Decision Curve Analysis Research Articles

Machine Learning for Predicting Risk and Prognosis of Acute Kidney Disease in Critically Ill Elderly Patients During Hospitalization: Internet-Based and Interpretable Model Study.

Acute kidney disease (AKD) affects more than half of critically ill elderly patients with acute kidney injury (AKI), which leads to worse short-term outcomes. We aimed to establish 2 machine learning models to predict the risk and prognosis of AKD in the elderly and to deploy the models as online apps. Data on elderly patients with AKI (n=3542) and AKD (n=2661) from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database were used to develop 2 models for predicting the AKD risk and in-hospital mortality, respectively. Data collected from Xiangya Hospital of Central South University were for external validation. A bootstrap method was used for internal validation to obtain relatively stable results. We extracted the indicators within 24 hours of the first diagnosis of AKI and the fluctuation range of some indicators, namely delta (day 3 after AKI minus day 1), as features. Six machine learning algorithms were used for modeling; the area under the receiver operating characteristic curve (AUROC), decision curve analysis, and calibration curve for evaluating; Shapley additive explanation (SHAP) analysis for visually interpreting; and the Heroku platform for deploying the best-performing models as web-based apps. For the model of predicting the risk of AKD in elderly patients with AKI during hospitalization, the Light Gradient Boosting Machine (LightGBM) showed the best overall performance in the training (AUROC=0.844, 95% CI 0.831-0.857), internal validation (AUROC=0.853, 95% CI 0.841-0.865), and external (AUROC=0.755, 95% CI 0.699-0.811) cohorts. In addition, LightGBM performed well for the AKD prognostic prediction in the training (AUROC=0.861, 95% CI 0.843-0.878), internal validation (AUROC=0.868, 95% CI 0.851-0.885), and external (AUROC=0.746, 95% CI 0.673-0.820) cohorts. The models deployed as online prediction apps allowed users to predict and provide feedback to submit new data for model iteration. In the importance ranking and correlation visualization of the model's top 10 influencing factors conducted based on the SHAP value, partial dependence plots revealed the optimal cutoff of some interventionable indicators. The top 5 factors predicting the risk of AKD were creatinine on day 3, sepsis, delta blood urea nitrogen (BUN), diastolic blood pressure (DBP), and heart rate, while the top 5 factors determining in-hospital mortality were age, BUN on day 1, vasopressor use, BUN on day 3, and partial pressure of carbon dioxide (PaCO2). We developed and validated 2 online apps for predicting the risk of AKD and its prognostic mortality in elderly patients, respectively. The top 10 factors that influenced the AKD risk and mortality during hospitalization were identified and explained visually, which might provide useful applications for intelligent management and suggestions for future prospective research.

The use of longitudinal CT-based radiomics and clinicopathological features predicts the pathological complete response of metastasized axillary lymph nodes in breast cancer

BackgroundAccurate assessment of axillary status after neoadjuvant therapy for breast cancer patients with axillary lymph node metastasis is important for the selection of appropriate subsequent axillary treatment decisions. Our objectives were to accurately predict whether the breast cancer patients with axillary lymph node metastases could achieve axillary pathological complete response (pCR).MethodsWe collected imaging data to extract longitudinal CT image features before and after neoadjuvant chemotherapy (NAC), analyzed the correlation between radiomics and clinicopathological features, and developed models to predict whether patients with axillary lymph node metastasis can achieve axillary pCR after NAC. The clinical utility of the models was determined via decision curve analysis (DCA). Subgroup analyses were also performed. Then, a nomogram was developed based on the model with the best predictive efficiency and clinical utility and was validated using the calibration plots.ResultsA total of 549 breast cancer patients with metastasized axillary lymph nodes were enrolled in this study. 42 independent radiomics features were selected from LASSO regression to construct a logistic regression model with clinicopathological features (LR radiomics-clinical combined model). The AUC of the LR radiomics-clinical combined model prediction performance was 0.861 in the training set and 0.891 in the testing set. For the HR + /HER2 − , HER2 + , and Triple negative subtype, the LR radiomics-clinical combined model yields the best prediction AUCs of 0.756, 0.812, and 0.928 in training sets, and AUCs of 0.757, 0.777 and 0.838 in testing sets, respectively.ConclusionsThe combination of radiomics features and clinicopathological characteristics can effectively predict axillary pCR status in NAC breast cancer patients.

