Brief introduction of medical database and data mining technology in big data era

Development of lactate-related gene signature and prediction of overall survival and chemosensitivity in patients with colorectal cancer

BACKGROUND

Colorectal cancer (CRC) is a malignant tumor of the digestive system that contains high levels of immune cells. Lactic acid, a major metabolite, plays a crucial role in tumor development, maintenance, and therapeutic response. However, the prognostic potential and therapeutic biomarker potential of lactate-related genes (LRGs) in CRC patients remain to be elucidated.

METHODS

We collected the mRNA expression profile and clinical data of CRC patients from the Cancer Genome Atlas (TCGA) database and the GSE59382 cohort. Univariate Cox regression, Lasso regression and multivariate Cox regression analysis were used to construct the prognosis model. Combined with the risk score and important clinicopathological features, the nomogram was established. In addition, the relationship between risk score and immune infiltration, immune checkpoint gene expression, and drug sensitivity was investigated.

RESULTS

We constructed lactate-related gene signatures (LRGS) based on four LRGs, which independently predicted the prognosis of CRC. Patients with different risk scores are found to have distinct immune status, tumor mutation load, immune response, and drug sensitivity. In addition, nomogram results determined by risk scores and clinical factors have higher predictive performance.

CONCLUSION

We found that LRGS is a reliable biomarker for predicting clinical outcomes, evaluating immune infiltration and efficacy, and predicting the sensitivity to drugs in patients with CRC.

The nomogram to predict the occurrence of sepsis-associated encephalopathy in elderly patients in the intensive care units: A retrospective cohort study

Background

Sepsis-associated encephalopathy (SAE) is a critical and common problem in elderly patients with sepsis, which is still short of efficient predictive tools. Therefore, this study aims to screen the risk factors and establish a useful predictive nomogram for SAE in elderly patients with sepsis in the intensive care unit (ICU).

Patients and methods

Elderly patients (age ≥ 65 years) with sepsis were selected from the Medical Information Mart for Intensive Care (MIMIC)-IV database. Data from demographics and laboratory examinations were collected on the first day of admission to the ICU. SAE was defined by two criteria in the presence of sepsis: ① a Glasgow Coma Scale (GCS) score of < 15 or ② delirium. Differences in demographics and laboratory tests were calculated between SAE and non-SAE groups. Participants were randomly divided into a training set and a validation set without replacement at a ratio of 6:4. A predictive nomogram was constructed in the training set by logistic regression analysis and then validated. The predictive capability of the nomogram was demonstrated by receiver operating characteristic (ROC) analysis and calibration curve analysis.

Results

A total of 22,361 patients were selected, of which 2,809 patients (12.7%) died in the hospital and 8,290 patients (37.1%) had SAE. In-hospital mortality in the SAE group was higher than that in the non-SAE group (18.8 vs. 8.9%, p < 0.001). Based on the results of logistic regression analysis, a nomogram integrating age, Na⁺, Sequential Organ Failure Assessment (SOFA) score, heart rate, and body temperature were constructed. The area under the curve (AUC) of the nomogram was 80.2% in the training set and 80.9% in the validation set. Calibration curve analysis showed a good predictive capacity of the nomogram.

Conclusion

SAE is an independent risk of in-hospital mortality in elderly patients in the intensive care unit. The nomogram has an excellent predictive capability of SAE and helps in clinical practice.

[Keyword Network Analysis and Topic Modeling of News Articles Related to Artificial Intelligence and Nursing]

PURPOSE

The purpose of this study was to identify the main keywords, network properties, and main topics of news articles related to artificial intelligence technology in the field of nursing.

METHODS

After collecting artificial intelligence-and nursing-related news articles published between January 1, 1991, and July 24, 2022, keywords were extracted via preprocessing. A total of 3,267 articles were searched, and 2,996 were used for the final analysis. Text network analysis and topic modeling were performed using NetMiner 4.4.

RESULTS

As a result of analyzing the frequency of appearance, the keywords used most frequently were education, medical robot, telecom, dementia, and the older adults living alone. Keyword network analysis revealed the following results: a density of 0.002, an average degree of 8.79, and an average distance of 2.43; the central keywords identified were 'education,' 'medical robot,' and 'fourth industry.' Five topics were derived from news articles related to artificial intelligence and nursing: 'Artificial intelligence nursing research and development in the health and medical field,' 'Education using artificial intelligence for children and youth care,' 'Nursing robot for older adults care,' 'Community care policy and artificial intelligence,' and 'Smart care technology in an aging society.'

CONCLUSION

The use of artificial intelligence may be helpful among the local community, older adult, children, and adolescents. In particular, health management using artificial intelligence is indispensable now that we are facing a super-aging society. In the future, studies on nursing intervention and development of nursing programs using artificial intelligence should be conducted.

Infections in Acute Pancreatitis: Organisms, Resistance-Patterns and Effect on Mortality

BACKGROUND

Acute pancreatitis (AP) is a common gastrointestinal disease in which infection is a serious complication. Understanding its bacterial spectrum and antibiotic resistance is of great significance for treatment.

OBJECTIVE

This retrospective study analyzed the Medical Information Mart for Intensive Care database with the aim of identifying the distribution characteristics of pathogenic bacteria in AP patients.

METHODS

First, 2089 AP patients were classified and analyzed statistically according to culture results. Second, the bacterial profile, antibiotic resistance, and antibiotic-use data of positive culture group were analyzed. Third, logistic regression analysis was used to identify the rick factors of culture results, and culture results correlations with mortality.

RESULTS

This study included 1486 patients in negative culture group, 603 patients in positive cultures. Enterococcus spp. (71%), Enterococcus faecium (61%), and Staphylococcus aureus coagulase-positive (54%) exhibited the highest proportions of drug resistance. Logistic regression revealed five factors related to positive culture (the number of antibiotics, length of stay in hospital, length of stay in intensive care unit, mechanical ventilation, and parenteral nutrition) and four factors related to distribution of multidrug-resistant bacterial infection (age, hemoglobin, length of stay in hospital, and duration on antibiotics).

CONCLUSIONS

This study found that enteric bacteria were the most common source of infection (26.7%). Carbapenems, penicillins containing enzyme inhibitors, fifth-generation cephalosporins, oxazolidinones, and some of the aminoglycoside antibiotics had high sensitivity, which can guide the use of clinical empiric antibiotics. Positive culture was an independent risk factor for in-hospital all-cause mortality of AP patients.

Validation of data mining models by comparing with conventional methods for dental age estimation in Korean juveniles and young adults

Teeth are known to be the most accurate age indicators of human body and are frequently applied in forensic age estimation. We aimed to validate data mining-based dental age estimation, by comparing the accuracy of the estimation and classification performance of 18-year thresholds with conventional methods and with data mining-based age estimation. A total of 2657 panoramic radiographs were collected from Koreans and Japanese populations aged 15 to 23 years. They were subdivided into a training and internal test set of 900 radiographs each from Koreans, and an external test set of 857 radiographs from Japanese. We compared the accuracy and classification performance of the test sets from conventional methods with those from the data mining models. The accuracy of the conventional method with the internal test set was slightly higher than that of the data mining models, with a slight difference (mean absolute error < 0.21 years, root mean square error < 0.24 years). The classification performance of the 18-year threshold was also similar between the conventional method and the data mining models. Thus, conventional methods can be replaced by data mining models in forensic age estimation using second and third molar maturity of Korean juveniles and young adults.

Association between glucocorticoid use and all-cause mortality in critically ill patients with heart failure: A cohort study based on the MIMIC-III database

Background: Heart failure (HF) is the terminal stage of various heart diseases. Conventional treatments have poor efficacy, and diuretic resistance can present. Previous studies have found that the use of glucocorticoids can enhance the diuretic effect of patients with heart failure and reduce heart failure symptoms. However, the relationship between glucocorticoid use and mortality in patients with heart failure in intensive care units is unclear. Objectives: The aim of this study was to determine the association between glucocorticoid use and all-cause mortality in critically ill patients with heart failure. Methods: The information on patients with heart failure in this study was extracted from the MIMIC-III (Medical Information Mart for Intensive Care-III) database. Patients in the glucocorticoid and non-glucocorticoid groups were matched using propensity scores. The Kaplan-Meier method was used to explore the difference in survival probability between the two groups. A Cox proportional-hazards regression model was used to analyze the hazard ratios (HRs) for the two patient groups. Subgroup analyses were performed with prespecified stratification variables to demonstrate the robustness of the results. Results: The study included 9,482 patients: 2,099 in the glucocorticoid group and 7,383 in the non-glucocorticoid group. There were 2,055 patients in each group after propensity-score matching. The results indicated that the non-glucocorticoid group was not significantly associated with reduced mortality in patients with heart failure during the 14-day follow-up period [HRs = .901, 95% confidence interval (CI) = .767-1.059]. During the follow-up periods of 15-30 and 15-90 days, the mortality risk was significantly lower in the non-glucocorticoid group than in the glucocorticoid group (HRs = .497 and 95% CI = .370-.668, and HRs = .400 and 95% CI = .310-.517, respectively). Subgroup analyses indicated no interaction among each stratification variable and glucocorticoid use. Conclusion: Glucocorticoid use was associated with an increased mortality risk in critically ill patients with heart failure.

New Personal Model for Forecasting the Outcome of Patients with Histological Grade III-IV Colorectal Cancer Based on Regional Lymph Nodes

Background

Metastases at regional lymph nodes could easily occur in patients with high-histological-grade colorectal cancer (CRC). However, few models were built on the basis of lymph nodes to predict the outcome of patients with histological grades III-IV CRC.

Methods

Data in the Surveillance, Epidemiology, and End Results databases were used. Univariate and multivariate analyses were performed. A personalized prediction model was built in accordance with the results of the analyses. A nomogram was tested in two datasets and assessed using a calibration curve, a consistency index (C-index), and an area under the curve (AUC).

Results

A total of 14,039 cases were obtained from the database. They were separated into two groups (9828 cases for constructing the model and 4211 cases for validation). Logistic and Cox regression analyses were then conducted. Factors such as log odds of positive lymph nodes (LODDS) were utilized. Then, a personalized prediction model was established. The C-index in the construction and validation groups was 0.770. The 1-, 3-, and 5-year AUCs were 0793, 0.828, and 0.830 in the construction group, respectively, and 0.796, 0.833, and 0.832 in the validation group, respectively. The calibration curves showed well consistency in the 1-, 3- and 5-year OS between prediction and reality in both groups.

