existing methods for population size estimation among key populations, and provide recommendations for future application of the existing methods. METHODS: Relevant studies published from 1st January 2000 to 4th August 2020 and related to key population size estimation were retrieved and 120 of 688 studies were assessed. After reading the full texts, 81 studies were further excluded. Therefore, 39 studies were included in this scoping review. Estimation methods included five digital methods, one in-person method, and four hybrid methods. FINDING: We summarized and organized the methods for population size estimateion into the following five categories: methods based on independent samples (including capture-recapture method and multiplier method), methods based on population counting (including Delphi method and mapping method), methods based on the official report (including workbook method), methods based on social network (including respondent-driven sampling method and network scale-up method) and methods based on data-driven technologies (Bayesian estimation method, Stochastic simulation method, and Laska, Meisner, and Siegel estimation method). Thirty-six (92%) articles were published after 2010 and 23 (59%) used multiple methods. Among the articles published after 2010, 11 in high-income countries and 28 in low-income countries. A total of 10 estimated the size of commercial sex workers, 14 focused on men who have sex with men, and 10 focused on people who inject drugs. CONCLUSIONS: There was no gold standard for population size estimation. Among 120 studies that were related to population size estimation of key populations, the most commonly used population estimation method is the multiplier method (26/120 studies). Every method has its strengths and biases. In recent years, novel methods based on data-driven technologies such as Bayesian estimation have been developed and applied in many surveys.