Identification of patients with internet gaming disorder via a radiomics-based machine learning model of subcortical structures in high-resolution T1-weighted MRI

It is of vital importance to establish an objective and reliable model to facilitate the early diagnosis and intervention of internet gaming disorder (IGD). A total of 133 patients with IGD and 110 healthy controls (HCs) were included. We extracted radiomic features of subcortical structures in high-resolution T1-weighted MRI. Different combinations of four feature selection methods (analysis of variance, Kruskal–Wallis, recursive feature elimination and relief) and ten classification algorithms were used to identify the most robust combined models for distinguishing IGD patients from HCs. Furthermore, a nomogram incorporating radiomic signatures and independent clinical factors was developed. Calibration curve and decision curve analyses were used to evaluate the nomogram. The combination of analysis of variance selector and logistic regression classifier identified that the radiomic model constructed with 20 features from the right caudate nucleus and amygdala showed better IGD screening performance. The radiomic model produced good areas under the curves (AUCs) in the training, validation and test cohorts (AUCs of 0.961, 0.903 and 0.895, respectively). In addition, sex, internet addiction test scores and radiomic scores were included in the nomogram as independent risk factors for IGD. Analysis of the correction curve and decision curve showed that the clinical-radiomic model has good reliability (C-index: 0.987). The nomogram incorporating radiomic features of subcortical structures and clinical characteristics achieved satisfactory classification performance and could serve as an effective tool for distinguishing IGD patients from HCs.

A nomogram to predict the risk of venous thromboembolism in patients with colon cancer in China.

To create a nomogram for predicting the likelihood of venous thromboembolism (VTE) in colon cancer patients from China. The data of colon cancer patients from Chongqing University Cancer Hospital between 2019 and 2022 were analyzed. Patients were divided into training set and internal validation set by random split-sample method in a split ratio of 7:3. The univariable and multivariable logistic analysis gradually identified the independent risk factors for VTE. A nomogram was created using all the variables that had a significance level of p < 0.05 in the multivariable logistic analysis and those with clinical significance. Calibration curves and clinical decision curve analysis (DCA) were used to assess model's fitting performance and clinical value. Harrell's C-index (concordance statistic) and the area under the receiver operating characteristic curves (AUC) were used to evaluate the predictive effectiveness of models. A total of 1996 patients were ultimately included. There were 1398 patients in the training set and 598 patients in the internal validation set. The nomogram included age, chemotherapy, targeted therapy, hypertension, activated partial thromboplastin time, prothrombin time, platelet, absolute lymphocyte count, and D-dimer. The C-index of nomogram and Khorana score were 0.754 (95% CI 0.711-0.798), 0.520 (95% CI 0.477-0.563) in the training cohort and 0.713 (95% CI 0.643-0.784), 0.542 (95% CI 0.473-0.612) in the internal validation cohort. We have established and validated a nomogram to predict the VTE risk of colon cancer patients in China, which encompasses a diverse age range, a significant population size, and various clinical factors. It facilitates the identification of high-risk groups and may enable the implementation of targeted preventive measures.

The relationship between intraoperative hypothermia and postoperative delirium: The PNDRFAP study.

Our study aimed to investigate the correlation between intraoperative hypothermia and postoperative delirium (POD) in patients undergoing general anesthesia for gastrointestinal surgery. The study comprised 750 participants from the Perioperative Neurocognitive Disorder Risk Factor and Prognosis (PNDRFAP) study database, which ultimately screened 510 individuals in the final analysis. Preoperative cognitive function was evaluated using the Mini-Mental State Examination (MMSE). The occurrence of POD was determined using the Confusion Assessment Method, and the severity of POD was evaluated using the Memorial Delirium Assessment Scale. Logistic regression was employed to scrutinize the association between intraoperative hypothermia and the incidence of POD, and the sensitivity analysis was conducted by introducing adjusted confounding variables. Decision curves and a nomogram model were utilized to assess the predictive efficacy of intraoperative hypothermia for POD. Mediation analysis involving 10,000 bootstrapped iterations was employed to appraise the suggested mediating effect of numeric rating scale (NRS) scores at 24 and 48h post-surgeries. The receiver-operating characteristic (ROC) was utilized to evaluate the effectiveness of intraoperative hypothermia in predicting POD. In the PNDRFAP study, the occurrence of POD was notably higher in the intraoperative hypothermia group (62.2%) compared to the intraoperative normal body temperature group (9.8%), with an overall POD incidence of 17.6%. Logistic regression analysis, adjusted for various confounding factors (age [40-90], gender, education, MMSE, smoking history, drinking history, hypertension, diabetes, and the presence of cardiovascular heart disease), demonstrated that intraoperative hypothermia significantly increased the risk of POD (OR=4.879, 95% CI=3.020-7.882, p<.001). Mediation analyses revealed that the relationship between intraoperative hypothermia and POD was partially mediated by NRS 24h after surgery, accounting for 14.09% of the association (p=.002). The area under the curve of the ROC curve was 0.685, which confirmed that intraoperative hypothermia could predict POD occurrence to a certain extent. Decision curve and nomogram analyses, conducted using the R package, further substantiated the predictive efficacy of intraoperative hypothermia on POD. Intraoperative hypothermia may increase the risk of POD, and this association may be partially mediated by NRS scores 24h after surgery.

