Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning

Osteoarthritis (OA) is increasing in prevalence and has a severe impact on patients' lives. However, our understanding of biomarkers driving OA risk remains limited. We developed a model predicting the five-year risk of OA diagnosis, integrating retrospective clinical, lifestyle and biomarker data from the UK Biobank (19,120 patients with OA, ROC-AUC: 0.72, 95%CI (0.71-0.73)). Higher age, BMI and prescription of non-steroidal anti-inflammatory drugs contributed most to increased OA risk prediction ahead of diagnosis. We identified 14 subgroups of OA risk profiles. These subgroups were validated in an independent set of patients evaluating the 11-year OA risk, with 88% of patients being uniquely assigned to one of the 14 subgroups. Individual OA risk profiles were characterised by personalised biomarkers. Omics integration demonstrated the predictive importance of key OA genes and pathways (e.g., GDF5 and TGF-β signalling) and OA-specific biomarkers (e.g., CRTAC1 and COL9A1). In summary, this work identifies opportunities for personalised OA prevention and insights into its underlying pathogenesis.

Comparison of the ADNEX and ROMA risk prediction models for the diagnosis of ovarian cancer: a multicentre external validation in patients who underwent surgery

BACKGROUND

Several diagnostic prediction models to help clinicians discriminate between benign and malignant adnexal masses are available. This study is a head-to-head comparison of the performance of the Assessment of Different NEoplasias in the adneXa (ADNEX) model with that of the Risk of Ovarian Malignancy Algorithm (ROMA).

METHODS

This is a retrospective study based on prospectively included consecutive women with an adnexal tumour scheduled for surgery at five oncology centres and one non-oncology centre in four countries between 2015 and 2019. The reference standard was histology. Model performance for ADNEX and ROMA was evaluated regarding discrimination, calibration, and clinical utility.

RESULTS

The primary analysis included 894 patients, of whom 434 (49%) had a malignant tumour. The area under the receiver operating characteristic curve (AUC) was 0.92 (95% CI 0.88-0.95) for ADNEX with CA125, 0.90 (0.84-0.94) for ADNEX without CA125, and 0.85 (0.80-0.89) for ROMA. ROMA, and to a lesser extent ADNEX, underestimated the risk of malignancy. Clinical utility was highest for ADNEX. ROMA had no clinical utility at decision thresholds <27%.

CONCLUSIONS

ADNEX had better ability to discriminate between benign and malignant adnexal tumours and higher clinical utility than ROMA.

CLINICAL TRIAL REGISTRATION

clinicaltrials.gov NCT01698632 and NCT02847832.

Persistent esophageal changes after histologic remission in eosinophilic esophagitis

BACKGROUND

Eosinophilic esophagitis (EoE) is characterized by persistent or relapsing allergic inflammation, and both clinical and histologic features of esophageal inflammation persist over time in most individuals. Mechanisms contributing to EoE relapse are not understood, and chronic EoE-directed therapy is therefore required to prevent long-term sequelae.

OBJECTIVE

We investigated whether EoE patients in histologic remission have persistent dysregulation of esophageal gene expression.

METHODS

Esophageal biopsy samples from 51 pediatric and 52 adult subjects with EoE in histopathologic remission (<15 eosinophils per high-power field [eos/hpf]) and control (48 pediatric and 167 adult) subjects from multiple institutions were subjected to molecular profiling by the EoE diagnostic panel, which comprises a set of 94 esophageal transcripts differentially expressed in active EoE.

RESULTS

Defining remission as <15 eos/hpf, we identified 51 and 32 differentially expressed genes in pediatric and adult EoE patients compared to control individuals, respectively (false discovery rate < 0.05). Using the stringent definition of remission (0 eos/hpf), the adult and pediatric cohorts continued to have 18 and 25 differentially expressed genes (false discovery rate < 0.05). Among 6 shared genes between adults and children, CDH26 was upregulated in both children and adults; immunohistochemistry demonstrated increased cadherin 26 staining in the epithelium of EoE patients in remission compared to non-EoE controls. In the adult cohort, POSTN expression correlated with the endoscopic reference system score (Spearman r = 0.35, P = .011), specifically correlating with the rings' endoscopic reference system subscore (r = 0.53, P = .004).

CONCLUSION

We have identified persistent EoE-associated esophageal gene expression in patients with disease in deep remission. These data suggest potential inflammation-induced epigenetic mechanisms may influence gene expression during remission in EoE and provide insight into possible mechanisms that underlie relapse in EoE.

Uncertainty-aware deep learning for trustworthy prediction of long-term outcome after endovascular thrombectomy

Acute ischemic stroke (AIS) is a leading global cause of mortality and morbidity. Improving long-term outcome predictions after thrombectomy can enhance treatment quality by supporting clinical decision-making. With the advent of interpretable deep learning methods in recent years, it is now possible to develop trustworthy, high-performing prediction models. This study introduces an uncertainty-aware, graph deep learning model that predicts endovascular thrombectomy outcomes using clinical features and imaging biomarkers. The model targets long-term functional outcomes, defined by the three-month modified Rankin Score (mRS), and mortality rates. A sample of 220 AIS patients in the anterior circulation who underwent endovascular thrombectomy (EVT) was included, with 81 (37%) demonstrating good outcomes (mRS ≤ 2). The performance of the different algorithms evaluated was comparable, with the maximum validation under the curve (AUC) reaching 0.87 using graph convolutional networks (GCN) for mRS prediction and 0.86 using fully connected networks (FCN) for mortality prediction. Moderate performance was obtained at admission (AUC of 0.76 using GCN), which improved to 0.84 post-thrombectomy and to 0.89 a day after stroke. Reliable uncertainty prediction of the model could be demonstrated.

A publicly available deep learning model and dataset for segmentation of breast, fibroglandular tissue, and vessels in breast MRI

Breast density, or the amount of fibroglandular tissue (FGT) relative to the overall breast volume, increases the risk of developing breast cancer. Although previous studies have utilized deep learning to assess breast density, the limited public availability of data and quantitative tools hinders the development of better assessment tools. Our objective was to (1) create and share a large dataset of pixel-wise annotations according to well-defined criteria, and (2) develop, evaluate, and share an automated segmentation method for breast, FGT, and blood vessels using convolutional neural networks. We used the Duke Breast Cancer MRI dataset to randomly select 100 MRI studies and manually annotated the breast, FGT, and blood vessels for each study. Model performance was evaluated using the Dice similarity coefficient (DSC). The model achieved DSC values of 0.92 for breast, 0.86 for FGT, and 0.65 for blood vessels on the test set. The correlation between our model's predicted breast density and the manually generated masks was 0.95. The correlation between the predicted breast density and qualitative radiologist assessment was 0.75. Our automated models can accurately segment breast, FGT, and blood vessels using pre-contrast breast MRI data. The data and the models were made publicly available.

Subjective feelings associated with expectations and rewards during risky decision-making in impulse control disorder

Impulse Control Disorder (ICD) in Parkinson's disease is a behavioral addiction induced by dopaminergic therapies, but otherwise unclear etiology. The current study investigates the interaction of reward processing variables, dopaminergic therapy, and risky decision-making and subjective feelings in patients with versus without ICD. Patients with (n = 18) and without (n = 12) ICD performed a risky decision-making task both 'on' and 'off' standard-of-care dopaminergic therapies (the task was performed on 2 different days with the order of on and off visits randomized for each patient). During each trial of the task, participants choose between two options, a gamble or a certain reward, and reported how they felt about decision outcomes. Subjective feelings of 'pleasure' are differentially driven by expectations of possible outcomes in patients with, versus without ICD. While off medication, the influence of expectations about risky-decisions on subjective feelings is reduced in patients with ICD versus without ICD. While on medication, the influence of expected outcomes in patients with ICD versus without ICD becomes similar. Computational modeling of behavior supports the idea that latent decision-making factors drive subjective feelings in patients with Parkinson's disease and that ICD status is associated with a change in the relationship between factors associated with risky behavior and subjective feelings about the experienced outcomes. Our results also suggest that dopaminergic medications modulate the impact expectations have on the participants' subjective reports. Altogether our results suggest that expectations about risky decisions may be decoupled from subjective feelings in patients with ICD, and that dopaminergic medications may reengage these circuits and increase emotional reactivity in patients with ICD.

