Area Under The Receiver Operating Characteristic Research Articles

Spatial Modelling Habitat Suitability of Javan Langur (<i>Trachypithecus auratus</i> É. Geoffroy Saint-Hilaire, 1812) in Bromo Tengger Semeru National Park (TNBTS), East Java

Javan Langur (T. auratus) is well-known as one of endemic primates from Java, Bali and Lombok Islands. The activities of land clearing, vegetation conversing, wild hunting and illegal wildlife trading are the main causes of the extinction of the Javan Langur. It can be used as an important issue for conservation action by making prediction maps of suitable habitat potential, especially for species facing a high risk of extinction in the wild. We were documenting an information about potential habitat for Javan langur using spatial suitability model in order to provide rigorous information as the basis for conservation activities of Javan langur in TNBTS. We used Landsat-8 TM image and geospatial data to support analysis as a representative of environmental parameters in order to develop the habitat model. We were using maximum entropy (MaxEnt) algorithm refers to Javan langur presence or absence. The results showed that the suitability of the Javan langur habitat in TNBTS has an excellent model accuracy level with an AUC (Area Under the Receiver Operating Characteristics) value of 0.964 and a standard deviation of 0.961. Parameters with the highest response values here are elevation, NDVI (Normalised Difference Vegetation Index) and temperatures.

Radiomics-Based Classification of Tumor and Healthy Liver on Computed Tomography Images.

Liver malignancies, particularly hepatocellular carcinoma and metastasis, stand as prominent contributors to cancer mortality. Much of the data from abdominal computed tomography images remain underused by radiologists. This study explores the application of machine learning in differentiating tumor tissue from healthy liver tissue using radiomics features. Preoperative contrast-enhanced images of 94 patients were used. A total of 1686 features classified as first-order, second-order, higher-order, and shape statistics were extracted from the regions of interest of each patient's imaging data. Then, the variance threshold, the selection of statistically significant variables using the Student's t-test, and lasso regression were used for feature selection. Six classifiers were used to identify tumor and non-tumor liver tissue, including random forest, support vector machines, naive Bayes, adaptive boosting, extreme gradient boosting, and logistic regression. Grid search was used as a hyperparameter tuning technique, and a 10-fold cross-validation procedure was applied. The area under the receiver operating curve (AUROC) assessed the performance. The AUROC scores varied from 0.5929 to 0.9268, with naive Bayes achieving the best score. The radiomics features extracted were classified with a good score, and the radiomics signature enabled a prognostic biomarker for hepatic tumor screening.

Using chest X-ray to predict tube thoracostomy in traumatic pneumothorax: A single-institution retrospective review.

Traumatic pneumothorax (PTX) is a common occurrence in thoracic trauma patients, with a majority requiring tube thoracostomy (TT) for management. Recently, the "35-mm" rule has advocated for observation of patients with PTX less than 35 mm on chest computed tomography (CT) scan. This rule has not been examined in chest x-ray (CXR). We hypothesize that a similar size cutoff can be determined in CXR predictive of need for tube thoracostomy. We performed a single-institution retrospective review of patients with traumatic PTX from 2018 to 2022, excluding those who underwent TT prior to CXR. Primary outcomes were size of pneumothorax on CXR and need for TT; secondary outcome was failed observation, defined as TT more than 4 hours after presentation. To determine the size cutoff on CXR to predict TT need, area under the receiver operating curve (AUROC) analyses were performed and Youden's index calculated (significance at p < 0.05). Predictors of failure were calculated using logistic regression. There were 341 pneumothoraces in 304 patients (94.4% blunt trauma, median injury severity score 14). Of these, 82 (24.0%) had a TT placed within the first 4 hours. Fifty-five of observed patients (21.2%) failed, and these patients had a larger PTX on CXR (8.6 mm [5.0-18.0 mm] vs. 0.0 mm [0.0-2.3 mm] ( p < 0.001)). Chest x-ray PTX size correlated moderately with CT size (r = 0.31, p < 0.001) and was highly predictive of need for TT insertion (AUC 0.75, p < 0.0001), with an optimal size cutoff predicting TT need of 38 mm. Chest x-ray imaging size was predictive of need for TT, with an optimal size cutoff on CXR of 38 mm, approaching the "35-mm rule." In addition to size, failed observation was predicted by presenting lactic acidosis and need for supplemental oxygen. This demonstrates this cutoff should be considered for prospective study in CXR. Therapeutic/Care Management; Level IV.

Association of EEG Background With Diffusion-Weighted Magnetic Resonance Neuroimaging and Short-Term Outcomes After Pediatric Cardiac Arrest.