Clinical study on forecasting the prognosis of patients with cerebellar hemorrhage based on CT radiomics models

Background and purposeWe created and verified a clinical–radiomics nomogram to forecast the prognosis of patients with cerebellar hemorrhage. Materials and MethodsEighty-eight patients with cerebellar hemorrhage from two different hospitals were retrospectively analyzed (62 cases from first hospital (training cohort) and 27 cases from second hospital (external validation cohort)). All patients underwent computed tomography scanning within 72 hours of the onset of cerebellar hemorrhage and were split into two categories based on their prognosis. Clinically independent risk factors were identified using the independent samples t-test and binary logistic regression analysis. Radiomic features were obtained using the SelectkBest, variance threshold, least absolute shrinkage, and selection operator. Radiomics scores (rad-scores) were produced from eight radiomics features of perihematomal edema (PHE) and hematoma associated with cerebellar hemorrhage prognosis. Logistic regression analyses of the rad-scores against clinical risk factors yielded clinical–radiomics column line plots. Predictive performance of the column line plots was validated using decision curve analysis analyses in both the training and external validation cohorts. ResultsTo construct rad-scores, eight radiomics traits and four clinical risk factors were combined to create clinical–radiomics models. In both the training cohort (area under the curve [AUC], 0.96 versus vs. 0.87/0.84) and the external validation cohort (AUC: 0.93 vs. 0.74/0.69), the clinical–radiomics nomogram achieved superior performance to the clinical and radiomics nomogram models. Clinical–radiomics column line plot demonstrated satisfactory calibration and clinical advantages in both cohorts. ConclusionsRadiomics features of PHE and hematoma can predict cerebellar hemorrhage prognosis. Clinical–radiomics models can help clinicians make individualized clinical treatment decisions and efficacy assessments for patients with cerebellar hemorrhage, help physicians prepare for possible deterioration, and promptly communicate treatment goals and efficacy expectations to patients and their families.

A real-world study was conducted to develop a nomogram that predicts the occurrence of anastomotic leakage in patients with esophageal cancer following esophagectomy.

The incidence of anastomotic leakage (AL) following esophagectomy is regarded as a noteworthy complication. There is a need for biomarkers to facilitate early diagnosis of AL in high-risk esophageal cancer (EC) patients, thereby minimizing its morbidity and mortality. We assessed the predictive abilities of inflammatory biomarkers for AL in patients after esophagectomy. In order to ascertain the predictive efficacy of biomarkers for AL, Receiver Operating Characteristic (ROC) curves were generated. Furthermore, univariate, LASSO, and multivariate logistic regression analyses were conducted to discern the risk factors associated with AL. Based on these identified risk factors, a diagnostic nomogram model was formulated and subsequently assessed for its predictive performance. Among the 438 patients diagnosed with EC, a total of 25 patients encountered AL. Notably, elevated levels of interleukin-6 (IL-6), IL-10, C-reactive protein (CRP), and procalcitonin (PCT) were observed in the AL group as compared to the non-AL group, demonstrating statistical significance. Particularly, IL-6 exhibited the highest predictive capacity for early postoperative AL, exhibiting a sensitivity of 92.00% and specificity of 61.02% at a cut-off value of 132.13 pg/ml. Univariate, LASSO, and multivariate logistic regression analyses revealed that fasting blood glucose ≥7.0mmol/L and heightened levels of IL-10, IL-6, CRP, and PCT were associated with an augmented risk of AL. Consequently, a nomogram model was formulated based on the results of multivariate logistic analyses. The diagnostic nomogram model displayed a robust discriminatory ability in predicting AL, as indicated by a C-Index value of 0.940. Moreover, the decision curve analysis provided further evidence supporting the clinical utility of this diagnostic nomogram model. This predictive instrument can serve as a valuable resource for clinicians, empowering them to make informed clinical judgments aimed at averting the onset of AL.

