Hosmer Lemeshow Research Articles

Value of Radiomic Analysis Combined With Diffusion Tensor Imaging in Early Diagnosis of HIV-Associated Neurocognitive Disorders.

The combination of radiomics and diffusion tensor imaging (DTI) may have potential clinical value in the early stage of HIV-associated neurocognitive disorders (HAND). To investigate the value of DTI-based radiomics in the early stage of HAND in people living with HIV (PLWH). Retrospective. A total of 138 male PLWH were included, including 68 with intact cognition (IC) and 70 with asymptomatic neurocognitive impairment (ANI). Seventy healthy controls (HCs) were recruited for tract-based spatial statistics (TBSS) analysis. All PLWHs were randomly divided into training and validation cohorts at a 7:3 ratio. A 3 T, single-shot spin-echo echo planar imaging (EPI). The differences between the PLWH groups were compared using TBSS and region of interest (ROI) analysis. Radiomic features were extracted from the corpus callosum (CC) on DTI postprocessed images, including fractional anisotropy (FA), axial diffusivity (AD), mean diffusivity (MD), and radial diffusivity (RD). The performance of the radiomic signatures was evaluated by ROC curve analysis. The radiomic signature with the highest area under the curve (AUC) was combined with clinical characteristics to construct a nomogram. Decision curve analysis (DCA) was performed to evaluate the ability of different methods in discriminating ANI. Chi-square test, independent-samples t test, Kruskal-Wallis test, Mann-Whitney U test, threshold-free cluster enhancement (TFCE), ROC curve analysis, DCA, multivariate logistic regression analysis, Hosmer-Lemeshow test. P < 0.05 with TFCE corrected and P < 0.0001 without TFCE corrected were considered statistically significant. The ANI group showed lower FA and higher AD than the IC group. In the validation cohort, the AUCs of the FA-, AD-, MD- and RD-based radiomic signatures and the clinicoradiomic nomogram were 0.829, 0.779, 0.790, 0.864, and 0.874, respectively. DCA revealed that the nomogram was of greater clinical value than TBSS analysis, the clinical models, and the RD-based radiomic signature. The combination of DTI and radiomics is correlated with early stage of HAND in PLWH. 3. Stage 2.

Nomogram to predict postpartum hemorrhage in cesarean delivery for twin pregnancies: a retrospective cohort study in China.

Postpartum hemorrhage (PPH) is the most common cause of maternal morbidity and mortality worldwide. A reliable risk assessment tool for PPH could optimize available interventions to reduce adverse maternal outcomes. The objective of this study was to explore a nomogram predicting the risk of postpartum hemorrhage after cesarean delivery for twin pregnancies. This single-center retrospective cohort study conducted twin pregnancies who underwent cesarean delivery between January 2014 and July 2021. Propensity score matching at baseline was used to match PPH (blood loss ≥1000 mL) and non-PPH group (blood loss <1000 mL). A nomogram was developed to predict the risk of PPH in cesarean delivery for twin pregnancies. The receiver operating characteristic curve (ROC), calibration plot, and decision curve analysis (DCA) were, respectively, used to evaluate the discrimination, calibration, and clinical utility of the prediction models. After propensity score matching, 186 twin pregnancies in the PPH group were matched with 186 controls in the non-PPH group. Seven independent prognostic variables, including antepartum albumin, assisted reproductive technology, hypertensive disorders of pregnancy, placenta previa, placenta accrete spectrum, intrapartum cesarean delivered, and estimated weights of twins, were used to build the nomogram. Based on the performance of the model, it appears that a good calibration (Hosmer-Lemeshow χ2 = 4.84, P > 0.05), an excellent predictive ability (area under the curve: 0.778, 95% CI: 0.732-0.825), and a good positive net benefit in the predictive model have been achieved. The nomogram was first generated to predict PPH in cesarean delivery for twin pregnancies, which could help clinicians to provide a reference for the preoperative surgical plan, choose optimal treatments, optimize healthcare resources, and thereby reduce the associated adverse maternal outcomes.