Conclusion

The nomogram built based on LODDS exhibited considerable reliability and accuracy.

Associations between dietary and blood inflammatory indices and their effects on cognitive function in elderly Americans

Objective

To determine the correlations between dietary and blood inflammation indices in elderly Americans and their effects on cognitive function.

Methods

This research extracted data from the 2011-2014 National Health and Nutrition Examination Survey for 2,479 patients who were ≥60 years old. Cognitive function was assessed as a composite cognitive function score (Z-score) calculated from the results of the Consortium to Establish a Registry for Alzheimer's Disease Word Learning and Delayed Recall tests, the Animal Fluency test, and the Digit Symbol Substitution Test. We used a dietary inflammatory index (DII) calculated from 28 food components to represent the dietary inflammation profile. Blood inflammation indicators included the white blood cell count (WBC), neutrophil count (NE), lymphocyte count (Lym), neutrophil-lymphocyte ratio (NLR), platelet-lymphocyte ratio (PLR), neutrophil-albumin ratio (NAR), systemic immune-inflammation index [SII, calculated as (peripheral platelet count) × NE/Lym], and systemic inflammatory response index [SIRI, calculated as (monocyte count) × NE/Lym]. WBC, NE, Lym, NLR, PLR, NAR, SII, SIRI, and DII were initially treated as continuous variables. For logistic regression, WBC, NE, Lym, NLR, PLR, NAR, SII, and SIRI were divided into quartile groups, and DII was divided into tertile groups.

Results

After adjusting for covariates, WBC, NE, NLR, NAR, SII, SIRI, and DII scores were markedly higher in the cognitively impaired group than in the normal group (p < 0.05). DII was negatively correlated with the Z-score when combined with WBC, NE, and NAR (p < 0.05). After adjusting for all covariates, DII was positively correlated with SII in people with cognitive impairment (p < 0.05). Higher DII with NLR, NAR, SII, and SIRI all increased the risk of cognitive impairment (p < 0.05).

Conclusion

DII was positively correlated with blood inflammation indicators, and higher DII and blood inflammation indicators increased the risk of developing cognitive impairment.

Prognosis of the Keratinizing Squamous Cell Carcinoma of the Tongue Based on Surveillance, Epidemiology, and End Results Database

BACKGROUND

The objective of this study is to determine the prognostic factors of keratinizing squamous cell carcinoma of the tongue (KTSCC) and to establish a prognostic nomogram of KTSCC to assist clinical diagnosis and treatment.

METHODS

This study identified 3874 patients with KTSCC from the Surveillance, Epidemiology, and End Results (SEER) database, and these patients were randomly divided into the training (70%, (n = 2711) and validation (30%, n = 1163) cohorts. Cox regression was then used to filter variables. Nomograms were then constructed based on meaningful variables. Finally, the concordance index (C-index), net reclassification index (NRI), integrated discrimination improvement (IDI), calibration charts, and decision-curve analysis (DCA), were used to evaluate the discrimination, accuracy and effectiveness of the model.

RESULTS

A nomogram model was established for predicting the 3-, 5-, and 8-year overall survival (OS) probabilities of patients with KTSCC. The model indicated that age, radiotherapy sequence, SEER stage, marital status, tumor size, American Joint Committee on Cancer (AJCC) stage, radiotherapy status, race, lymph node dissection status, and sex were factors influencing the OS of patients with KTSCC. Verified by C-index, NRI, IDI, calibration curve, and DCA curve, our model has better discrimination, calibration, accuracy and net benefit compared to the AJCC system.

CONCLUSIONS

This study identified the factors that affect the survival of KTSCC patients and established a prognostic nomogram that can help clinicians predict the 3-, 5-, and 8-year survival rates of KTSCC patients.

Cardiopulmonary prognosis of prophylactic endotracheal intubation in patients with upper gastrointestinal bleeding undergoing endoscopy

BACKGROUND

It is controversial whether prophylactic endotracheal intubation (PEI) protects the airway before endoscopy in critically ill patients with upper gastrointestinal bleeding (UGIB). The study aimed to explore the predictive value of PEI for cardiopulmonary outcomes and identify high-risk patients with UGIB undergoing endoscopy.

METHODS

Patients undergoing endoscopy for UGIB were retrospectively enrolled in the eICU Collaborative Research Database (eICU-CRD). The composite cardiopulmonary outcomes included aspiration, pneumonia, pulmonary edema, shock or hypotension, cardiac arrest, myocardial infarction, and arrhythmia. The incidence of cardiopulmonary outcomes within 48 h after endoscopy was compared between the PEI and non-PEI groups. Logistic regression analyses and propensity score matching analyses were performed to estimate effects of PEI on cardiopulmonary outcomes. Moreover, restricted cubic spline plots were used to assess for any threshold effects in the association between baseline variables and risk of cardiopulmonary outcomes (yes/no) in the PEI group.

RESULTS

A total of 946 patients were divided into the PEI group (108/946, 11.4%) and the non-PEI group (838/946, 88.6%). After propensity score matching, the PEI group (n=50) had a higher incidence of cardiopulmonary outcomes (58.0% vs. 30.3%, P=0.001). PEI was a risk factor for cardiopulmonary outcomes after adjusting for confounders (odds ratio [OR] 3.176, 95% confidence interval [95% CI] 1.567-6.438, P=0.001). The subgroup analysis indicated the similar results. A shock index >0.77 was a predictor for cardiopulmonary outcomes in patients undergoing PEI (P=0.015). The probability of cardiopulmonary outcomes in the PEI group depended on the Charlson Comorbidity Index (OR 1.465, 95% CI 1.079-1.989, P=0.014) and shock index >0.77 (compared with shock index ≤0.77 [OR 2.981, 95% CI 1.186-7.492, P=0.020, AUC=0.764]).

CONCLUSION

PEI may be associated with cardiopulmonary outcomes in elderly and critically ill patients with UGIB undergoing endoscopy. Furthermore, a shock index greater than 0.77 could be used as a predictor of a worse prognosis in patients undergoing PEI.

A nomogram for predicting survival in patients with skin non-keratinizing large cell squamous cell carcinoma: A study based on the Surveillance, Epidemiology, and End Results database

Introduction

This study aimed to develop and validate a nomogram for predicting cancer-specific survival (CSS) in patients with non-keratinized large cell squamous cell carcinoma (NKLCSCC) at 3, 5, and 8 years after diagnosis.

Methods

Data on SCC patients were collected from the Surveillance, Epidemiology, and End Results database. Training (70%) and validation (30%) cohorts were generated using random selection of patients. Independent prognostic factors were selected using the backward stepwise Cox regression model. To predict the CSS rates in patients with NKLCSCC at 3, 5, and 8 years after diagnosis, all of the factors were incorporated into the nomogram. Indicators such as the concordance index (C-index), area under the time-dependent receiver operating characteristic curve (AUC), net reclassification index (NRI), integrated discrimination improvement (IDI), calibration curve, and decision-curve analysis (DCA) were then used to validate the performance of the nomogram.

Results

This study included 9,811 patients with NKLCSCC. Twelve prognostic factors were identified by Cox regression analysis in the training cohort, which were age, number of regional nodes examined, number of positive regional nodes, sex, race, marital status, American Joint Committee on Cancer (AJCC) stage, surgery status, chemotherapy status, radiotherapy status, summary stage, and income. The constructed nomogram was validated both internally and externally. The nomogram had good discrimination ability, as indicated by the comparatively high C-indices and AUC values. The nomogram was properly calibrated, as indicated by the calibration curves. Our nomogram was superior to the AJCC model, as illustrated by its superior NRI and IDI values. DCA curves indicated the clinical usability of the nomogram.

Conclusion

The first nomogram for prognosis predictions of patients with NKLCSCC has been developed and verified. Its performance and usability demonstrated that the nomogram could be utilized in clinical settings. However, additional external verification is still required.

Crafting a prognostic nomogram for the overall survival rate of cutaneous verrucous carcinoma using the surveillance, epidemiology, and end results database

BACKGROUND

The aim of this study was to establish and verify a predictive nomogram for patients with cutaneous verrucous carcinoma (CVC) who will eventually survive and to determine the accuracy of the nomogram relative to the conventional American Joint Committee on Cancer (AJCC) staging system.

METHODS

Assessments were performed on 1125 patients with CVC between 2004 and 2015, and the results of those examinations were recorded in the Surveillance, Epidemiology, and End Results (SEER) database. Patients were randomly divided at a ratio of 7:3 into the training (n = 787) and validation (n = 338) cohorts. Predictors were identified using stepwise regression analysis in the COX regression model for create a nomogram to predict overall survival of CVC patients at 3-, 5-, and 8-years post-diagnosis. We compared the performance of our model with that of the AJCC prognosis model using several evaluation metrics, including C-index, NRI, IDI, AUC, calibration plots, and DCAs.

RESULTS

Multivariate risk factors including sex, age at diagnosis, marital status, AJCC stage, radiation status, and surgery status were employed to determine the overall survival (OS) rate (P<0.05). The C-index nomogram performed better than the AJCC staging system variable for both the training (0.737 versus 0.582) and validation cohorts (0.735 versus 0.573), which AUC (> 0.7) revealed that the nomogram exhibited significant discriminative ability. The statistically significant NRI and IDI values at 3-, 5-, and 8-year predictions for overall survival (OS) in the validation cohort (55.72%, 63.71%, and 78.23%, respectively and 13.65%, 20.52%, and 23.73%, respectively) demonstrate that the established nomogram outperforms the AJCC staging system (P < 0.01) in predicting OS for patients with cutaneous verrucous carcinoma (CVC). The calibration plots indicate good performance of the nomogram, while decision curve analyses (DCAs) show that the predictive model could have a favorable clinical impact.

CONCLUSION

This study constructed and validated a nomogram for predicting the prognosis of patients with CVC in the SEER database and assessed it using several variables. This nomogram model can assist clinical staff in making more-accurate predictions than the AJCC staging method about the 3-, 5-, and 8-year OS probabilities of patients with CVC.

Development and validation of a simple nomogram for predicting the short-term prognosis of patients with pulmonary embolism

BACKGROUND

Pulmonary embolism (PE) is a disease caused by blood clots, tumor embolism, and other emboli within the pulmonary arteries. Various scoring scales are used for PE. One such same is the PESI, but it has 12 variables, making it inconvenient for clinical application.

