Survival analysis and regression models

Evaluating machine learning model bias and racial disparities in non-small cell lung cancer using SEER registry data

BACKGROUND

Despite decades of pursuing health equity, racial and ethnic disparities persist in healthcare in America. For cancer specifically, one of the leading observed disparities is worse mortality among non-Hispanic Black patients compared to non-Hispanic White patients across the cancer care continuum. These real-world disparities are reflected in the data used to inform the decisions made to alleviate such inequities. Failing to account for inherently biased data underlying these observations could intensify racial cancer disparities and lead to misguided efforts that fail to appropriately address the real causes of health inequity.

OBJECTIVE

Estimate the racial/ethnic bias of machine learning models in predicting two-year survival and surgery treatment recommendation for non-small cell lung cancer (NSCLC) patients.

METHODS

A Cox survival model, and a LOGIT model as well as three other machine learning models for predicting surgery recommendation were trained using SEER data from NSCLC patients diagnosed from 2000-2018. Models were trained with a 70/30 train/test split (both including and excluding race/ethnicity) and evaluated using performance and fairness metrics. The effects of oversampling the training data were also evaluated.

RESULTS

The survival models show disparate impact towards non-Hispanic Black patients regardless of whether race/ethnicity is used as a predictor. The models including race/ethnicity amplified the disparities observed in the data. The exclusion of race/ethnicity as a predictor in the survival and surgery recommendation models improved fairness metrics without degrading model performance. Stratified oversampling strategies reduced disparate impact while reducing the accuracy of the model.

CONCLUSION

NSCLC disparities are complex and multifaceted. Yet, even when accounting for age and stage at diagnosis, non-Hispanic Black patients with NSCLC are less often recommended to have surgery than non-Hispanic White patients. Machine learning models amplified the racial/ethnic disparities across the cancer care continuum (which are reflected in the data used to make model decisions). Excluding race/ethnicity lowered the bias of the models but did not affect disparate impact. Developing analytical strategies to improve fairness would in turn improve the utility of machine learning approaches analyzing population-based cancer data.

Mechanism of CXCL8 regulation of methionine metabolism to promote angiogenesis in gliomas

BACKGROUND

Gliomas are the most common malignant brain tumors characterized by angiogenesis and invasive growth. A detailed understanding of its molecular characteristics could provide potential therapeutic targets. In the present study, we sought to explore the key gene CXCL8 in methionine metabolism in gliomas and its potential role in angiogenesis.

METHODS

U251 glioma cells were divided into control and methionine-restriction tolerant (constructed with 1/4 of the standard level of methionine in the culture medium) groups for transcriptome and metabolome analysis. To confirm the functions and mechanism of CXCL8 in glioma, heat map, volcano map, Go enrichment, gene set enrichment analysis (GSEA), protein-protein interaction network analysis, RT-PCR, western blotting assays, chicken embryo chorioallantoic membrane (CAM) test, chicken embryo yolk sac membrane (YSM) test and transplantation tumor nude mice model were performed. The TCGA database, CGGA database and clinical tissue samples were used to analyze CXCL8's significance on prognosis for patients with glioma.

RESULTS

CXCL8 expression was significantly up-regulated in methionine-restricted tolerance cells, it also activated vascular system development and triggered angiogenesis. CXCL8 expression is negatively correlated with survival prognosis in gliomas.

CONCLUSIONS

Glioma cells promote angiogenesis in methionine-restricted environments through the activation of CXCL8, compensating for nutrient deprivation, and possibly contributing to the failure of antiangiogenic therapy.

Discover overlooked complications after preeclampsia from three real-world medical record datasets of over 100,000 pregnancies

Structured Abstract Importance Preeclampsia poses a significant threat to women's long-term health. However, what diseases are affected and at what level they are affected by PE needs a thorough investigation. Objective To conduct the first large-scale, non-hypothesis-driven study using EHR data from multiple medical centers to comprehensively explore adverse health outcomes after preeclampsia Design: Retrospective multi-cohort case-control study Participants: We analyzed 3,592 preeclampsia patients and 23,040 non-preeclampsia controls from the University of Michigan Healthcare System. We externally validated the findings using UK Biobank data (443 cases, 14,870 controls) and Cedar Sinai data(2755 cases, 60,305 controls). Main outcomes: We showed that six complications are significantly affected by PE. We demonstrate the effect of race as well as preeclampsia severity on these complications. Results PE significantly increases the risk of later hypertension, uncomplicated and complicated diabetes, renal failure and obesity, after careful confounder adjustment. We also identified that hypothyroidism risks are significantly reduced in PE patients, particularly among African Americans. Severe PE affects hypertension, renal failure, uncomplicated diabetes and obesity more than mild PE, as expected. Caucasians are affected more negatively than African Americans by PE on future hypertension, uncomplicated and complicated diabetes and obesity. Conclusion This study fills a gap in the comprehensive assessment of preeclampsia's long-term effects using large-scale EHR data and rigorous statistical methods. Our findings emphasize the need for extended monitoring and tailored interventions for women with a history of preeclampsia, by considering pre-existing conditions, preeclampsia severity, and racial differences.

Letter: Allostatic Load and Adverse Prognosis in Inflammatory Bowel Disease-Need More Evidence. Authors' Reply

This article is linked to Zhao et al papers. To view these articles, visit https://doi.org/10.1111/apt.18217 and https://doi.org/10.1111/apt.18249.

AI-driven innovations in Alzheimer's disease: Integrating early diagnosis, personalized treatment, and prognostic modelling

Alzheimer's disease (AD) presents a significant challenge in neurodegenerative research and clinical practice due to its complex etiology and progressive nature. The integration of artificial intelligence (AI) into the diagnosis, treatment, and prognostic modelling of AD holds promising potential to transform the landscape of dementia care. This review explores recent advancements in AI applications across various stages of AD management. In early diagnosis, AI-enhanced neuroimaging techniques, including MRI, PET, and CT scans, enable precise detection of AD biomarkers. Machine learning models analyze these images to identify patterns indicative of early cognitive decline. Additionally, AI algorithms are employed to detect genetic and proteomic biomarkers, facilitating early intervention. Cognitive and behavioral assessments have also benefited from AI, with tools that enhance the accuracy of neuropsychological tests and analyze speech and language patterns for early signs of dementia. Personalized treatment strategies have been revolutionized by AI-driven approaches. In drug discovery, virtual screening and drug repurposing, guided by predictive modelling, accelerate the identification of effective treatments. AI also aids in tailoring therapeutic interventions by predicting individual responses to treatments and monitoring patient progress, allowing for dynamic adjustment of care plans. Prognostic modelling, another critical area, utilizes AI to predict disease progression through longitudinal data analysis and risk prediction models. The integration of multi-modal data, combining clinical, genetic, and imaging information, enhances the accuracy of these predictions. Deep learning techniques are particularly effective in fusing diverse data types to uncover new insights into disease mechanisms and progression. Despite these advancements, challenges remain, including ethical considerations, data privacy, and the need for seamless integration of AI tools into clinical workflows. This review underscores the transformative potential of AI in AD management while highlighting areas for future research and development. By leveraging AI, the healthcare community can improve early diagnosis, personalize treatments, and predict disease outcomes more accurately, ultimately enhancing the quality of life for individuals with AD.

Development of a prognostic model for predicting long-term visual acuity after cataract surgery in children with bilateral congenital cataracts: a single centre retrospective, observational study

BACKGROUND

To evaluate factors influencing best corrected visual acuity (BCVA) in paediatric patients with bilateral congenital cataracts (CC) after cataract extraction and intraocular lens (IOL) implantation, as well as develop a robust model for predicting long-term visual acuity.

METHODS

This retrospective study followed 194 paediatric patients with bilateral CC from January 2008 to December 2021. The endpoint event was defined as a final BCVA < 0.22 Log MAR at the last follow-up, which indicated good outcome. The probability of reaching this endpoint event was modelled using Cox proportional hazards regression analysis and internally validated through 200 iteration of 5-fold cross-validation.

RESULTS

A prognostic model for long-term visual acuity in bilateral CC after surgical treatment was established as follows: ln h(t) = -0.009 × "age at cataract extraction" - 0.015 × "age at IOL implantation" - 2.934 × "without nystagmus at last follow - up" + ln h0(0), in which h0(t) represents the baseline risk equation that can be any non-negative equation for time (t); h(t) represents the probability of the endpoint event occurring at time (t) without any endpoint event occurring before it. The model was visualized using a nomogram and contour plot to facilitate clinical practice. The model demonstrated reasonably accurate discrimination with an area under the receiver operating characteristic curve of 0.712 (95% confidence interval [CI]: 0.589-0.835) and a C-index of 0.797 (95% CI: 0.683-0.911). According to the model, children with bilateral CC had a higher likelihood of achieving a good outcome (BCVA < 0.22 Log MAR) if they underwent cataract extraction before the age of six months (hazard ratio [HR] 1.80, 95% CI: 0.92-3.70), received IOL implantation before the age of thirty-one months (HR 3.70, 95% CI: 1.77-7.80), and presented without nystagmus during their last follow-up visit (HR 11.20, 95% CI: 3.96-31.80).

CONCLUSIONS

This long-term visual acuity prognostic model demonstrates adequate performance for individualized prediction and assists in clinical decision-making. The risk stratification index guides optimal timing for surgery.

A predictive model and rapid multi-dynamic algorithm developed based on tumor-stroma percentage in gastric cancer: a retrospective, observational study

Background

Tumor-stroma percentage (TSP) is a prognostic risk factor in numerous solid tumors. Despite this, the prognostic significance of TSP in gastric cancer (GC) remains underexplored. Through the development of a personalized predictive model and a semi-automatic identification system, our study aimed to fully unlock the predictive potential of TSP in GC.

Methods

We screened GC patients from Shanghai General Hospital (SGH) between 2012 and 2019 to develop and validate a nomogram. Univariate and multivariate Cox proportional hazards regression analyses were employed to identify independent prognostic factors influencing the prognosis for GC patients. The nomogram was further validated externally by using a cohort from Bengbu Medical College (BMC). All patients underwent radical gastrectomy, with those diagnosed with locally advanced GC receiving adjuvant chemotherapy. The primary outcome measured was overall survival (OS). The semi-automatic identification of the TSP was achieved through a computer-aided detection (CAD) system, denoted as TSP-cad, while TSP identified by pathologists was labeled as TSP-visual.