Muscle stiffness indicating mission crew health in space

Muscle function is compromised by gravitational unloading in space affecting overall musculoskeletal health. Astronauts perform daily exercise programmes to mitigate these effects but knowing which muscles to target would optimise effectiveness. Accurate inflight assessment to inform exercise programmes is critical due to lack of technologies suitable for spaceflight. Changes in mechanical properties indicate muscle health status and can be measured rapidly and non-invasively using novel technology. A hand-held MyotonPRO device enabled monitoring of muscle health for the first time in spaceflight (> 180 days). Greater/maintained stiffness indicated countermeasures were effective. Tissue stiffness was preserved in the majority of muscles (neck, shoulder, back, thigh) but Tibialis Anterior (foot lever muscle) stiffness decreased inflight vs. preflight (p < 0.0001; mean difference 149 N/m) in all 12 crewmembers. The calf muscles showed opposing effects, Gastrocnemius increasing in stiffness Soleus decreasing. Selective stiffness decrements indicate lack of preservation despite daily inflight countermeasures. This calls for more targeted exercises for lower leg muscles with vital roles as ankle joint stabilizers and in gait. Muscle stiffness is a digital biomarker for risk monitoring during future planetary explorations (Moon, Mars), for healthcare management in challenging environments or clinical disorders in people on Earth, to enable effective tailored exercise programmes.

An artificial intelligence based abdominal aortic aneurysm prognosis classifier to predict patient outcomes

Abdominal aortic aneurysms (AAA) have been rigorously investigated to understand when their clinically-estimated risk of rupture-an event that is the 13th leading cause of death in the US-exceeds the risk associated with repair. Yet the current clinical guideline remains a one-size-fits-all "maximum diameter criterion" whereby AAA exceeding a threshold diameter is thought to make the risk of rupture high enough to warrant intervention. However, between 7 and 23.4% of smaller-sized AAA have been reported to rupture with diameters below the threshold. In this study, we train and assess machine learning models using clinical, biomechanical, and morphological indices from 381 patients to develop an aneurysm prognosis classifier to predict one of three outcomes for a given AAA patient: their AAA will remain stable, their AAA will require repair based as currently indicated from the maximum diameter criterion, or their AAA will rupture. This study represents the largest cohort of AAA patients that utilizes the first available medical image and clinical data to classify patient outcomes. The APC model therefore represents a potential clinical tool to striate specific patient outcomes using machine learning models and patient-specific image-based (biomechanical and morphological) and clinical data as input. Such a tool could greatly assist clinicians in their management decisions for patients with AAA.

Survival prediction of glioblastoma patients using modern deep learning and machine learning techniques

In this study, we utilized data from the Surveillance, Epidemiology, and End Results (SEER) database to predict the glioblastoma patients' survival outcomes. To assess dataset skewness and detect feature importance, we applied Pearson's second coefficient test of skewness and the Ordinary Least Squares method, respectively. Using two sampling strategies, holdout and five-fold cross-validation, we developed five machine learning (ML) models alongside a feed-forward deep neural network (DNN) for the multiclass classification and regression prediction of glioblastoma patient survival. After balancing the classification and regression datasets, we obtained 46,340 and 28,573 samples, respectively. Shapley additive explanations (SHAP) were then used to explain the decision-making process of the best model. In both classification and regression tasks, as well as across holdout and cross-validation sampling strategies, the DNN consistently outperformed the ML models. Notably, the accuracy were 90.25% and 90.22% for holdout and five-fold cross-validation, respectively, while the corresponding R2 values were 0.6565 and 0.6622. SHAP analysis revealed the importance of age at diagnosis as the most influential feature in the DNN's survival predictions. These findings suggest that the DNN holds promise as a practical auxiliary tool for clinicians, aiding them in optimal decision-making concerning the treatment and care trajectories for glioblastoma patients.

Optimized network based natural language processing approach to reveal disease comorbidities in COVID-19

A novel virus emerged from Wuhan, China, at the end of 2019 and quickly evolved into a pandemic, significantly impacting various industries, especially healthcare. One critical lesson from COVID-19 is the importance of understanding and predicting underlying comorbidities to better prioritize care and pharmacological therapies. Factors like age, race, and comorbidity history are crucial in determining disease mortality. While clinical data from hospitals and cohorts have led to the identification of these comorbidities, traditional approaches often lack a mechanistic understanding of the connections between them. In response, we utilized a deep learning approach to integrate COVID-19 data with data from other diseases, aiming to detect comorbidities with mechanistic insights. Our modified algorithm in the mpDisNet package, based on word-embedding deep learning techniques, incorporates miRNA expression profiles from SARS-CoV-2 infected cell lines and their target transcription factors. This approach is aligned with the emerging field of network medicine, which seeks to define diseases based on distinct pathomechanisms rather than just phenotypes. The main aim is discovery of possible unknown comorbidities by connecting the diseases by their miRNA mediated regulatory interactions. The algorithm can predict the majority of COVID-19's known comorbidities, as well as several diseases that have yet to be discovered to be comorbid with COVID-19. These potentially comorbid diseases should be investigated further to raise awareness and prevention, as well as informing the comorbidity research for the next possible outbreak.

Improving the precision of shock resuscitation by predicting fluid responsiveness with machine learning and arterial blood pressure waveform data

Fluid bolus therapy (FBT) is fundamental to the management of circulatory shock in critical care but balancing the benefits and toxicities of FBT has proven challenging in individual patients. Improved predictors of the hemodynamic response to a fluid bolus, commonly referred to as a fluid challenge, are needed to limit non-beneficial fluid administration and to enable automated clinical decision support and patient-specific precision critical care management. In this study we retrospectively analyzed data from 394 fluid boluses from 58 pigs subjected to either hemorrhagic or distributive shock. All animals had continuous blood pressure and cardiac output monitored throughout the study. Using this data, we developed a machine learning (ML) model to predict the hemodynamic response to a fluid challenge using only arterial blood pressure waveform data as the input. A Random Forest binary classifier referred to as the ML fluid responsiveness algorithm (MLFRA) was trained to detect fluid responsiveness (FR), defined as a ≥ 15% change in cardiac stroke volume after a fluid challenge. We then compared its performance to pulse pressure variation, a commonly used metric of FR. Model performance was assessed using the area under the receiver operating characteristic curve (AUROC), confusion matrix metrics, and calibration curves plotting predicted probabilities against observed outcomes. Across multiple train/test splits and feature selection methods designed to assess performance in the setting of small sample size conditions typical of large animal experiments, the MLFRA achieved an average AUROC, recall (sensitivity), specificity, and precision of 0.82, 0.86, 0.62. and 0.76, respectively. In the same datasets, pulse pressure variation had an AUROC, recall, specificity, and precision of 0.73, 0.91, 0.49, and 0.71, respectively. The MLFRA was generally well-calibrated across its range of predicted probabilities and appeared to perform equally well across physiologic conditions. These results suggest that ML, using only inputs from arterial blood pressure monitoring, may substantially improve the accuracy of predicting FR compared to the use of pulse pressure variation. If generalizable, these methods may enable more effective, automated precision management of critically ill patients with circulatory shock.

Comparison of seven surrogate insulin resistance indexes for prediction of incident coronary heart disease risk: a 10-year prospective cohort study

Background

There were seven novel and easily accessed insulin resistance (IR) surrogates established, including the Chinese visceral adiposity index (CVAI), the visceral adiposity index (VAI), lipid accumulation product (LAP), triglyceride glucose (TyG) index, TyG-body mass index (TyG-BMI), TyG-waist circumference (TyG-WC) and TyG-waist to height ratio (TyG-WHtR). We aimed to explore the association between the seven IR surrogates and incident coronary heart disease (CHD), and to compare their predictive powers among Chinese population.