Development and validation of a nomogram for predicting severe respiratory syncytial virus-associated bronchiolitis

BackgroundRespiratory syncytial virus (RSV) is the most common cause of bronchiolitis and is related to the severity of the disease. This study aimed to develop and validate a nomogram for predicting severe bronchiolitis in infants and young children with RSV infection.MethodsA total of 325 children with RSV-associated bronchiolitis were enrolled, including 125 severe cases and 200 mild cases. A prediction model was built on 227 cases and validated on 98 cases, which were divided by random sampling in R software. Relevant clinical, laboratory and imaging data were collected. Multivariate logistic regression models were used to determine optimal predictors and to construct nomograms. The performance of the nomogram was evaluated by the area under the characteristic curve (AUC), calibration ability and decision curve analysis (DCA).ResultsThere were 137 (60.4%) mild and 90 (39.6%) severe RSV-associated bronchiolitis cases in the training group (n = 227) and 63 (64.3%) mild and 35 (35.7%) severe cases in the validation group (n = 98). Multivariate logistic regression analysis identified 5 variables as significant predictive factors to construct the nomogram for predicting severe RSV-associated bronchiolitis, including preterm birth (OR = 3.80; 95% CI, 1.39–10.39; P = 0.009), weight at admission (OR = 0.76; 95% CI, 0.63–0.91; P = 0.003), breathing rate (OR = 1.11; 95% CI, 1.05–1.18; P = 0.001), lymphocyte percentage (OR = 0.97; 95% CI, 0.95–0.99; P = 0.001) and outpatient use of glucocorticoids (OR = 2.27; 95% CI, 1.05–4.9; P = 0.038). The AUC value of the nomogram was 0.784 (95% CI, 0.722–0.846) in the training set and 0.832 (95% CI, 0.741–0.923) in the validation set, which showed a good fit. The calibration plot and Hosmer‒Lemeshow test indicated that the predicted probability had good consistency with the actual probability both in the training group (P = 0.817) and validation group (P = 0.290). The DCA curve shows that the nomogram has good clinical value.ConclusionA nomogram for predicting severe RSV-associated bronchiolitis in the early clinical stage was established and validated, which can help physicians identify severe RSV-associated bronchiolitis and then choose reasonable treatment.

External validation of the ACC/AHA ASCVD risk score in a Colombian population cohort

No cardiovascular risk score has included Latin American patients in its development. The ACC/AHA ASCVD risk score has not been validated in Latin America; consequently, its predictive capacity in the population of the region is unknown. The aim of this study is to evaluate the discrimination capacity and calibration of the ACC/AHA ASCVD score to predict the 10-year risk of a cardiovascular event in a primary prevention cohort followed in a Colombian hospital. A retrospective cohort study was conducted in primary prevention patients belonging to an intermediate/high-risk and low-risk cohort without established atherosclerotic disease. Cardiovascular risk was calculated at inclusion. The calibration was analyzed by comparing observed and expected events in the different risk categories. A discrimination analysis was made using the area under the ROC curve and C statistic. A total of 918 patients were included—202 from the intermediate/high-risk and 716 from the low-risk cohort. The median cardiovascular risk was 3.6% (IQR 1.7–8.5%). At the 10-year follow-up, 40 events (4,4%) occurred. The area under the ROC curve was 0.782 (95% CI 0.71–0.85). The Hosmer–Lemeshow test did not show differences between expected and observed events. The ACC/AHA ASCVD score is calibrated and has good discrimination capacity in predicting 10-year risk of cardiovascular events in a Colombian population.

Predicting therapeutic efficacy of oral rehydration salts in children with vasovagal syncope.

This study was designed to develop an easy-to-perform and inexpensive measure to predict efficacy of the oral rehydration salts (ORS) in children with vasovagal syncope (VVS). Children diagnosed with VVS and treated with ORS for a median of 3 months at the Peking University First Hospital, China, were enrolled and followed up. Demographic data, clinical hemodynamic parameters, and variables related to red blood cells were collected at the baseline. On the basis of changes in symptom scores after treatment, participants were divided into effective or ineffective groups at the end of the follow-up. Logistic regression analysis was used to investigate parameters related to therapeutic efficacy of ORS and a predictive model of ORS effectiveness was created. The predictive efficiency was evaluated using the receiver operating characteristic curve. The accuracy/consistency was evaluated by the Hosmer-Lemeshow test and calibration curve. Internal validation was done using the bootstrap approach. Totally 97 pediatric participants were included in the study and 4 (4.1%) were lost during the follow-up. ORS therapy was effective in 46 children and ineffective in 47 children. Children in the effective group had higher baseline red blood cell count, hemoglobin, and hematocrit than those in the ineffective group (p < 0.01). Through logistic regression analysis, the baseline hematocrit and body mass index (BMI) were included in predictive model for the response to ORS treatment. The predictive efficacy of the model showed an area under the curve of 0.77 (p < 0.01). The predicted probability cut-off value of 0.5 was found to be optimal, with a resulting sensitivity of 67.4% and specificity of 80.9%. In the Hosmer-Lemeshow test, p-value was 0.75, and the calibration plot showed a good model fitness. Internal validation was performed using the bootstrap approach (n = 1,000), showing 95% confidence interval of 0.67-0.86. Hemoglobin combined with BMI was useful for predicting the therapeutic efficacy of ORS in children with VVS.

