Receiver Operating Characteristic Curve Area Research Articles

Enhancing Missense Variant Pathogenicity Prediction with MissenseNet: Integrating Structural Insights and ShuffleNet-Based Deep Learning Techniques

The classification of missense variant pathogenicity continues to pose significant challenges in human genetics, necessitating precise predictions of functional impacts for effective disease diagnosis and personalized treatment strategies. Traditional methods, often compromised by suboptimal feature selection and limited generalizability, are outpaced by the enhanced classification model, MissenseNet (Missense Classification Network). This model, advancing beyond standard predictive features, incorporates structural insights from AlphaFold2 protein predictions, thus optimizing structural data utilization. MissenseNet, built on the ShuffleNet architecture, incorporates an encoder-decoder framework and a Squeeze-and-Excitation (SE) module designed to adaptively adjust channel weights and enhance feature fusion and interaction. The model’s efficacy in classifying pathogenicity has been validated through superior accuracy compared to conventional methods and by achieving the highest areas under the Receiver Operating Characteristic (ROC) and Precision-Recall (PR) curves (Area Under the Curve and Area Under the Precision-Recall Curve) in an independent test set, thus underscoring its superiority.

Gender and contemporary risk of adverse events in atrial fibrillation.

The role of gender in decision-making for oral anticoagulation in patients with atrial fibrillation (AF) remains controversial. The population cohort study used electronic healthcare records of 16 587 749 patients from UK primary care (2005-2020). Primary (composite of all-cause mortality, ischaemic stroke, or arterial thromboembolism) and secondary outcomes were analysed using Cox hazard ratios (HR), adjusted for age, socioeconomic status, and comorbidities. 78 852 patients were included with AF, aged 40-75 years, no prior stroke, and no prescription of oral anticoagulants. 28 590 (36.3%) were women, and 50 262 (63.7%) men. Median age was 65.7 years (interquartile range 58.5-70.9), with women being older and having other differences in comorbidities. During a total follow-up of 431 086 patient-years, women had a lower adjusted primary outcome rate with HR 0.89 vs. men (95% confidence interval [CI] 0.87-0.92; P < .001) and HR 0.87 after censoring for oral anticoagulation (95% CI 0.83-0.91; P < .001). This was driven by lower mortality in women (HR 0.86, 95% CI 0.83-0.89; P < .001). No difference was identified between women and men for the secondary outcomes of ischaemic stroke or arterial thromboembolism (adjusted HR 1.00, 95% CI 0.94-1.07; P = .87), any stroke or any thromboembolism (adjusted HR 1.02, 95% CI 0.96-1.07; P = .58), and incident vascular dementia (adjusted HR 1.13, 95% CI 0.97-1.32; P = .11). Clinical risk scores were only modest predictors of outcomes, with CHA2DS2-VA (ignoring gender) superior to CHA2DS2-VASc for primary outcomes in this population (receiver operating characteristic curve area 0.651 vs. 0.639; P < .001) and no interaction with gender (P = .45). Removal of gender from clinical risk scoring could simplify the approach to which patients with AF should be offered oral anticoagulation.

Decoding Patient Heterogeneity Influencing Radiation-Induced Brain Necrosis.

In radiotherapy (RT) for brain tumors, patient heterogeneity masks treatment effects, complicating the prediction and mitigation of radiation-induced brain necrosis. Therefore, understanding this heterogeneity is essential for improving outcome assessments and reducing toxicity. We developed a clinically practical pipeline to clarify the relationship between dosimetric features and outcomes by identifying key variables. We processed data from a cohort of 130 patients treated with proton therapy for brain and head and neck tumors, utilizing an expert-augmented Bayesian network to understand variable interdependencies and assess structural dependencies. Critical evaluation involved a 3-level grading system for each network connection and a Markov blanket analysis to identify variables directly impacting necrosis risk. Statistical assessments included log-likelihood ratio (LLR), integrated discrimination index (IDI), net reclassification index (NRI), and receiver operating characteristic (ROC). The analysis highlighted tumor location and proximity to critical structures like white matter and ventricles as major determinants of necrosis risk. The majority of network connections were clinically supported, with quantitative measures confirming the significance of these variables in patient stratification (LLR=12.17, p=0.016; IDI=0.15; NRI=0.74). The ROC curve area was 0.66, emphasizing the discriminative value of non-dosimetric variables. Key patient variables critical to understanding brain necrosis post-RT were identified, aiding the study of dosimetric impacts and providing treatment confounders and moderators. This pipeline aims to enhance outcome assessments by revealing at-risk patients, offering a versatile tool for broader applications in RT to improve treatment personalization in different disease sites.