OBJECTIVES

The aim of this study was to develop a new simple nomogram model to assess 30-day survival in PE patients. The new nomogram makes it easier and faster for clinicians to assess the prognosis of patients with PE.

METHODS

We collected data about the patients with PE from the Medical Information Mart for Intensive Care-III (MIMIC-III) database and used the receiver operating characteristic (ROC) curve, area under the ROC curve (AUROC), calibration plot, integrated discrimination improvement (IDI), and decision curve analysis (DCA) to evaluate the predictive power of the new model, and compared these with the PESI.

RESULTS

According to the multivariable Cox regression model results, alongside the actual clinical conditions, we included the following seven variables: race, bicarbonate, age, tumor, systolic blood pressure (SBP), body temperature, and oxygen saturation (Spo2). The AUROC of the new model was greater than 0.70. Its IDI exceeded 0, but with P-value>0.05.

CONCLUSION

The predictive performance of the new model was not worse than the PESI, but the new model only has seven variables, and is therefore more convenient for clinicians to use.

Establishment of a Prognostic Nomogram for Cancer-Specific Survival in Patients With Base-of-Tongue Squamous Cell Carcinoma: A Retrospective Study Based on the SEER Database

BACKGROUND

The aim of this retrospective study was to construct and clinically apply a nomogram for cancer-specific survival (CSS) in patients diagnosed with base-of-tongue squamous cell carcinoma (BOTSCC) to predict their survival prognosis.

METHODS

We collected 8448 patients diagnosed with BOTSCC during 2004-2015 from the Surveillance, Epidemiology, and End Results (SEER) database and divided 30% and 70% of them into validation and training cohorts, respectively. We utilized backward stepwise regression in the Cox model to select variables. Predictive variables were subsequently identified from the variables selected above by using multivariate Cox regression. The new survival model was compared with the American Joint Committee on Cancer (AJCC) prognosis model using the following variables: calibration curve, time-dependent area under the receiver operating characteristic curve (AUC), concordance index (C-index), integrated discrimination improvement (IDI), decision-curve analysis (DCA), and net reclassification improvement (NRI).

RESULTS

A nomogram was established for predicting the CSS probability in patients with BOTSCC. Factors including sex, race, age at diagnosis, marital status, radiotherapy status, chemotherapy status, TNM AJCC stage, surgery status, tumor size, and months from diagnosis to treatment were selected through multivariate Cox regression as independent predictors of CSS. Calibration plots indicated that the model we established had satisfactory calibration ability. The AUC, C-index, IDI, DCA, and NRI results illustrated that the nomogram performed explicit prognoses more accurately than did the AJCC system alone.

CONCLUSION

We identified the relevant factors affecting the survival of BOTSCC patients and analyzed the data on patients suffering from BOTSCC in the SEER database. These factors were used to construct a new nomogram to give clinical staff a more-visual prediction model for the 3-, 5-, and 8-year probabilities of CSS for patients newly diagnosed with BOTSCC, thereby aiding clinical decision making.

Association of thiamine administration and prognosis in critically ill patients with heart failure

Background: Thiamine deficiency is common in patients with heart failure, and thiamine supplement can benefit these patients. However, the association between thiamine administration and prognosis among critically ill patients with heart failure remains unclear. Thus, this study aims to prove the survival benefit of thiamine use in critically ill patients with heart failure. Methods: A retrospective cohort analysis was performed on the basis of the Medical Information Mart of Intensive Care-Ⅳ database. Critically ill patients with heart failure were divided into the thiamine and non-thiamine groups depending on whether they had received thiamine therapy or not during hospitalization. The association between thiamine supplement and in-hospital mortality was assessed by using the Kaplan-Meier (KM) method and Cox proportional hazard models. A 1:1 nearest propensity-score matching (PSM) and propensity score-based inverse probability of treatment weighting (IPW) were also performed to ensure the robustness of the findings. Results: A total of 7,021 patients were included in this study, with 685 and 6,336 in the thiamine and non-thiamine groups, respectively. The kaplan-meier survival curves indicated that the thiamine group had a lower in-hospital mortality than the none-thiamine group. After adjusting for various confounders, the Cox regression models showed significant beneficial effects of thiamine administration on in-hospital mortality among critically ill patients with heart failure with a hazard ratio of 0.78 (95% confidence interval: 0.67-0.89) in the fully adjusted model. propensity-score matching and probability of treatment weighting analyses also achieved consistent results. Conclusion: Thiamine supplement is associated with a decreased risk of in-hospital mortality in critically ill patients with heart failure who are admitted to the ICU. Further multicenter and well-designed randomized controlled trials with large sample sizes are necessary to validate this finding.

Machine learning for the prediction of cognitive impairment in older adults

Objective

The purpose of this study was to develop and validate a predictive model of cognitive impairment in older adults based on a novel machine learning (ML) algorithm.

Methods

The complete data of 2,226 participants aged 60-80 years were extracted from the 2011-2014 National Health and Nutrition Examination Survey database. Cognitive abilities were assessed using a composite cognitive functioning score (Z-score) calculated using a correlation test among the Consortium to Establish a Registry for Alzheimer's Disease Word Learning and Delayed Recall tests, Animal Fluency Test, and the Digit Symbol Substitution Test. Thirteen demographic characteristics and risk factors associated with cognitive impairment were considered: age, sex, race, body mass index (BMI), drink, smoke, direct HDL-cholesterol level, stroke history, dietary inflammatory index (DII), glycated hemoglobin (HbA1c), Patient Health Questionnaire-9 (PHQ-9) score, sleep duration, and albumin level. Feature selection is performed using the Boruta algorithm. Model building is performed using ten-fold cross-validation, machine learning (ML) algorithms such as generalized linear model (GLM), random forest (RF), support vector machine (SVM), artificial neural network (ANN), and stochastic gradient boosting (SGB). The performance of these models was evaluated in terms of discriminatory power and clinical application.

Results

The study ultimately included 2,226 older adults for analysis, of whom 384 (17.25%) had cognitive impairment. After random assignment, 1,559 and 667 older adults were included in the training and test sets, respectively. A total of 10 variables such as age, race, BMI, direct HDL-cholesterol level, stroke history, DII, HbA1c, PHQ-9 score, sleep duration, and albumin level were selected to construct the model. GLM, RF, SVM, ANN, and SGB were established to obtain the area under the working characteristic curve of the test set subjects 0.779, 0.754, 0.726, 0.776, and 0.754. Among all models, the GLM model had the best predictive performance in terms of discriminatory power and clinical application.

Conclusions

ML models can be a reliable tool to predict the occurrence of cognitive impairment in older adults. This study used machine learning methods to develop and validate a well performing risk prediction model for the development of cognitive impairment in the elderly.

Low- or high-dose preventive aspirin use and risk of death from all-cause, cardiovascular disease, and cancer: A nationally representative cohort study

Background and aim: For a long time, aspirin has been recommended for the prevention of cardiovascular disease (CVD). However, results of long-term effects of aspirin use on the risk of CVD and all-cause death as well as cause-specific mortality are not consistent. This study aims to investigate the relationship between low- or high-dose preventive aspirin use and the risk of death from all-cause, CVD, and cancer among US adults aged 40 years and older. Methods: A prospective cohort study was conducted by utilizing four cycles of the National Health and Nutrition Examination Survey (NHANES) and linked 2019 mortality files. Cox proportional hazard models accounting for multiple covariates were used to calculate hazard ratio (HR) and 95% confidence interval (CI) for the associations between low- or high-dose aspirin use and risk of death. Results: A total of 10,854 individuals (5,364 men and 5,490 women) were enrolled in the study. During a median follow-up of 4.8 years, 924 death events including 294 CVD death and 223 cancer death were documented. We found no evidence that taking low-dose aspirin decreased the chance of dying from any cause (HR: 0.92, 95% CI: 0.79-1.06), CVD (HR: 1.03, 95% CI: 0.79-1.33), or cancer (HR: 0.80, 95% CI: 0.60-1.08). High-dose aspirin users had a higher risk of CVD death compared to participants who had never used aspirin (HR: 1.63, 95% CI: 1.11-2.41). Conclusion: Using low-dose aspirin has no effect on the risk of death from any causes, whereas taking high doses of aspirin increases the risk of CVD death.

Predicting Risk Factors of Acute Kidney Injury in the First 7 Days after Admission: Analysis of a Group of Critically Ill Patients

Background

Acute kidney injury (AKI) is a common complication in critically ill patients. Some predictive models have been reported, but the conclusions are controversial. The aim of this study was the formation of nomograms to predict risk factors for AKI in critically ill patients within the first 7 days after admission to the intensive care unit (ICU).

Methods

Data were extracted from the Medical Information Mart for Intensive Care- (MIMIC-) III database. The random forest method was used to fill in the missing values, and least absolute shrinkage and selection operator (Lasso) regression analysis was performed to screen for possible risk factors.

Results

A total of 561 patients were enrolled. Complication with AKI is significantly associated with a longer length of stay (LOS). For all patients, the predictors contained in the prediction nomogram included hypertension, coronary artery disease (CAD), cardiopulmonary bypass (CPB), coronary artery bypass grafting (CABG), Simplified Acute Physiology Score II (SAPS II), central venous pressure (CVP) measured for the first time after admission, and maximum and minimum mean artery pressure (MAP). The model showed good discrimination (C - index = 0.818, 95% CI: 0.779-0.857). In the subgroup of patients with well-controlled blood glucose levels, the significant predictors included hypertension, CABG, CPB, SAPS II, and maximum and minimum MAP. Good discrimination was also present before (C - index = 0.785, 95% CI: 0.736-0.834) and after adjustment (adjusted C - index = 0.770).

Conclusion

Hypertension, CAD, CPB, CABG, SAPS II, CVP measured for the first time after admission, and maximum and minimum MAP were independent risk factors for AKI in critically ill patients.

Development and validation of prognostic nomograms in patients with gallbladder mucinous adenocarcinoma: A population-based study

Background

Gallbladder mucinous adenocarcinoma (GBMAC) is an uncommon malignant gallbladder tumor. There are few studies on its prognosis, with the majority consisting of small series or individual cases. We sought to develop and validate nomograms for predicting overall survival (OS) and cancer-specific survival (CSS) in GBMAC patients.