A prediction nomogram for postoperative gastroparesis syndrome in right colon cancer: a retrospective study.

To investigate the high-risk factors for postoperative gastroparesis syndrome (PGS) in right colon cancer and to build a prediction nomogram for personalized prediction of PGS. Our study retrospectively analyzed 361 patients with right colon cancer who underwent right hemicolectomy at The First Hospital of Putian City in Fujian Province, China and who were hospitalized between January 2012 and July 2022. Multivariate logistic regression was used to determine the independent risk factors for PGS and to establish a nomogram model. Furthermore, discrimination, calibration, and clinical benefits were used to evaluate the model. The multivariate logistic regression revealed that dissection of the subpyloric lymph nodes (No. 206 lymph node) (OR 5.242, P = 0.005), preoperative fasting blood glucose level (OR 3.708, P = 0.024), preoperative albumin level (OR 3.503, P = 0.020), and total operative time (OR 4.648, P = 0.014) were independent risk factors for PGS. Based on the above four factors, the area under the ROC curve (AUC) and C-index of the nomogram were 0.831. The prediction nomogram's calibration curve was closer to the ideal diagonal, and the Hosmer‒Lemeshow test indicated that the nomogram fit well (P = 0.399). Moreover, the decision curve analysis revealed that the model can present better clinical benefits when the threshold probability was between 1 and 28%, and the internal validation verified the dependability of the model (C-index = 0.948). A risk prediction nomogram based on perioperative factors provided the physician with a simple, visual, and efficient tool for the prediction and management of PGS in right colon cancer.

The BIDIAP index: a clinical, analytical and ultrasonographic score for the diagnosis of acute appendicitis in children

BackgroundPediatric acute appendicitis (PAA) continues to be a diagnostic challenge today. The diagnostic performance of classical indices is only moderate, especially in pediatric population. This study aimed to define a clinical, radiological and analytical index for the diagnosis of PAA.Materials and methodsThis prospective study included 151 patients divided into two groups: (1) 53 patients with non-surgical abdominal pain (NSAP) and (2) 98 patients with a confirmed PAA. Sociodemographic and clinical characteristics were compared between groups using the Mann–Whitney U test and the Fisher exact test. To identify the predictors of PAA, we performed a multivariable logistic regression using a forward stepwise analysis and we assigned multiples of integer values to the selected variables. The diagnostic performance of the index was assessed by calculating the area under the receiver operating characteristic curve. Intra-cohort calibration was assessed with the Hosmer–Lemeshow test. ResultsWe developed the BIDIAP index (BIomarkers for the DIagnosis of Appendicitis in Pediatrics), which included three variables that independently predicted higher odds of PAA: appendiceal caliber (≥ 6.9 mm), systemic immune-inflammation index (≥ 890) and peritoneal irritation, which scored 4, 3 and 2 points, respectively. Mean (SD) score of the participants was 2.38 (2.06) in group 1 and 7.89 (1.50) in group 2. The area under the ROC was 0.97 (95% CI 0.95–0.99). The cut-off point was established at 4 points, resulting in a sensitivity of 98.98% and a specificity of 77.78%.ConclusionsThe BIDIAP index has an exceptional diagnostic performance in PAA. The importance of these results lies in its novelty and in the simplicity of the index. Although external validation will be necessary, initial results look promising.

Determinants of premature rupture of membrane among pregnant women in Harar town, Eastern Ethiopia: A case-control study

IntroductionPremature rupture of membrane is a disruption of fetal membrane followed by passage of watery fluid gush before the onset of labor any time beyond 28 weeks of gestation. It is a significant cause of perinatal morbidity and mortality. Many studies were conducted on the pre-labor rupture of membrane, yet limited evidence is available on its determinants. This study aimed to identify determinants of premature rupture of membrane among pregnant women in Harar Town, Eastern Ethiopia. Methodology; Hospital-based unmatched case control study was conducted on 115 cases and 230 controls from 15th May to 15th July 2021. The study was conducted on two public hospitals in Harar town. All cases admitted at the time of data collection were included until the required sample size was obtained and controls were selected by using simple random sampling among all non -cases. Data were collected using interviewer-guided semi-structured questionnaires. The data were entered into Epi-data version 3.1 and exported to SPSS version 22.0 for analysis. Binary logistic regression was used to identify determinants of the outcome and statistical significance declared at a p-value less than 0.05. Multicollinearity was checked and model fitness was assessed by Hosmer Lemeshow test. ResultsAbnormal vaginal discharge [AOR 2.15 (2.53, 22.46), history of cesarean delivery [AOR 2.06 (1.11, 6.78)], history of premature rupture of membrane [AOR 4.62 (2.06, 11.52)] and history of abortion [AOR 2.81 (1.04, 6.23)] increase the odds of the outcome. ConclusionBad obstetric histories in the current and previous pregnancies are related to premature rupture of membrane. Therefore, it is recommended that health care providers should take attention to women with past and current pregnancy complications.