Prostate-Specific Membrane Antigen Positron Emission Tomography/Computed Tomography-Derived Radiomic Models in Prostate Cancer Prognostication.

The clinical integration of prostate membrane specific antigen (PSMA) positron emission tomography and computed tomography (PET/CT) scans represents potential for advanced data analysis techniques in prostate cancer (PC) prognostication. Among these tools is the use of radiomics, a computer-based method of extracting and quantitatively analyzing subvisual features in medical imaging. Within this context, the present review seeks to summarize the current literature on the use of PSMA PET/CT-derived radiomics in PC risk stratification. A stepwise literature search of publications from 2017 to 2023 was performed. Of 23 articles on PSMA PET/CT-derived prostate radiomics, PC diagnosis, prediction of biopsy Gleason score (GS), prediction of adverse pathology, and treatment outcomes were the primary endpoints of 4 (17.4%), 5 (21.7%), 7 (30.4%), and 7 (30.4%) studies, respectively. In predicting PC diagnosis, PSMA PET/CT-derived models performed well, with receiver operator characteristic curve area under the curve (ROC-AUC) values of 0.85-0.925. Similarly, in the prediction of biopsy and surgical pathology results, ROC-AUC values had ranges of 0.719-0.84 and 0.84-0.95, respectively. Finally, prediction of recurrence, progression, or survival following treatment was explored in nine studies, with ROC-AUC ranging 0.698-0.90. Of the 23 studies included in this review, 2 (8.7%) included external validation. While explorations of PSMA PET/CT-derived radiomic models are immature in follow-up and experience, these results represent great potential for future investigation and exploration. Prior to consideration for clinical use, however, rigorous validation in feature reproducibility and biologic validation of radiomic signatures must be prioritized.

Predischarge Peak Atrial Longitudinal Strain and Plasma N-terminal Pro-hormone Brain Natriuretic Peptide as a Predictor of Short-term Rehospitalization and Cardiovascular Mortality in Patients with Acute Heart Failure.

The postacute heart failure (AHF) rehospitalization rate is attributed to persistent hemodynamic congestion despite clinical improvement. Peak atrial longitudinal strain (PALS), utilizing speckle tracking echocardiography technology, shows potential in post-AHF prognosis. Meanwhile, N-terminal pro-hormone brain natriuretic peptide (NT-proBNP) remains a known biomarker of intracardiac congestion. This study aimed to determine the relationship between predischarge PALS and NT-proBNP as predictors of major adverse cardiac event (MACE) in patients after AHF hospitalization. This study is a prospective cohort study, conducted in Prof. Dr. I G.N.G Ngoerah Hospital, Bali, Indonesia. The study included hospitalized AHF patients, collecting demographic data, comorbidities, therapies, and echocardiographic measures before discharge. Predischarge PALS and NT-proBNP were taken within 24 h before discharge. The main outcome was MACE, defined as rehospitalization and cardiovascular mortality within 90 days. Comparative statistical analyses was done using independent t-test for continuous variables (Mann-Whitney U test for variables with abnormal distribution) and Chi-squared tests. Receiver operating characteristic (ROC) used in determining optimal threshold values of predischarge PALS and NT-proBNP as a predictor of MACE. Kaplan-Meier curves were employed to gauge event-free survival differences between these cohorts. Then, independent Cox regression was used to identify the predictors of MACE. The study enrolled 67 patients with varying ejection fraction (EF) (16 - heart failure with preserved ejection fraction, 10 - heart failure with mildly reduced ejection fraction, and 41 - heart failure with reduced ejection fraction; mean age: 56.88 ± 14.57 years). Over the 90-day follow-up, 21 patients (31.3%) encountered MACE. Both PALS (area under the curve [AUC] 0.816) and NT-proBNP (AUC 0.856) before discharge served as predictors of MACE. There was no significant AUC difference between ROC curves (area difference: 0.039, P = 0.553). The regression model highlighted that PALS and NT-proBNP level before discharge acted as independent predictors of MACE, irrespective of EF, average E/e', or estimated predischarge pulmonary capillary wedge pressure. Predischarge PALS is comparable to NT-proBNP levels as independent predictors of short-term MACE after AHF hospitalization.