Methods

This is a 10-year prospective cohort study conducted in China including 6393 participants without cardiovascular disease (CVD) at baseline. We developed Cox regression analyses to examine the association of IR surrogates with CHD (hazard ratio [HR], 95% confidence intervals [CI]). Moreover, the receiver operating characteristic (ROC) curve was performed to compare the predictive values of these indexes for incident CHD by the areas under the ROC curve (AUC).

Results

During a median follow-up period of 10.25 years, 246 individuals newly developed CHD. Significant associations of the IR surrogates (excepted for VAI) with incident CHD were found in our study after fully adjustment, and the fifth quintile HRs (95% CIs) for incident CHD were respectively 2.055(1.216-3.473), 1.446(0.948-2.205), 1.753(1.099-2.795), 2.013(1.214-3.339), 3.169(1.926-5.214), 2.275(1.391-3.719) and 2.309(1.419-3.759) for CVAI, VAI, LAP, TyG, TyG-BMI, TyG-WC and TyG-WHtR, compared with quintile 1. Furthermore, CVAI showed maximum predictive capacity for CHD among these seven IR surrogates with the largest AUC: 0.632(0.597,0.667).

Conclusion

The seven IR surrogates (excepted for VAI) were independently associated with higher prevalence of CHD, among which CVAI is the most powerful predictor for CHD incidence in Chinese populations.

Prediction of sentinel lymph node metastasis in breast cancer patients based on preoperative features: a deep machine learning approach

Sentinel lymph node (SLN) biopsy is the standard surgical approach to detect lymph node metastasis in breast cancer. Machine learning is a novel tool that provides better accuracy for predicting positive SLN involvement in breast cancer patients. This study obtained data from 2890 surgical cases of breast cancer patients from two referral hospitals in Iran from 2000 to 2021. Patients whose SLN involvement status was identified were included in our study. The dataset consisted of preoperative features, including patient features, gestational factors, laboratory data, and tumoral features. In this study, TabNet, an end-to-end deep learning model, was proposed to predict SLN involvement in breast cancer patients. We compared the accuracy of our model with results from logistic regression analysis. A total of 1832 patients with an average age of 51 ± 12 years were included in our study, of which 697 (25.5%) had SLN involvement. On average, the TabNet model achieved an accuracy of 75%, precision of 81%, specificity of 70%, sensitivity of 87%, and AUC of 0.74, while the logistic model demonstrated an accuracy of 70%, precision of 73%, specificity of 65%, sensitivity of 79%, F1 score of 73%, and AUC of 0.70 in predicting the SLN involvement in patients. Vascular invasion, tumor size, core needle biopsy pathology, age, and FH had the most contributions to the TabNet model. The TabNet model outperformed the logistic regression model in all metrics, indicating that it is more effective in predicting SLN involvement in breast cancer patients based on preoperative data.

Mechanism-centric regulatory network identifies NME2 and MYC programs as markers of Enzalutamide resistance in CRPC

Heterogeneous response to Enzalutamide, a second-generation androgen receptor signaling inhibitor, is a central problem in castration-resistant prostate cancer (CRPC) management. Genome-wide systems investigation of mechanisms that govern Enzalutamide resistance promise to elucidate markers of heterogeneous treatment response and salvage therapies for CRPC patients. Focusing on the de novo role of MYC as a marker of Enzalutamide resistance, here we reconstruct a CRPC-specific mechanism-centric regulatory network, connecting molecular pathways with their upstream transcriptional regulatory programs. Mining this network with signatures of Enzalutamide response identifies NME2 as an upstream regulatory partner of MYC in CRPC and demonstrates that NME2-MYC increased activities can predict patients at risk of resistance to Enzalutamide, independent of co-variates. Furthermore, our experimental investigations demonstrate that targeting MYC and its partner NME2 is beneficial in Enzalutamide-resistant conditions and could provide an effective strategy for patients at risk of Enzalutamide resistance and/or for patients who failed Enzalutamide treatment.

Quick model-based viscoelastic clot strength predictions from blood protein concentrations for cybermedical coagulation control

Cybermedical systems that regulate patient clotting in real time with personalized blood product delivery will improve treatment outcomes. These systems will harness popular viscoelastic assays of clot strength such as thromboelastography (TEG), which help evaluate coagulation status in numerous conditions: major surgery (e.g., heart, vascular, hip fracture, and trauma); liver cirrhosis and transplants; COVID-19; ICU stays; sepsis; obstetrics; diabetes; and coagulopathies like hemophilia. But these measurements are time-consuming, and thus impractical for urgent care and automated coagulation control. Because protein concentrations in a blood sample can be measured in about five minutes, we develop personalized, phenomenological, quick, control-oriented models that predict TEG curve outputs from input blood protein concentrations, to facilitate treatment decisions based on TEG curves. Here, we accurately predict, experimentally validate, and mechanistically justify curves and parameters for common TEG assays (Functional Fibrinogen, Citrated Native, Platelet Mapping, and Rapid TEG), and verify results with trauma patient clotting data.

Discovering biomarkers associated and predicting cardiovascular disease with high accuracy using a novel nexus of machine learning techniques for precision medicine

Personalized interventions are deemed vital given the intricate characteristics, advancement, inherent genetic composition, and diversity of cardiovascular diseases (CVDs). The appropriate utilization of artificial intelligence (AI) and machine learning (ML) methodologies can yield novel understandings of CVDs, enabling improved personalized treatments through predictive analysis and deep phenotyping. In this study, we proposed and employed a novel approach combining traditional statistics and a nexus of cutting-edge AI/ML techniques to identify significant biomarkers for our predictive engine by analyzing the complete transcriptome of CVD patients. After robust gene expression data pre-processing, we utilized three statistical tests (Pearson correlation, Chi-square test, and ANOVA) to assess the differences in transcriptomic expression and clinical characteristics between healthy individuals and CVD patients. Next, the recursive feature elimination classifier assigned rankings to transcriptomic features based on their relation to the case-control variable. The top ten percent of commonly observed significant biomarkers were evaluated using four unique ML classifiers (Random Forest, Support Vector Machine, Xtreme Gradient Boosting Decision Trees, and k-Nearest Neighbors). After optimizing hyperparameters, the ensembled models, which were implemented using a soft voting classifier, accurately differentiated between patients and healthy individuals. We have uncovered 18 transcriptomic biomarkers that are highly significant in the CVD population that were used to predict disease with up to 96% accuracy. Additionally, we cross-validated our results with clinical records collected from patients in our cohort. The identified biomarkers served as potential indicators for early detection of CVDs. With its successful implementation, our newly developed predictive engine provides a valuable framework for identifying patients with CVDs based on their biomarker profiles.

Rod Angulation Relationship with Thoracic Kyphosis after Adolescent Idiopathic Scoliosis Posterior Instrumentation

INTRODUCTION

Surgery to correct spinal deformities in scoliosis involves the use of contoured rods to reshape the spine and correct its curvatures. It is crucial to bend these rods appropriately to achieve the best possible correction. However, there is limited research on how the rod bending process relates to spinal shape in adolescent idiopathic scoliosis surgery.

METHODS

A retrospective study was conducted using a prospective multicenter scoliosis database. This study included adolescent idiopathic scoliosis patients from the database who underwent surgery with posterior instrumentation covering the T4 to T12 segments. Standing global spine X-rays were used in the analysis. The sagittal Cobb angles between T5 and T11 were measured on the spine. Additionally, the curvature of the rods between T5 and T11 was measured using the tangent method. To assess the relationship between these measurements, the difference between the dorsal kyphosis (TK) and the rod kyphosis (RK) was calculated (ΔK = TK - RK). This study aimed to analyze the correlation between ΔK and various patient characteristics. Both descriptive and statistical analyses were performed to achieve this goal.