The model for predicting the central lymph node metastasis in cN0 papillary thyroid microcarcinoma with Hashimoto's thyroiditis.

The relationship between Hashimoto's thyroiditis (HT) and papillary thyroid microcarcinoma (PTMC) is controversial. These include central lymph node metastasis (CLNM), which affects the prognosis of PTMC patients. This study aimed to establish a predictive model combining ultrasonography and clinicopathological features to accurately evaluate latent CLNM in PTMC patients with HT at the clinical lymph node-negative (cN0) stage. In this study, 1102 PTMC patients who received thyroidectomy and central cervical lymph node dissection (CLND) from the First Affiliated Hospital of Shandong First Medical University from January 2021 to December 2022 and the 960th Hospital of PLA from January 2021 to December 2022 were jointly collected. The clinical differences between PTMCs with HT and those without HT were compared. A total of 373 PTMCs with HT in cN0 were randomly divided into a training cohort and a validation cohort. By analyzing and screening the risk factors of CLNM, a nomogram model was established and verified. The predictive performance was measured by the receiver operating characteristic (ROC) curve, calibration curve, and clinical decision curve analysis (DCA). The ratio of central lymph node metastasis (CLNMR) in PTMCs with HT was 0.0% (0.0%, 15.0%) and 7.7% (0.0%, 40.0%) in the non-HT group (P<0.001). Multivariate logistic regression analysis showed that age, gender, calcification, adjacent to trachea or capsule, and TPOAB were predictors of CLNM in PTMCs with HT. The areas under the curve (AUC) of the prediction models in the training cohort and the validation cohort were 0.835 and 0.825, respectively, which showed good differentiation ability. DCA indicates that the prediction model also has high net benefit and clinical practical value. This study found that CLN involvement was significantly reduced in PTMC patients with HT, suggesting that different methods should be used to predict CLNM in PTMC patients with HT and without HT, to more accurately assist preoperative clinical evaluation. The actual CLNM situation of PTMCs with HT in cN0 can be accurately predicted by the combination of ultrasonography and clinicopathological features.

Identification of prognostic signatures in remnant gastric cancer through an interpretable risk model based on machine learning: a multicenter cohort study

ObjectiveThe purpose of this study was to develop an individual survival prediction model based on multiple machine learning (ML) algorithms to predict survival probability for remnant gastric cancer (RGC).MethodsClinicopathologic data of 286 patients with RGC undergoing operation (radical resection and palliative resection) from a multi-institution database were enrolled and analyzed retrospectively. These individuals were split into training (80%) and test cohort (20%) by using random allocation. Nine commonly used ML methods were employed to construct survival prediction models. Algorithm performance was estimated by analyzing accuracy, precision, recall, F1-score, area under the receiver operating characteristic curve (AUC), confusion matrices, five-fold cross-validation, decision curve analysis (DCA), and calibration curve. The best model was selected through appropriate verification and validation and was suitably explained by the SHapley Additive exPlanations (SHAP) approach.ResultsCompared with the traditional methods, the RGC survival prediction models employing ML exhibited good performance. Except for the decision tree model, all other models performed well, with a mean ROC AUC above 0.7. The DCA findings suggest that the developed models have the potential to enhance clinical decision-making processes, thereby improving patient outcomes. The calibration curve reveals that all models except the decision tree model displayed commendable predictive performance. Through CatBoost-based modeling and SHAP analysis, the five-year survival probability is significantly influenced by several factors: the lymph node ratio (LNR), T stage, tumor size, resection margins, perineural invasion, and distant metastasis.ConclusionsThis study established predictive models for survival probability at five years in RGC patients based on ML algorithms which showed high accuracy and applicative value.

Survival nomogram and risk classification system for patients with adrenocortical carcinoma: a study based on the SEER database and external validation in China.