Adaption and validation of the Rwandese version of the Mood Disorder Questionnaire for the screening of bipolar disorder

BackgroundBipolar disorder is challenging to diagnose. In Rwanda, a sub-Saharan country with a limited number of psychiatrists, the number of people with an undetected diagnosis of bipolar disorder could be high. Still, no screening tool for the disorder is available in the country. This study aimed to adapt and validate the Mood Disorder Questionnaire in the Rwandan population. MethodsThe Mood Disorder Questionnaire was translated into Kinyarwanda. The process involved back-translation, cross-cultural adaptation, field testing of the pre-final version, and final adjustments. A total of 331 patients with either bipolar disorder or unipolar major depression from two psychiatric outpatient hospitals were included. The statistical analysis included reliability and validity analyses and receiver operating characteristic curve (ROC) analysis. The optimal cut-off was chosen by maximizing Younden's index. ResultsThe Rwandese version of The Mood Disorder Questionnaire had adequate internal consistency (Cronbach's alpha =0.91). The optimal threshold value was at least six positive items, which yielded excellent sensitivity (94.7%), and specificity (97.3%). The ROC area under the curve (AUC) was 0.99. ConclusionThe adapted tool showed good psychometric properties in terms of reliability and validity for the screening of bipolar disorder, with a recommended cutoff value of six items on the symptom checklist for a positive score and an exclusion of items 14 and 15. The tool has the potential to be a crucial instrument to identify otherwise undetected cases of bipolar disorder in Rwanda, improving access to mental health treatment, thus enhancing the living conditions of people with bipolar disorder.

Empirical exploration of predictive maintenance in concrete manufacturing: Harnessing machine learning for enhanced equipment reliability in construction project management

Predictive Maintenance (PdM) in concrete manufacturing, particularly when viewed through construction project management, is increasingly recognized for its significance. The potential for PdM to predict structural inadequacies and systematically plan machinery maintenance is essential for seamless construction workflows. The primary focus of this investigation was to utilize the capabilities of machine learning (ML) algorithms to develop and validate sophisticated predictive models. These models aimed to proactively identify maintenance needs, thereby reducing unexpected downtimes and malfunctions in machinery vital to the concrete production cycle. Within this investigation, seven predominant ML classifiers –Logistic Regression (LR), Decision Tree (DT), Artificial Neural Networks (ANN), Light Gradient Boosting Machine (LightGBM), Extreme Gradient Boosting (XGBoost) and Categorical boosting (CatBoost) and Random Forest (RF) – were thoroughly evaluated. Their effectiveness was gauged using several metrics, including the F1-score, Accuracy, Recall, and the area under the Receiver Operating Characteristic (ROC) curve. Notably, the CatBoost surpassed its peers, achieving an F1-score of 0.985, an accuracy of 0.984, a Recall of 0.983, and an ROC curve area of 0.984. Additionally, certain indicators (i.e., 24-hour mean vibration, 24-hour mean pressure, Error3-count, and 24-hour mean volt) were identified as pivotal in predicting machinery failure within the concrete manufacturing framework. Insights from this empirical study further underscore the importance of embedding ML into construction project management methodologies. Insights from this empirical study further emphasize the importance of embedding ML into construction project management methodologies. Such integration enhances the precision of maintenance forecasts and bolsters equipment dependability, ensuring that the concrete manufacturing paradigm is both effective and optimally beneficial.