Machine learning-based radiomics model to predict benign and malignant PI-RADS v2.1 category 3 lesions: a retrospective multi-center study

PurposeTo develop machine learning-based radiomics models derive from different MRI sequences for distinction between benign and malignant PI-RADS 3 lesions before intervention, and to cross-institution validate the generalization ability of the models.MethodsThe pre-biopsy MRI datas of 463 patients classified as PI-RADS 3 lesions were collected from 4 medical institutions retrospectively. 2347 radiomics features were extracted from the VOI of T2WI, DWI and ADC images. The ANOVA feature ranking method and support vector machine classifier were used to construct 3 single-sequence models and 1 integrated model combined with the features of three sequences. All the models were established in the training set and independently verified in the internal test and external validation set. The AUC was used to compared the predictive performance of PSAD with each model. Hosmer–lemeshow test was used to evaluate the degree of fitting between prediction probability and pathological results. Non-inferiority test was used to check generalization performance of the integrated model.ResultsThe difference of PSAD between PCa and benign lesions was statistically significant (P = 0.006), with the mean AUC of 0.701 for predicting clinically significant prostate cancer (internal test AUC = 0.709 vs. external validation AUC = 0.692, P = 0.013) and 0.630 for predicting all cancer (internal test AUC = 0.637 vs. external validation AUC = 0.623, P = 0.036). T2WI-model with the mean AUC of 0.717 for predicting csPCa (internal test AUC = 0.738 vs. external validation AUC = 0.695, P = 0.264) and 0.634 for predicting all cancer (internal test AUC = 0.678 vs. external validation AUC = 0.589, P = 0.547). DWI-model with the mean AUC of 0.658 for predicting csPCa (internal test AUC = 0.635 vs. external validation AUC = 0.681, P = 0.086) and 0.655 for predicting all cancer (internal test AUC = 0.712 vs. external validation AUC = 0.598, P = 0.437). ADC-model with the mean AUC of 0.746 for predicting csPCa (internal test AUC = 0.767 vs. external validation AUC = 0.724, P = 0.269) and 0.645 for predicting all cancer (internal test AUC = 0.650 vs. external validation AUC = 0.640, P = 0.848). Integrated model with the mean AUC of 0.803 for predicting csPCa (internal test AUC = 0.804 vs. external validation AUC = 0.801, P = 0.019) and 0.778 for predicting all cancer (internal test AUC = 0.801 vs. external validation AUC = 0.754, P = 0.047).ConclusionsThe radiomics model based on machine learning has the potential to be a non-invasive tool to distinguish cancerous, noncancerous and csPCa in PI-RADS 3 lesions, and has relatively high generalization ability between different date set.

Development and evaluation of nomograms for predicting osteoarthritis progression based on MRI cartilage parameters: data from the FNIH OA biomarkers Consortium