Methods

The clinicopathological data of GBMAC patients from 1975 to 2019 was extracted from the Surveillance, Epidemiology, and End Results (SEER) database, and all patients were randomly divided into a training cohort (70%) and a validation cohort (30%). Using multivariate Cox regression analyses based on Akaike information criterion (AIC), prognostic and important variables for GBMAC were determined. On the basis of these factors, nomograms were developed to predict the 1-, 3-, and 5-year OS and CSS rates of patients with GBMAC. Multiple parameters, including the area under the subject operating characteristic curve (AUC), the calibration plots, and the decision curve analysis (DCA), were then used to evaluate the accuracy of nomograms.

Results

Following exclusion, a total of 707 GBMAC patients were enrolled, and the training cohort (490, 70%) and validation cohort (217, 30%) were randomly assigned. Grade, surgery, radiation, and SEER stage were predictive factors for patients with GBMAC, as indicated by univariate and multivariate Cox regression analyses based on AIC. We created nomograms for predicting OS and CSS in GBMAC using the four factors. The calibration curves and area under the curves (AUCs) indicated that our nomograms have a moderate degree of predictive accuracy and capability. The results of the DCA revealed that the nomogram has a high predictive value.

Conclusion

We established the first nomograms for predicting 1-, 3-, and 5-year OS and CSS in GBMAC patients, thereby contributing to the prognostication of patients and clinical management.

Recent estimates and predictions of 5-year survival rate in patients with pancreatic cancer: A model-based period analysis

Background

The 5-year survival rate for pancreatic cancer (PC) is incredibly low, resulting in this often being a fatal disease. Timely and accurate assessment of the survival rate and prognosis of patients with PC is of great significance for the development of new programs for prevention, monitoring, and treatment.

Methods

Period analysis and further stratified analysis were used to determine the 5-year relative survival rate (RSR) of patients with PC from 2002 to 2016 using the Surveillance, Epidemiology, and End Results (SEER) project database of the National Cancer Institute. Based on this, a generalized linear model was created to predict the survival rate of patients from 2017 to 2021.

Result

During 2002-2016, the 5-year RSR of patients with PC increased from 7.9 to 23.7%. The generalized linear model predicted that the survival rate had increased to 33.9% during 2017-2021, and hence, it was still unacceptably low. The survival rate of patients aged ≥75 years at diagnosis was the lowest among all age groups and was predicted to be only 21.4% during 2017-2021. Notably, the survival rate of patients with differentiation grade III at diagnosis remains particularly low at 7.6%.

Conclusion

The survival rates of patients with PC, although slightly improved, remain extremely low. Timely assessment of the trend of survival rate changes in patients with PC further improves the prognosis of tumor patients and provides data support for relevant medical works to formulate effective tumor prevention and control policies.

Prognostic nomogram based on the lymph node metastasis indicators for patients with bladder cancer: A SEER population‐based study and external validation

PURPOSE

This study aimed to compare the prognostic value of multiple lymph node metastasis (LNM) indicators and to develop optimal prognostic nomograms for bladder cancer (BC) patients.

METHODS

BC patients were obtained from the Surveillance, Epidemiology, and End Results (SEER) database between 2004 and 2015, and randomly partitioned into training and internal validation cohorts. Genomic and clinical data were collected from The Cancer Genome Atlas (TCGA) as external validation cohort. The predictive efficiency of LNM indicators was compared by constructing multivariate Cox regression models. We constructed nomograms on basis of the optimal models selected for overall survival (OS) and cause-specific survival (CSS). The performance of nomograms was evaluated with calibration plot, time-dependent area under the curve (AUC) and decision curve analysis (DCA) in three cohorts. We subsequently estimated the difference of biological function and tumor immunity between two risk groups stratified by nomograms in TCGA cohort.

RESULTS

Totally, 10,093 and 107 BC patients were screened from the SEER and TCGA databases. N classification, positive lymph nodes (PLNs), lymph node ratio (LNR) and log odds of positive lymph nodes (LODDS) were all independent predictors for OS and CSS. The filtered models containing LODDS had minimal Akaike Information Criterion, maximal concordance indexes and AUCs. Age, LODDS, T and M classification were integrated into nomogram for OS, while nomogram for CSS included gender, tumor grade, LODDS, T and M classification. The nomograms were successfully validated in predictive accuracy and clinical utility in three cohorts. Additionally, the tumor microenvironment was different between two risk groups.

CONCLUSIONS

LODDS demonstrated superior prognostic performance over N classification, PLN and LNR for OS and CSS of BC patients. The nomograms incorporating LODDS provided appropriate prediction of BC, which could contribute to the tumor assessment and clinical decision-making.

Impact of Falls Within 3 Months on the Short-Term Prognoses of Elderly Patients in Intensive Care Units: A Retrospective Cohort Study Using Stabilized Inverse Probability Treatment Weighting

Background

Falls are a major public health problem in the older adults that can lead to poor clinical outcomes. There have been few reports on the short-term prognoses of older critically ill patients, and so we sought to determine the impact of falls on elderly patients in intensive care units (ICUs).

Patients and Methods

This retrospective study of 4503 patients (aged 65 years or older) analyzed data in the Medical Information Mart for Intensive Care-III critical care database. Of those, 2459 (54.6%) older adults are males, and 2044 (45.4%) older adults are females. Based on information from the medical care record assessment forms, patients were classified into the following two groups based on whether they had a fall within the previous 3 months: falls (n=1142) and nonfalls (n=3361). The primary outcomes of this study were 30- and 90-day mortality. Associations between the results of the Kaplan-Meier (KM) survival analysis, Cox proportional-hazards regression models, and subgroup analysis and its outcomes were analyzed using stabilized inverse probability treatment weighting (IPTW).

Results

KM survival curves with stabilized IPTW indicated that 30- and 90-day survival rates were significantly lower in elderly critically ill patients with a history of falls within the previous 3 months than in those patients without a history of falls (all p<0.001). Multivariate Cox proportional-hazards regression analysis indicated that 30- and 90-day mortality rates were 1.35 times higher (95% confidence interval [CI]=1.16-1.57, p<0.001) and 1.36 times higher (95% CI=1.19-1.55, p<0.001), respectively, in elderly critically ill patients with a history of falls within the previous 3 months than in those patients without a history of falls.

Conclusion

Experience of falls within the previous 3 months prior to ICU admission increased the risk of short-term mortality and affected the prognoses of elderly patients. Falls should therefore receive adequate attention from clinical healthcare providers and management decision-makers.

Establishment and verification of a nomogram to predict tumor-specific mortality risk in triple-negative breast cancer: a competing risk model based on the SEER cohort study

Background

Triple-negative breast cancer (TNBC) is the subtype of breast cancer with the worst prognosis, and traditional survival analysis methods are biased when predicting mortality. To predict the risk of death in patients with triple-negative breast cancer more precisely, a competing risk model was developed.

Methods

The clinicopathological data of the TNBC patients from 2010 to 2015 were collected from the Surveillance, Epidemiology, and End Results (SEER) database. The data were assigned into a training set and testing set at a ratio of 7:3 in a randomized pattern. Univariate and multivariate competing risk models were applied to find the independent prognostic factors. A prediction nomogram for cancer-specific mortality (CSM) risk was constructed. The accuracy and discrimination of the nomogram were assessed using receiver operating characteristic (ROC) area under the curve (AUC), concordance index (C-index), and a calibration curve using the training and testing sets, respectively.

Results

A total of 28,430 TNBC patients were randomly grouped into the training set (n=19,900) and the testing set (n=8,530). The median time for follow-up was 59 [1-107] months. A total of 7,014 (24.67%) patients died, among whom 4,801 (68.45%) died from breast cancer and 2,213 (31.55%) due to non-breast cancer events. The independent risk factors were primary site of tumor, grade, tumor-node-metastasis (TNM) stage, T stage, approach of surgery, chemotherapy, axillary lymph node metastases, brain metastases, and liver metastases. The prediction nomogram was constructed by using the aforementioned variables. The 1-, 3-, and 5-year AUC of CSM were predicted to be 0.856, 0.81, and 0.782, respectively, in the training set, and 0.856, 0.81, and 0.782 in the testing set, respectively. The C-index of the nomogram was 0.801 and 0.799 in the training and testing sets, respectively. As confirmed by the validation and training calibration curves, the nomogram was consistent with the results.

Conclusions

We used competing risk models to identify risk factors for CSM and constructed a CSM risk prediction nomogram for TNBC patients, that may be utilized to predict CSM risk in TNBC patients clinically and assist in the creation of individualised clinical treatment options.

Heterogeneous catalysis mediated by light, electricity and enzyme via machine learning: Paradigms, applications and prospects

Energy crisis and environmental pollution have become the bottleneck of human sustainable development. Therefore, there is an urgent need to develop new catalysts for energy production and environmental remediation. Due to the high cost caused by blind screening and limited valuable computing resources, the traditional experimental methods and theoretical calculations are difficult to meet with the requirements. In the past decades, computer science has made great progress, especially in the field of machine learning (ML). As a new research paradigm, ML greatly accelerates the theoretical calculation methods represented by first principal calculation and molecular dynamics, and establish the physical picture of heterogeneous catalytic processes for energy and environment. This review firstly summarized the general research paradigms of ML in the discovery of catalysts. Then, the latest progresses of ML in light-, electricity- and enzyme-mediated heterogeneous catalysis were reviewed from the perspective of catalytic performance, operating conditions and reaction mechanism. The general guidelines of ML for heterogeneous catalysis were proposed. Finally, the existing problems and future development trend of ML in heterogeneous catalysis mediated by light, electricity and enzyme were summarized. We highly expect that this review will facilitate the interaction between ML and heterogeneous catalysis, and illuminate the development prospect of heterogeneous catalysis.

Association of lactate to albumin ratio and bicarbonate with short-term mortality risk in patients with acute myocardial infarction

Background

Previous studies have indicated that the ratio of lactate/albumin (L/A) has predictive value for the prognosis of critically ill patients with heart failure. Some studies have also indicated that a low serum bicarbonate concentration is inversely related to the mortality risk of patients with cardiogenic shock. However, the value of bicarbonate and the L/A ratio for predicting the mortality risk of patients with acute myocardial infarction (AMI) is still unclear. We therefore conducted a retrospective study to research this problem.

Methods

The subjects of this study were patients with AMI, and the data source was the Medical Information Mart for Intensive Care III database. The primary endpoint was 30-day all-cause mortality after admission. The Receiver operating characteristic (ROC) curve was used to compare the predictive value of L/A ratio, lactate and albumin for end-point events. The effects of different L/A ratio levels and different bicarbonate concentrations on 7-day and 30-day all-cause mortality were compared using Kaplan–Meier (K-M) curves. Hazard ratios for different L/A ratio and different bicarbonate concentrations were investigated using COX proportional hazards models.