The cutoff value of small airway dysfunction in children with bronchial asthma

Objective: To explore the cutoff value for assessing small airway dysfunction in children with asthma. Methods: A total of 364 asthmatic children aged 5 to 14 years, with normal ventilatory function, followed up at the Asthma Clinic of the Children's Hospital of Capital Institute of Pediatrics from January 2017 to January 2018, were selected as the case group. Concurrently, 403 healthy children of the same age range and without any symptoms in the community were chosen as the control group, and pulmonary function tests were conducted. The values of forced expiratory volume in 1 second (FEV1), forced vital capacity (FVC), forced expiratory flow at 50% of FVC (FEF50), forced expiratory flow at 75% of FVC (FEF75) and maximum mid-expiratory flow (MMEF) were compared between case group and control group. Statistical tests such as t-test, χ2 test, or Mann-Whitney U test were used to analyze the differences between the groups. Receiver operating characteristic (ROC) curves were constructed, and the maximum Youden Index was utilized to determine the optimal cutoff values and thresholds for identifying small airway dysfunction in asthmatic children. Results: This study comprised 364 children in the case group (220 boys and 144 girls) and 403 children in the control group (198 boys and 205 girls). The small airway parameters (FEF50%pred, FEF75%pred, MMEF%pred) in the asthmatic group were significantly lower than in the control group (77% (69%, 91%) vs. 95% (83%, 109%), 67% (54%, 82%) vs. 84% (70%, 102%), 76% (66%, 90%) vs. 97% (86%, 113%), Z=12.03, 11.35, 13.66, all P<0.001). The ROC curve area under the curve for FEF50%pred, FEF75%pred, MMEF%pred was 0.75, 0.74, and 0.79, respectively. Using a cutoff value of 80% for FEF50%pred achieved a sensitivity of 56.9% and specificity of 81.4%. A cutoff value of 74% for FEF75%pred resulted in a sensitivity of 67.3% and specificity of 69.2%. Finally, using a cutoff value of 84% for MMEF%pred achieved a sensitivity of 67.9% and specificity of 77.2%. Conclusion: In the presence of normal ventilatory function, utilizing FEF50<80% predicted or MMEF<84% predicted can accurately serve as criteria for identifying small airway dysfunction in children with controlled asthma.

Prospective validity of a clinical prediction rule for response to non-surgical multidisciplinary management of knee osteoarthritis in tertiary care: a multisite prospective longitudinal study

ObjectivesWe tested a previously developed clinical prediction tool—a nomogram consisting of four patient measures (lower patient-expected benefit, lower patient-reported knee function, greater knee varus angle and severe medial knee radiological...

Diagnostic value of liver stiffness measurement combined with risk scores for esophagogastric variceal bleeding in patients with hepatitis B cirrhosis

PurposeTo assess the predictive value of liver stiffness measurement (LSM) and three bleeding risk scoring systems for esophagogastric varices bleeding (EGVB) in patients with hepatitis B cirrhosis during hospitalization. MethodsIn this study, 210 patients who had hepatitis B cirrhosis were selected as the subjects. They were categorized into two groups based on whether EGVB occurred during hospitalization: a bleeding group (70 cases) and a non-bleeding group (140 cases). Logistic regression was used to analyze the factors related to the occurrence of EGVB, and the diagnostic performance was evaluated using a receiver operating characteristic (ROC) curve. ResultsSignificant differences were observed between the two groups in systolic blood pressure, platelet count, albumin, urea nitrogen, LSM, pre-endoscopic Rockall score (PRS), Glasgow-Blatchford score (GBS), and AIMS65 score (P < 0.05). The correlation analysis showed that LSM had significant positive relationship with PRS, GBS and AIMS65 score. Logistic regression analysis revealed that LSM and GBS score were independent risk factors for EGVB occurrence during hospitalization. ROC curve analysis showed that the combined prediction model of LSM and GBS score had the best prediction performance for EGVB occurrence, with an ROC curve area of 0.811, which was significantly better than the three risk scoring systems (P < 0.05), but similar to the predicted value of LSM (P = 0.335). ConclusionsThe combination of LSM and GBS score can significantly improve the predictive efficacy of EGVB occurrence in patients with hepatitis B cirrhosis during hospitalization, which has important clinical significance for patients' prognosis.

Nomogram model based on preoperative clinical characteristics of unilateral papillary thyroid carcinoma to predict contralateral medium-volume central lymph node metastasis.