Results

A total of 813 GC patients from SGH and 59 from BMC were enrolled in our study. TSP-visual was identified as an adverse prognostic factor for OS in GC and was found to be associated with pathological Tumor Node Metastasis staging system (pTNM) stage, T stage, N stage, perineural invasion (PNI), lymphovascular invasion (LVI), TSP-visual, tumor size, and other factors. Multivariate Cox regression using the training cohort revealed that TSP-visual (hazard ratio [HR], 2.042; 95% confidential interval [CI], 1.485-2.806; P < 0.001), N stage (HR, 2.136; 95% CI, 1.343-3.397; P = 0.010), PNI (HR , 1.791; 95% CI, 1.270-2.526; P = 0.001), and LVI (HR, 1.482; 95% CI, 1.021-2.152; P = 0.039) were independent predictors. These factors were incorporated into a novel nomogram, which exhibited strong predictive accuracy for 5-year OS in the training, internal validation, and external validation cohorts (area under the curve = 0.744, 0.759, and 0.854, respectively). The decision curve analysis of the nomogram and concordance indexes across the three cohorts outperformed the traditional pTNM (P < 0.05). Additionally, TSP-cad assessment using a rapid multi-dynamic algorithm demonstrated good agreement with TSP-visual.

Conclusions

The novel nomogram based on TSP could effectively identify individuals at risk of a poor prognosis among patients with GC. TSP-cad is anticipated to enhance the evaluation process of TSP.

Effects of Prenatal Breastfeeding Education on Breastfeeding Duration Beyond 12 Weeks: A Systematic Review

The proportion of infants in the United States who are breastfed at 1 year remains well below the Healthy People 2030 target. The health implications of suboptimal breastfeeding durations are significant, including increased risk of childhood leukemia and maternal Type 2 diabetes. Prenatal breastfeeding education provides an opportunity to improve breastfeeding self-efficacy among pregnant individuals and to establish their coping skills in case future breastfeeding problems arise. Although prenatal breastfeeding education is known to improve breastfeeding self-efficacy, characteristics of prenatal breastfeeding education interventions that are successful at increasing breastfeeding duration have not been well defined. Using Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and the Health Action Process Approach, we conducted a systematic review of the literature examining the impact of prenatal breastfeeding education interventions on breastfeeding duration measured at least 12 weeks postpartum. Twenty-one studies were identified. Prenatal breastfeeding education was most likely to increase breastfeeding duration when education interventions integrated psychological components (Health Action Process Approach coping planning) or were paired with in-person postpartum breastfeeding support. Additional research is needed to examine the role of psychological components in breastfeeding education interventions in diverse populations and to determine the specific psychological intervention components with the greatest impact on breastfeeding duration.

Factors affecting the survival of prediabetic patients: comparison of Cox proportional hazards model and random survival forest method

BACKGROUND

The worldwide prevalence of type 2 diabetes mellitus in adults is experiencing a rapid increase. This study aimed to identify the factors affecting the survival of prediabetic patients using a comparison of the Cox proportional hazards model (CPH) and the Random survival forest (RSF).

METHOD

This prospective cohort study was performed on 746 prediabetics in southwest Iran. The demographic, lifestyle, and clinical data of the participants were recorded. The CPH and RSF models were used to determine the patients' survival. Furthermore, the concordance index (C-index) and time-dependent receiver operating characteristic (ROC) curve were employed to compare the performance of the Cox proportional hazards (CPH) model and the random survival forest (RSF) model.

RESULTS

The 5-year cumulative T2DM incidence was 12.73%. Based on the results of the CPH model, NAFLD (HR = 1.74, 95% CI: 1.06, 2.85), FBS (HR = 1.008, 95% CI: 1.005, 1.012) and increased abdominal fat (HR = 1.02, 95% CI: 1.01, 1.04) were directly associated with diabetes occurrence in prediabetic patients. The RSF model suggests that factors including FBS, waist circumference, depression, NAFLD, afternoon sleep, and female gender are the most important variables that predict diabetes. The C-index indicated that the RSF model has a higher percentage of agreement than the CPH model, and in the weighted Brier Score index, the RSF model had less error than the Kaplan-Meier and CPH model.

CONCLUSION

Our findings show that the incidence of diabetes was alarmingly high in Iran. The results suggested that several demographic and clinical factors are associated with diabetes occurrence in prediabetic patients. The high-risk population needs special measures for screening and care programs.

Determinants of survival time for HIV/AIDS patients in the pastoralist region of Borena: a study at Yabelo General Hospital, South East Ethiopia

INTRODUCTION

HIV/AIDS is one of the most dangerous diseases globally, impacting public health, economics, society, political issues, and communities. As of 2023, the World Health Organization estimates that 40.4 million people are living with HIV/AIDS. This study aimed to identify the determinants of survival time for HIV/AIDS patients in the pastoralist region of Borena at Yabelo General Hospital.

METHOD

The study design was a retrospective cohort study, with a sample size of 293 individuals living with HIV/AIDS, based on recorded data. This research utilized survival model analysis, employing Kaplan-Meier plots, the log-rank test, and Cox proportional hazard model analysis.

RESULT

Out of the total sample size, 179 (61.1%) were female and 114 (38.1%) were male. Among these males, 36 (31.6%) were deceased. The analysis using the Cox proportional hazard model revealed that the following variables were significantly associated with the survival time of HIV/AIDS patients: gender, educational status, area of residence, tuberculosis (TB), and opportunistic infections.

CONCLUSIONS

We concluded that individuals living with HIV/AIDS in urban areas have a lower risk of death compared to those in rural areas, indicating that rural residents have a reduced survival probability. Therefore, the Borena zone administration should focus on adult patients to enhance life expectancy.

Multiple omics integrative analysis identifies GARS1 as a novel prognostic and immunological biomarker: from pan-cancer to bladder cancer

Glycyl-tRNA synthetase (GARS1) is differentially expressed across cancers. In this study, the value of GARS1 in the diagnosis and prognosis of various cancers was comprehensively evaluated by multiple omics integrative pan-cancer analysis and experimental verification. Through Kaplan-Meier, ROC and multiple databases, we explored GARS1 expression and prognostic and diagnostic patterns across cancers. The GARS1 relative reaction network was identified in PPI, GO, KEGG, methylation models and the genetic mutation atlas. Further research on the GARS1 value in bladder urothelial carcinoma (BLCA) was conducted by regression and nomogram models. We further analyzed the correlation between GARS1 and immune markers and cells in BLCA. Finally, in vitro experiments were used to validate GARS1 the oncogenic function of GARS1 in BLCA. We found that GARS1 was highly expressed across cancers, especially in BLCA. GARS1 expression was correlated with poor survival and had high diagnostic value in most tumor types. GARS1 is significantly associated with tRNA-related pathways whose mutation sites are mainly located on tRNA synthetase. In addition, Upregulation of GARS1 was connected with immune cell infiltration and five key MMR genes. M2 macrophages, TAMs, Th1 and T-cell exhaustion, and marker sets associated with GARS1 expression indicated specific immune infiltration in BLCA. Finally, in vitro experiments validated that GARS1 expression promotes BLCA cell proliferation and metastasis and inhibits apoptosis. Overall, GARS1 can be a novel prognostic and immunological biomarker through multiple omics integrative pan-cancer analysis. The expression of GARS1 in BLCA was positively correlated with specific immune infiltration, indicating that GARS1 might be related to the tumor immune microenvironment.

Prognostic Value of High-Density Lipoprotein Cholesterol in Patients with Overt Hepatic Encephalopathy

Overt hepatic encephalopathy (OHE), a serious complication of liver cirrhosis, is associated with alterations in lipid and lipoprotein metabolism. We evaluated the correlation between high-density lipoprotein cholesterol (HDL-C) levels and transplant-free (TF) mortality in patients with OHE. Patients with OHE admitted to Beijing Ditan Hospital between January 2010 and August 2016 (n = 821) and between September 2016 and December 2020 (n = 480) were included in the training and validation sets, respectively. Independent predictors were explored by a multivariate Cox regression analysis, and the area under the receiver operating characteristic curve (AUC) was used to assess the prognostic value of these factors. The prognostic value of HDL-C was good (AUC at 1 year: 0.745) and was equivalent to that of the Model for End-Stage Liver Disease (MELD) score (AUC at 1 year: 0.788). The optimal threshold values for HDL-C and MELD were 0.5 mmol/L and 17, respectively. The 1-year TF mortality rates in the low-risk (HDL-C ≥ 0.5 mmol/L and MELD < 17) and high-risk (HDL-C < 0.5 mmol/L and MELD ≥ 17) groups were 7.5% and 51.5% in the training set and 10.1% and 48.2% in the validation set, respectively. HDL-C level < 0.5 mmol/L and MELD score > 17 can facilitate the identification of high-risk patients and provide a basis for timely treatment.

Mortality in French people with polyhandicap/profound intellectual and multiple disabilities

BACKGROUND

In recent decades, progress has been made in the care of people with polyhandicap/profound intellectual and multiple disabilities (PIMD) through a better understanding of the pathophysiology and the development of new care management and rehabilitation strategies adapted to these extreme pathologies. Although there is a lack of knowledge about the health status and care management of the oldest people, a better understanding of the natural course of life of people with polyhandicap/PIMD would consequently allow the optimisation of preventive and curative care management strategies. Few robust data on mortality and life expectancy have been documented for this population in France. Our aims are to estimate the median survival time and assess the factors associated with mortality in people with polyhandicap/PIMD receiving care in France.

METHODS

This study included people with polyhandicap/PIMD, followed by the French national cohort 'Eval-PLH' since 2015. These individuals were included in specialised rehabilitation centres and residential institutions. The people included in the first wave of the cohort (2015-2016) were eligible for the present study. Vital status on 1 January 2022 (censoring date) was collected in two ways: (1) spontaneous reporting by the participating centre to the coordinating team and (2) systematic checking on the French national death platform. According to the vital status, survival was calculated in years from the date of birth to the date of death or from the date of birth to the censoring date. The factors associated with mortality were evaluated using the Cox proportional regression hazards model.

RESULTS

Data from 780 individuals aged between 3 and 67 years were analysed. At the censoring date, 176 (22.6%) had died, and the mean survival was 52.8 years (95% confidence interval: 51.1-54.5). Mortality was significantly associated with a progressive aetiology, recurrent pulmonary infections, drug-resistant epilepsy and a higher number of medical devices.

CONCLUSIONS

This study shows for the first time the survival and impact of factors associated with mortality in people with polyhandicap/PIMD in France.