RESULTS

This study encompassed a cohort of 99 patients, resulting in a total of 198 ΔK measurements for analysis. A linear regression analysis was conducted, revealing a statistically significant positive correlation between the kyphosis of the rods and that of the spine (r = 0.77, p = 0.0001). On average, the disparity between spinal and rod kyphosis averaged 5.5°. However, it is noteworthy that despite this modest mean difference, there was considerable variability among the patients. In particular, in 84% of cases, the concave rod exhibited less kyphosis than the spine, whereas the convex rod displayed greater kyphosis than the spine in 64% of cases. It was determined that the primary factor contributing to the flattening of the left rod was the magnitude of the coronal Cobb angle, both before and after the surgical procedure. These findings emphasize the importance of considering individual patient characteristics when performing rod bending procedures, aiming to achieve the most favorable outcomes in corrective surgery.

CONCLUSIONS

Although there is a notable and consistent correlation between the curvature of the spine and the curvature of the rods, it is important to acknowledge the substantial heterogeneity observed in this study. This heterogeneity suggests that individual patient factors play a significant role in shaping the outcome of spinal corrective surgery. Furthermore, this study highlights that more severe spinal curvatures in the frontal plane have an adverse impact on the shape of the rods in the sagittal plane. In other words, when the scoliosis curve is more pronounced in the frontal plane, it tends to influence the way the rods are shaped in the sagittal plane. This underscores the complexity of spinal deformities and the need for a tailored approach in surgical interventions to account for these variations among patients.

Prediction and stratification of longitudinal risk for chronic obstructive pulmonary disease across smoking behaviors

Smoking is the leading risk factor for chronic obstructive pulmonary disease (COPD) worldwide, yet many people who never smoke develop COPD. We perform a longitudinal analysis of COPD in the UK Biobank to derive and validate the Socioeconomic and Environmental Risk Score which captures additive and cumulative environmental, behavioral, and socioeconomic exposure risks beyond tobacco smoking. The Socioeconomic and Environmental Risk Score is more predictive of COPD than smoking status and pack-years. Individuals in the highest decile of the risk score have a greater risk for incident COPD compared to the remaining population. Never smokers in the highest decile of exposure risk are more likely to develop COPD than previous and current smokers in the lowest decile. In general, the prediction accuracy of the Social and Environmental Risk Score is lower in non-European populations. While smoking status is often considered in screening COPD, our finding highlights the importance of other non-smoking environmental and socioeconomic variables.

Deep learning of cell spatial organizations identifies clinically relevant insights in tissue images

Recent advancements in tissue imaging techniques have facilitated the visualization and identification of various cell types within physiological and pathological contexts. Despite the emergence of cell-cell interaction studies, there is a lack of methods for evaluating individual spatial interactions. In this study, we introduce Ceograph, a cell spatial organization-based graph convolutional network designed to analyze cell spatial organization (for example,. the cell spatial distribution, morphology, proximity, and interactions) derived from pathology images. Ceograph identifies key cell spatial organization features by accurately predicting their influence on patient clinical outcomes. In patients with oral potentially malignant disorders, our model highlights reduced structural concordance and increased closeness in epithelial substrata as driving features for an elevated risk of malignant transformation. In lung cancer patients, Ceograph detects elongated tumor nuclei and diminished stroma-stroma closeness as biomarkers for insensitivity to EGFR tyrosine kinase inhibitors. With its potential to predict various clinical outcomes, Ceograph offers a deeper understanding of biological processes and supports the development of personalized therapeutic strategies.

Identification of CT-based non-invasive radiomic biomarkers for overall survival prediction in oral cavity squamous cell carcinoma

This study addresses the limited non-invasive tools for Oral Cavity Squamous Cell Carcinoma (OSCC) survival prediction by identifying Computed Tomography (CT)-based biomarkers to improve prognosis prediction. A retrospective analysis was conducted on data from 149 OSCC patients, including CT radiomics and clinical information. An ensemble approach involving correlation analysis, score screening, and the Sparse-L1 algorithm was used to select functional features, which were then used to build Cox Proportional Hazards models (CPH). Our CPH achieved a 0.70 concordance index in testing. The model identified two CT-based radiomics features, Gradient-Neighboring-Gray-Tone-Difference-Matrix-Strength (GNS) and normalized-Wavelet-LLL-Gray-Level-Dependence-Matrix-Large-Dependence-High-Gray-Level-Emphasis (HLE), as well as stage and alcohol usage, as survival biomarkers. The GNS group with values above 14 showed a hazard ratio of 0.12 and a 3-year survival rate of about 90%. Conversely, the GNS group with values less than or equal to 14 had a 49% survival rate. For normalized HLE, the high-end group (HLE > - 0.415) had a hazard ratio of 2.41, resulting in a 3-year survival rate of 70%, while the low-end group (HLE ≤ - 0.415) had a 36% survival rate. These findings contribute to our knowledge of how radiomics can be used to predict the outcome so that treatment plans can be tailored for patients people with OSCC to improve their survival.

Effective gestational weight gain advice to optimize infant birth weight in Japan based on quantile regression analysis

The optimal range of gestational weight gain (GWG) was recently raised in Japan. This may help reduce small-for-gestational-age (SGA) infants, but may also increase large-for-gestational-age (LGA) infants. This study performed hypothetical experiments to determine effective GWG advice based on quantile regression analysis. In a total of 354,401 singleton pregnancies registered in the perinatal database of the Japan Society of Obstetrics and Gynecology (2013-2017), the proportions of SGA and LGA were 9.33% and 11.13%, respectively. Using regression coefficients of GWG across the birth weight-for-gestational-age quantile distribution, we analyzed changes in their proportions by simulating a uniform 3-kg extra increase in GWG or an increase or decrease based on GWG adequacy. A hypothetical experiment of a uniform increase in GWG resulted in SGA and LGA proportions of 7.26% (95% confidence interval 7.15-7.36) and 14.51% (14.37-14.66), respectively. By contrast, assuming a 3-kg increase in women with inadequate GWG and a 3-kg decrease in women with excessive GWG resulted in SGA and LGA proportions of 8.42% (8.31-8.54) and 11.50% (11.37-11.62), respectively. Our real-world data analysis suggests that careful adjustment of GWG based on GWG adequacy will be effective in optimizing infant birth weight in Japan.

Nucleotide excision repair deficiency is a targetable therapeutic vulnerability in clear cell renal cell carcinoma

Due to a demonstrated lack of DNA repair deficiencies, clear cell renal cell carcinoma (ccRCC) has not benefitted from targeted synthetic lethality-based therapies. We investigated whether nucleotide excision repair (NER) deficiency is present in an identifiable subset of ccRCC cases that would render those tumors sensitive to therapy targeting this specific DNA repair pathway aberration. We used functional assays that detect UV-induced 6-4 pyrimidine-pyrimidone photoproducts to quantify NER deficiency in ccRCC cell lines. We also measured sensitivity to irofulven, an experimental cancer therapeutic agent that specifically targets cells with inactivated transcription-coupled nucleotide excision repair (TC-NER). In order to detect NER deficiency in clinical biopsies, we assessed whole exome sequencing data for the presence of an NER deficiency associated mutational signature previously identified in ERCC2 mutant bladder cancer. Functional assays showed NER deficiency in ccRCC cells. Some cell lines showed irofulven sensitivity at a concentration that is well tolerated by patients. Prostaglandin reductase 1 (PTGR1), which activates irofulven, was also associated with this sensitivity. Next generation sequencing data of the cell lines showed NER deficiency-associated mutational signatures. A significant subset of ccRCC patients had the same signature and high PTGR1 expression. ccRCC cell line-based analysis showed that NER deficiency is likely present in this cancer type. Approximately 10% of ccRCC patients in the TCGA cohort showed mutational signatures consistent with ERCC2 inactivation associated NER deficiency and also substantial levels of PTGR1 expression. These patients may be responsive to irofulven, a previously abandoned anticancer agent that has minimal activity in NER-proficient cells.