To explore the preoperative high-risk clinical factors for contralateral medium-volume central lymph node metastasis (conMVCLNM) in unilateral papillary thyroid carcinoma (uPTC) and the indications for dissection of contralateral central lymph nodes (conCLN). Clinical and pathological data of 204 uPTC patients who underwent thyroid surgery at the Hangzhou First People's Hospital from September 2010 to October 2022 were collected. Univariate and multivariate logistic regression analyses were conducted to determine the independent risk factors for contralateral central lymph node metastasis (conCLNM) and conMVCLNM in uPTC patients based on the preoperative clinical data. Predictive models for conCLNM and conMVCLNM were constructed using logistic regression analyses and validated using receiver operating characteristic (ROC) curves, concordance index (C-index), calibration curves, and decision curve analysis (DCA). Univariate and multivariate logistic regression analyses showed that gender (P < 0.001), age (P < 0.001), tumor diameter (P < 0.001), and multifocality (P = 0.008) were independent risk factors for conCLNM in uPTC patients. Gender(P= 0.026), age (P = 0.010), platelet-to-lymphocyte ratio (PLR) (P =0.003), and tumor diameter (P = 0.036) were independent risk factors for conMVCLNM in uPTC patients. A predictive model was established to assess the risk of conCLNM and conMVCLNM, with ROC curve areas of 0.836 and 0.845, respectively. The C-index, the calibration curve, and DCA demonstrated that the model had good diagnostic value. Gender, age, tumor diameter, and multifocality are high-risk factors for conCLNM in uPTC patients. Gender, age, tumor diameter, and PLR are high-risk factors for conMVCLNM in uPTC patients, and preventive conCLN dissection should be performed.

Evaluating aspartate–platelet ratio index: A potential non-invasive tool for liver cirrhosis and esophageal varices detection

Evaluating aspartate–platelet ratio index: A potential non-invasive tool for liver cirrhosis and esophageal varices detection

Analysis of the impact of radiotherapy and surgical treatment regimens based on the SEER database on the survival outcomes of rectal cancer patients over 70 years.

This study evaluates the impact of different combinations of treatment regimens, such as additional radiation, chemotherapy, and surgical treatments, on the survival of elderly rectal cancer patients ≥ 70 years of age to support physicians' clinical decision-making. Data from a sample of elderly rectal cancer patients aged ≥ 70 years diagnosed from 2005-2015 from the US surveillance, epidemiology, and end results (SEER) database were retrospectively analyzed. The best cut-off point was selected using the x-tile software for the three continuity indices: age, tumor size, and number of regional lymph nodes. All patients were categorized into either the neoadjuvant radiotherapy and surgery group (R_S group), the surgical treatment group (S group), or the surgery and adjuvant radiotherapy group (S_R group). The propensity score allocation was used to match each included study subject in a 1:1 ratio, and the restricted mean survival time method (RMST) was used to predict the mean survival of rectal cancer patients within 5 and 10 years. The prognostic risk factors for rectal cancer patients were determined using univariate and multivariate Cox regression analyses, and nomograms were constructed. A subgroup stratification analysis of patients with different treatment combination regimens was performed using the Kaplan-Meier method, and log-rank tests were used for between-group comparisons. The model's predictive accuracy was assessed by receiver operating characteristic (ROC) curves, correction curves, and a clinical decision curve analysis (DCA). A total of 7556 cases of sample data from 2005 to 2015 were included, which were categorized into 6639 patients (87.86%) in the S group, 408 patients (5.4%) in the R_S group, and 509 patients (6.74%) in the S_R group, according to the relevant order of radiotherapy and surgery. After propensity score matching (PSM), the primary clinical characteristics of the groups were balanced and comparable. The difference in the mean survival time before and after PSM was not statistically significant in both R_S and S groups (P value > 0.05), and the difference in the mean survival time after PSM was statistically substantial in S_R and S groups (P value < 0.05). In the multifactorial Cox analysis, the M1 stage and Nodes ≥ 9 were independent risk factors. An age between 70-75 was an independent protective factor for patients with rectal cancer in the R_S and S groups. The Marital_status, T4 stage, N2 stage, M1 stage, and Nodes ≥ 9 were independent risk factors for patients with rectal cancer in the S_R and S groups, and an age between 70-81 was an independent protective factor. The ROC curve area, the model C index, and the survival calibration curve suggested good agreement between the actual and predicted values of the model. The DCA for 3-year, 5-year, and 10-year survival periods indicated that the model had some potential for application. The results of the study showed no significant difference in the overall survival (OS) between elderly patients who received neoadjuvant radiotherapy and surgery and those who received surgery alone; elderly patients who received surgery and adjuvant radiotherapy had some survival benefits compared with those who received surgery alone, though the benefit of adjuvant radiotherapy was not significant. Therefore, radiotherapy for rectal cancer patients older than 70 years old should be based on individual differences in condition, and a precise treatment plan should be developed.