BackgroundOsteoarthritis (OA) is a leading cause of disability worldwide. However, the existing methods for evaluating OA patients do not provide enough comprehensive information to make reliable predictions of OA progression. This retrospective study aimed to develop prediction nomograms based on MRI cartilage that can predict disease progression of OA.MethodsA total of 600 subjects with mild-to-moderate osteoarthritis from the Foundation for National Institute of Health (FNIH) project of osteoarthritis initiative (OAI). The MRI cartilage parameters of the knee at baseline were measured, and the changes in cartilage parameters at 12- and 24-month follow-up were calculated. The least absolute shrinkage and selection operator (LASSO) regression analysis was used to extract the valuable characteristic parameters at different time points including cartilage thickness, cartilage volume, subchondral bone exposure area and uniform cartilage thickness in different sub regions of the knee, and the MRI cartilage parameters score0, scoreΔ12, and scoreΔ24 at baseline, 12 months, and 24 months were constructed. ScoreΔ12, and scoreΔ24 represent changes between 12 M vs. baseline, and 24 M vs. baseline, respectively. Logistic regression analysis was used to construct the nomogram0, nomogramΔ12, and nomogramΔ24, including MRI-based score and risk factors. The area under curve (AUC) was used to evaluate the differentiation of nomograms in disease progression and subgroup analysis. The calibration curve and Hosmer-Lemeshow (H-L) test were used to verify the calibration of the nomograms. Clinical usefulness of each prediction nomogram was verified by decision curve analysis (DCA). The nomograms with predictive efficacy were analyzed by secondary analysis. Internal verification was assessed using bootstrapping validation.ResultsEach nomogram included cartilage score, KL grade, WOMAC pain score, WOMAC disability score, and minimum joint space width. The AUC of nomogram0, nomogramΔ12, and nomogramΔ24 in predicing the progression of radiology and pain were 0.69, 0.64, and 0.71, respectively. All three nomograms had good calibration. Analysis by DCA showed that the clinical effectiveness of nomogramΔ24 was higher than others. Secondary analysis showed that nomogram0 and nomogramΔ24 were more capable of predicting OA radiologic progression than pain progression.ConclusionNomograms based on MRI cartilage change were useful for predicting the progression of mild to moderate OA.

Development and validation of a risk prediction model for frailty in patients with diabetes

BackgroundFrailty is the third most common complication of diabetes after macrovascular and microvascular complications. The aim of this study was to develop a validated risk prediction model for frailty in patients with diabetes.MethodsThe research used data from the China Health and Retirement Longitudinal Study (CHARLS), a dataset representative of the Chinese population. Twenty-five indicators, including socio-demographic variables, behavioral factors, health status, and mental health parameters, were analyzed in this study. The study cohort was randomly divided into a training set and a validation set at a ratio of 70 to 30%. LASSO regression analysis was used to screen the variables for the best predictors of the model based on a 10-fold cross-validation. The logistic regression model was applied to explore the associated factors of frailty in patients with diabetes. A nomogram was constructed to develop the prediction model. Calibration curves were applied to evaluate the accuracy of the nomogram model. The area under the receiver operating characteristic curve and decision curve analysis were conducted to assess predictive performance.ResultsOne thousand four hundred thirty-six patients with diabetes from the CHARLS database collected in 2013 (n = 793) and 2015 (n = 643) were included in the final analysis. A total of 145 (10.9%) had frailty symptoms. Multivariate logistic regression analysis showed that marital status, activities of daily living, waist circumference, cognitive function, grip strength, social activity, and depression as predictors of frailty in people with diabetes. These factors were used to construct the nomogram model, which showed good concordance and accuracy. The AUC values of the predictive model and the internal validation set were 0.912 (95%CI 0.887–0.937) and 0.881 (95% CI 0.829–0.934). Hosmer–Lemeshow test values were P = 0.824 and P = 0.608 (both > 0.05). Calibration curves showed significant agreement between the nomogram model and actual observations. ROC and DCA indicated that the nomogram had a good predictive performance.ConclusionsComprehensive nomogram constructed in this study was a promising and convenient tool to evaluate the risk of frailty in patients with diabetes, and contributed clinicians to screening the high-risk population.

External validation of geriatric influenza death score: A multicenter study.

The Geriatric Influenza Death (GID) score was developed to help decision making in older patients with influenza in the emergency department (ED), but external validation is unavailable. Thus, we conducted a study was to fill the data gap. We recruited all older patients (≥65 years) who visited the ED of three hospitals between 2009 and 2018. Demographic data and clinical characteristics were retrospectively collected. Discrimination, goodness of fit, and performance of the GID score were evaluated. Of the 5,508 patients (121 died) with influenza, the mean age was 76.6±7.4 (standard deviation) years, and 49.3% were males. The GID score was higher in the mortality group (1.7±1.1 vs. 0.8±0.8, p <0.01). With 0 as the reference, the odds ratio for morality with score of 1, 2 and ≥3 was 3.08 (95% confidence interval [CI]: 1.66-5.71), 6.69 (95% CI: 3.52-12.71), and 23.68 (95% CI: 11.95-46.93), respectively. The area under the curve was 0.722 (95% CI: 0.677-0.766), and the Hosmer-Lemeshow goodness of fit test was 1.000. The GID score had excellent negative predictive values with different cut-offs. The GID score had good external validity, and further studies are warranted for wider application.

A nomogram for predicting the risk of poor prognosis in patients with poor-grade aneurysmal subarachnoid hemorrhage following microsurgical clipping.