Results

The Area Under Curve (AUC) of L/A ratio, lactate, and albumin were 0.736, 0.718, and 0.620, respectively. (1) L/A ratio: The patients were divided into three groups according to their L/A ratio: tertile T1 (L/A ratio ≤ 0.47), tertile T2 (L/A ratio ≤ 0.97), and tertile T3 (L/A ratio > 0.97). The T2 and T3 groups had higher 30-day all-cause mortality risks than the T1 group. The restricted cubic spline (RCS) model indicated that there was a nonlinear relationship between L/A ratio and 30-day mortality (P < 0.05). (2) Bicarbonate concentration: The patients were also divided into three groups based on their bicarbonate concentration: G1 (22–27 mmol/L), G2 (< 22 mmol/L), and G3 (> 27 mmol/L). The G2 and G3 groups had higher 30-day all-cause mortality risks than the G1 group. The RCS model indicated that there was a nonlinear relationship between bicarbonate concentration and 30-day mortality (P < 0.05). The RCS model indicated that there was a nonlinear relationship between hemoglobin level and 30-day all-cause mortality (P < 0.05).

Conclusion

L/A ratio and bicarbonate concentration and hemoglobin level have predictive value for predicting 30-day mortality in patients with acute myocardial infarction.

Effects of different types of radiation therapy on cardiac-specific death in patients with thyroid malignancy

Radiation therapy (RT) is one of the common and widely used treatment method for thyroid tumors. Considering that the thyroid is located close to the heart, the radiation generated during the treatment of thyroid tumors may have an adverse greater impact on the heart. This study is to explore the influencing factors, especially additional effects of RT, on cardiac-specific death among patients with malignant thyroid tumors. Collecting information from the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) database using SEER*Stat. Patients with malignant thyroid tumors were searched, whether receiving RT or not. Ultimately, 201, 346 eligible patients were included. Propensity Score Matching (PSM) was used to minimize bias of baseline characteristics by adjusting for confounding factors. COX (proportional hazards) and fine-gray (competing risk) model regression analysis were used to explore the effects of various influencing factors on cardiac-specific death. The present analysis showed that, compared with non-RT, RT based upon radioactive implants and beam radiation were associated with lower risk of cardiac-specific death in patients with thyroid malignancy, beam radiation therapy may had a similar effect. Besides, the remaining RT methods did not significantly increase the risk of cardiac-specific death. In addition, Asian or Pacific Islander ethnicity, female sex, marital status, combined summary stage (localized, regional, and distant), high-income, and later year of diagnosis were associated with lower risk of cardiac-specific death. While older age of diagnosis, African ethnicity, non-Hispanic ancestry, and derived AJCC stage (IV) were risk factors for cardiac-specific death. These results help to identify the factors influencing cardiac-specific death among patients with thyroid malignancies. Furthermore, it may helps to improve the clinical application of RT without too much concern about adverse cardiac effects.

Development and validation of novel risk prediction models of breast cancer based on stanniocalcin‐1 level

PURPOSE

The function of stanniocalcin-1 (STC-1) in the oncogenesis and progression of tumors has been extensively studied. The purpose of this study was to investigate the relationship between secreted STC-1 and prognosis in patients with breast cancer (BC) and to determine whether STC-1 could be a key prognostic factor in BC.

METHODS

The STC-1 level was measured by ELISA and clinical data from 1210 female patients with BC were used to develop and validate nomograms. We then verified the models through the plotting of ROC curves and calibration curves, calculating the C-index, and performing decision curve analyses (DCA).

RESULTS

The level of STC-1 in the peripheral plasma was significantly correlated with the T stage, N stage, clinical stage, grade, hormone receptors, HER-2 status, and tumor subtype. Cox regression analyses revealed that estrogen receptor(ER) status, N stage, and STC-1 level were risk factors for overall survival (OS), whereas T stage, N stage, and STC-1 level were independent prognostic factors for distant disease-free survival (DDFS) and disease-free survival (DFS). Both the ROC curve and the C-index confirmed the high resolution of these models, while the DCA identified the feasibility of their practical application. In addition, the calibration curves indicated good consistency between the predicted and actual survival rates.

CONCLUSION

Nomograms were created based on STC-1 levels for 3-, 5-, and 7-year OS, DDFS, and DFS of patients with BC respectively. As a key prognostic factor for BC, peripheral blood STC-1 level can be used clinically as a liquid biopsy indicator.

Factors Associated With Lymph Node Yield and Effects of Lymph Node Density on Survival of Patients With Pulmonary Sarcomatoid Carcinoma

OBJECTIVE

The objective of this study was to identify factors associated with lymph node yield (LNY) during surgeries for pulmonary sarcomatoid carcinoma (PSC) and to determine effects of lymph node density (LND) on the overall survival (OS) of patients with PSC.

MATERIALS AND METHODS

The SEER Research Plus database was searched for data on patients with PSC from 1988 to 2018. Poisson regression was used of all patients with PSC to identify relevant factors associated with LNY. Univariate and multivariate Cox regression analyses were adopted for lymph node (LN)-positive patients to evaluate the impact of LND on OS. The 5-year OS rates of patients with PSC were compared based on their LN status and LND.

RESULTS

There were 545 eligible patients in the study sample, 175 of which were LN-positive. These patients had significantly lower 5-year OS than those with no positive LNs ( P <0.001). Poisson regression analysis indicated relevant factors increasing LNY included higher diagnosis age, non-Hispanic American Indian or Alaska Native races, larger tumor, pleomorphic carcinoma histology, and more advanced disease stages. The Cox regression analysis indicated higher LND ( P =0.022) was probably associated with a worse prognosis for LN-positive patients. The group with LND ≥0.12 had a higher risk of death than the group with LND <0.12 ( P <0.001) among LN-positive patients with PSC.

CONCLUSIONS

Patients with PSC with high LND experienced worse outcomes than those with low LND. Further risk stratification of patients with PSC may help to improve survival benefits based on prognostic indicators of LND.

Efficacy and safety of monoclonal antibodies in neuromyelitis optica spectrum disorders: A survival meta-analysis of randomized controlled trials

Background

Monoclonal antibodies such as rituximab (RTX), eculizumab, inebilizumab, satralizumab, and tocilizumab have been found to be effective therapies for neuromyelitis optica spectrum disease ​(NMOSD) in several clinical randomized controlled trials.

Objective

The purpose of this meta-analysis of randomized controlled trials was to assess the efficacy and safety of monoclonal antibodies in the treatment of NMOSD.

Methods

We searched the following databases for relevant English language literature from the establishment of the database to June 2021: PubMed, Embase, Cohorane Library, the Central Register of Controlled Trials (CENTRAL), and Web of Science. Randomized controlled trials of monoclonal antibodies were the targets of the review.

Results

We included seven trials containing 775 patients (485 in the monoclonal antibody group and 290 in the control group). Patients in the monoclonal group (HR 0.24, 95% CI: 0.14 to 0.40, P ​< ​0.00001), as well as patients with seropositive AQP4-IgG (HR 0.18, 95% CI: 0.11 to 0.29, P ​< ​0.00001), both had a higher free recurrence rate than that in the control group. In the first year (HR 0.25, 95% CI: 0.09 to 0.71, P ​= ​0.009) and the second year (HR 0.32, 95% CI: 0.13 to 0.81, P ​= ​0.02), no relapses were documented. The average changes of the expanded disability status scale (EDSS) score decreased by 0.29 (95% CI: -0.09 to 0.51, P ​= ​0.005). Upper respiratory tract infection (OR 1.52, 95% CI: 0.76 to 3.04, P ​= ​0.24), urinary tract infection(OR 0.79, 95% CI: 0.51 to 1.21, P ​= ​0.27), and headache (OR 1.30, 95% CI: 0.78 to 2.17, P ​= ​0.31) were three most frequent adverse reactions.

Conclusions

Monoclonal antibodies are particularly effective treatments in avoiding recurrence for NMOSD patients, according to this meta-analysis. The associated adverse responses are not significantly different from those seen with traditional immunosuppressants.

Evaluation of factors that influenced the length of hospital stay using data mining techniques

Background

length of stay (LOS) is the time between hospital admission and discharge. LOS has an impact on hospital management and hospital care functions.

Methods

A descriptive, retrospective study was designed on about 27,500 inpatients between March 2019 and 2020. Required data were collected from six wards (CCU, ICU, NICU, General, Maternity, and Women) in a teaching hospital. Clinical data such as demographic characteristics (age, sex), type of ward, and duration of hospital stay were analyzed by the R-studio program. Violin plots, bar charts, mosaic plots, and tree-based models were used to demonstrate the results.

Results

The mean age of the population was 40.8 ± 19.2 years. The LOS of the study population was 2.43 ± 4.13 days. About 60% of patients were discharged after staying one day in the hospital. After staying one day in the hospital, 67% of women were discharged. However, 23% of men were discharged within this time frame. The majority of LOS in the CCU, ICU, and NICU ranged from 5 to 9 days.; In contrast, LOS was one day in General, Maternity, and Woman wards. Due to the tree plot, there was a different LOS pattern between Maternity-Women and the CCU-General-ICU-NICU wards group.

Conclusion

We observed that patients with more severe diseases hospitalized in critical care wards had a longer LOS than those not admitted to critical care wards. The older patient had longer hospital LOS than the younger. By excluding Maternity and Woman wards, LOS in the hospital was comparable between males and females and demonstrated a similar pattern.

A Novel Prognostic Model for Patients with Primary Gastric Diffuse Large B-Cell Lymphoma

Objectives. Primary gastric diffuse large B-cell lymphoma (PG-DLBCL) is a common phenotype of extranodal non-Hodgkin’s lymphoma (NHL). This research aims to identify a model for predicting overall survival (OS) and cancer-specific survival (CSS) in PG-DLBCL. Methods. A total of 1716 patients diagnosed with PG-DLBCL between 1975 and 2017 were obtained from the SEER database and further randomly divided into the training and validating cohorts at a ratio of 7 : 3. Univariate and multivariate cox analyses were conducted to determine significant variables for the construction of nomogram. The performance of the model was then assessed by the concordance index (C-index), the calibration plot, and the area under the receiver operating characteristic (ROC) curve (AUC). Results. Multivariate analysis revealed that age, race, insurance status, Ann Arbor stage, marital status, chemotherapy, and radiation therapy all showed a significant association with OS and CSS. These characteristics were applied to build a nomogram. In the training cohort, the discrimination of nomogram for OS and CSS prediction was excellent (C-index = 0.764, 95% CI, 0.744–0.784 and C-index = 0.756, 95% CI, 0.732–0.780). The AUC of the nomogram for predicting 3- and 5-year OS was 0.779 and 0.784 and CSS was 0.765 and 0.772. Similar results were also observed in the internal validation set. Conclusions. We have successfully established a novel nomogram for predicting OS and CSS in PG-DLBCL patients with good accuracy, which can help physicians to quickly and accurately complete the evaluation of survival probability, risk stratification, and therapeutic strategy at diagnosis.