Multi-modality risk prediction of cardiovascular diseases for breast cancer cohort in the All of Us Research Program

OBJECTIVE

This study leverages the rich diversity of the All of Us Research Program (All of Us)'s dataset to devise a predictive model for cardiovascular disease (CVD) in breast cancer (BC) survivors. Central to this endeavor is the creation of a robust data integration pipeline that synthesizes electronic health records (EHRs), patient surveys, and genomic data, while upholding fairness across demographic variables.

MATERIALS AND METHODS

We have developed a universal data wrangling pipeline to process and merge heterogeneous data sources of the All of Us dataset, address missingness and variance in data, and align disparate data modalities into a coherent framework for analysis. Utilizing a composite feature set including EHR, lifestyle, and social determinants of health (SDoH) data, we then employed Adaptive Lasso and Random Forest regression models to predict 6 CVD outcomes. The models were evaluated using the c-index and time-dependent Area Under the Receiver Operating Characteristic Curve over a 10-year period.

RESULTS

The Adaptive Lasso model showed consistent performance across most CVD outcomes, while the Random Forest model excelled particularly in predicting outcomes like transient ischemic attack when incorporating the full multi-model feature set. Feature importance analysis revealed age and previous coronary events as dominant predictors across CVD outcomes, with SDoH clustering labels highlighting the nuanced impact of social factors.

DISCUSSION

The development of both Cox-based predictive model and Random Forest Regression model represents the extensive application of the All of Us, in integrating EHR and patient surveys to enhance precision medicine. And the inclusion of SDoH clustering labels revealed the significant impact of sociobehavioral factors on patient outcomes, emphasizing the importance of comprehensive health determinants in predictive models. Despite these advancements, limitations include the exclusion of genetic data, broad categorization of CVD conditions, and the need for fairness analyses to ensure equitable model performance across diverse populations. Future work should refine clinical and social variable measurements, incorporate advanced imputation techniques, and explore additional predictive algorithms to enhance model precision and fairness.

CONCLUSION

This study demonstrates the liability of the All of Us's diverse dataset in developing a multi-modality predictive model for CVD in BC survivors risk stratification in oncological survivorship. The data integration pipeline and subsequent predictive models establish a methodological foundation for future research into personalized healthcare.

Identification of a pro-protein synthesis osteosarcoma subtype for predicting prognosis and treatment

Osteosarcoma (OS) is a heterogeneous malignant spindle cell tumor that is aggressive and has a poor prognosis. Although combining surgery and chemotherapy has significantly improved patient outcomes, the prognosis for OS patients with metastatic or recurrent OS has remained unsatisfactory. Therefore, it is imperative to gain a fresh perspective on OS development mechanisms and treatment strategies. After studying single-cell RNA sequencing (scRNA-seq) data in public databases, we identified seven OS subclonal types based on intra-tumor heterogeneity. Subsequently, we constructed a prognostic model based on pro-protein synthesis osteosarcoma (PPS-OS)-associated genes. Correlation analysis showed that the prognostic model performs extremely well in predicting OS patient prognosis. We also demonstrated that the independent risk factors for the prognosis of OS patients were tumor primary site, metastatic status, and risk score. Based on these factors, nomograms were constructed for predicting the 3- and 5-year survival rates. Afterward, the investigation of the tumor immune microenvironment (TIME) revealed the vital roles of γδ T-cell and B-cell activation. Drug sensitivity analysis and immune checkpoint analysis identified drugs that have potential application value in OS. Finally, the jumping translocation breakpoint (JTB) gene was selected for experimental validation. JTB silencing suppressed the proliferation, migration, and invasion of OS cells. Therefore, our research suggests that PPS-OS-related genes facilitate the malignant progression of OS and may be employed as prognostic indicators and therapeutic targets in OS.

Time to first remission and survival in patients with acromegaly: Evidence from the UK Acromegaly Register Study (UKAR)

OBJECTIVE

This study aimed to understand the effect of time to remission of acromegaly on survival in people living with acromegaly.

DESIGN, PATIENTS AND MEASUREMENT

This cross-sectional study used data from the UK Acromegaly Register. We considered remission of acromegaly growth hormone controlled at ≤2 μg/L following the diagnosis of acromegaly. We used the accelerated failure time model to assess the effect of time to remission on survival in acromegaly.

RESULTS

The study population comprises 3569 individuals with acromegaly, with a median age of diagnosis of 47.3 (36.5-57.8) years, 48% females and a majority white population (61%). The number of individuals with the first remission of acromegaly was 2472, and the median time to first remission was 1.92 (0.70-6.58) years. In this study, time to first remission in acromegaly was found to have a significant effect on survival (p < .001); for every 1-year increase in time to first remission, there was a median 1% reduction in survival in acromegaly. In an analysis adjusted for covariates, the survival rate was 52% higher (p < .001) in those who underwent surgery as compared to those who did not have surgery, 18% higher (p = .01) in those who received treatment with somatostatin analogues (SMA) as compared to those with dopamine agonists and 21% lower (p < .001) in those who received conventional radiotherapy as compared to those who did not receive radiotherapy.

CONCLUSION

In conclusion, this population-based study conducted in patients with acromegaly revealed that faster remission time, surgical intervention and treatment with SMA are linked to improved survival outcomes.

Early-onset alcohol, tobacco, and illicit drug use with age at onset of hypertension: a survival analysis

PURPOSE

To examine the associations of age when first substance use and early-onset substance use before age 18 with age at onset (AAO) of hypertension.

METHODS

This study included 19,270 individuals with AAO of hypertension from the 2015-2019 National Survey on Drug Use and Health. Age when first use of 10 substance use variables included alcohol, daily cigarettes, cigars, smokeless tobacco, marijuana, cocaine, hallucinogens, lysergic acid diethylamide (LSD), inhalants, and methamphetamine use. The outcome was AAO of hypertension and variable cluster analysis was used to classify the exposures and outcome. Substance use status was classified into three categories: early-onset substance use (first used substance before age 18), late-onset substance use (first used substance after age 18), and never used.

RESULTS

The mean AAO of hypertension was 42.7 years. Age when first use of 10 substance use variables had significant correlations with AAO of hypertension (all p values < 0.001). Individuals with early-onset alcohol, cigars, smokeless tobacco, marijuana, hallucinogens, inhalants, cocaine, LSD, and methamphetamine use revealed significantly earlier onset of hypertension than those never used. Compared with never used substances, the Cox regression model showed that early-onset alcohol, smokeless tobacco, marijuana, inhalants, and methamphetamine use had an increased risk of AAO of hypertension [hazard ratio (HR) (95%CI) = 1.22 (1.13, 1.31), 1.36 (1.24, 1.49), 1.85 (1.75, 1.95), 1.41 (1.30, 1.52), and 1.27 (1.07,1.50), respectively].

CONCLUSION

These findings suggest that intervention strategies or programs focusing on preventing early-onset substance use before age 18 may delay the onset of adult hypertension.

Deep learning models for predicting the survival of patients with hepatocellular carcinoma based on a surveillance, epidemiology, and end results (SEER) database analysis

Hepatocellular carcinoma (HCC) is a common malignancy with poor survival and requires long-term follow-up. Hence, we collected information on patients with Primary Hepatocellular Carcinoma in the United States from the Surveillance, Epidemiology, and EndResults (SEER) database. We used this information to establish a deep learning with a multilayer neural network (the NMTLR model) for predicting the survival rate of patients with Primary Hepatocellular Carcinoma. HCC patients pathologically diagnosed between January 2011 and December 2015 in the SEER (Surveillance, Epidemiology, and End Results) database of the National Cancer Institute of the United States were selected as study subjects. We utilized two deep learning-based algorithms (DeepSurv and Neural Multi-Task Logistic Regression [NMTLR]) and a machine learning-based algorithm (Random Survival Forest [RSF]) for model training. A multivariable Cox Proportional Hazards (CoxPH) model was also constructed for comparison. The dataset was randomly divided into a training set and a test set in a 7:3 ratio. The training dataset underwent hyperparameter tuning through 1000 iterations of random search and fivefold cross-validation. Model performance was assessed using the concordance index (C-index), Brier score, and Integrated Brier Score (IBS). The accuracy of predicting 1-year, 3-year, and 5-year survival rates was evaluated using Receiver Operating Characteristic (ROC) curves, calibration plots, and Area Under the Curve (AUC). The primary outcomes were the 1-year, 3-year, and 5-year overall survival rates. Models were developed using DeepSurv, NMTLR, RSF, and Cox Proportional Hazards regression. Model differentiation was evaluated using the C-index, calibration with concordance plots, and risk stratification capability with the log-rank test. The study included 2197 HCC patients, randomly divided into a training cohort (70%, n = 1537) and a testing cohort (30%, n = 660). Clinical characteristics between the two cohorts showed no significant statistical difference (p > 0.05). The deep learning models outperformed both RSF and CoxPH models, with C-indices of 0.735 (NMTLR) and 0.731 (DeepSurv) in the test dataset. The NMTLR model demonstrated enhanced accuracy and well-calibrated survival estimates, achieving an Area Under the Curve (AUC) of 0.824 for 1-year survival predictions, 0.813 for 3-year, and 0.803 for 5-year survival rates. This model's superior calibration and discriminative ability enhance its utility for clinical prognostication in Primary Hepatocellular Carcinoma. We deployed the NMTLR model as a web application for clinical practice. The NMTLR model have potential advantages over traditional linear models in prognostic assessment and treatment recommendations. This novel analytical approach may provide reliable information on individual survival and treatment recommendations for patients with primary liver cancer.

ResDeepSurv: A Survival Model for Deep Neural Networks Based on Residual Blocks and Self-attention Mechanism

Survival analysis, as a widely used method for analyzing and predicting the timing of event occurrence, plays a crucial role in the medicine field. Medical professionals utilize survival models to gain insight into the effects of patient covariates on the disease, and the correlation with the effectiveness of different treatment strategies. This knowledge is essential for the development of treatment plans and the enhancement of treatment approaches. Conventional survival models, such as the Cox proportional hazards model, require a significant amount of feature engineering or prior knowledge to facilitate personalized modeling. To address these limitations, we propose a novel residual-based self-attention deep neural network for survival modeling, called ResDeepSurv, which combines the benefits of neural networks and the Cox proportional hazards regression model. The model proposed in our study simulates the distribution of survival time and the correlation between covariates and outcomes, but does not impose strict assumptions on the basic distribution of survival data. This approach effectively accounts for both linear and nonlinear risk functions in survival data analysis. The performance of our model in analyzing survival data with various risk functions is on par with or even superior to that of other existing survival analysis methods. Furthermore, we validate the superior performance of our model in comparison to currently existing methods by evaluating multiple publicly available clinical datasets. Through this study, we prove the effectiveness of our proposed model in survival analysis, providing a promising alternative to traditional approaches. The application of deep learning techniques and the ability to capture complex relationships between covariates and survival outcomes without relying on extensive feature engineering make our model a valuable tool for personalized medicine and decision-making in clinical practice.

p16 and p53 can Serve as Prognostic Markers for Head and Neck Squamous Cell Carcinoma

OBJECTIVE

The present study aimed to explore the expression and clinical significance of human papilloma virus-related pathogenic factors (p16, cyclin D1, p53) in patients with head and neck squamous cell carcinoma (HNSCC) and construct a predictive model.