Adrenocortical carcinoma (ACC) is an extremely rare and highly invasive malignant tumor. However, there is currently no reliable method to predict the prognosis of ACC. Our objective is to construct a nomogram and a risk classification system to predict the 1-year, 3-year, and 5-year overall survival (OS) of ACC. We retrieved clinicopathological data of patients diagnosed with ACC in The Surveillance, Epidemiology, and End Results (SEER) database and divided them into training and validation cohorts with a 7:3 ratio. Simultaneously, we collected an external validation cohort from The First Affiliated Hospital of Naval Medical University (Shanghai, China). Univariate and multivariate Cox analyses were performed to identify relevant risk factors, which were then combined to develop a correlation nomogram. The predictive performance of the nomogram was evaluated using the concordance index (C-index), receiver-operating characteristic curve (ROC), and calibration curves. Decision curve analysis (DCA) was applied to assess the clinical utility of the nomogram. In addition, Kaplan-Meier survival curves were generated to demonstrate the variation in OS between groups. The final nomogram consisted of five factors: age, T, N, M, and history of chemotherapy. Our prognostic model demonstrated significant discriminative ability, with C-index and the area under the receiver operating characteristic (AUC) values exceeding 0.70. Additionally, DCA validated the clinical utility of the nomogram. In the entire cohort, the median OS for patients in the low- and high-risk groups was 70 and 10 months, respectively. A nomogram and a corresponding risk classification system were developed in order to predict the OS of patients diagnosed with ACC. These tools have the potential to provide valuable support for patient counseling and assist in the decision-making process related to treatment options.

RETFound-enhanced community-based fundus disease screening: real-world evidence and decision curve analysis.

Visual impairments and blindness are major public health concerns globally. Effective eye disease screening aided by artificial intelligence (AI) is a promising countermeasure, although it is challenged by practical constraints such as poor image quality in community screening. The recently developed ophthalmic foundation model RETFound has shown higher accuracy in retinal image recognition tasks. This study developed an RETFound-enhanced deep learning (DL) model for multiple-eye disease screening using real-world images from community screenings. Our results revealed that our DL model improved the sensitivity and specificity by over 15% compared with commercial models. Our model also shows better generalisation ability than AI models developed using traditional processes. Additionally, decision curve analysis underscores the higher net benefit of employing our model in both urban and rural settings in China. These findings indicate that the RETFound-enhanced DL model can achieve a higher net benefit in community-based screening, advocating its adoption in low- and middle-income countries to address global eye health challenges.

A nomogram to predict cryptococcal meningitis in patients with pulmonary cryptococcosis

BackgroundThe most serious manifestation of pulmonary cryptococcosis is complicated with cryptococcal meningitis, while its clinical manifestations lack specificity with delayed diagnosis and high mortality. The early prediction of this complication can assist doctors to carry out clinical interventions in time, thus improving the cure rate. This study aimed to construct a nomogram to predict the risk of cryptococcal meningitis in patients with pulmonary cryptococcosis through a scoring system. MethodsThe clinical data of 525 patients with pulmonary cryptococcosis were retrospectively analyzed, including 317 cases (60.38 %) with cryptococcal meningitis and 208 cases (39.62 %) without cryptococcal meningitis. The risk factors of cryptococcal meningitis were screened by univariate analysis, LASSO regression analysis and multivariate logistic regression analysis. Then the risk factors were incorporated into the nomogram scoring system to establish a prediction model. The model was validated by receiver operating characteristic (ROC) curve, decision curve analysis (DCA) and clinical impact curve. ResultsFourteen risk factors for cryptococcal meningitis in patients with pulmonary cryptococcosis were screened out by statistical method, including 6 clinical manifestations (fever, headache, nausea, psychiatric symptoms, tuberculosis, hematologic malignancy) and 8 clinical indicators (neutrophils, lymphocytes, glutamic oxaloacetic transaminase, T cells, helper T cells, killer T cells, NK cells and B cells). The AUC value was 0.978 (CI 96.2 %∼98.9 %), indicating the nomogram was well verified. ConclusionThe nomogram scoring system constructed in this study can accurately predict the risk of cryptococcal meningitis in patients with pulmonary cryptococcosis, which may provide a reference for clinical diagnosis and treatment of patients with cryptococcal meningitis.

Establishment and validation of a nomogram for predicting overall survival of upper-tract urothelial carcinoma with bone metastasis: a population-based study