Aneurysmal subarachnoid hemorrhage (aSAH) is a common and potentially fatal cerebrovascular disease. Poor-grade aSAH (Hunt-Hess grades IV and V) accounts for 20-30% of patients with aSAH, with most patients having a poor prognosis. This study aimed to develop a stable nomogram model for predicting adverse outcomes at 6 months in patients with aSAH, and thus, aid in improving the prognosis. The clinical data and imaging findings of 150 patients with poor-grade aSAH treated with microsurgical clipping of intracranial aneurysms on admission from December 2015 to October 2021 were retrospectively analyzed. Least absolute shrinkage and selection operator (LASSO), logistic regression analyses, and a nomogram were used to develop the prognostic models. Receiver operating characteristic (ROC) curves and Hosmer-Lemeshow tests were used to assess discrimination and calibration. The bootstrap method (1,000 repetitions) was used for internal validation. Decision curve analysis (DCA) was performed to evaluate the clinical validity of the nomogram model. LASSO regression analysis showed that age, Hunt-Hess grade, Glasgow Coma Scale (GCS), aneurysm size, and refractory hyperpyrexia were potential predictors for poor-grade aSAH. Logistic regression analyses revealed that age (OR: 1.107, 95% CI: 1.056-1.116, P < 0.001), Hunt-Hess grade (OR: 8.832, 95% CI: 2.312-33.736, P = 0.001), aneurysm size (OR: 6.871, 95% CI: 1.907-24.754, P = 0.003) and refractory fever (OR: 3.610, 95% CI: 1.301-10.018, P < 0.001) were independent predictors of poor outcome. The area under the ROC curve (AUC) was 0.909. The calibration curve and Hosmer-Lemeshow tests showed that the nomogram had good calibration ability. Furthermore, the DCA curve showed better clinical utilization of the nomogram. This study provides a reliable and valuable nomogram that can accurately predict the risk of poor prognosis in patients with poor-grade aSAH after microsurgical clipping. This tool is easy to use and can help physicians make appropriate clinical decisions to significantly improve patient prognosis.

Construction of pain prediction model for patients undergoing hepatic arterial chemoembolization

ObjectiveTo construct a predictive model for pain in patients undergoing hepatic arterial chemoembolization (TACE) in interventional operating room.MethodsThrough literature review and expert interviews, a questionnaire was prepared for the assessment of pain factors in patients with hepatic arterial chemoembolization. A prospective cohort study was used to select 228 patients with hepatic arterial chemoembolization in a tertiary and first-class hospital. The data of the patients in the pain group and the non-pain group were compared, and a rapid screening prediction model was constructed by univariate analysis and logistic regression analysis, and its prediction effect was tested.ResultsTumor size, liver cancer stage, and chemoembolization with drug-loaded microspheres and pirarubicin hydrochloride (THP) mixed with lipiodol were independent predictors of pain in patients after hepatic arterial chemoembolization. Finally, the pain prediction model after TACE was obtained. The results of Hosmer–Lemeshow test showed that the model fit was good (χ2 = 13.540, p = 0.095). The area under the receiver operating characteristic curve was 0.798, p < 0.001.ConclusionThe rapid screening and prediction model of pain in patients undergoing hepatic arterial chemoembolization has certain efficacy, which is helpful for clinical screening of patients with high risk of pain, and can provide reference for predictive pain management decision-making.

Predictive role of arterial lactate in acute kidney injury associated with off-pump coronary artery bypass grafting.

This observational study aims to explore the predictive role of postoperative arterial lactate in off-pump coronary artery bypass grafting (CABG)-associated acute kidney injury (AKI). A total of 500 consecutive patients who underwent off-pump CABG from August 2020 to August 2021 at the Department of Cardiovascular Surgery, Qilu Hospital of Shandong University, were included. Logistic regression analysis was used to confirm the independent risk factors of off-pump CABG-associated AKI. Receiver operating characteristic (ROC) curve was performed to evaluate the discrimination ability and Hosmer-Lemeshow goodness of fit test was performed to evaluate the calibration ability. The incidence of off-pump CABG-associated AKI was 20.6%. Female gender, preoperative albumin, baseline serum creatinine, 12 h postoperative arterial lactate and duration of mechanical ventilation were independent risk factors. The area under the ROC curve (AUC) of 12 h postoperative arterial lactate for predicting off-pump CABG-associated AKI was 0.756 and the cutoff value was 1.85. The prediction model that incorporated independent risk factors showed reliable predictive ability (AUC = 0.846). Total hospital stay, intensive care unit stay, occurrence of other postoperative complications, and 28-day mortality were all significantly higher in AKI group compared to non-AKI group. 12 h postoperative arterial lactate was a validated predictive biomarker for off-pump CABG-associated AKI. We constructed a predictive model that facilitates the early recognition and management of off-pump CABG-associated AKI.