Toward MR protocol-agnostic, unbiased brain age predicted from clinical-grade MRIs

The difference between the estimated brain age and the chronological age ('brain-PAD') could become a clinical biomarker. However, most brain age models were developed for research-grade high-resolution T1-weighted MRIs, limiting their applicability to clinical-grade MRIs from various protocols. We adopted a dual-transfer learning strategy to develop a model agnostic to modality, resolution, or slice orientation. We retrained a convolutional neural network (CNN) using 6281 clinical MRIs from 1559 patients, among 7 modalities and 8 scanner models. The CNN was trained to estimate brain age from synthetic research-grade magnetization-prepared rapid gradient-echo MRIs (MPRAGEs) generated by a 'super-resolution' method. The model failed with T2-weighted Gradient-Echo MRIs. The mean absolute error (MAE) was 5.86-8.59 years across the other modalities, still higher than for research-grade MRIs, but comparable between actual and synthetic MPRAGEs for some modalities. We modeled the "regression bias" in brain age, for its correction is crucial for providing unbiased summary statistics of brain age or for personalized brain age-based biomarkers. The bias model was generalizable as its correction eliminated any correlation between brain-PAD and chronological age in new samples. Brain-PAD was reliable across modalities. We demonstrate the feasibility of brain age predictions from arbitrary clinical-grade MRIs, thereby contributing to personalized medicine.

Genome interpretation in a federated learning context allows the multi-center exome-based risk prediction of Crohn's disease patients

High-throughput sequencing allowed the discovery of many disease variants, but nowadays it is becoming clear that the abundance of genomics data mostly just moved the bottleneck in Genetics and Precision Medicine from a data availability issue to a data interpretation issue. To solve this empasse it would be beneficial to apply the latest Deep Learning (DL) methods to the Genome Interpretation (GI) problem, similarly to what AlphaFold did for Structural Biology. Unfortunately DL requires large datasets to be viable, and aggregating genomics datasets poses several legal, ethical and infrastructural complications. Federated Learning (FL) is a Machine Learning (ML) paradigm designed to tackle these issues. It allows ML methods to be collaboratively trained and tested on collections of physically separate datasets, without requiring the actual centralization of sensitive data. FL could thus be key to enable DL applications to GI on sufficiently large genomics data. We propose FedCrohn, a FL GI Neural Network model for the exome-based Crohn's Disease risk prediction, providing a proof-of-concept that FL is a viable paradigm to build novel ML GI approaches. We benchmark it in several realistic scenarios, showing that FL can indeed provide performances similar to conventional ML on centralized data, and that collaborating in FL initiatives is likely beneficial for most of the medical centers participating in them.

Inflammation in the tumor-adjacent lung as a predictor of clinical outcome in lung adenocarcinoma

Approximately 30% of early-stage lung adenocarcinoma patients present with disease progression after successful surgical resection. Despite efforts of mapping the genetic landscape, there has been limited success in discovering predictive biomarkers of disease outcomes. Here we performed a systematic multi-omic assessment of 143 tumors and matched tumor-adjacent, histologically-normal lung tissue with long-term patient follow-up. Through histologic, mutational, and transcriptomic profiling of tumor and adjacent-normal tissue, we identified an inflammatory gene signature in tumor-adjacent tissue as the strongest clinical predictor of disease progression. Single-cell transcriptomic analysis demonstrated the progression-associated inflammatory signature was expressed in both immune and non-immune cells, and cell type-specific profiling in monocytes further improved outcome predictions. Additional analyses of tumor-adjacent transcriptomic data from The Cancer Genome Atlas validated the association of the inflammatory signature with worse outcomes across cancers. Collectively, our study suggests that molecular profiling of tumor-adjacent tissue can identify patients at high risk for disease progression.

Altered Gut Microbiota and Short-chain Fatty Acids in Chinese Children with Constipated Autism Spectrum Disorder

Gastrointestinal symptoms are more prevalent in children with autism spectrum disorder (ASD) than in typically developing (TD) children. Constipation is a significant gastrointestinal comorbidity of ASD, but the associations among constipated autism spectrum disorder (C-ASD), microbiota and short-chain fatty acids (SCFAs) are still debated. We enrolled 80 children, divided into the C-ASD group (n = 40) and the TD group (n = 40). In this study, an integrated 16S rRNA gene sequencing and gas chromatography-mass spectrometry-based metabolomics approach was applied to explore the association of the gut microbiota and SCFAs in C-ASD children in China. The community diversity estimated by the Observe, Chao1, and ACE indices was significantly lower in the C-ASD group than in the TD group. We observed that Ruminococcaceae_UCG_002, Erysipelotrichaceae_UCG_003, Phascolarctobacterium, Megamonas, Ruminiclostridium_5, Parabacteroides, Prevotella_2, Fusobacterium, and Prevotella_9 were enriched in the C-ASD group, and Anaerostipes, Lactobacillus, Ruminococcus_gnavus_group, Lachnospiraceae_NK4A136_group, Ralstonia, Eubacterium_eligens_group, and Ruminococcus_1 were enriched in the TD group. The propionate levels, which were higher in the C-ASD group, were negatively correlated with the abundance of Lactobacillus taxa, but were positively correlated with the severity of ASD symptoms. The random forest model, based on the 16 representative discriminant genera, achieved a high accuracy (AUC = 0.924). In conclusion, we found that C-ASD is related to altered gut microbiota and SCFAs, especially decreased abundance of Lactobacillus and excessive propionate in faeces, which provide new clues to understand C-ASD and biomarkers for the diagnosis and potential strategies for treatment of the disorder. This study was registered in the Chinese Clinical Trial Registry ( www.chictr.org.cn ; trial registration number ChiCTR2100052106; date of registration: October 17, 2021).

Machine learning-based prediction models for parathyroid carcinoma using pre-surgery cognitive function and clinical features

Patients with parathyroid carcinoma (PC) are often diagnosed postoperatively, due to incomplete resection during the initial surgery, resulting in poor outcomes. The aim of our study was to investigate the pre-surgery indicators of PC and try to develop a predictive model for PC utilizing machine learning. Evaluation of pre-surgery neuropsychological function and confirmation of pathology were carried out in 133 patients with primary hyperparathyroidism in Beijing Chaoyang Hospital from December 2019 to January 2023. Patients were randomly divided into a training cohort (n = 93) and a validating cohort (n = 40). Analysis of the clinical dataset, two machine learning including the extreme gradient boosting (XGBoost) and the least absolute shrinkage and selection operator (LASSO) regression were utilized to develop the prediction model for PC. Logistic regression analysis was also conducted for comparison. Significant differences in elevated parathyroid hormone and decreased serum phosphorus in PC compared to (BP). The lower score of MMSE and MOCA was observed in PC and a cutoff of MMSE < 24 was the optimal threshold to stratify PC from BP (area under the curve AUC 0.699 vs 0.625). The predicted probability of PC by machine learning was similar to the observed probability in the test set, whereas the logistic model tended to overpredict the possibility of PC. The XGBoost model attained a higher AUC than the logistic algorithms and LASSO models. (0.835 vs 0.683 vs 0.607). Preoperative cognitive function may be a probable predictor for PC. The cognitive function-based prediction model based on the XGBoost algorithm outperformed LASSO and logistic regression, providing valuable preoperative assistance to surgeons in clinical decision-making for patients suspected PC.