Prediction Of Neonatal Jaundice(Hyperbilirubinemia) By Using Logistic Regression

The present study is undertaken with the aim to evaluate the utility of Logistic Regression classification in pre- identifying Neonatal Jaundice. In this research by using Logistic Regression classification we have predicted Neonatal Jaundice by using various parameters of neonates. Here our aim is to provide a benchmark and improve earlier ones in the field of Neonatal Jaundice diagnostics with the help of Logistic Regression classification technique. We have collected data of neonates by observing them from birth up to 72-96 hours of postnatal life. We have predicted Neonatal Jaundice by using Logistic regression. Then we have obtained accuracy, precision, recall, F1 score as evaluative measures. Also by using Receiver Operating Characteristic curve (ROC) area under curve (AUC) is 89.99%. This indicates that our model has higher discriminating power in prediction of Neonatal Jaundice.

The Neutrophil-to-Lymphocyte Ratio as a Predictor of Acute Exacerbations Among Patients With COPD in Uganda.

The neutrophil-to-lymphocyte ratio (NLR) is an inexpensive biomarker that potentially predicts acute exacerbations of chronic obstructive pulmonary disease (AECOPDs). We evaluated the association of baseline NLR and respiratory hospitalization risk within one year among chronic obstructive pulmonary disease (COPD) patients in Uganda, a low- and middle-income country. A total of 312 COPD patients were followed for one year. Clinical characteristics and exacerbation rates were collected. Poisson regression with robust variance estimators was used to measure the association between NLR and hospital admissions due to COPD exacerbations. Receiver-operator characteristic (ROC) curves and the area under the curve were used to assess the ability of NLR to predict AECOPDs. The median (Q 1, Q 3) age was 64 years (53, 71). Females comprised 50.96% (n=159) of the cohort, and 71.2% (n=222) of participants had moderate or severe COPD. A total of 9.9% (n=31) of participants experienced a COPD exacerbation during the period of follow-up. At baseline, the median (Q 1, Q 3) NLR ratio among participants who experienced an exacerbation was 1.46 (0.92, 2.33) compared to 1.03 (0.72,1.42) among those who did not experience one during the follow-up period (p=0.002). Using Youden and Liu's methods, the optimal NLR cutoff for predicting COPD exacerbation was 1.17. This cutoff resulted in a ROC curve area of 0.64 (95% confidence interval: 0.56, 0.73). The NLR could be used as a risk predictor, in low- and middle-income countries, for hospital admissions due to COPD exacerbations. A cutoff of 1.17 was an independent predictor of hospitalization due to acute exacerbations of COPD within one year.

A nomogram for predicting prognosis in patients with transjugular intrahepatic portosystemic shunt creation based on deep learning-derived spleen volume-to-platelet ratio.

The objective of our study was to develop a nomogram to predict post-transjugular intrahepatic portosystemic shunt (TIPS) survival in patients with cirrhosis based on CT images. This retrospective cohort study included patients who had received TIPS operation at the Wenzhou Medical University First Affiliated Hospital between November 2013 and April 2017. To predict prognosis, a nomogram and Web-based probability were developed to assess the overall survival (OS) rates at 1, 3, and 5 years based on multivariate analyses. With deep learning algorithm, the automated measurement of liver and spleen volumes can be realized. We assessed the predictive accuracy and discriminative ability of the nomogram using the concordance index (C-index), receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA). Age, total bilirubin, and spleen volume-to-platelet ratio (SVPR) were identified as the independent risk factors for OS. The nomogram was constructed based on the above risk factors. The C-index (0.80, 0.74, 0.70), ROC curve (area under curve: 0.828, 0.761, 0.729), calibration curve, and DCA showed that nomogram good at predictive value, stability, and clinical benefit in the prediction of 1-, 3-, 5-year OS in patients with TIPS creation. We constructed a nomogram for predicting prognosis in patients with TIPS creation based on risk factors. The nomogram can help clinicians in identifying patients with poor prognosis, eventually facilitating earlier treatment and selecting suitable patients before TIPS. This study developed the first nomogram based on SVPR to predict the prognosis of patients treated with TIPS. The nomogram could help clinician in non-invasive decision-making.