Prognostic risk factor analysis and nomogram construction for primary liver cancer in elderly patients based on SEER database

OBJECTIVE

To evaluate the risk factors and construct a nomogram model for the prognosis of primary liver cancer in the elderly based on the data from the US SEER database.

METHODS

The latest data of patients with primary liver cancer were extracted from the SEER database using SEER*STAT software, and the required variables were included. The data were screened and then divided into a training cohort and a validation cohort. A nomogram model was constructed by screening the variables through univariate and multivariate Cox analysis. The C-Index, ROC and calibration curves were used for model evaluation.

RESULTS

A total of 10 824 eligible cases from 2004 to 2017 were extracted, among which, 7757 cases were included in the training cohort and 3247 in the validation cohort. The C-Index of the model was 0.747 (in the training cohort) and 0.773 (in the validation cohort). The 3-year area under the curve (AUCs) of the training and the validation cohorts were 0.760 and 0.750, and the 5-year AUCs of the two cohorts were 0.761 and 0.748. The calibration curves showed an ideal calibration of the constructed model.

CONCLUSIONS

The nomogram model constructed followed by Cox regression analysis showed moderate calibration and discrimination property, and can provide reference to a certain extent for furture clinical application of primary liver cancer in the elderly.

Analysis of Prognostic Factors and Construction of Prognostic Models for Invasive Micropapillary Carcinoma of the Breast

OBJECTIVE

To compare and analyze the clinical characteristics of invasive micropapillary carcinoma (IMPC) of the breast (IMPC-B) and invasive ductal carcinoma (IDC) of the breast (IDC-B) and establish a prognostic model of IMPC-B.

METHODS

We retrospectively analyzed data for patients diagnosed with breast cancer in the Surveillance, Epidemiology, and End Results (SEER) database between 2010 and 2018 and screened 581 patients with IMPC and 1325 patients with IDC. We compared age, race, laterality, tumor site, histological grade, type of surgery, radiation, chemotherapy, whether the first primary tumor, T stage, N stage, M stage, and molecular type between IMPC-B and IDC-B and draw survival curves of IMPC-B and IDC-B. The relationship between clinical factors and prognosis was investigated by univariate analysis using the Log-rank test and multivariate analysis of the Cox proportional hazards regression model. A risk scoring model was constructed based on independent risk factors to distinguish high-risk and low-risk patients; in addition, a nomogram was created to predict patient survival.

RESULTS

There were differences between the two groups in the age of onset, race, tumor site, histological grade, type of surgery, N stage, and molecular type (p < 0.05). Overall survival was decreased in IMPC-B compared with IDC-B (p < 0.05). The prognosis of IMPC-B was significantly correlated with histological grade, whether the first primary tumor, type of surgery, radiotherapy, chemotherapy, T stage, and N stage. Based on the relationship between the above factors and overall survival prognosis, the risk score model we constructed can effectively distinguish high-risk and low-risk patients (p < 0.05). The established nomogram had better performance in predicting survival in patients with IMPC-B (C - index = 0.78).

CONCLUSION

IMPC-B has a worse prognosis than IDC-B, with earlier age of onset, higher histological grade, and later N stage, and luminal breast cancer is the main type. The nomogram can well predict the prognosis of patients with IMPC-B, which has a high clinical reference value and provides a scientific basis for clinical treatment.

Nomogram for predicting overall survival in patients with invasive micropapillary carcinoma after breast-conserving surgery: A population-based analysis

BACKGROUND

Due to the loss of prediction of overall survival (OS) for patients with invasive micropapillary carcinoma (IMPC) after breast-conserving surgery (BCS), this study aimed to construct a nomogram for predicting OS in IMPC patients after BCS.

METHODS

In total, 481 eligible cases staged 0-III IMPC from 2000 to 2016 were retrieved from the SEER database. A nomogram was built based on the variables selected by LASSO regression to predict the 3-year and 5-year probabilities of OS.

RESULTS

A total of 336 patients were randomly assigned to the training cohort and 145 cases in the validation cohort. The LASSO regression revealed that six variables (age at diagnosis, AJCC stage, marital status, ER status, PR status, and chemotherapy) were predictive variables of OS, and then a nomogram model and an easy-to-use online tool were constructed. The C-indices 0.771 in the training cohort and 0.715 in the validation cohort suggested the robustness of the model. The AUC values for 3-year and 5-year OS in the training cohort were 0.782, 0.790, and 0.674, and 0.682 in the validation cohort, respectively. Based on the cutoff values of 147.23 and 222.44 scores calculated by X-tile analysis, participants in the low-risk group (≤147.23 scores) had a more favorable OS in comparison with those in the medium (>147.23, but <222.44 scores)- and high-risk groups (≥222.44 scores).

CONCLUSIONS

By risk stratification, this model is expected to provide a precise and personalized prediction of the cumulative risk and guide treatment decision-making in improving OS strategies for IMPC patients.

The Incidence Trend of Papillary Thyroid Carcinoma in the United States During 2003-2017

Background

The rapid increase in the detection rate of thyroid cancer over the past few decades has caused some unexpected economic burdens. However, that of papillary thyroid carcinoma (PTC) seems to have had the opposite trend, which is worthy of further comprehensive exploration.

Methods

The Surveillance, Epidemiology, and End Results 18 database was used to identify patients with PTC diagnosed during 2003-2017. The incidence trends were analyzed using joinpoint analysis and an age-period-cohort model.

Results

The overall PTC incidence rate increased from 9.9 to 16.1 per 100 000 between 2003 and 2017. The joinpoint analysis indicated that the incidence growth rate began to slow down in 2009 (annual percentage change [APC] = 3.1%, 95% confidence interval [CI] = 1.9%–4.4%). After reaching its peak in 2015, it began to decrease by 2.8% (95% CI = −4.6% to −1.0%) per year. The stratified analysis indicated that the incidence patterns of different sexes, age groups, races, and tumor stages and sizes had similar downward trends, including for the localized (APC = −4.5%, 95% CI = −7% to −1.9%) and distant (APC = −1.3%, 95% CI = −2.7% to −.1%) stages, and larger tumors (APC = −4%, 95% CI = −12% to 4.7%). The age-period-cohort model indicated a significant period effect on PTC, which gradually weakened after 2008-2012. The cohort effect indicates that the risk of late birth cohorts is gradually stabilizing and lower than that of early birth cohorts.

Conclusion

The analysis results of the recent downward trend and period effect for the incidence of each subgroup further support the important role of correcting overdiagnosis in reducing the prevalence of PTC. Future research needs to analyze more-recent data to verify these downward trends.

Exploring the Risk Factors of Conjunctival Squamous Cell Carcinoma and Establishing a Prognostic Model: Retrospective Study

Objective

Exploring the risk factors of conjunctival squamous cell carcinoma (CSCC) and establishing a prognostic model.

Methods

Information on patients with CSCC was extracted from the SEER database, conducting a retrospective study. 650 patients with CSCC were finally included in the model. Descriptive analysis was performed by Chi-square test and T-test. The risk factors of CSCC were explored by COX multivariate analysis, and the corresponding prognostic model was established as a result.

Results

The all-cause mortality rate of CSCC was 38.3%, and the risk factors were age (HR = 1.077), sex (HR = 0.691), grade (HR = 7.857), laterality (HR = 1.403), N (HR = 7.195), M (HR = 0.217), and surgery (HR = 1.618), all P < 0.05. The new model had C index and area under curve ROC (AUC) value greater than 0.7. Calibration curve, Net Reclassification Index (NRI), Integrated Discrimination Improvement (IDI), and Decision Curve Analysis (DCA) indicate the new model has better predictive performance than the American Joint Committee on Cancer (AJCC-TNM).

Conclusions

Compared with the clinical guidance of AJCC (TNM) for patients with CSCC, the established model exhibits good performance and can provide guidance for clinical decision-making.

Effect of radiotherapy on cardiac-specific death in patients with non-malignant tumors of central nervous system and related clinical features

Importance

Cardiac-specific death from radiation caused by radiation therapy (RT) in patients with malignant tumors has received extensive attention, however, little is known regarding the potential cardiotoxic effects of RT in patients with non-malignant tumors.

Objectives and methods

In this study, we used the SEER data to explore the incidence of post-radiation cardiovascular complications in patients with non-malignant tumors of central nervous system (CNS), and identify the influencing factors of cardiac-specific death.

Results

Ultimately 233, 306 patients were included (97.8% of patients had brain tumors and 2.2% had spinal cord tumors). For patients with non-malignant tumors of CNS, RT {yes (odds ratio [OR] 0.851, 95% confidence interval [CI] 0.774–0.936, p = 0.001, before propensity score matching (PSM); OR 0.792, 95% CI 0.702–0.894, p < 0.001, after PSM) vs. no} was associated with lower risk of cardiac-specific death, other clinical features affecting cardiac death similar to those in patients with non-malignant tumors of CNS receiving RT. For patients with non-malignant tumors of CNS receiving RT, female, married status, Hispanic ethnicity, surgery, and tumor site (brain exclude nerve and endocrine, nervous system) were associated with lower risks of cardiac-specific death, while earlier year of diagnosis, older age of diagnosis, Black, larger tumor and bilateral tumor were risk factors for cardiac-specific death.

Conclusions

Our study shows the influencing factors for cardiac-specific death in patients with non-malignant tumors of CNS, and found RT is associated with lower risk of cardiac-specific death. These results can facilitate the identification of patients with non-malignant tumors of CNS who can benefit from RT while avoiding cardiovascular events. In addition, this study helps to enhance the clinical use of RT in these populations, especially in patients who may have impaired cardiac function due to CNS tumors.