An artificial intelligence approach for predicting death or organ failure after hospitalization for COVID-19: development of a novel risk prediction tool and comparisons with ISARIC-4C, CURB-65, qSOFA, and MEWS scoring systems

BackgroundWe applied machine learning (ML) algorithms to generate a risk prediction tool [Collaboration for Risk Evaluation in COVID-19 (CORE-COVID-19)] for predicting the composite of 30-day endotracheal intubation, intravenous administration of vasopressors, or death after COVID-19 hospitalization and compared it with the existing risk scores.MethodsThis is a retrospective study of adults hospitalized with COVID-19 from March 2020 to February 2021. Patients, each with 92 variables, and one composite outcome underwent feature selection process to identify the most predictive variables. Selected variables were modeled to build four ML algorithms (artificial neural network, support vector machine, gradient boosting machine, and Logistic regression) and an ensemble model to generate a CORE-COVID-19 model to predict the composite outcome and compared with existing risk prediction scores. The net benefit for clinical use of each model was assessed by decision curve analysis.ResultsOf 1796 patients, 278 (15%) patients reached primary outcome. Six most predictive features were identified. Four ML algorithms achieved comparable discrimination (P > 0.827) with c-statistics ranged 0.849–0.856, calibration slopes 0.911–1.173, and Hosmer–Lemeshow P > 0.141 in validation dataset. These 6-variable fitted CORE-COVID-19 model revealed a c-statistic of 0.880, which was significantly (P < 0.04) higher than ISARIC-4C (0.751), CURB-65 (0.735), qSOFA (0.676), and MEWS (0.674) for outcome prediction. The net benefit of the CORE-COVID-19 model was greater than that of the existing risk scores.ConclusionThe CORE-COVID-19 model accurately assigned 88% of patients who potentially progressed to 30-day composite events and revealed improved performance over existing risk scores, indicating its potential utility in clinical practice.

CT-based radiomics with various classifiers for histological differentiation of parotid gland tumors.

This study assessed whether radiomics features could stratify parotid gland tumours accurately based on only noncontrast CT images and validated the best classifier of different radiomics models. In this single-centre study, we retrospectively recruited 249 patients with a diagnosis of pleomorphic adenoma (PA), Warthin tumour (WT), basal cell adenoma (BCA) or malignant parotid gland tumours (MPGTs) from June 2020 to August 2022. Each patient was randomly classified into training and testing cohorts at a ratio of 7:3, and then, pairwise comparisons in different parotid tumour groups were performed. CT images were transferred to 3D-Slicer software and the region of interest was manually drawn for feature extraction. Feature selection methods were performed using the intraclass correlation coefficient, t test and least absolute shrinkage and selection operator. Five common classifiers, namely, random forest (RF), support vector machine (SVM), logistic regression (LR), K-nearest neighbours (KNN) and general Bayesian network (Gnb), were selected to build different radiomics models. The receiver operating characteristic curve, area under the curve (AUC), accuracy, sensitivity, specificity and F-1 score were used to assess the prediction performances of these models. The calibration of the model was calculated by the Hosmer-Lemeshow test. DeLong's test was utilized for comparing the AUCs. The radiomics model based on the RF, SVM, Gnb, LR, LR and RF classifiers obtained the highest AUC in differentiating PA from MPGTs, WT from MPGTs, BCA from MPGTs, PA from WT, PA from BCA, and WT from BCA, respectively. Accordingly, the AUC and the accuracy of the model for each classifier were 0.834 and 0.71, 0.893 and 0.79, 0.844 and 0.79, 0.902 and 0.88, 0.602 and 0.68, and 0.861 and 0.94, respectively. Our study demonstrated that noncontrast CT-based radiomics could stratify refined pathological types of parotid tumours well but could not sufficiently differentiate PA from BCA. Different classifiers had the best diagnostic performance for different parotid tumours. Our study findings add to the current knowledge on the differential diagnosis of parotid tumours.

Prediction model for gestational diabetes mellitus using the XG Boost machine learning algorithm.