METHODS

The Cancer Genome Atlas was used to obtain clinical data for 112 patients with HNSCC. Expression of p16, p53, and cyclin D1 was quantified. We used the survival package of the R program to set the cut-off value. Values above the cut-off were considered positive, while values below the cut-off were negative. Kaplan-Meier analysis and univariate and multivariate Cox regression analyses were performed to investigate prognostic clinicopathological indicators and the expression of p16, p53, and cyclin D1. A predictive model was constructed based on the results of multifactor Cox regression analysis, and the accuracy of the predictive model was verified through final calibration analysis. Follow-up of patients with HNSCC at the Affiliated Hospital of Binzhou Medical University was conducted from 2015 to 2017, and reliability of the predictive model was validated based on follow-up data and molecular expression levels.

RESULTS

According to the results, expression of p16 and p53 was significantly associated with prognosis (P < .05). The predictive model constructed based on the expression levels of p16 and p53 was useful for evaluating the prognosis of patients with HNSCC. The predictive model was validated using follow-up data obtained from the hospital, and the trend of the follow-up results was consistent with the predictive model.

CONCLUSION

p16 and p53 can be used as key indicators to predict the prognosis of HNSCC patients and as critical immunohistochemical indicators in clinical practice. The survival model constructed based on p16 and p53 expression levels reliably predicts patient prognosis.

An interpretable knee replacement risk assessment system for osteoarthritis patients

Objective

Knee osteoarthritis (OA) is a complex disease with heterogeneous representations. Although it is modifiable to prevention and early treatment, there still lacks a reliable and accurate prognostic tool. Hence, we aim to develop a quantitative and self-administrable knee replacement (KR) risk stratification system for knee osteoarthritis (KOA) patients with clinical features.

Method

A total of 14 baseline features were extracted from 9592 cases in the Osteoarthritis Initiative (OAI) cohort. A survival model was constructed using the Random Survival Forests algorithm. The prediction performance was evaluated with the concordance index (C-index) and average receiver operating characteristic curve (AUC). A three-class KR risk stratification system was built to differentiate three distinct KR-free survival groups. Thereafter, Shapley Additive Explanations (SHAP) was introduced for model explanation.

Results

KR incidence was accurately predicted by the model with a C-index of 0.770 (±0.0215) and an average AUC of 0.807 (±0.0181) with 14 clinical features. Three distinct survival groups were observed from the ten-point KR risk stratification system with a four-year KR rate of 0.79%, 5.78%, and 16.2% from the low, medium, and high-risk groups respectively. KR is mainly caused by pain medication use, age, surgery history, diabetes, and a high body mass index, as revealed by SHAP.

Conclusion

A self-administrable and interpretable KR survival model was developed, underscoring a KR risk scoring system to stratify KOA patients. It will encourage regular self-assessments within the community and facilitate personalised healthcare for both primary and secondary prevention of KOA.

Predictors of divorce and duration of marriage among first marriage women in Dejne administrative town

Divorce is a common occurrence in the marital lives of spouses. Consequently, numerous divorced spouses and their children face various social, economic, physiological, and health problems after breaking their marriage. This study aimed to identify the predictors of divorce and the duration of marriage. We conducted a community-based cross-sectional study among 423 randomly selected residents of Dejen Township in April 2020, of which only 369 respondents met the study inclusion criteria. We used structured questionnaires to collect data. The predictors of divorce and duration of marriage were analyzed using binary logistic regression and the Gompertz regression model, respectively. A p value less than 0.05 was used to express statistical significance. The prevalence of divorce was 21.14% [95% CI (19.01-23.27%)]. Half of these women broke up their marriage after 11 years. A high age difference (7 or more years) between spouses, an early marriage, infertility among women, the presence of third parties, women without formal education, women in the workforce, sexually dissatisfied women, women who did not live together with their husbands at the same address, partner violence, marital control behaviour of husbands, drug-abused husbands, spouses without children, and women who knew multiple sexual partners were the significant predictors of divorce. Partner violence, sexually dissatisfied women, women who made their own marriage decisions, marital control behaviour of husbands, women who did not live together with their husbands at the same address, drug-abused husbands and spouses without children were significant predictors of shorter marriage durations. In this study, the prevalence of divorce was high. Therefore, a community-based, integrated strategy is needed to minimize the divorce rate.

Segmentation of patients with small cell lung cancer into responders and non-responders using the optimal cross-validation technique

BACKGROUND

The timing of treating cancer patients is an essential factor in the efficacy of treatment. So, patients who will not respond to current therapy should receive a different treatment as early as possible. Machine learning models can be built to classify responders and nonresponders. Such classification models predict the probability of a patient being a responder. Most methods use a probability threshold of 0.5 to convert the probabilities into binary group membership. However, the cutoff of 0.5 is not always the optimal choice.

METHODS

In this study, we propose a novel data-driven approach to select a better cutoff value based on the optimal cross-validation technique. To illustrate our novel method, we applied it to three clinical trial datasets of small-cell lung cancer patients. We used two different datasets to build a scoring system to segment patients. Then the models were applied to segment patients into the test data.

RESULTS

We found that, in test data, the predicted responders and non-responders had significantly different long-term survival outcomes. Our proposed novel method segments patients better than the standard approach using a cutoff of 0.5. Comparing clinical outcomes of responders versus non-responders, our novel method had a p-value of 0.009 with a hazard ratio of 0.668 for grouping patients using the Cox proportion hazard model and a p-value of 0.011 using the accelerated failure time model which approved a significant difference between responders and non-responders. In contrast, the standard approach had a p-value of 0.194 with a hazard ratio of 0.823 using the Cox proportion hazard model and a p-value of 0.240 using the accelerated failure time model indicating the responders and non-responders do not differ significantly in survival.

CONCLUSION

In summary, our novel prediction method can successfully segment new patients into responders and non-responders. Clinicians can use our prediction to decide if a patient should receive a different treatment or stay with the current treatment.

Timing to First Whole Blood Transfusion and Survival Following Severe Hemorrhage in Trauma Patients

Importance

Civilian trauma centers have revived interest in whole-blood (WB) resuscitation for patients with life-threatening bleeding. However, there remains insufficient evidence that the timing of WB transfusion when given as an adjunct to a massive transfusion protocol (MTP) is associated with a difference in patient survival outcome.

Objective

To evaluate whether earlier timing of first WB transfusion is associated with improved survival at 24 hours and 30 days for adult trauma patients presenting with severe hemorrhage.

Design, Setting, and Participants

This retrospective cohort study used the American College of Surgeons Trauma Quality Improvement Program databank from January 1, 2019, to December 31, 2020, for adult patients presenting to US and Canadian adult civilian level 1 and 2 trauma centers with systolic blood pressure less than 90 mm Hg, with shock index greater than 1, and requiring MTP who received a WB transfusion within the first 24 hours of emergency department (ED) arrival. Patients with burns, prehospital cardiac arrest, deaths within 1 hour of ED arrival, and interfacility transfers were excluded. Data were analyzed from January 3 to October 2, 2023.

Exposure

Patients who received WB as an adjunct to MTP (earlier) compared with patients who had yet to receive WB as part of MTP (later) at any given time point within 24 hours of ED arrival.

Main Outcomes and Measures

Primary outcomes were survival at 24 hours and 30 days.

Results

A total of 1394 patients met the inclusion criteria (1155 male [83%]; median age, 39 years [IQR, 25-51 years]). The study cohort included profoundly injured patients (median Injury Severity Score, 27 [IQR, 17-35]). A survival curve demonstrated a difference in survival within 1 hour of ED presentation and WB transfusion. Whole blood transfusion as an adjunct to MTP given earlier compared with later at each time point was associated with improved survival at 24 hours (adjusted hazard ratio, 0.40; 95% CI, 0.22-0.73; P = .003). Similarly, the survival benefit of earlier WB transfusion remained present at 30 days (adjusted hazard ratio, 0.32; 95% CI, 0.22-0.45; P < .001).

Conclusions and Relevance

In this cohort study, receipt of a WB transfusion earlier at any time point within the first 24 hours of ED arrival was associated with improved survival in patients presenting with severe hemorrhage. The survival benefit was noted shortly after transfusion. The findings of this study are clinically important as the earlier timing of WB administration may offer a survival advantage in actively hemorrhaging patients requiring MTP.

Discover overlooked complications after preeclampsia using electronic health records

Background

Preeclampsia (PE) is a severe pregnancy complication characterized by hypertension and end-organ damage such as proteinuria. PE poses a significant threat to women's long-term health, including an increased risk of cardiovascular and renal diseases. Most previous studies have been hypothesis-based, potentially overlooking certain significant complications. This study conducts a comprehensive, non-hypothesis-based analysis of PE-complicated diagnoses after pregnancies using multiple large-scale electronic health records (EHR) datasets.

Method

From the University of Michigan (UM) Healthcare System, we collected 4,348 PE patients for the cases and 27,377 patients with pregnancies not complicated by PE or related conditions for the controls. We first conducted a non-hypothesis-based analysis to identify any long-term adverse health conditions associated with PE using logistic regression with adjustments to demographics, social history, and medical history. We confirmed the identified complications with UK Biobank data which contain 443 PE cases and 14,870 non-PE controls. We then conducted a survival analysis on complications that exhibited significance in more than 5 consecutive years post-PE. We further examined the potential racial disparities of identified complications between Caucasian and African American patients.