Early and fair COVID-19 outcome risk assessment using robust feature selection

Personalized medicine plays an important role in treatment optimization for COVID-19 patient management. Early treatment in patients at high risk of severe complications is vital to prevent death and ventilator use. Predicting COVID-19 clinical outcomes using machine learning may provide a fast and data-driven solution for optimizing patient care by estimating the need for early treatment. In addition, it is essential to accurately predict risk across demographic groups, particularly those underrepresented in existing models. Unfortunately, there is a lack of studies demonstrating the equitable performance of machine learning models across patient demographics. To overcome this existing limitation, we generate a robust machine learning model to predict patient-specific risk of death or ventilator use in COVID-19 positive patients using features available at the time of diagnosis. We establish the value of our solution across patient demographics, including gender and race. In addition, we improve clinical trust in our automated predictions by generating interpretable patient clustering, patient-level clinical feature importance, and global clinical feature importance within our large real-world COVID-19 positive patient dataset. We achieved 89.38% area under receiver operating curve (AUROC) performance for severe outcomes prediction and our robust feature ranking approach identified the presence of dementia as a key indicator for worse patient outcomes. We also demonstrated that our deep-learning clustering approach outperforms traditional clustering in separating patients by severity of outcome based on mutual information performance. Finally, we developed an application for automated and fair patient risk assessment with minimal manual data entry using existing data exchange standards.

Deconvolution of clinical variance in CAR-T cell pharmacology and response

Chimeric antigen receptor T cell (CAR-T) expansion and persistence vary widely among patients and predict both efficacy and toxicity. However, the mechanisms underlying clinical outcomes and patient variability are poorly defined. In this study, we developed a mathematical description of T cell responses wherein transitions among memory, effector and exhausted T cell states are coordinately regulated by tumor antigen engagement. The model is trained using clinical data from CAR-T products in different hematological malignancies and identifies cell-intrinsic differences in the turnover rate of memory cells and cytotoxic potency of effectors as the primary determinants of clinical response. Using a machine learning workflow, we demonstrate that product-intrinsic differences can accurately predict patient outcomes based on pre-infusion transcriptomes, and additional pharmacological variance arises from cellular interactions with patient tumors. We found that transcriptional signatures outperform T cell immunophenotyping as predictive of clinical response for two CD19-targeted CAR-T products in three indications, enabling a new phase of predictive CAR-T product development.

Single-cell characterization of macrophages in uveal melanoma uncovers transcriptionally heterogeneous subsets conferring poor prognosis and aggressive behavior

Uveal melanoma (UM) is the most frequent primary intraocular malignancy with high metastatic potential and poor prognosis. Macrophages represent one of the most abundant infiltrating immune cells with diverse functions in cancers. However, the cellular heterogeneity and functional diversity of macrophages in UM remain largely unexplored. In this study, we analyzed 63,264 single-cell transcriptomes from 11 UM patients and identified four transcriptionally distinct macrophage subsets (termed MΦ-C1 to MΦ-C4). Among them, we found that MΦ-C4 exhibited relatively low expression of both M1 and M2 signature genes, loss of inflammatory pathways and antigen presentation, instead demonstrating enhanced signaling for proliferation, mitochondrial functions and metabolism. We quantified the infiltration abundance of MΦ-C4 from single-cell and bulk transcriptomes across five cohorts and found that increased MΦ-C4 infiltration was relevant to aggressive behaviors and may serve as an independent prognostic indicator for poor outcomes. We propose a novel subtyping scheme based on macrophages by integrating the transcriptional signatures of MΦ-C4 and machine learning to stratify patients into MΦ-C4-enriched or MΦ-C4-depleted subtypes. These two subtypes showed significantly different clinical outcomes and were validated through bulk RNA sequencing and immunofluorescence assays in both public multicenter cohorts and our in-house cohort. Following further translational investigation, our findings highlight a potential therapeutic strategy of targeting macrophage subsets to control metastatic disease and consistently improve the outcome of patients with UM.

Augmentation strategies for an imbalanced learning problem on a novel COVID-19 severity dataset

Since the beginning of the COVID-19 pandemic, many different machine learning models have been developed to detect and verify COVID-19 pneumonia based on chest X-ray images. Although promising, binary models have only limited implications for medical treatment, whereas the prediction of disease severity suggests more suitable and specific treatment options. In this study, we publish severity scores for the 2358 COVID-19 positive images in the COVIDx8B dataset, creating one of the largest collections of publicly available COVID-19 severity data. Furthermore, we train and evaluate deep learning models on the newly created dataset to provide a first benchmark for the severity classification task. One of the main challenges of this dataset is the skewed class distribution, resulting in undesirable model performance for the most severe cases. We therefore propose and examine different augmentation strategies, specifically targeting majority and minority classes. Our augmentation strategies show significant improvements in precision and recall values for the rare and most severe cases. While the models might not yet fulfill medical requirements, they serve as an appropriate starting point for further research with the proposed dataset to optimize clinical resource allocation and treatment.

Predicting acute pancreatitis severity with enhanced computed tomography scans using convolutional neural networks

This study aimed to evaluate acute pancreatitis (AP) severity using convolutional neural network (CNN) models with enhanced computed tomography (CT) scans. Three-dimensional DenseNet CNN models were developed and trained using the enhanced CT scans labeled with two severity assessment methods: the computed tomography severity index (CTSI) and Atlanta classification. Each labeling method was used independently for model training and validation. Model performance was evaluated using confusion matrices, areas under the receiver operating characteristic curve (AUC-ROC), accuracy, precision, recall, F1 score, and respective macro-average metrics. A total of 1,798 enhanced CT scans met the inclusion criteria were included in this study. The dataset was randomly divided into a training dataset (n = 1618) and a test dataset (n = 180) with a ratio of 9:1. The DenseNet model demonstrated promising predictions for both CTSI and Atlanta classification-labeled CT scans, with accuracy greater than 0.7 and AUC-ROC greater than 0.8. Specifically, when trained with CT scans labeled using CTSI, the DenseNet model achieved good performance, with a macro-average F1 score of 0.835 and a macro-average AUC-ROC of 0.980. The findings of this study affirm the feasibility of employing CNN models to predict the severity of AP using enhanced CT scans.

Inherently interpretable position-aware convolutional motif kernel networks for biological sequencing data

Artificial neural networks show promising performance in detecting correlations within data that are associated with specific outcomes. However, the black-box nature of such models can hinder the knowledge advancement in research fields by obscuring the decision process and preventing scientist to fully conceptualize predicted outcomes. Furthermore, domain experts like healthcare providers need explainable predictions to assess whether a predicted outcome can be trusted in high stakes scenarios and to help them integrating a model into their own routine. Therefore, interpretable models play a crucial role for the incorporation of machine learning into high stakes scenarios like healthcare. In this paper we introduce Convolutional Motif Kernel Networks, a neural network architecture that involves learning a feature representation within a subspace of the reproducing kernel Hilbert space of the position-aware motif kernel function. The resulting model enables to directly interpret and evaluate prediction outcomes by providing a biologically and medically meaningful explanation without the need for additional post-hoc analysis. We show that our model is able to robustly learn on small datasets and reaches state-of-the-art performance on relevant healthcare prediction tasks. Our proposed method can be utilized on DNA and protein sequences. Furthermore, we show that the proposed method learns biologically meaningful concepts directly from data using an end-to-end learning scheme.

An Overview of Current Biomarkers, the Therapeutic Implications, and the Emerging Role of hERG1 Expression in Gastric Cancer: A Literature Review

Gastric cancer remains one of the most commonly diagnosed cancers in the world. It carries a high mortality rate, with cases being more prevalent in the developing world, and has been linked to diet and Helicobacter pylori infection. It is a highly heterogeneous disease, with most cases being of a sporadic nature. Most patients present at an advanced stage due to the asymptomatic nature of the early stages of the disease. A multidisciplinary approach is often best implemented to help decide how to best manage individual cases. However, the overall clinical outcome and survival of patients with advanced gastric cancer remain poor. Recent therapeutic advancements focus on the identification of molecular biomarkers associated with gastric cancer that have predictive, diagnostic, and prognostic implications. This enables the development of specific targeted therapies that have shown efficacy in numerous trials, either as monotherapy or in combination with standard chemotherapy. Despite this, tumour heterogeneity and treatment resistance are still issues leading to poor survival outcomes. An emerging approach is focusing efforts on the bidirectional crosstalk between tumour cells and the microenvironment through targeting ion channels. A key player in this is human ether-á-go-go-related gene 1 (hERG1). This voltage-gated potassium ion channel has been shown to have predictive, diagnostic, and prognostic significance, enabling the stratification of high-risk individuals. In addition, targeting hERG1 in combination with chemotherapy has been shown to potentiate tumour regression. This comprehensive literature review will aim to consolidate our understanding of current biomarkers in gastric cancer. The relevance of hERG1 in gastric cancer as a useful novel biomarker and the potential therapeutic implications as targeted therapy will be explored. This offers a new and personalised approach to helping to manage patients with gastric cancer.