BackgroundBone metastasis (BM) carries a poor prognosis for patients with upper-tract urothelial carcinoma (UTUC). This study aims to identify survival predictors and develop a prognostic nomogram for overall survival (OS) in UTUC patients with BM.MethodsThe Surveillance, Epidemiology, and End Results database was used to select patients with UTUC between 2010 and 2019. The chi-square test was used to assess the baseline differences between the groups. Kaplan–Meier analysis was employed to assess OS. Univariate and multivariate analyses were conducted to identify prognostic factors for nomogram establishment. An independent cohort was used for external validation of the nomogram. The discrimination and calibration of the nomogram were evaluated using concordance index (C-index), area under receiver operating characteristic curve (AUC), calibration curve, and decision curve analysis (DCA). All statistical analyses were performed using SPSS 23.0 and R software 4.2.2.ResultsThe mean OS for UTUC patients with BM was 10 months (95% CI: 8.17 to 11.84), with 6-month OS, 1-year OS, and 3-year OS rates of 41%, 21%, and 3%, respectively. Multi-organ metastases (HR = 2.21, 95% CI: 1.66 to 2.95, P < 0.001), surgery (HR = 0.72, 95% CI: 0.56 to 0.91, P = 0.007), and chemotherapy (HR = 0.37, 95% CI: 0.3 to 0.46, P < 0.001) were identified as independent prognostic factors. The C-index was 0.725 for the training cohort and 0.854 for the validation cohort, and all AUC values were > 0.679. The calibration curve and DCA curve showed the accuracy and practicality of the nomogram.ConclusionsThe OS of UTUC patients with BM was poor. Multi-organ metastases was a risk factor for OS, while surgery and chemotherapy were protective factors. Our nomogram was developed and validated to assist clinicians in evaluating the OS of UTUC patients with BM.

Prognostic value of preoperative combined neutrophil, monocyte, and lymphocyte scores in patients with renal cell carcinoma undergoing laparoscopic nephrectomy: A retrospective study.

In a multi-institutional clinical study, we assessed the prognostic significance of a novel indicator preoperative peripheral blood immune (PBIS) scores that combined ratios of preoperative lymphocyte, monocyte, and neutrophil of renal cell carcinoma (RCC) patients undergoing laparoscopic nephrectomy. Between January 2014 and December 2019, 438 patients with RCC were retrospectively analyzed in three centers. We used X-tile software to obtain the optimum cut-off values for neutrophils, monocytes, and lymphocytes to classify the patients. To assess the relationship between PBIS score and overall survival (OS), and cancer-specific survival (CSS) in patients with RCC by Kaplan-Meier survival curves and Cox regression analyses. In addition, predictive OS and CSS nomograms were constructed. The discriminative ability of nomogram and predictive performance accuracy were verified with consistency index (C-index), calibration curves, receiver operating curve (ROC) curves, decision curve analysis (DCA) curves, and time-dependent ROC curves. The optimum cutoff values for monocytes, lymphocytes, and neutrophils were 0.46, 1.01, and 4.50, respectively. We divided patients into four subgroups according to PBIS scores, which were significantly associated with M-stage (p = 0.008), T-stage (p < 0.001), N-stage (p = 0.006), and AJCC stage (p < 0.001). Multivariate Cox regression analysis revealed that RCC patients with lower PBIS scores showed a worse postoperative prognosis and served as an independent predictor of OS (p = 0.002) and CSS (p < 0.001). Ultimately, the nomograms based on PBIS scores demonstrated excellent predictive performance for OS (C-index: 0.770) and CSS (C-index: 0.828) through the analysis of calibration curves, ROC curves, DCA curves, and time-dependent ROC curves. PBIS score served as novel and effective predictor to accurately predict OS and CSS in patients with RCC receiving laparoscopic nephrectomy.

A nomogram for predicting postoperative delirium in pediatric patients following cardiopulmonary bypass: A prospective observational study

ObjectivesTo create a nomogram for early delirium detection in pediatric patients following cardiopulmonary bypass. Research Methodology/DesignThis prospective, observational study was conducted in the Cardiac Intensive Care Unit at a Children's Hospital, enrolling 501 pediatric patients from February 2022 to January 2023. Perioperative data were systematically collected through the hospital information system. Postoperative delirium was assessed using the Cornell Assessment of Pediatric Delirium (CAPD). For model development, Least Absolute Shrinkage and Selection Operator (LASSO) regression was employed to identify the most relevant predictors. These selected predictors were then incorporated into a multivariable logistic regression model to construct the predictive nomogram. The performance of the model was evaluated by Harrell's concordance index, receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis. External validity of the model was confirmed through the C-index and calibration plots. ResultsFive independent predictors were identified: age, SpO2 levels, lymphocyte count, diuretic use, and midazolam administration, integrated into a predictive nomogram. This nomogram demonstrated strong predictive capacity (AUC 0.816, concordance index 0.815) with good model fit (Hosmer-Lemeshow test p = 0.826) and high accuracy. Decision curve analysis showed a significant net benefit, and external validation confirmed the nomogram's reliability. ConclusionsThe study successfully developed a precise and effective nomogram for identifying pediatric patients at high risk of post-cardiopulmonary bypass delirium, incorporating age, SpO2 levels, lymphocyte counts, diuretic use, and midazolam medication. Implications for Clinical PracticeThis nomogram aids early delirium detection and prevention in critically ill children, improving clinical decisions and treatment optimization. It enables precise monitoring and tailored medication strategies, significantly contributes to reducing the incidence of delirium, thereby enhancing the overall quality of patient care.