Findings

Uncomplicated hypertension, complicated diabetes, congestive heart failure, renal failure, and obesity exhibited significantly increased risks whereas hypothyroidism showed decreased risks, in 5 consecutive years after PE in the UM discovery data. UK Biobank data confirmed the increased risks of uncomplicated hypertension, complicated diabetes, congestive heart failure, renal failure, and obesity. Further survival analysis using UM data indicated significantly increased risks in uncomplicated hypertension, complicated diabetes, congestive heart failure, renal failure, and obesity, and significantly decreased risks in hypothyroidism. There exist racial differences in the risks of developing hypertension and hypothyroidism after PE. PE protects against hypothyroidism in African American postpartum women but not Cacausians; it also increases the risks of uncomplicated hypertension but less severely in African American postpartum women as compared to Cacausians.

Interpretation

This study addresses the lack of a comprehensive examination of PE's long-term effects utilizing large-scale EHR and advanced statistical methods. Our findings underscore the need for long-term monitoring and interventions for women with a history of PE, emphasizing the importance of personalized postpartum care. Notably, the racial disparities observed in the impact of PE on hypertension and hypothyroidism highlight the necessity of tailored aftercare based on race.

Whole-transcriptome defines novel glucose metabolic subtypes in colorectal cancer

Colorectal cancer (CRC) is the most prevalent malignancy of the digestive system. Glucose metabolism plays a crucial role in CRC development. However, the heterogeneity of glucose metabolic patterns in CRC is not well characterized. Here, we classified CRC into specific glucose metabolic subtypes and identified the key regulators. 2228 carbohydrate metabolism-related genes were screened out from the GeneCards database, 202 of them were identified as prognosis genes in the TCGA database. Based on the expression patterns of the 202 genes, three metabolic subtypes were obtained by the non-negative matrix factorization clustering method. The C1 subtype had the worst survival outcome and was characterized with higher immune cell infiltration and more activation in extracellular matrix pathways than the other two subtypes. The C2 subtype was the most prevalent in CRC and was characterized by low immune cell infiltration. The C3 subtype had the smallest number of individuals and had a better prognosis, with higher levels of NRF2 and TP53 pathway expression. Secreted frizzled-related protein 2 (SFRP2) and thrombospondin-2 (THBS2) were confirmed as biomarkers for the C1 subtype. Their expression levels were elevated in high glucose condition, while their knockdown inhibited migration and invasion of HCT 116 cells. The analysis of therapeutic potential found that the C1 subtype was more sensitive to immune and PI3K-Akt pathway inhibitors than the other subtypes. To sum up, this study revealed a novel glucose-related CRC subtype, characterized by SFRP2 and THBS2, with poor prognosis but possible therapeutic benefits from immune and targeted therapies.

Long-term Outcomes of Robot-assisted Versus Laparoscopic Surgery for Colon Cancer: A Nationwide Register-based Cohort Study

OBJECTIVE

To determine long-term survival in patients undergoing robot-assisted surgery (RAS) or laparoscopic surgery (LAS) for colon cancer.

BACKGROUND

The potential long-term benefits of RAS compared with LAS for colon cancer are not well examined. Using a register-based approach, we aimed to compare these 2 surgical platforms in an analysis of long-term outcomes, including recurrence-free survival and all-cause- and colon cancer-specific mortality.

METHODS

A nationwide register-based cohort study of patients with Union for International Cancer Control stage I-III colon cancer undergoing planned RAS or LAS from 2010 through 2018. Patient demographic, clinical, and pathological data were retrieved from Danish national registers. Survival and recurrence rates were estimated by Cox proportional hazard multivariate regression analysis adjusting for baseline covariates.

RESULTS

A total of 7565 patients [LAS=6905 (91%) and RAS=660 (9%)] were included in the complete case survival analysis. Patients undergoing LAS had a significantly increased risk of cancer recurrence [LAS=1178 (17.1%), RAS=82 (12.4%), P =0.002] with a mean follow-up time of 4.93 years (standard deviation 2.47). The survival analysis of recurrence-free survival favored RAS [hazard ratio adjusted =0.80, 95% CI (0.64-1.00), P =0.049]. No associations between the 2 surgical platforms were evident regarding all-cause [hazard ratio adjusted =0.98, 95% CI (0.82-1.17), P =0.783] or colon cancer-specific mortality [hazard ratio adjusted =0.89, 95% CI (0.67-1.18), P =0.405].

CONCLUSIONS

Adopting RAS for colon cancer was associated with improved recurrence-free survival. However, it did not cause a lower all-cause- or colon cancer-specific mortality.

The Effects of Exercise-Based Injury Prevention Programmes on Injury Risk in Adult Recreational Athletes: A Systematic Review and Meta-Analysis

BACKGROUND

Injuries are common in adult recreational athletes. Exercise-based injury prevention programmes offer the potential to reduce the risk of injury and have been a popular research topic. Yet, syntheses and meta-analyses on the effects of exercise-based injury prevention programmes for adult recreational athletes are lacking.

OBJECTIVES

We aimed to synthesise and quantify the pooled intervention effects of exercise-based injury prevention programmes delivered to adults who participate in recreation sports.

METHODS

Studies were eligible for inclusion if they included adult recreational athletes (aged > 16 years), an exercise-based intervention and used a randomised controlled trial design. Exclusion criteria were studies without a control group, studies using a non-randomised design and studies including participants who were undertaking activity mandatory for their occupation. Eleven literature databases were searched from earliest record, up to 9 June, 2022. The Physiotherapy Evidence Database (PEDro) scale was used to assess the risk of bias in all included studies. Reported risk statistics were synthesised in a random-effects meta-analysis to quantify pooled treatment effects and associated 95% confidence intervals and prediction intervals.

RESULTS

Sixteen studies met the criteria. Risk statistics were reported as risk ratios [RRs] (n = 12) or hazard ratios [HRs] (n = 4). Pooled estimates of RRs and HRs were 0.94 (95% confidence interval 0.80-1.09) and 0.65 (95% confidence interval 0.39-1.08), respectively. Prediction intervals were 0.80-1.09 and 0.16-2.70 for RR and HR, respectively. Heterogeneity was very low for RR studies, but high for HR studies (tau = 0.29, I2 = 81%). There was evidence of small study effects for RR studies, evidenced by funnel plot asymmetry and Egger's test for small study bias: - 0.99 (CI - 2.08 to 0.10, p = 0.07).

CONCLUSIONS

Pooled point estimates were suggestive of a reduced risk of injury in intervention groups. Nevertheless, these risk estimates were insufficiently precise, too heterogeneous and potentially compromised by small study effects to arrive at any robust conclusion. More large-scale studies are required to clarify whether exercise-based injury prevention programmes are effective in adult recreational athletes.

CLINICAL TRIAL REGISTRATION

The protocol for this review was prospectively registered in the PROSPERO database (CRD42021232697).

Interaction between body mass index and family history of cancer on the risk of female breast cancer

Both body mass index (BMI) and family history of cancer are established risk factors for female breast cancer. However, few studies explored the potential interaction between both factors. We assessed the association of BMI and its interaction with family cancer history on the risk of female breast cancer in Shanghai, China. Based on a population-based prospective cohort study started from 2008 to 2012 with 15,055 Chinese female participants in Minhang district, Shanghai. Cox regression models were used to estimate the association of BMI and its interaction with a family history of cancer on breast cancer risk. The additive interaction was evaluated by the relative excess risk due to interaction (RERI) and the attributable proportion due to interaction (AP), and the multiplicative interaction was assessed by the product term (BMI* family history of cancer) in the Cox regression model. Compared with BMI of < 24 kg/m2 and no family history of cancer, women with BMI of ≥ 24 kg/m2 and a family history of cancer had a higher risk for breast cancer with HR 2.06 (95% CI 1.39, 3.06). There was an additive interaction between BMI and family history of cancer on breast cancer incidence, with the RERI being 0.29 (95% CI 0.08, 0.51) and the AP being 0.37 (95% CI 0.08, 0.66). The coexistence of obesity and cancer family history may exacerbate breast cancer incidence risk, highlighting the importance of weight management in women with a family history of cancer.

Establishment and Validation of a Four-stress Granule-related Gene Signature in Hepatocellular Carcinoma

Background and Aims

Stress granules (SGs) as membrane-less cytoplasmic foci formed in response to unfavorable external stimuli could promote cancer cells to adapt to hostile environments. Hepatocellular carcinoma (HCC) is prone to be highly aggressive once diagnosed, which markedly reduces patient survival time. Therefore, it is crucial to develop valid diagnostic markers to prognosticate HCC patient prognosis, which promotes individualized precision therapeutics in HCC. Considering the pro-tumorigenic activity of SGs, it is of great potential value to construct a prognostic tool for HCC based on the expression profiles of SG-related genes (SGGs).

Methods

Bioinformatic analysis was employed to establish an SGG-based prognostic signature. Western blotting and real-time polymerase chain reaction assays were used to assess the expression patterns of the related SGGs. Loss-of-function experiments were performed to analyze the effect of the SGGs on SG formation and cell survival.

Results

A four-SGG signature (KPNA2, MEX3A, WDR62, and SFN) targeting HCC was established and validated to exhibit a robust performance in predicting HCC prognosis. Consistently, all four genes were further found to be highly expressed in human HCC tissues. More important, we demonstrated that individually knocking down the four SGGs significantly reduced HCC cell proliferation and metastasis by compromising the SG formation process.

Conclusions

We developed an SGG-based predictive signature that can be used as an independent prognostic tool for HCC. The strong predictive power of this signature was further elucidated by the carcinogenic activity of KPNA2, MEX3A, WDR62, and SFN in HCC cells by regulating SG formation.

Deep Survival Analysis for Interpretable Time-Varying Prediction of Preeclampsia Risk

Objective

Survival analysis is widely utilized in healthcare to predict the timing of disease onset. Traditional methods of survival analysis are usually based on Cox Proportional Hazards model and assume proportional risk for all subjects. However, this assumption is rarely true for most diseases, as the underlying factors have complex, non-linear, and time-varying relationships. This concern is especially relevant for pregnancy, where the risk for pregnancy-related complications, such as preeclampsia, varies across gestation. Recently, deep learning survival models have shown promise in addressing the limitations of classical models, as the novel models allow for non-proportional risk handling, capturing nonlinear relationships, and navigating complex temporal dynamics.

Methods

We present a methodology to model the temporal risk of preeclampsia during pregnancy and investigate the associated clinical risk factors. We utilized a retrospective dataset including 66,425 pregnant individuals who delivered in two tertiary care centers from 2015-2023. We modeled the preeclampsia risk by modifying DeepHit, a deep survival model, which leverages neural network architecture to capture time-varying relationships between covariates in pregnancy. We applied time series k-means clustering to DeepHit's normalized output and investigated interpretability using Shapley values.