A novel nomogram and risk classification system predicting the overall survival of patients with papillary renal cell carcinoma after nephrectomy: A population-based study

Background

Papillary renal cell carcinoma (pRCC) is the largest histologic subtype of non-clear-cell RCC. To date, there is no reliable nomogram to predict the prognosis of patients with pRCC after nephrectomy. We aimed to first establish an effective nomogram to predict the overall survival (OS) of patients with pRCC after nephrectomy.

Methods

A total of 3,528 eligible patients with pRCC after nephrectomy were identified from the Surveillance, Epidemiology, and End Results (SEER) database between 2010 and 2015. The patients were randomized into the training cohort (n = 2,472) and the validation cohort (n = 1,056) at a 7:3 ratio. In total, 122 real-world samples from our institute (titled the AHMU-pRCC cohort) were used as the external validation cohort. Univariate and subsequent multivariate Cox regression analyses were conducted to identify OS-related prognostic factors, which were further used to establish a prognostic nomogram for predicting 1-, 3-, and 5-year OS probabilities. The performance of the nomogram was evaluated by using the concordance index (C-index), receiver operating characteristic curve (ROC), calibration plot, and decision curve analysis (DCA).

Results

Multivariate Cox analysis showed that age, race, marital status, TNM stage, tumor size, and surgery were significant OS-related prognostic factors. A prognostic model consisting of these clinical parameters was developed and virtualized by a nomogram. High C-index and area under the ROC curve (AUC) values of the nomogram at 1, 3, and 5 years were found in the training, validation, and AHMU-pRCC cohorts. The calibration plot and DCA also showed that the nomogram had a satisfactory clinical application value. A risk classification system was established to risk-stratify patients with pRCC.

Conclusion

Based on a large cohort from the public SEER database, a reliable nomogram predicting the OS of patients with pRCC after nephrectomy was constructed, which could optimize the survival assessment and clinical treatment.

How to conduct well-designed clinical research

Clinicians and healthcare decision-makers conduct their clinical practice based on the results of clinical trials. However, some health problems remain unresolved; in such cases, further research is required. To ensure reliable research results, it is important to understand the study design and conduct well-designed clinical trials. Many study designs can be chosen within the two broad categories of observational and interventional. Clinical studies have a variety of designs, including case series, case-control, cross-sectional, and prospective and retrospective cohort studies. Well-designed clinical studies can clarify important differences between treatment options and provide data on long-term drug efficacy and safety. Interpreting the results of clinical trials can be difficult because weaknesses in research design, data collection methods, analytic methods, and reporting can compromise their value and usefulness. However, although randomized controlled trials are limited owing to ethical and practical issues, they are optimal for investigating the effects of therapy and establishing causality. Here we present an overview of different clinical research designs and review their advantages and limitations.

Analysis and prediction of relative survival trends in patients with non-Hodgkin lymphoma in the United States using a model-based period analysis method

Background

Survival rates are usually used to evaluate the effect of cancer treatment and prevention. This study aims to analyze the 5-year relative survival of non-Hodgkin lymphoma (NHL) in United States using population-based cancer registry data.

Methods

A period analysis was used to evaluate the improvement in long-term prognosis of patients with NHL from 2004 to 2018, and a generalized linear model was developed to predict the 5-year relative survival rates of patients during 2019–2023 based on data from the SEER database stratified by age, sex, race and subtype.

Results

In this study, relative survival improved for all NHL, although the extent of improvement varied by sex, age group and lymphoma subtype. Survival improvement was also noted for NHL subtypes, although the extent varied, with marginal-zone lymphoma having the highest 5-year relative survival rate (92.5%) followed by follicular lymphoma (91.6%) and chronic lymphocytic leukemia/small lymphocytic lymphoma (87.3%). Across all subtypes, survival rates were slightly higher in females than in males. Survival rates are lower in the elderly than in the young. Furthermore, the study demonstrated that black patients had lower NHL survival rates than white patients. Survival rates for NHL were higher in rural areas than in urban areas. Patients with extra-nodal NHL had a higher survival rate than patients with nodal NHL.

Conclusion

Overall, patient survival rates for NHL gradually improved during 2004–2018. The trend continues with a survival rate of 75.2% for the period 2019–2023. Analysis by NHL subtype and subgroups indicating that etiology and risk factors may differ by subtype. Identification of population-specific prevention strategies and treatments for each subtype can be aided by understanding these variations.

Association between age and the presence and mortality of breast cancer synchronous brain metastases in the United States: A neglected SEER analysis

Background

The extent of the relationship between age and the presence of breast cancer synchronous brain metastases (BCSBMs) and mortality has not yet been well-identified or sufficiently quantified. We aimed to examine the association of age with the presence of BCSBMs and all-cause and cancer-specific mortality outcomes using the SEER database.

Methods

Age-associated risk of the presence and survival of BCSBMs were evaluated on a continuous scale (restricted cubic spline, RCS) with logistic or Cox regression models. The main endpoints were the presence of BCSBMs and all-cause mortality or cancer-specific mortality. Cox proportional hazards regression and competing risk models were used in survival analysis.

Results

Among 374,132 adult breast cancer patients, 1,441 (0.38%) had BMs. The presence of BCSBMs displayed a U-shaped relationship with age, with the highest point of the curve occurring at the age of 62. In both the younger (age ≤ 61) and older (age ≥ 62) groups, the observed curve showed a nearly linear relationship between age and the presence of BCSBMs. The relationship between age and all-cause mortality (ASM) and cancer-specific mortality (CSM) was linear. Older age at diagnosis was associated with a higher risk of ASM (HR 1.019, 95% CI: 1.013-1.024, p < 0.001) and CSM (HR 1.016, 95% CI: 1.010-1.023, p < 0.001) in multivariable Cox models. Age (sHR 1.007, 95% CI 1-1.013, p = 0.049) was substantially related to a significantly increased risk of CSM in competing risk models.

Conclusion

Age had a non-linear U-shaped relationship with the presence of BCSBMs and a linear relationship with BCSBMs mortality.

Dynamic nomogram for predicting acute kidney injury in patients with acute ischemic stroke: A retrospective study

Background

This study sought to develop and validate a dynamic nomogram chart to assess the risk of acute kidney injury (AKI) in patients with acute ischemic stroke (AIS).

Methods

These data were drawn from the Medical Information Mart for Intensive Care III (MIMIC-III) database, which collects 47 clinical indicators of patients after admission to the hospital. The primary outcome indicator was the occurrence of AKI within 48 h of intensive care unit (ICU) admission. Independent risk factors for AKI were screened from the training set using univariate and multifactorial logistic regression analyses. Multiple logistic regression models were developed, and nomograms were plotted and validated in an internal validation set. Based on the receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA) to estimate the performance of this nomogram.

Results

Nomogram indicators include blood urea nitrogen (BUN), creatinine, red blood cell distribution width (RDW), heart rate (HR), Oxford Acute Severity of Illness Score (OASIS), the history of congestive heart failure (CHF), the use of vancomycin, contrast agent, and mannitol. The predictive model displayed well discrimination with the area under the ROC curve values of 0.8529 and 0.8598 for the training set and the validator, respectively. Calibration curves revealed favorable concordance between the actual and predicted incidence of AKI (p > 0.05). DCA indicates the excellent net clinical benefit of nomogram in predicting AKI.

Conclusion

In summary, we explored the incidence of AKI in patients with AIS during ICU stay and developed a predictive model to help clinical decision-making.

Radix Salvia miltiorrhiza for Ankylosing Spondylitis: Determining Potential Inflammatory Molecular Targets and Mechanism Using Network Pharmacology

Radix Salvia miltiorrhiza (RSM) is widely used for the clinical improvement of inflammatory diseases. However, the actions of RSM in the treatment of ankylosing spondylitis (AS) have not been fully explored. Therefore, this study was designed to use retrospective clinical data mining approach to understand the effects of RSM on AS-related immuno-inflammatory processes, use network pharmacology to predict therapeutic targets of RSM, and to further investigate the pharmacological molecular mechanism in vitro. RSM treatment has a long-term correlation with the improvement of AS-related immuno-inflammatory indicators through computational models. We established protein-protein interaction networks, conducted KEGG analysis to enrich significant TNF pathways, and finally obtained three core targets of RSM in the treatment of AS, namely, prostaglandin endoperoxide synthase 2 (PTGS2), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-alpha). Screening of RSM active ingredients with node degree greater than 20 yielded cryptotanshinone and tanshinone IIA, and previous studies have reported their anti-inflammatory effects. In vitro, both cryptotanshinone and tanshinone IIA significantly inhibited the expressions of PTGS2, IL-6, and TNF-α in peripheral blood mononuclear cells in AS patients. In conclusion, cryptotanshinone and tanshinone IIA, which are the active components of RSM, may inhibit the activation of TNF signaling pathway in AS patients by downregulating the expression of PTGS2, IL-6, and TNF-α. These findings illustrate that RSM may be a promising therapeutic candidate for AS, but further validation is required.

Application of miRNA Biomarkers in Predicting Overall Survival Outcomes for Lung Adenocarcinoma

Background

With the development of research, the importance of microRNAs (miRNAs) in the occurrence, metastasis, and prognosis of lung adenocarcinoma (LUAD) has attracted extensive attention. This study is aimed at predicting overall survival (OS) results through bioinformatics to identify novel miRNA biomarkers and hub genes.

Materials and Methods

The data of LUAD-related miRNA and mRNA samples was downloaded from The Cancer Genome Atlas (TCGA) database. Upon screening and pretreatment of initial data, TCGA data were analyzed using R platform and a series of analytical tools to identify biomarkers with high specificity and sensitivity.

Results

7 miRNAs and 13 hub genes that had strong relation to the overall surviving status were identified in patients with LUAD. The expression of seven miRNAs (hsa-miR-19a-3p, hsa-miR-126-5p, hsa-miR-556-3p, hsa-miR-671-5p, hsa-miR-937-3p, hsa-miR-4664-3p, and hsa-miR-4746-5p) could apparently improve the OS rate of patient with LUAD. The 13 hub genes, namely, CCT6A, CDK5R1, CEP55, DNAJB4, EGLN3, HDGF, HOXC8, LIMD1, MKI67, PCP4L1, PPIL1, SCAI, and STK32A, showed a correlation with the OS status.

Conclusion

7 miRNAs were identified as novel biomarkers for the prognosis of patients with LUAD. This study offered a deeper comprehension of LUAD treatment and prognosis from the molecular level and helped enhance the understanding of the pathogenesis and potential molecular events of LUAD.