Predictive value of lactate dehydrogenase for Mycoplasma pneumoniae necrotizing pneumonia in children based on decision curve analysis and dose–response analysis

Mycoplasma pneumoniae necrotizing pneumonia (MPNP) has a long and severe disease course, which seriously threatens to jeopardize patients' lives and health. Early prediction is essential for good recovery and prognosis. In the present study, we retrospect 128 children with MPNP and 118 children with Mycoplasma pneumoniae pneumonia combined with pulmonary consolidation to explore the predictive value of lactate dehydrogenase (LDH) in children with MPNP by propensity score matching method, multiple logistic regression analysis, dose–response analysis and decision curve analysis. The WBC count, PLT count and percentage of neutrophils were significantly higher in necrosis group than consolidation group. The serum CRP, PCT, ESR, D-D, FIB, ALT, LDH, IgG and IgM were significantly higher in necrosis group. Compared to consolidation group, necrosis group is more severe in chest pain and dyspnea. Multivariate logistic regression analysis showed that duration of LDH levels, high fever, d-dimer, and fibrinogen were independent predictive factors for the incidence of MPNP. Restricted cubic spline analysis showed that a non-linear dose–response relationship between the continuous changes of LDH level and the incidence of MPNP. Decision curve analysis revealed that LDH had an important clinical value in predicting MPNP. This study provides a potential serologic indicator for early diagnosis of MPNP.

Live-donor kidney transplant outcome prediction (L-TOP) using artificial intelligence.

Outcome prediction for live-donor kidney transplantation improves clinical and patient decisions and donor selection. However, the concurrently used models are of limited discriminative or calibration power and there is a critical need to improve the selection process. We aimed to assess the value of various artificial intelligence (AI) algorithms to improve the risk stratification index. We evaluated pre-transplant variables among 66914 live-donor kidney transplants (performed between 01/12/2007-01/06/2021) from the United Network of Organ Sharing database, randomized into training (80%) and test (20%) sets. The primary outcome measure was death-censored graft survival. We tested four machine learning models for discrimination (time-dependent concordance index, CTD, and area under the ROC curve) and calibration (integrated Brier score, IBS). We used decision curve analysis to assess the potential clinical utility. Among the models, the deep Cox mixture model showed the best discriminative performance (AUC=0.70, 0.68, and 0.68 at 5, 10, and 13 years post-transplant, respectively). CTD reached 0.70, 0.67, and 0.66 at 5, 10, and 13 years post-transplant. The IBS score was 0.09, indicating good calibration. In comparison, applying the Living Kidney Donor Profile Index (LKDPI) on the same cohort produced a CTD of 0.56 and an AUC of 0.55-0.58 only. Decision curve analysis showed an additional net benefit compared to the LKDPI, 'Treat all' and 'Treat None' approaches. Our AI-based deep Cox mixture model, termed Live-Donor Kidney Transplant Outcome Prediction outperforms existing prediction models, including the LKDPI, with the potential to improve decisions for optimum live donor selection by ranking potential transplant pairs based on graft survival. This model could be adopted to improve the outcomes of paired exchange programs.

Development and validation of a nomogram for predicting poor operative visibility during FESS in Chinese adult patients with CRS.