Results

We demonstrate that DeepHit can effectively handle high-dimensional data and evolving risk hazards over time with performance similar to the Cox Proportional Hazards model, achieving an area under the curve (AUC) of 0.78 for both models. The deep survival model outperformed traditional methodology by identifying time-varied risk trajectories for preeclampsia, providing insights for early and individualized intervention. K-means clustering resulted in patients delineating into low-risk, early-onset, and late-onset preeclampsia groups- notably, each of those has distinct risk factors.

Conclusion

This work demonstrates a novel application of deep survival analysis in time-varying prediction of preeclampsia risk. Our results highlight the advantage of deep survival models compared to Cox Proportional Hazards models in providing personalized risk trajectory and demonstrating the potential of deep survival models to generate interpretable and meaningful clinical applications in medicine.

Improving Clostridioides difficile Infectious Disease Treatment Response via Adherence to Clinical Practice Guidelines

Clostridioides difficile (C. difficile) is a predominant nosocomial infection, and guidelines for improving diagnosis and treatment were published in 2017. We conducted a single-center, retrospective 10-year cohort study of patients with primary C. difficile infectious disease (CDID) at the largest referral Lithuanian university hospital, aiming to evaluate the clinical and laboratory characteristics of CDID and their association with the outcomes, as well as implication of concordance with current Clinical Practice Guidelines. The study enrolled a total of 370 patients. Cases with non-concordant CDID treatment resulted in more CDID-related Intensive Care Unit (ICU) admissions (7.5 vs. 1.8%) and higher CDID-related mortality (13.0 vs. 1.8%) as well as 30-day all-cause mortality (61.0 vs. 36.1%) and a lower 30-day survival compared with CDID cases with concordant treatment (p < 0.05). Among cases defined by two criteria for severe CDID, only patients with non-concordant metronidazole treatment had refractory CDID (68.8 vs. 0.0%) compared with concordant vancomycin treatment. In the presence of non-concordant metronidazole treatment for severe CDID, only cases defined by two severity criteria had more CDID-related ICU admissions (18.8 vs. 0.0%) and higher CDID-related mortality (25.0 vs. 2.0%, p < 0.05) compared with cases defined by one criterion. Severe comorbidities and the continuation of concomitant antibiotics administered at CDID onset reduced (p < 0.05) the 30-day survival and increased (p = 0.053) 30-day all-cause mortality, with 57.6 vs. 10.7% and 52.0 vs. 25.0%, respectively. Conclusions: CDID treatment non-concordant with the guidelines was associated with various adverse outcomes. In CDID with leukocytes ≥ 15 × 109/L and serum creatinine level > 133 µmol/L (>1.5 mg/dL), enteral vancomycin should be used to avoid refractory response, as metronidazole use was associated with CDID-related ICU admission and CDID-related mortality. Severe comorbidities worsened the outcomes as they were associated with reduced 30-day survival. The continuation of concomitant antibiotic therapy increased 30-day all-cause mortality; thus, it needs to be reasonably justified, deescalated or stopped.

Identification of Metabolism-Related Prognostic Biomarkers and Immune Features of Head and Neck Squamous Cell Carcinoma

We aimed to identify an effective metabolic subtype and risk score to predict survival and immunotherapy response in head and neck squamous cell carcinoma (HNSCC). Data were obtained from an online database. We screened significant prognostic metabolism-related genes between the normal and tumor groups using a series of bioinformatics methods. Based on the selected prognostic genes, we conducted a subtype analysis to identify significantly different subtypes in HNSCC. We then investigated survival, immune features, and hallmark differences among different subtypes. LASSO was utilized to identify optimal genes for the risk score model construction. Finally, distribution of the risk score samples was analyzed for different subtypes. A total of 32 significantly prognostic metabolism-related genes were screened, and all samples were grouped into two subtypes: cluster 1 and cluster 2. Cluster 1 had worse survival. Different immune cell infiltration (CD8 T cells, macrophages, and regulatory T cells) and immune checkpoint gene expression (PD-1 and CLAT-4) were observed between the two clusters. Twelve optimal genes were involved in risk score model, and high-risk group had poorer survival. Cluster 1 contained more high-risk samples (60%). Finally, four genes CAV1, GGT6, PYGL, and HS3ST1 were identified as significantly related to immune cells, and these genes were differentially expressed in the normal oral epithelial cells and HNSCC cells. The subtypes and risk score model in the study provide a promising biomarker for prognosis and immunotherapy response.

Interpretable machine learning for predicting chronic kidney disease progression risk

Objective

Chronic kidney disease (CKD) poses a major global health burden. Early CKD risk prediction enables timely interventions, but conventional models have limited accuracy. Machine learning (ML) enhances prediction, but interpretability is needed to support clinical usage with both in diagnostic and decision-making.

Methods

A cohort of 491 patients with clinical data was collected for this study. The dataset was randomly split into an 80% training set and a 20% testing set. To achieve the first objective, we developed four ML algorithms (logistic regression, random forests, neural networks, and eXtreme Gradient Boosting (XGBoost)) to classify patients into two classes-those who progressed to CKD stages 3-5 during follow-up (positive class) and those who did not (negative class). For the classification task, the area under the receiver operating characteristic curve (AUC-ROC) was used to evaluate model performance in discriminating between the two classes. For survival analysis, Cox proportional hazards regression (COX) and random survival forests (RSFs) were employed to predict CKD progression, and the concordance index (C-index) and integrated Brier score were used for model evaluation. Furthermore, variable importance, partial dependence plots, and restrict cubic splines were used to interpret the models' results.

Results

XGBOOST demonstrated the best predictive performance for CKD progression in the classification task, with an AUC-ROC of 0.867 (95% confidence interval (CI): 0.728-0.100), outperforming the other ML algorithms. In survival analysis, RSF showed slightly better discrimination and calibration on the test set compared to COX, indicating better generalization to new data. Variable importance analysis identified estimated glomerular filtration rate, age, and creatinine as the most important predictors for CKD survival analysis. Further analysis revealed non-linear associations between age and CKD progression, suggesting higher risks in patients aged 52-55 and 65-66 years. The association between cholesterol levels and CKD progression was also non-linear, with lower risks observed when cholesterol levels were in the range of 5.8-6.4 mmol/L.

Conclusions

Our study demonstrated the effectiveness of interpretable ML models for predicting CKD progression. The comparison between COX and RSF highlighted the advantages of ML in survival analysis, particularly in handling non-linearity and high-dimensional data. By leveraging interpretable ML for unraveling risk factor relationships, contrasting predictive techniques, and exposing non-linear associations, this study significantly advances CKD risk prediction to enable enhanced clinical decision-making.

Evaluating the Sick Quitting Hypothesis for Frailty Status and Reducing Alcohol Use Among People With HIV in a Longitudinal Clinical Cohort Study

ABSTRACT

"Sick quitting," a phenomenon describing reductions in alcohol consumption following poor health, may explain observations that alcohol appears protective for frailty risk. We examined associations between frailty and reductions in drinking frequency among people with HIV (PWH). At six Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) sites between January 2012 and August 2021, we assessed whether frailty, measured through validated modified frailty phenotype, precedes reductions in drinking frequency. We associated time-updated frailty with quitting and reducing frequency of any drinking and heavy episodic drinking (HED), adjusted for demographic and clinical characteristics in Cox models. Among 5,654 PWH reporting drinking, 60% reported >monthly drinking and 18% reported ≥monthly HED. Over an average of 5.4 years, frail PWH had greater probabilities of quitting (HR: 1.56, 95% confidence interval [95% CI] [1.13-2.15]) and reducing (HR: 1.35, 95% CI [1.13-1.62]) drinking frequency, as well as reducing HED frequency (HR: 1.58, 95% CI [1.20-2.09]) versus robust PWH. Sick quitting likely confounds the association between alcohol use and frailty risk, requiring investigation for control.

Evaluating the Sick Quitting Hypothesis for Frailty Status and Reducing Alcohol Use Among People With HIV in a Longitudinal Clinical Cohort Study

“Sick quitting”, a phenomenon describing reductions in alcohol consumption following poor health, may explain observations that alcohol appears protective for frailty risk. We examined associations between frailty and reductions in drinking frequency among people with HIV (PWH). At six Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) sites between January 2012 and August 2021, we assessed whether frailty, measured via validated modified frailty phenotype, precedes reductions in drinking frequency. We associated time-updated frailty with quitting and reducing frequency of any drinking and heavy episodic drinking (HED), adjusted for demographic and clinical characteristics in Cox models. Among 5,654 PWH reporting drinking, 60% reported >monthly drinking and 18% reported ≥monthly HED. Over an average of 5.4 years, frail PWH had greater probabilities of quitting (HR:1.56, 95%CI:1.13–2.15) and reducing (HR:1.35, 95%CI:1.13–1.62) drinking frequency, as well as reducing HED frequency (HR:1.58, 95%CI:1.20–2.09) vs. robust PWH. Sick quitting likely confounds the association between alcohol use and frailty risk, requiring investigation for control.

Patient Readmission for Ischemic Stroke: Risk Factors and Impact on Mortality

Patient readmission for ischemic stroke significantly strains the healthcare and medical insurance systems. Current understanding of the risk factors associated with these readmissions, as well as their subsequent impact on mortality within China, remains insufficient. This is particularly evident in the context of comprehensive, contemporary population studies. This 4-year retrospective cohort study included 125 397 hospital admissions for ischemic stroke from 838 hospitals located in 22 regions (13 urban and 9 rural) of a major city in western China, between January 1, 2015 and December 31, 2018. The Chi-square tests were used in univariate analysis. Accounting for intra-subject correlations of patients' readmissions, accelerated failure time (AFT) shared frailty models were used to examine readmission events and pure AFT models for mortality. Risk factors for patient readmission after ischemic stroke include frequent admission history, male gender, employee's insurance, advanced age, residence in urban areas, index hospitalization in low-level hospitals, extended length of stay (LOS) during index hospitalization, specific comorbidities and subtypes of ischemic stroke. Furthermore, our findings indicated that an additional admission for ischemic stroke increased patient mortality by 16.4% (P < .001). Stroke readmission contributed to an increased risk of hospital mortality. Policymakers can establish more effective and targeted policies to reduce readmissions for stroke by controlling these risk factors.

Self-Reported Sleep Duration and Bedtime Are Associated with Dementia Risk in Community-Dwelling People Aged 40-74 Years: The Murakami Cohort Study

Background

Sleep is a potentially modifiable factor associated with dementia, including Alzheimer's disease, but current evidence supporting this is insufficient.