Endobronchial ultrasound: A novel screening test for pulmonary hypertension prior to major pulmonary surgery

ObjectivesPulmonary hypertension (PH) is an important physiologic variable in the assessment of patients undergoing major thoracic operations but all too often neglected because of the need for right heart catheterization (RHC) due to the inaccuracy of transthoracic echocardiography. Patients with lung cancer often require endobronchial ultrasound (EBUS) as part of the staging of the cancer. We sought to investigate whether EBUS can be used to screen these patients for PH. MethodsPatients undergoing a major thoracic operation requiring EBUS for staging were included prospectively in the study. All patients had also a RHC (gold standard). We aimed to compare the pulmonary artery pressure measurements by EBUS with the RHC values. ResultsA total of 20 patients were enrolled in the study. The prevalence of abnormal pulmonary artery pressure was 65% based on RHC. All patients underwent measurement of the pulmonary vascular acceleration time (PVAT) by EBUS with no adverse events. Linear regression analysis comparing PVAT and RHC showed a correlation (r = −0.059, −0.010 to −0.018, P = .007). A receiver operator characteristic curve (area under the curve = 0.736) was used to find the optimal PVAT threshold (140 milliseconds) to predict PH; this was used to calculate a positive and negative likelihood ratio following a positive diagnosis of 2.154 and 0.538, respectively. ConclusionsEBUS interrogation of pulmonary artery hemodynamic is safe and feasible. EBUS may be used as a screening test for PH in high-risk individuals.

The association of maternal serum uric acid with the risk of small for gestational age newborn: a retrospective cohort study

Problem Prior results on the association between serum uric acid (UA) levels in the early trimester and the risk of small for gestational age (SGA) remain unclear. This study evaluated the association of maternal first-, second-, and third-trimester UA levels with the risk of SGA infants. Method of study A total of 23, 194 singleton mothers from the International Peace Maternity and Child Health Hospital between January 2014 and January 2017 were included. Maternal UA levels were measured at 12.1 ± 1.08th (UA1) and 32.2 ± 1.03th (UA2) gestational weeks. △UA was calculated as the difference between UA2 and UA1. Logistic regression and restricted cubic spline (RCS) were performed to evaluate the association between maternal UA and △UA during pregnancy and SGA. Receiver operating characteristic (ROC) analysis was employed to assess the serum uric acid prediction value. Results Women in the higher quartiles of UA1 had a significantly higher risk of SGA. A clear increased risk for SGA was observed with higher quartiles for UA2 (p for trend <0.05). An approximately “J-shaped” relationship was observed between UA2 and △UA, and the risk of SGA was observed. When compared with those with a lower level of UA in the first trimester, those with a higher level of UA1, the more increase in the later UA levels were associated with a higher risk of SGA [adjusted odds ratio (aOR) = 1.67, 95% CI:1.37–2.05]. The ROC curve areas were 0.525 for UA1, 0.582 for UA2 and 0.576 for △UA. Conclusions The findings suggested that non-preeclamptic and non-hypertensive women who experienced early pregnancy with high UA levels had an elevated risk of SGA. Moreover, a high maternal UA level in the earlier trimester may be an early predictor of SGA.

Can machine learning models improve early detection of brain metastases using diffusion weighted imaging-based radiomics?