To develop the extreme gradient boosting (XG Boost) machine learning (ML) model for predicting gestational diabetes mellitus (GDM) compared with a model using the traditional logistic regression (LR) method. A case-control study was carried out among pregnant women, who were assigned to either the training set (these women were recruited from August 2019 to November 2019) or the testing set (these women were recruited in August 2020). We applied the XG Boost ML model approach to identify the best set of predictors out of a set of 33 variables. The performance of the prediction model was determined by using the area under the receiver operating characteristic (ROC) curve (AUC) to assess discrimination, and the Hosmer-Lemeshow (HL) test and calibration plots to assess calibration. Decision curve analysis (DCA) was introduced to evaluate the clinical use of each of the models. A total of 735 and 190 pregnant women were included in the training and testing sets, respectively. The XG Boost ML model, which included 20 predictors, resulted in an AUC of 0.946 and yielded a predictive accuracy of 0.875, whereas the model using a traditional LR included four predictors and presented an AUC of 0.752 and yielded a predictive accuracy of 0.786. The HL test and calibration plots show that the two models have good calibration. DCA indicated that treating only those women whom the XG Boost ML model predicts are at risk of GDM confers a net benefit compared with treating all women or treating none. The established model using XG Boost ML showed better predictive ability than the traditional LR model in terms of discrimination. The calibration performance of both models was good.

Development and validation of risk prediction models for large for gestational age infants using logistic regression and two machine learning algorithms.

Large for gestational age (LGA) is one of the adverse outcomes during pregnancy that endangers the life and health of mothers and offspring. We aimed to establish prediction models for LGA at late pregnancy. Data were obtained from an established Chinese pregnant women cohort of 1285 pregnant women. LGA was diagnosed as >90th percentile of birth weight distribution of Chinese corresponding to gestational age of the same-sex newborns. Women with gestational diabetes mellitus (GDM) were classified into three subtypes according to the indexes of insulin sensitivity and insulin secretion. Models were established by logistic regression and decision tree/random forest algorithms, and validated by the data. A total of 139 newborns were diagnosed as LGA after birth. The area under the curve (AUC) for the training set is 0.760 (95% confidence interval [CI] 0.706-0.815), and 0.748 (95% CI 0.659-0.837) for the internal validation set of the logistic regression model, which consisted of eight commonly used clinical indicators (including lipid profile) and GDM subtypes. For the prediction models established by the two machine learning algorithms, which included all the variables, the training set and the internal validation set had AUCs of 0.813 (95% CI 0.786-0.839) and 0.779 (95% CI 0.735-0.824) for the decision tree model, and 0.854 (95% CI 0.831-0.877) and 0.808 (95% CI 0.766-0.850) for the random forest model. We established and validated three LGA risk prediction models to screen out the pregnant women with high risk of LGA at the early stage of the third trimester, which showed good prediction power and could guide early prevention strategies.

Development and Validation of a Risk Scoring Tool for Bronchopulmonary Dysplasia in Preterm Infants Based on a Systematic Review and Meta-Analysis.

Background: Bronchopulmonary dysplasia (BPD) is the most common serious pulmonary morbidity in preterm infants with high disability and mortality rates. Early identification and treatment of BPD is critical. Objective: This study aimed to develop and validate a risk scoring tool for early identification of preterm infants that are at high-risk for developing BPD. Methods: The derivation cohort was derived from a systematic review and meta-analysis of risk factors for BPD. The statistically significant risk factors with their corresponding odds ratios were utilized to construct a logistic regression risk prediction model. By scoring the weights of each risk factor, a risk scoring tool was established and the risk stratification was divided. External verification was carried out by a validation cohort from China. Results: Approximately 83,034 preterm infants with gestational age < 32 weeks and/or birth weight < 1500 g were screened in this meta-analysis, and the cumulative incidence of BPD was about 30.37%. The nine predictors of this model were Chorioamnionitis, Gestational age, Birth weight, Sex, Small for gestational age, 5 min Apgar score, Delivery room intubation, and Surfactant and Respiratory distress syndrome. Based on the weight of each risk factor, we translated it into a simple clinical scoring tool with a total score ranging from 0 to 64. External validation showed that the tool had good discrimination, the area under the curve was 0.907, and that the Hosmer-Lemeshow test showed a good fit (p = 0.3572). In addition, the results of the calibration curve and decision curve analysis suggested that the tool showed significant conformity and net benefit. When the optimal cut-off value was 25.5, the sensitivity and specificity were 0.897 and 0.873, respectively. The resulting risk scoring tool classified the population of preterm infants into low-risk, low-intermediate, high-intermediate, and high-risk groups. This BPD risk scoring tool is suitable for preterm infants with gestational age < 32 weeks and/or birth weight < 1500 g. Conclusions: An effective risk prediction scoring tool based on a systematic review and meta-analysis was developed and validated. This simple tool may play an important role in establishing a screening strategy for BPD in preterm infants and potentially guide early intervention.