The purpose of this study is to develop and evaluate a nomogram that is capable of predicting poor operative visibility during functional endoscopic sinus surgery. To identify potential risk factors, patients with chronic rhinosinusitis who underwent functional endoscopic sinus surgery (FESS) between January 2019 and December 2022 were selected from our hospital's electronic medical record system. Data on general patient information, clinical manifestations, clotting-related test indices, Lund-Machay score of sinuses CT scanning, Lund-kennedy score of nasal endoscopies, anesthesia methods, intraoperative blood pressure and heart rate, and Boezaart bleeding score were collected. Minimum absolute convergence and selection operator (LASSO) regression, as well as multivariate logistic regression, were used to determine the risk factors. A nomogram was developed in order to predict poor operating visibility during FESS, and its performance was evaluated utilizing both the training and verification datasets via various measures including receiver operating characteristic (ROC) curve analysis, area under the curve (AUC), Hosmer-Lemeshow goodness-of-fit test, calibration curve, and decision curve analysis. Of the 369 patients who met the inclusion criteria, 88 of them exhibited POV during FESS. By deploying LASSO and multivariate logistic regression analyses, six risk factors were identified and used to construct a nomogram for predicting POV during FESS. These factors include prothrombin time (PT), prothrombin activity (PTA), Lund-Mackay score (LMS), Lund-Kennedy score (LKS), anesthetic method, and intraoperative hypertension. The AUC of the training set was found to be 0.820 while that of the verification set was 0.852. The Hosmer-Lemeshow goodness-of-fit test and calibration curve analysis revealed good consistency between predicted and actual probabilities. Also, the decision curve demonstrated that the nomogram had a high degree of clinical usefulness and net benefit. The constructed nomogram has a strong ability to predict the poor intraoperative field in patients with chronic rhinosinusitis, which can help preoperative judgment of high-risk patients and provide evidence for perioperative management and preoperative plan formulation.

Analysis of risk factors of short-term prognosis in patients with severe Budd-Chiari syndrome

Objective: To explore the risk factors of short-term prognosis of severe BCS patients,established and verified the nomogram prediction model for these BCS patients and evaluated its clinical application value. Methods: This study is a retrospective cohort study. The clinical data of 171 patients with severe BCS diagnosed were retrospectively analyzed in the Department of Hepatopancreatobiliary Surgery First Affiliated Hospital of Zhengzhou University from January 2018 to December 2023. There were 105 males and 66 females, aged (52.1±12.8) years (range: 18 to 79 years). The patients were divided into two groups based on whether they died within 28 days: the death group (n=38) and the survival group (n=133). The risk factors for short-term death of patients were analyzed,and independent risk factors were screened by univariate and multivariate analysis. Furthermore,these factors were used to establish the nomogram prediction model. The area under the curve(AUC),the Bootstrap Resampling,the Hosmer-Lemeshow test and the Decision Curve Analysis(DCA) were used to verify the model's differentiation,internal verification,calibration degree and clinical effectiveness,respectively. Results: Univariate and multivariate Logistics regression analysis showed that the history of hepatic encephalopathy,white blood cell,glomerular filtration rate and prothrombin time are independent risk factors (P<0.05). The above factors were used to successfully establish the prediction model with 0.908 of AUC and 0.895 of the internal verification of AUC,indicating that the predictive model was valuable. The 0.663 P-values in the Hosmer-Lemeshow test indicated the high calibration degree of the model. The clinical effectiveness of the model was proved by the 18% clinical benefit population using the DCA curve with the 17% probability threshold. Conclusions: The independent risk factors are the history of hepatic encephalopathy,white blood cell,glomerular filtration rate and prothrombin time. An adequate basis was acquired by establishing a nomogram prediction model of the short-term prognosis of severe BCS,which was helpful for early clinical screening and identification of high-risk patients with severe BCS who could die in the short term and timely providing timely intervention measures for improving the prognosis.

Development of a nomogram for predicting in-hospital mortality in patients with liver cirrhosis and sepsis

In this study, we aimed to investigate the risk factors associated with in-hospital mortality in patients with cirrhosis and sepsis, establish and validate the nomogram. This retrospective study included patients diagnosed with liver cirrhosis and sepsis in the Medical Information Mart for Intensive Care IV (MIMIC-IV). Models were compared by the area under the curve (AUC), integrated discriminant improvement (IDI), net reclassification index (NRI) and decision curve analysis (DCA). A total of 1,696 patients with cirrhosis and sepsis were included in the final cohort. Our final model included the following 9 variables: age, heartrate, total bilirubin (TBIL), glucose, sodium, anion gap (AG), fungal infections, mechanical ventilation, and vasopressin. The nomogram were constructed based on these variables. The AUC values of the nomograms were 0.805 (95% CI 0.776–0.833), which provided significantly higher discrimination compared to that of SOFA score [0.684 (95% CI 0.647–0.720)], MELD-Na [0.672 (95% CI 0.636–0.709)] and ABIC [0.674(95% CI 0.638–0.710)]. We established the first nomogram for predicting in-hospital mortality in patients with liver cirrhosis and sepsis based on these factors. This nomogram can performs well and facilitates clinicians to identify people at high risk of in-hospital mortality.