Cancer-associated fibroblast-related prognostic signature predicts prognosis and immunotherapy response in pancreatic adenocarcinoma based on single-cell and bulk RNA-sequencing

Cancer-associated fibroblasts (CAFs) influence many aspects of pancreatic adenocarcinoma (PAAD) carcinogenesis, including tumor cell proliferation, angiogenesis, invasion, and metastasis. A six-gene prognostic signature was constructed for PAAD based on the 189 CAF marker genes identified in single-cell RNA-sequencing data. Multivariate analyses showed that the risk score was independently prognostic for survival in the TCGA (P < 0.001) and ICGC (P = 0.004) cohorts. Tumor infiltration of CD8 T (P = 0.005) cells and naïve B cells (P = 0.001) was greater in the low-risk than in the high-risk group, with infiltration of these cells negatively correlated with risk score. Moreover, the TMB score was lower in the low-risk than in the high-risk group (P = 0.0051). Importantly, patients in low-risk group had better immunotherapy responses than in the high-risk group in an independent immunotherapy cohort (IMvigor210) (P = 0.039). The CAV1 and SOD3 were highly expressed in CAFs of PAAD tissues, which revealed by immunohistochemical staining. In summary, this comprehensive analysis resulted in the development of a novel prognostic signature, which was associated with immune cell infiltration, drug sensitivity, and TMB, and could predict the prognosis and immunotherapy response of patients with PAAD.

PC3T: a signature-driven predictor of chemical compounds for cellular transition

Cellular transitions hold great promise in translational medicine research. However, therapeutic applications are limited by the low efficiency and safety concerns of using transcription factors. Small molecules provide a temporal and highly tunable approach to overcome these issues. Here, we present PC3T, a computational framework to enrich molecules that induce desired cellular transitions, and PC3T was able to consistently enrich small molecules that had been experimentally validated in both bulk and single-cell datasets. We then predicted small molecule reprogramming of fibroblasts into hepatic progenitor-like cells (HPLCs). The converted cells exhibited epithelial cell-like morphology and HPLC-like gene expression pattern. Hepatic functions were also observed, such as glycogen storage and lipid accumulation. Finally, we collected and manually curated a cell state transition resource containing 224 time-course gene expression datasets and 153 cell types. Our framework, together with the data resource, is freely available at http://pc3t.idrug.net.cn/ . We believe that PC3T is a powerful tool to promote chemical-induced cell state transitions.

C-reactive protein-to-albumin ratio as a biomarker in patients with sepsis: a novel LASSO-COX based prognostic nomogram

To develop a C-reactive protein-to-albumin ratio (CAR)-based nomogram for predicting the risk of in-hospital death in sepsis patients. Sepsis patients were selected from the MIMIC-IV database. Independent predictors were determined by multiple Cox analysis and then integrated to predict survival. The performance of the model was evaluated using the concordance index (C-index), receiver operating characteristic curve (ROC) analysis, and calibration curve. The risk stratifications analysis and subgroup analysis of the model in overall survival (OS) were assessed by Kaplan-Meier (K-M) curves. A total of 6414 sepsis patients were included. C-index of the CAR-based model was 0.917 [standard error (SE): 0.112] for the training set and 0.935 (SE: 0.010) for the validation set. The ROC curve analysis showed that the area under the curve (AUC) of the nomogram was 0.881 in the training set and 0.801 in the validation set. And the calibration curve showed that the nomogram performs well in both the training and validation sets. K-M curves indicated that patients with high CAR had significantly higher in-hospital mortality than those with low CAR. The CAR-based model has considerably high accuracy for predicting the OS of sepsis patients.

Optimal controlling of anti-TGF-[Formula: see text] and anti-PDGF medicines for preventing pulmonary fibrosis

In the repair of injury, some transforming growth factor-[Formula: see text]s (TGF-[Formula: see text]s) and platelet-derived growth factors (PDGFs) bind to fibroblast receptors as ligands and cause the differentiation of fibroblasts into myofibroblasts. When the injury repair is repeated, the myofibroblasts proliferate excessively, forming fibrotic tissue. We goal to control myofibroblasts proliferation and apoptosis with anti-transforming growth factor-[Formula: see text] (anti-TGF-[Formula: see text]) and anti-platelet-derived growth factor (anti-PDGF) medicines. The novel optimal regulator control problem with two controls (medicines) is proposed to simulate how to the preventing pulmonary fibrosis. Idiopathic pulmonary fibrosis (IPF) consists of restoring a system of cells, protein, and tissue networks with injury and scar. Myofibroblasts proliferation back to its equilibrium position after it has been disturbed by abnormal repair. Thus, the optimal regulator control problem with a parabolic partial differential equation as a constraint, zero flux boundary, and given specific initial conditions, is considered. The myofibroblast diffusion equation stands as a governing dynamic system while the objective function is the summation of myofibroblast, anti-TGF-[Formula: see text] and anti-PDGF medicines for the fixed final time. Here, myofibroblast is a nonlinear state of time while anti-TGF-[Formula: see text] and anti-PDGF are two unknown control functions. In order to solve the corresponding problem a weighted Galerkin method is used. Firstly, we convert the myofibroblast diffusion equation to a system of ordinary differential equations using the Lagrangian interpolation polynomials defined at Gauss-Lobatto integration points. Secondly, by the calculus of variations, the optimal control problem is solved successfully using canonical Hamiltonian and extended Riccati equations. Numerical results are given, and the plots are depicted. Moreover, solutions to the problem in which there is no control are compared. Numerical results show that, over time, the myofibroblast increases and then remains constant when there is no control. In contrast, the current solution decreases and vanishes after 300 days by prescribing controller medicines for anti-TGF-[Formula: see text] and anti-PDGF. The optimal strategy proposed in this paper helps practitioners to reduce myofibroblasts by controlling both anti-TGF-[Formula: see text] and anti-PDGF medicines.

Metabolomics Applied to Pediatric Asthma: What Have We Learnt in the Past 10 Years?

Background: Asthma is the most common chronic condition in children. It is a complex non-communicable disease resulting from the interaction of genetic and environmental factors and characterized by heterogeneous underlying molecular mechanisms. Metabolomics, as with the other omic sciences, thanks to the joint use of high-throughput technologies and sophisticated multivariate statistical methods, provides an unbiased approach to study the biochemical-metabolic processes underlying asthma. The aim of this narrative review is the analysis of the metabolomic studies in pediatric asthma published in the past 10 years, focusing on the prediction of asthma development, endotype characterization and pharmaco-metabolomics. Methods: A total of 43 relevant published studies were identified searching the MEDLINE/Pubmed database, using the following terms: "asthma" AND "metabolomics". The following filters were applied: language (English), age of study subjects (0-18 years), and publication date (last 10 years). Results and Conclusions: Several studies were identified within the three areas of interest described in the aim, and some of them likely have the potential to influence our clinical approach in the future. Nonetheless, further studies are needed to validate the findings and to assess the role of the proposed biomarkers as possible diagnostic or prognostic tools to be used in clinical practice.