Objective

This study aimed to determine whether sleep duration and bedtime patterns are associated with the risk of dementia among middle-aged and older people.

Methods

This cohort study had an eight-year follow-up period. Participants were 13,601 community-dwelling people aged 40-74 years living in Murakami (Niigata, Japan). Data were collected using a self-administered questionnaire. Predictors were self-reported sleep duration and bedtime, and the outcome was newly-diagnosed dementia determined using the long-term care insurance database. Covariates were demographic characteristics, body mass index, smoking, alcohol consumption, total physical activity, insomnia symptoms, disease history, and either bedtime or sleep duration. Cox proportional hazard models were used to calculate hazard ratios (HRs).

Results

The mean age of participants at baseline was 59.2 years. Over a mean follow-up period of 8.0 years, 319 cases of dementia were observed. A long self-reported sleep duration relative to the reference sleep duration (7 hours) was associated with increased dementia risk, with the "8 hours" group (adjusted HR = 1.30, 95% CI:0.99-1.73) and "≥9 hours" group (adjusted HR = 1.46, 95% CI:1.00-2.15) having an increased risk (marginally significant) relative to the reference group. Early bedtime was associated with increased dementia risk (adjusted p for trend = 0.0010), with the "21 : 00 or earlier" group (adjusted HR = 1.61, 95% CI:1.14-2.28) having an increased risk relative to the reference ("23 : 00").

Conclusions

A long self-reported sleep duration and early bedtime are both associated with increased dementia risk in middle-aged and older people.

Extensive prediction of drug response in mutation-subtype-specific LUAD with machine learning approach

Background

Lung cancer is the most prevalent cancer diagnosis and the leading cause of cancer death worldwide. Therapeutic failure in lung cancer (LUAD) is heavily influenced by drug resistance. This challenge stems from the diverse cell populations within the tumor, each having unique genetic, epigenetic, and phenotypic profiles. Such variations lead to varied therapeutic responses, thereby contributing to tumor relapse and disease progression.

Methods

The Genomics of Drug Sensitivity in Cancer (GDSC) database was used in this investigation to obtain the mRNA expression dataset, genomic mutation profile, and drug sensitivity information of NSCLS. Machine Learning (ML) methods, including Random Forest (RF), Artificial Neurol Network (ANN), and Support Vector Machine (SVM), were used to predict the response status of each compound based on the mRNA and mutation characteristics determined using statistical methods. The most suitable method for each drug was proposed by comparing the prediction accuracy of different ML methods, and the selected mRNA and mutation characteristics were identified as molecular features for the drug-responsive cancer subtype. Finally, the prognostic influence of molecular features on the mutational subtype of LUAD in publicly available datasets.

Results

Our analyses yielded 1,564 gene features and 45 mutational features for 46 drugs. Applying the ML approach to predict the drug response for each medication revealed an upstanding performance for SVM in predicting Afuresertib drug response (area under the curve [AUC] 0.875) using CIT, GAS2L3, STAG3L3, ATP2B4-mut, and IL15RA-mut as molecular features. Furthermore, the ANN algorithm using 9 mRNA characteristics demonstrated the highest prediction performance (AUC 0.780) in Gefitinib with CCL23-mut.

Conclusion

This work extensively investigated the mRNA and mutation signatures associated with drug response in LUAD using a machine-learning approach and proposed a priority algorithm to predict drug response for different drugs.

Targeting eIF4A triggers an interferon response to synergize with chemotherapy and suppress triple-negative breast cancer

Protein synthesis is frequently dysregulated in cancer and selective inhibition of mRNA translation represents an attractive cancer therapy. Here, we show that therapeutically targeting the RNA helicase eIF4A with zotatifin, the first-in-class eIF4A inhibitor, exerts pleiotropic effects on both tumor cells and the tumor immune microenvironment in a diverse cohort of syngeneic triple-negative breast cancer (TNBC) mouse models. Zotatifin not only suppresses tumor cell proliferation but also directly repolarizes macrophages toward an M1-like phenotype and inhibits neutrophil infiltration, which sensitizes tumors to immune checkpoint blockade. Mechanistic studies revealed that zotatifin reprograms the tumor translational landscape, inhibits the translation of Sox4 and Fgfr1, and induces an interferon (IFN) response uniformly across models. The induction of an IFN response is partially due to the inhibition of Sox4 translation by zotatifin. A similar induction of IFN-stimulated genes was observed in breast cancer patient biopsies following zotatifin treatment. Surprisingly, zotatifin significantly synergizes with carboplatin to trigger DNA damage and an even heightened IFN response, resulting in T cell-dependent tumor suppression. These studies identified a vulnerability of eIF4A in TNBC, potential pharmacodynamic biomarkers for zotatifin, and provide a rationale for new combination regimens consisting of zotatifin and chemotherapy or immunotherapy as treatments for TNBC.

Application of parametric survival analysis to women patients with breast cancer at Jimma University Medical Center

Public health systems in both industrialized and undeveloped countries continue to struggle with the worldwide problem of breast cancer. In sub-Saharan African countries, notably Ethiopia, it is the form of cancer that strikes women the most commonly. Despite the extreme difficulties, the causes of mortality in Ethiopia have not yet been identified. In addition, little study has been done in this area. Therefore, the major objective of this analysis was to pinpoint the factors that were most responsible for the decreased life expectancy of breast cancer patients at the University of Jimma Medical Center. 552 women who had been treated for breast cancer at Jimma University Medical Center between October 2018 and December 2022 were included in this study, which used a retrospective cohort study design and five-year follow-up data. The most frequent and widely used test for comparing the probability of survival curves between several categorical independent variables was the log-rank test. Next, semi-parametric methods for multivariable analysis using the Cox proportional hazards model were used. Furthermore, a parametric strategy that includes fully parametric survival models better achieves the goal of the analysis. Among covariate, age of patient (ϕ = 254.06; 95% CI (3.95, 7.13), P-value = 0.000), patient live in urban (ϕ = 0.84; 95% CI (-0.35,-0.00), P-value = 0.047), preexisting comorbidity (ϕ = 2.46; 95% CI (0.39, 1.41), P-value = 0.001), overweight women cancer patient (ϕ = 0.05; 95% CI(-4.41,-1.57), P-value = 0.000, positive Axillary Node status cancer patient (ϕ = 0.04; 95% CI(-4.45,-1.88), P-value = 0.000), both surgery and chemotropic baseline treatment patient (ϕ = 0.53; 95% CI(-1.12,-0.16), P-value = 0.009) significantly affected the survival of women breast cancer. Age of breast cancer patient, patient education level, place of residence, marital status, pre-existing comorbidity, axillary node status, estrogen receptor, tumor size, body mass index at diagnosis, stage of cancer, and baseline treatment were found to have a significant effect on time to survive for women with breast cancer at the University of Jimma Medical Center, Oromia region, Ethiopia. However, the covariate histologic grade, number of positive lymph nodes involved, and type of hormone used were insignificant to the survival of breast cancer patients.

Recombinant human erythropoietin protects against immature brain damage induced by hypoxic/ischemia insult

To investigate the neuroprotection of recombinant human erythropoietin (rhEPO) against hypoxic/ischemic (HI) insult in three-day-old rats. Postnatal day 3 (PD3) rats were randomly divided into three groups: Sham group, HI group and HI+rhEPO group. Ligation of the right common carotid artery and hypoxia to induce HI brain injury. After HI insult, the rats received intraperitoneal injection of rhEPO (5000 IU/Kg, qod) in HI+rhEPO group or equal saline in other groups. On PD10, damage of brain tissue was examined by hematoxylin-eosin (HE) staining, observation of neuronal apoptosis in the hippocampus and cortex using immunofluorescence assay (marker: TUNEL). Immunohistochemical staining or western blotting was performed to detect the expression of cyclooxygenase-2 (COX-2), Caspase-3 and phosphorylated Akt (p-Akt) protein. On PD28, cognitive ability of rats was assessed by Morris water maze test. HI injury causes brain pathological morphology and cognitive function damage in PD3 rats, which can be alleviated by rhEPO intervention. Compared with the HI group, the HI+rhEPO group showed an increase in platform discovery rate and cross platform frequency, while the search platform time was shortened (P < 0.05). The proportion of TUNEL positive neurons and the expression of COX-2 and Caspase-3 proteins in brain tissue in the hippocampus and cortex was decreased, while the expression of p-Akt protein was upregulated (P < 0.05). RhEPO could protect against the pathological and cognitive impairment of immature brain induced by HI insult. This neuroprotective activity may involve in inhibiting inflammatory and apoptosis by activation of PI3K/Akt signaling pathway.

Estimation of the onset time of diabetic complications in type 2 diabetes patients in Thailand: a survival analysis

BACKGROUND

This study aimed to identify factors associated with the onset time of diabetic complications in patients with type 2 diabetes mellitus (T2DM) and determine the best-fitted survival model.

METHODS

A retrospective cohort study was conducted among T2DM patients enrolled from October 1, 2016 to July 15, 2020 at the National Health Security Office (NHSO). In total, 388 T2DM patients were included. Cox proportional-hazard and parametric models were used to identify factors related to the onset time of diabetic complications. The Akaike information criterion, Bayesian information criterion, and Cox-Snell residual were compared to determine the best-fitted survival model.

RESULTS

Thirty diabetic complication events were detected among the 388 patients (7.7%). A 90% survival rate for the onset time of diabetic complications was found at 33 months after the first T2DM diagnosis. According to multivariate analysis, a duration of T2DM ≥42 months (time ratio [TR], 0.56; 95% confidence interval [CI], 0.33-0.96; p=0.034), comorbid hypertension (TR, 0.30; 95% CI, 0.15-0.60; p=0.001), mildly to moderately reduced levels of the estimated glomerular filtration rate (eGFR) (TR, 0.43; 95% CI, 0.24-0.75; p=0.003) and an eGFR that was severely reduced or indicative of kidney failure (TR, 0.38; 95% CI, 0.16-0.88; p=0.025) were significantly associated with the onset time of diabetic complications (p<0.05).

CONCLUSION

Patients with T2DM durations of more than 42 months, comorbid hypertension, and decreased eGFR were at risk of developing diabetic complications. The NHSO should be aware of these factors to establish a policy to prevent diabetic complications after the diagnosis of T2DM.

Professionalism in Physician Assistant Education as a Predictor of Future Licensing Board Disciplinary Actions

PURPOSE

The purpose of this study was to evaluate associations between postgraduate disciplinary actions (PGDA) by state licensing boards and physician assistant (PA) school documented professionalism violations (DPV) and academic probation.