Metastatic complications are a major cause of cancer-related morbidity, with up to 40% of cancer patients experiencing at least one brain metastasis. Earlier detection may significantly improve patient outcomes and overall survival. We investigated machine learning (ML) models for early detection of brain metastases based on diffusion weighted imaging (DWI) radiomics. Longitudinal diffusion imaging from 116 patients previously treated with stereotactic radiosurgery (SRS) for brain metastases were retrospectively analyzed. Clinical contours from 600 metastases were extracted from radiosurgery planning computed tomography, and rigidly registered to corresponding contrast enhanced-T1 and apparent diffusion coefficient (ADC) maps. Contralateral contours located in healthy brain tissue were used as control. The dataset consisted of (I) radiomic features using ADC maps, (II) radiomic feature change calculated using timepoints before the metastasis manifested on contrast enhanced-T1, (III) primary cancer, and (IV) anatomical location. The dataset was divided into training and internal validation sets using an 80/20 split with stratification. Four classification algorithms [Linear Support Vector Machine (SVM), Random Forest (RF), AdaBoost, and XGBoost] underwent supervised classification training, with contours labeled either 'control' or 'metastasis'. Hyperparameters were optimized towards balanced accuracy. Various model metrics (receiver operating characteristic curve area scores, accuracy, recall, and precision) were calculated to gauge performance. The radiomic and clinical data set, feature engineering, and ML models developed were able to identify metastases with an accuracy of up to 87.7% on the training set, and 85.8% on an unseen test set. XGBoost and RF showed superior accuracy (XGBoost: 0.877±0.021 and 0.833±0.47, RF: 0.823±0.024 and 0.858±0.045) for training and validation sets, respectively. XGBoost and RF also showed strong area under the receiver operating characteristic curve (AUC) performance on the validation set (0.910±0.037 and 0.922±0.034, respectively). AdaBoost performed slightly lower in all metrics. SVM model generalized poorly with the internal validation set. Important features involved changes in radiomics months before manifesting on contrast enhanced-T1. The proposed models using diffusion-based radiomics showed encouraging results in differentiating healthy brain tissue from metastases using clinical imaging data. These findings suggest that longitudinal diffusion imaging and ML may help improve patient care through earlier diagnosis and increased patient monitoring/follow-up. Future work aims to improve model classification metrics, robustness, user-interface, and clinical applicability.

1510. Evaluation of Quantitative (1,3)-β-D-glucan in Prediction of Pneumocystis jirovecii Pneumonia in Patients Living with HIV in New York

Abstract Background (1,3)-β-D-glucan (BDG) has been utilized as an indirect marker for Pneumocystis jirovecii pneumonia (PJP). The clinical utility of BDG, however, is limited because of uncertainty in interpreting quantitative values. A correlation was drawn between quantitative BDG and the diagnosis of PJP in hospitalized patients living with human immunodeficiency virus (HIV) at four affiliate campuses of the Montefiore Health System in the Bronx, NY. Methods All Pneumocystis direct fluorescent assays (DFA), polymerase chain reactions (PCR) either from bronchoalveolar lavage (BAL) fluid or sputum, and quantitative BDG assays either from serum or BAL fluid were retrieved from the Montefiore microbiology laboratory (date range: 8/1/18 - 8/31/22). Results from hospitalized patients with HIV whose CD4 count within 3 months were available, with either PJP DFA or PCR result within 7 days of serum BDG and lactate dehydrogenase (LDH) test were selected (n = 146). CD4 count was grouped into &amp;lt; 50 (n = 106), 50-100 (n = 15), 100-200 (n = 16), and &amp;gt;200 (n = 9). BDG, LDH, and CD4 were considered as input variables in a logistic regression model for true PJP status. The area under the curve (AUC) of the empirical receiver operating characteristic (ROC) curve with 95% confidence intervals (CI) was estimated based on trapezoidal area. Comparisons between ROC curve areas were performed via contrast matrix to take differences of the AUC of the empirical ROC curves and chi-square testing. Results Of the 146, 29 had PJP detected and 117 did not. When BDG only was considered, the AUC of the ROC curve was 0.78. When LDH quartiles (cutoffs: 232, 334, and 480 U/L) were considered with BDG, there was a statistically significant increase in AUC (0.84 vs. 0.78; p = 0.025). Adding CD4 group did not influence AUC. Figure 1 ROC curves for BDG only, BDG with CD4 count, BDG with LDH quartile, and BDG with CD4 group and LDH quartile combined Table 1 Area under the curve calculated for the ROC curves for BDG only, BDG with CD4 count, BDG with LDH quartile, and BDG with CD4 group and LDH quartile combined Conclusion The test characteristics (AUC) improved when LDH quartiles were used in conjunction with BDG, suggesting a better prediction of PJP diagnosis when BDG and LDH were considered together. This result supports further investigation of clinical markers of PJP given the limitations of establishing a microbiologic diagnosis. Further steps of the investigation will involve analyzing if other clinical factors including HIV viral load, oxygen saturation, and radiographical reading, influence predictability of PJP. Disclosures All Authors: No reported disclosures