Body mass index, genetic susceptibility, and Alzheimer's disease: a longitudinal study based on 475,813 participants from the UK Biobank

BACKGROUND

The association between body mass index (BMI) and Alzheimer's disease (AD) remains controversial. Genetic and environmental factors are now considered contributors to AD risk. However, little is known about the potential interaction between genetic risk and BMI on AD risk.

OBJECTIVE

To study the causal relationship between BMI and AD, and the potential interaction between AD genetic risk and BMI on AD risk.

METHODS AND RESULTS

Using the UK Biobank database, 475,813 participants were selected for an average follow-up time of more than 10 years.

MAIN FINDINGS

1) there was a nonlinear relationship between BMI and AD risk in participants aged 60 years or older (p for non-linear < 0.001), but not in participants aged 37-59 years (p for non-linear = 0.717) using restricted cubic splines; 2) for participants aged 60 years and older, compared with the BMI (23-30 kg/m²) group, the BMI (< 23 kg/m²) group was associated with a higher AD risk (HR = 1.585; 95% CI 1.304-1.928, p < 0.001) and the BMI (> 30 kg/m²) group was associated with a lower AD risk (HR = 0.741; 95% CI 0.618-0.888, p < 0.01) analyzed using the Cox proportional risk model; 3) participants with a combination of high AD genetic risk score (AD-GRS) and BMI (< 23 kg/m²) were associated with the highest AD risk (HR = 3.034; 95% CI 2.057-4.477, p < 0.001). In addition, compared with the BMI (< 23 kg/m²), the higher BMI was associated with a lower risk of AD in participants with the same intermediate or high AD-GRS; 4) there was a reverse causality between BMI and AD when analyzed using bidirectional Mendelian randomization (MR).

CONCLUSION

There was a reverse causality between BMI and AD analyzed using MR. For participants aged 60 years and older, the higher BMI was associated with a lower risk of AD in participants with the same intermediate or high AD genetic risk. BMI (23-30 kg/m²) may be a potential intervention for AD.

Using a machine learning model to predict the development of acute kidney injury in patients with heart failure

Background

Heart failure (HF) is a life-threatening complication of cardiovascular disease. HF patients are more likely to progress to acute kidney injury (AKI) with a poor prognosis. However, it is difficult for doctors to distinguish which patients will develop AKI accurately. This study aimed to construct a machine learning (ML) model to predict AKI occurrence in HF patients.

Materials and methods

The data of HF patients from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database was retrospectively analyzed. A ML model was established to predict AKI development using decision tree, random forest (RF), support vector machine (SVM), K-nearest neighbor (KNN), and logistic regression (LR) algorithms. Thirty-nine demographic, clinical, and treatment features were used for model establishment. Accuracy, sensitivity, specificity, and the area under the receiver operating characteristic curve (AUROC) were used to evaluate the performance of the ML algorithms.

Results

A total of 2,678 HF patients were engaged in this study, of whom 919 developed AKI. Among 5 ML algorithms, the RF algorithm exhibited the highest performance with the AUROC of 0.96. In addition, the Gini index showed that the sequential organ function assessment (SOFA) score, partial pressure of oxygen (PaO₂), and estimated glomerular filtration rate (eGFR) were highly relevant to AKI development. Finally, to facilitate clinical application, a simple model was constructed using the 10 features screened by the Gini index. The RF algorithm also exhibited the highest performance with the AUROC of 0.95.

Conclusion

Using the ML model could accurately predict the development of AKI in HF patients.

Machine learning models for predicting survival in patients with ampullary adenocarcinoma

Objective

The aim of this study was to predict the long-term survival probability of patients with ampullary adenocarcinoma (AAC), which would provide a theoretical basis for the long-term care of these patients.

Methods

Data on patients with AAC during 2004–2015 were obtained from the Surveillance, Epidemiology, and End Results database, which were split at a 7:3 ratio into two independent cohorts: training and testing cohorts. Differences in survival between the two groups were tested using the Kaplan–Meier estimator and log-rank test methods. We constructed six survival analysis methods: the American Joint Committee on Cancer TNM stage, Cox Proportional Hazards regression, CoxTime, DeepSurv, XGBoost Survival Embeddings, and Random Survival Forest. The performances of these models were evaluated using the C-index, receiver operating characteristic (ROC), and calibration curves.

Results

This study included 2,935 patients with AAC. Univariate Cox regression analyses of the training cohort indicated that race, marital status at diagnosis, scope of regional lymph node surgery, tumor grade, summary stage, American Joint Committee on Cancer stage, TNM stage T, and TNM stage N were important factors affecting survival (P ​< ​0.05). The results of the C-index indicated that DeepSurv performed the best among the six models, with the highest C-index of 0.731. The areas under the ROC curves of the DeepSurv model at the 1-year, 3-year, 5-year, and 10-year time points were 0.823, 0.786, 0.803, and 0.813, respectively. The calibration curve indicated that DeepSurv performed well, with good calibration.

Conclusions

Machine learning models such as DeepSurv have a stronger performance in the survival analysis of patients with AAC.

Developing an ensemble machine learning model for early prediction of sepsis-associated acute kidney injury

Sepsis-associated acute kidney injury (S-AKI) is very common and early prediction is beneficial. This study aiming to develop an accurate ensemble model to predict the risk of S-AKI based on easily available clinical information. Patients with sepsis from the United States (US) database Medical Information Mart for Intensive Care-IV were used as a modeling cohort to predict the occurrence of AKI by combining Support Vector Machine, Random Forest, Neural Network, and Extreme Gradient Boost as four first-level learners via stacking algorithm. The external validation databases were the eICU Collaborative Research Database from US and Critical Care Database comprising infection patients at Zigong Fourth People’s Hospital from China, whose AUROC values for the ensemble model 48–12 h before the onset of AKI were 0.774–0.788 and 0.756–0.813, respectively. In this study, an ensemble model for early prediction of S-AKI onset was developed and it demonstrated good performance in multicenter external datasets.

•

We developed an ensemble model to predict sepsis-associated AKI early.

•

The model was constructed by stacking algorithm with high discriminative power.

•

External tests from US and China showed evidence of generalizability.

•

We constructed easy-to-access website calculator for clinicians.

Medication Regularity of Traditional Chinese Medicine in the Treatment of Aplastic Anemia Based on Data Mining

Objective

Aplastic anemia (AA) is an uncommon disease, characterized by pancytopenia and hypocellular bone marrow, but it is common in the blood system. The medication rules of traditional Chinese medicine (TCM) in the treatment of AA are not clear, for which it is worth exploring the medication rules by data mining methods.

Methods

This study used SPSS Modeler 18.0 and SPSS statistics to analyze the cases of AA from Zhejiang Provincial Hospital of Chinese Medicine (ZJHCM) from March 1, 2019, to March 1, 2022. Data mining methods, including frequency analysis, cluster analysis, and association rule learning, were performed in order to explore the medication rules for AA.

Results

(1) A total of 859 prescriptions, which met the inclusion criteria, consisted of 255 herbs. In descending order of the frequency of herbal medicine, we have Danggui, Huangqi, Shudihuang, Fuling, Gancao, Shanyao, Shanzhuyu, Baizhu, Dangshen, and Xianhecao. (2) Frequency analysis of herb properties: the Four Qi of 255 kinds of TCMs are mainly warm and neutral medicines. The Five Flavors are mainly sweet medicines, followed by bitter medicines. The main meridians are the liver, spleen, and kidney. (3) Clustering of medications: TCMs with the top 20 frequencies are classified into 9 groups by cluster analysis. (4) Association rule analysis of high-frequency herbs: using the Apriori algorithm, the results showed that there were 3 herb pairs with support of over 0.3 and 12 herb pairs with confidence above 0.85.

Conclusion

The basic pathogenesis of AA (Sui Lao) is spleen and kidney essence deficiency, Qi deficiency, and blood stasis. The main herbs have warm and neutral properties, sweet tastes, and liver, spleen, and kidney meridian tropisms, whose purpose is to tonify the kidney and invigorate the spleen, tonify Qi, and promote blood circulation.

Prediction of low cardiac output syndrome in patients following cardiac surgery using machine learning

Background

This study aimed to develop machine learning models to predict Low Cardiac Output Syndrome (LCOS) in patients following cardiac surgery using machine learning algorithms.

Methods

The clinical data of cardiac surgery patients in Nanjing First Hospital between June 2019 and November 2020 were retrospectively extracted from the electronic medical records. Six conventional machine learning algorithms, including logistic regression, support vector machine, decision tree, random forest, extreme gradient boosting and light gradient boosting machine, were employed to construct the LCOS predictive models with all predictive features (full models) and selected predictive features (reduced models). The discrimination of these models was evaluated by the area under the receiver operating characteristic curve (AUC) and the calibration of the models was assessed by the calibration curve. Shapley Additive explanation (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME) were used to interpret the predictive models.

Results

Data from 1,585 patients [982 (62.0%) were male, aged 18 to 88, 212 (13.4%) with LCOS] were employed to train and validate the LCOS models. Among the full models, the RF model (AUC: 0.909, 95% CI: 0.875–0.943; Sensitivity: 0.849, 95% CI: 0.724–0.933; Specificity: 0.835, 95% CI: 0.796–0.869) and the XGB model (AUC: 0.897, 95% CI: 0.859–0.935; Sensitivity: 0.830, 95% CI: 0.702–0.919; Specificity: 0.809, 95% CI: 0.768–0.845) exhibited well predictive power for LCOS. Eleven predictive features including left ventricular ejection fraction (LVEF), first post-operative blood lactate (Lac), left ventricular diastolic diameter (LVDd), cumulative time of mean artery blood pressure (MABP) lower than 65 mmHg (MABP &lt; 65 time), hypertension history, platelets level (PLT), age, blood creatinine (Cr), total area under curve above threshold central venous pressure (CVP) 12 mmHg and 16 mmHg, and blood loss during operation were used to build the reduced models. Among the reduced models, RF model (AUC: 0.895, 95% CI: 0.857–0.933; Sensitivity: 0.830, 95% CI: 0.702–0.919; Specificity: 0.806, 95% CI: 0.765–0.843) revealed the best performance. SHAP and LIME plot showed that LVEF, Lac, LVDd and MABP &lt; 65 time significantly contributed to the prediction model.

Conclusion

In this study, we successfully developed several machine learning models to predict LCOS after surgery, which may avail to risk stratification, early detection and management of LCOS after cardiac surgery.