METHODS

This was a retrospective cohort study comprising PA graduates from 2001 to 2011 at 3 institutions (n = 1364) who were evaluated for the main outcome of PGDA and independent variable of DPV and academic probation. Random-effects multiple logistic regression and accelerated failure time parametric survival analysis were used to investigate the association of PGDA with DPV and academic probation.

RESULTS

Postgraduate disciplinary action was statistically significant and positively associated with DPV when unadjusted (odds ratio [OR] = 5.15; 95% CI: 1.62-16.31; P = .01) and when adjusting for age, sex, overall PA program GPA (GPA), and Physician Assistant National Certifying Exam Score (OR = 5.39; 95% CI: 1.54-18.85; P = .01) (fully adjusted). Academic probation increased odds to 8.43 times (95% CI: 2.85-24.92; P < .001) and 9.52 times (95% CI: 2.38-38.01; P < .001) when fully adjusted.

CONCLUSION

Students with professionalism violation or academic probation while in the PA school had significant higher odds of receiving licensing board disciplinary action compared with those who did not. Academic probation had a greater magnitude of effect and could represent an intersection of professionalism and academic performance.

Treatment Outcomes and Risk Factors for Incomplete Treatment after Definitive Chemoradiotherapy for Non-Resectable or Metastatic Esophageal Cancer

Among patients with unresectable or metastatic esophageal cancer who receive definitive chemotherapy or chemoradiotherapy, the rates of treatment-related adverse events and incomplete treatment remain high. We conducted this study to investigate survival after definitive treatments and identify predicting factors for incomplete treatment. The data of patients who received definitive chemotherapy or chemoradiotherapy for esophageal cancer were retrospectively examined. The patients were assigned to Group 1: incomplete definitive treatment; Group 2: complete definitive treatment; or Group 3: complete definitive treatment with additional salvage surgery. The data of 273 patients (90, 166, and 17 in Groups 1, 2, and 3, respectively) were analyzed. In the survival analysis, the median overall survival of Groups 1, 2, and 3 were 2.6, 10.3, and 29.5 months, respectively. A significant difference in 3-year overall survival was observed among the groups (2.2%, 12.4%, and 48.5%, p < 0.001). In multivariable analysis, the independent risk factors for incomplete definitive treatment included poor performance score (hazard ratio (HR): 5.23, p = 0.001), bone metastasis (HR: 2.18, p = 0.024), airway invasion (HR: 2.90, p = 0.001), and liver cirrhosis (HR: 3.20, p = 0.026). Incomplete definitive treatment is associated with a far worse prognosis. Poor performance, bone metastasis, airway invasion, and liver cirrhosis are risk factors for incomplete treatment.

Classification of patients with osteoarthritis through clusters of comorbidities using 633 330 individuals from Spain

OBJECTIVES

To explore clustering of comorbidities among patients with a new diagnosis of OA and estimate the 10-year mortality risk for each identified cluster.

METHODS

This is a population-based cohort study of individuals with first incident diagnosis of OA of the hip, knee, ankle/foot, wrist/hand or 'unspecified' site between 2006 and 2020, using SIDIAP (a primary care database representative of Catalonia, Spain). At the time of OA diagnosis, conditions associated with OA in the literature that were found in ≥1% of the individuals (n = 35) were fitted into two cluster algorithms, k-means and latent class analysis. Models were assessed using a range of internal and external evaluation procedures. Mortality risk of the obtained clusters was assessed by survival analysis using Cox proportional hazards.

RESULTS

We identified 633 330 patients with a diagnosis of OA. Our proposed best solution used latent class analysis to identify four clusters: 'low-morbidity' (relatively low number of comorbidities), 'back/neck pain plus mental health', 'metabolic syndrome' and 'multimorbidity' (higher prevalence of all studied comorbidities). Compared with the 'low-morbidity' cluster, the 'multimorbidity' cluster had the highest risk of 10-year mortality (adjusted hazard ratio [HR]: 2.19 [95% CI: 2.15, 2.23]), followed by the 'metabolic syndrome' cluster (adjusted HR: 1.24 [95% CI: 1.22, 1.27]) and the 'back/neck pain plus mental health' cluster (adjusted HR: 1.12 [95% CI: 1.09, 1.15]).

CONCLUSION

Patients with a new diagnosis of OA can be clustered into groups based on their comorbidity profile, with significant differences in 10-year mortality risk. Further research is required to understand the interplay between OA and particular comorbidity groups, and the clinical significance of such results.

Construction and validation of a immune-related prognostic gene DHRS1 in hepatocellular carcinoma based on bioinformatic analysis

A member of the short-chain dehydrogenase/reductase superfamily (DHRS1, SDR19C1) is a member of the short-chain dehydrogenase/reductase superfamily and a potential predictor of hepatocellular carcinoma (HCC). However, the role of DHRS1 in HCC immunity remains unclear. We systematically analyzed the association between DHRS1 and HCC immunity with transcriptional and clinical data from the Tumor Immune Estimation Resource, an integrated repository portal for tumor immune system interactions, and cBioPortal databases. Six DHRS1-associated immunomodulators strongly correlated with survival and were uncovered by exploiting univariate and multivariate Cox analyses. We created a risk score for each patient by adding the points from each immunomodulator and then classified them into high and low risk categories. Survival analysis were used to compare the overall survival between the 2 groups, and the receiver operating characteristic curve was applied to assess the accuracy of the risk score. Data from our center were adopted as the external validation set, the risk score was calculated using the risk coefficient of the 6 genes in the training cohort, and survival analysis were executed to verify the experimental group results. A nomogram was ultimately constructed with the R package. Our data revealed a correlation between the levels of immune cell infiltration and either the DHRS1 gene copy numbers or mRNA levels in HCC. Second, we generated a signature based on the 6 DHRS1-related immunomodulators (KDR, TNFRSF4, CD276, TNFSF4, SLAMF6, and SIGLEC9). We postulate that the generated risk scores would serve as an independent indicator of HCC prognosis, with an area under the receiver operating characteristic curve for the risk score of 0.743. We further established external validation sets to reconfirm the predictive validity of the risk score. Finally, a prognostic nomogram and calibration curve were created. The DHRS1 gene may exert an impact on HCC immunity. We posit that the nominated immune signature based on DHRS1-associated immunomodulators could constitute a promising prognostic biomarker in HCC.

Injury risk functions for the four primary knee ligaments

The purpose of this study was to develop injury risk functions (IRFs) for the anterior and posterior cruciate ligaments (ACL and PCL, respectively) and the medial and lateral collateral ligaments (MCL and LCL, respectively) in the knee joint. The IRFs were based on post-mortem human subjects (PMHSs). Available specimen-specific failure strains were supplemented with statistically generated failure strains (virtual values) to accommodate for unprovided detailed experimental data in the literature. The virtual values were derived from the reported mean and standard deviation in the experimental studies. All virtual and specimen-specific values were thereafter categorized into groups of static and dynamic rates, respectively, and tested for the best fitting theoretical distribution to derive a ligament-specific IRF. A total of 10 IRFs were derived (three for ACL, two for PCL, two for MCL, and three for LCL). ACL, MCL, and LCL received IRFs in both dynamic and static tensile rates, while a sufficient dataset was achieved only for dynamic rates of the PCL. The log-logistic and Weibull distributions had the best fit (p-values: >0.9, RMSE: 2.3%-4.7%) to the empirical datasets for all the ligaments. These IRFs are, to the best of the authors' knowledge, the first attempt to generate injury prediction tools based on PMHS data for the four knee ligaments. The study has summarized all the relevant literature on PHMS experimental tensile tests on the knee ligaments and utilized the available empirical data to create the IRFs. Future improvements require upcoming experiments to provide comparable testing and strain measurements. Furthermore, emphasis on a clear definition of failure and transparent reporting of each specimen-specific result is necessary.

Tobacco Smoking and Pack-Years Are Associated With Frailty Among People With HIV

BACKGROUND

Tobacco smoking increases frailty risk among the general population and is common among people with HIV (PWH) who experience higher rates of frailty at younger ages than the general population.

METHODS

We identified 8608 PWH across 6 Centers for AIDS Research Network of Integrated Clinical Systems sites who completed ≥2 patient-reported outcome assessments, including a frailty phenotype measuring unintentional weight loss, poor mobility, fatigue, and inactivity, and scored 0-4. Smoking was measured as baseline pack-years and time-updated never, former, or current use with cigarettes/day. We used Cox models to associate smoking with risk of incident frailty (score ≥3) and deterioration (frailty score increase by ≥2 points), adjusted for demographics, antiretroviral medication, and time-updated CD4 count.

RESULTS

The mean follow-up of PWH was 5.3 years (median: 5.0), the mean age at baseline was 45 years, 15% were female, and 52% were non-White. At baseline, 60% reported current or former smoking. Current (HR: 1.79; 95% confidence interval: 1.54 to 2.08) and former (HR: 1.31; 95% confidence interval: 1.12 to 1.53) smoking were associated with higher incident frailty risk, as were higher pack-years. Current smoking (among younger PWH) and pack-years, but not former smoking, were associated with higher risk of deterioration.

CONCLUSIONS

Among PWH, smoking status and duration are associated with incident and worsening frailty.

Quantitative Analysis on Molecular Characteristics Evolution of Gastric Cancer Progression and Prognosis

The dynamic changes of key biological characteristics from gastric low-grade intraepithelial neoplasia (LGIN) to high-grade intraepithelial neoplasia (HGIN) to early gastric cancer (EGC) are still unclear, which greatly affect the accurate diagnosis and treatment of EGC and prognosis evaluation of gastric cancer (GC). In this study, bioinformatics methods/tools are applied to quantitatively analyze molecular characteristics evolution of GC progression, and a prognosis model is constructed. This study finds that some dysregulated differentially expressed mRNAs (DEmRNAs) in the LGIN stage may continue to promote the occurrence and development of EGC. Among the LGIN, HGIN, and EGC stages, there are differences and relevance in the transcription expression patterns of DEmRNAs, and the activation related to immune cells is very different. The biological functions continuously changed during the progression from LGIN to HGIN to EGC. The COX model constructed based on the three EGC-related DEmRNAs has GC prognostic risk prediction ability. The evolution of biological characteristics during the development of EGC mined by the authors provides new insight into understanding the molecular mechanism of EGC occurrence and development. The three-gene prognostic risk model provides a new method for assisting GC clinical treatment decisions.