Predicting the severity of postoperative scars using artificial intelligence based on images and clinical data

Evaluation of scar severity is crucial for determining proper treatment modalities; however, there is no gold standard for assessing scars. This study aimed to develop and evaluate an artificial intelligence model using images and clinical data to predict the severity of postoperative scars. Deep neural network models were trained and validated using images and clinical data from 1283 patients (main dataset: 1043; external dataset: 240) with post-thyroidectomy scars. Additionally, the performance of the model was tested against 16 dermatologists. In the internal test set, the area under the receiver operating characteristic curve (ROC-AUC) of the image-based model was 0.931 (95% confidence interval 0.910‒0.949), which increased to 0.938 (0.916‒0.955) when combined with clinical data. In the external test set, the ROC-AUC of the image-based and combined prediction models were 0.896 (0.874‒0.916) and 0.912 (0.892‒0.932), respectively. In addition, the performance of the tested algorithm with images from the internal test set was comparable with that of 16 dermatologists. This study revealed that a deep neural network model derived from image and clinical data could predict the severity of postoperative scars. The proposed model may be utilized in clinical practice for scar management, especially for determining severity and treatment initiation.

Development of a machine learning-based prediction model for sepsis-associated delirium in the intensive care unit

Septic patients in the intensive care unit (ICU) often develop sepsis-associated delirium (SAD), which is strongly associated with poor prognosis. The aim of this study is to develop a machine learning-based model for the early prediction of SAD. Patient data were extracted from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database and the eICU Collaborative Research Database (eICU-CRD). The MIMIC-IV data were divided into a training set and an internal validation set, while the eICU-CRD data served as an external validation set. Feature variables were selected using least absolute shrinkage and selection operator regression, and prediction models were built using logistic regression, support vector machines, decision trees, random forests, extreme gradient boosting (XGBoost), k-nearest neighbors and naive Bayes methods. The performance of the models was evaluated in the validation set. The model was also applied to a group of patients who were not assessed or could not be assessed for delirium. The MIMIC-IV and eICU-CRD databases included 14,620 and 1723 patients, respectively, with a median time to diagnosis of SAD of 24 and 30 h. Compared with Non-SAD patients, SAD patients had higher 28-days ICU mortality rates and longer ICU stays. Among the models compared, the XGBoost model had the best performance and was selected as the final model (internal validation area under the receiver operating characteristic curves (AUROC) = 0.793, external validation AUROC = 0.701). The XGBoost model outperformed other models in predicting SAD. The establishment of this predictive model allows for earlier prediction of SAD compared to traditional delirium assessments and is applicable to patients who are difficult to assess with traditional methods.

Predicting ward transfer mortality with machine learning

In order to address a long standing challenge for internal medicine physicians we developed artificial intelligence (AI) models to identify patients at risk of increased mortality. After querying 2,425 records of patients transferred from non-intensive care units to intensive care units from the Veteran Affairs Corporate Data Warehouse (CDW), we created two datasets. The former used 22 independent variables that included "Length of Hospital Stay" and "Days to Intensive Care Transfer," and the latter lacked these two variables. Since these two variables are unknown at the time of admission, the second set is more clinically relevant. We trained 16 machine learning models using both datasets. The best-performing models were fine-tuned and evaluated. The LightGBM model achieved the best results for both datasets. The model trained with 22 variables achieved a Receiver Operating Characteristics Curve-Area Under the Curve (ROC-AUC) of 0.89 and an accuracy of 0.72, with a sensitivity of 0.97 and a specificity of 0.68. The model trained with 20 variables achieved a ROC-AUC of 0.86 and an accuracy of 0.71, with a sensitivity of 0.94 and a specificity of 0.67. The top features for the former model included "Total length of Stay," "Admit to ICU Transfer Days," and "Lymphocyte Next Lab Value." For the latter model, the top features included "Lymphocyte First Lab Value," "Hemoglobin First Lab Value," and "Hemoglobin Next Lab Value." Our clinically relevant predictive mortality model can assist providers in optimizing resource utilization when managing large caseloads, particularly during shift changes.

Integrating a Virtual ICU with Cardiac and Cardiovascular ICUs: Managing the Needs of a Complex and High-Acuity Specialty ICU Cohort

A long-standing shortage of critical care intensivists and nurses, exacerbated by the coronavirus disease (COVID-19) pandemic, has led to an accelerated adoption of tele-critical care in the United States (US). Due to their complex and high-acuity nature, cardiac, cardiovascular, and cardiothoracic intensive care units (ICUs) have generally been limited in their ability to leverage tele-critical care resources. In early 2020, Houston Methodist Hospital (HMH) launched its tele-critical care program called Virtual ICU, or vICU, to improve its ICU staffing efficiency while providing high-quality, continuous access to in-person and virtual intensivists and critical care nurses. This article provides a roadmap with prescriptive specifications for planning, launching, and integrating vICU services within cardiac and cardiovascular ICUs-one of the first such integrations among the leading academic US hospitals. The success of integrating vICU depends upon the (1) recruitment of intensivists and RNs with expertise in managing cardiac and cardiovascular patients on the vICU staff as well as concerted efforts to promote mutual trust and confidence between in-person and virtual providers, (2) consultations with the bedside clinicians to secure their buy-in on the merits of vICU resources, and (3) collaborative approaches to improve workflow protocols and communications. Integration of vICU has resulted in the reduction of monthly night-call requirements for the in-person intensivists and an increase in work satisfaction. Data also show that support of the vICU is associated with a significant reduction in the rate of Code Blue events (denoting a situation where a patient requires immediate resuscitation, typically due to a cardiac or respiratory arrest). As the providers become more comfortable with the advances in artificial intelligence and big data-driven technology, the Cardiac ICU Cohort continues to improve methods to predict and track patient trends in the ICUs.

Interpretable deep learning for improving cancer patient survival based on personal transcriptomes

Precision medicine chooses the optimal drug for a patient by considering individual differences. With the tremendous amount of data accumulated for cancers, we develop an interpretable neural network to predict cancer patient survival based on drug prescriptions and personal transcriptomes (CancerIDP). The deep learning model achieves 96% classification accuracy in distinguishing short-lived from long-lived patients. The Pearson correlation between predicted and actual months-to-death values is as high as 0.937. About 27.4% of patients may survive longer with an alternative medicine chosen by our deep learning model. The median survival time of all patients can increase by 3.9 months. Our interpretable neural network model reveals the most discriminating pathways in the decision-making process, which will further facilitate mechanistic studies of drug development for cancers.

The Utility of Predictive Modeling and a Systems Process Approach to Reduce Emergency Department Crowding: A Position Paper

Emergency department (ED) crowding and its main causes, exit block and boarding, continue to threaten the quality and safety of ED care. Most interventions to reduce crowding have not been comprehensive or system solutions, only focusing on part of the care procession and not directly affecting boarding reduction. This position paper proposes that the ED crowding problem can be optimally addressed by applying a systems approach using predictive modeling to identify patients at risk of being admitted to the hospital and uses that information to initiate the time-consuming bed management process earlier in the care continuum, shortening the time during which patients wait in the ED for an inpatient bed assignment, thus removing the exit block that causes boarding and subsequently reducing crowding.

Real time monitoring of COVID-19 intervention effectiveness through contact tracing data

Communities worldwide have used vaccines and facemasks to mitigate the COVID-19 pandemic. When an individual opts to vaccinate or wear a mask, they may lower their own risk of becoming infected as well as the risk that they pose to others while infected. The first benefit-reducing susceptibility-has been established across multiple studies, while the second-reducing infectivity-is less well understood. Using a new statistical method, we estimate the efficacy of vaccines and facemasks at reducing both types of risks from contact tracing data collected in an urban setting. We find that vaccination reduced the risk of onward transmission by 40.7% [95% CI 25.8-53.2%] during the Delta wave and 31.0% [95% CI 19.4-40.9%] during the Omicron wave and that mask wearing reduced the risk of infection by 64.2% [95% CI 5.8-77.3%] during the Omicron wave. By harnessing commonly-collected contact tracing data, the approach can broadly provide timely and actionable estimates of intervention efficacy against a rapidly evolving pathogen.