EEG and MRI features are independently associated with pediatric cardiac arrest (CA) outcomes, but it is unclear whether their combination improves outcome prediction. We aimed to assess the association of early EEG background category with MRI ischemia after pediatric CA and determine whether addition of MRI ischemia to EEG background features and clinical variables improves short-term outcome prediction. This was a single-center retrospective cohort study of pediatric CA with EEG initiated ≤24 hours and MRI obtained ≤7 days of return of spontaneous circulation. Initial EEG background was categorized as normal, slow/disorganized, discontinuous/burst-suppression, or attenuated-featureless. MRI ischemia was defined as percentage of brain tissue with apparent diffusion coefficient (ADC) <650 × 10-6 mm2/s and categorized as high (≥10%) or low (<10%). Outcomes were mortality and unfavorable neurologic outcome (Pediatric Cerebral Performance Category increase ≥1 from baseline resulting in ICU discharge score ≥3). The Kruskal-Wallis test evaluated the association of EEG with MRI. Area under the receiver operating characteristic (AUROC) curve evaluated predictive accuracy. Logistic regression and likelihood ratio tests assessed multivariable outcome prediction. We evaluated 90 individuals. EEG background was normal in 16 (18%), slow/disorganized in 42 (47%), discontinuous/burst-suppressed in 12 (13%), and attenuated-featureless in 20 (22%) individuals. The median percentage of MRI ischemia was 5% (interquartile range 1-18); 32 (36%) individuals had high MRI ischemia burden. Twenty-eight (31%) individuals died, and 58 (64%) had unfavorable neurologic outcome. Worse EEG background category was associated with more MRI ischemia (p < 0.001). The combination of EEG background and MRI ischemia burden had higher predictive accuracy than EEG alone (AUROC: mortality: 0.92 vs 0.87, p = 0.03) or MRI alone (AUROC: mortality: 0.92 vs 0.84, p = 0.02; unfavorable: 0.83 vs 0.73, p < 0.01). Addition of percentage of MRI ischemia to clinical variables and EEG background category improved prediction for mortality (χ2 = 19.1, p < 0.001) and unfavorable neurologic outcome (χ2 = 4.8, p = 0.03) and achieved high predictive accuracy (AUROC: mortality: 0.97; unfavorable: 0.92). Early EEG background category was associated with MRI ischemia after pediatric CA. Combining EEG and MRI data yielded higher outcome predictive accuracy than either modality alone. The addition of MRI ischemia to clinical variables and EEG background improved short-term outcome prediction.

Machine learning study using 2020 SDHS data to determine poverty determinants in Somalia

Extensive research has been conducted on poverty in developing countries using conventional regression analysis, which has limited prediction capability. This study aims to address this gap by applying advanced machine learning (ML) methods to predict poverty in Somalia. Utilizing data from the first-ever 2020 Somalia Demographic and Health Survey (SDHS), a cross-sectional study design is considered. ML methods, including random forest (RF), decision tree (DT), support vector machine (SVM), and logistic regression, are tested and applied using R software version 4.1.2, while conventional methods are analyzed using STATA version 17. Evaluation metrics, such as confusion matrix, accuracy, precision, sensitivity, specificity, recall, F1 score, and area under the receiver operating characteristic (AUROC), are employed to assess the performance of predictive models. The prevalence of poverty in Somalia is notable, with approximately seven out of ten Somalis living in poverty, making it one of the highest rates in the region. Among nomadic pastoralists, agro-pastoralists, and internally displaced persons (IDPs), the poverty average stands at 69%, while urban areas have a lower poverty rate of 60%. The accuracy of prediction ranged between 67.21% and 98.36% for the advanced ML methods, with the RF model demonstrating the best performance. The results reveal geographical region, household size, respondent age group, husband employment status, age of household head, and place of residence as the top six predictors of poverty in Somalia. The findings highlight the potential of ML methods to predict poverty and uncover hidden information that traditional statistical methods cannot detect, with the RF model identified as the best classifier for predicting poverty in Somalia.

A BERT-based model for the prediction of lncRNA subcellular localization in Homo sapiens

Understanding the subcellular localization of lncRNAs is crucial for comprehending their regulation activities. The conventional detection of lncRNA subcellular location usually uses in situ detection techniques, which are resource intensive. Some machine learning-based algorithms have been proposed for lncRNA subcellular location prediction in mammals. However, due to the low level of conservation of lncRNA sequence, the performance of cross-species models remains unsatisfactory. In this study, we curated a novel dataset containing subcellular location information of lncRNAs in Homo sapiens. Subsequently, based on the BERT pre-trained language algorithm, we developed a model for lncRNA subcellular location prediction. Our model achieved a micro-average area under the receiver operating characteristic (AUROC) of 0.791 on the training set and an AUROC of 0.700 on the testing nucleus set. Additionally, we conducted cross-species validation and motif discovery to further investigate underlying patterns. In summary, our study provides valuable guidance and computational analysis tools for exploring the mechanisms of lncRNA subcellular localization and the dynamic spatial changes of RNA in abnormal physiological states.

Machine Learning-Based Prediction of Glioma IDH Gene Mutation Status Using Physio-Metabolic MRI of Oxygen Metabolism and Neovascularization (A Bicenter Study).

The mutational status of the isocitrate dehydrogenase (IDH) gene plays a key role in the treatment of glioma patients because it is known to affect energy metabolism pathways relevant to glioma. Physio-metabolic magnetic resonance imaging (MRI) enables the non-invasive analysis of oxygen metabolism and tissue hypoxia as well as associated neovascularization and microvascular architecture. However, evaluating such complex neuroimaging data requires computational support. Traditional machine learning algorithms and simple deep learning models were trained with radiomic features from clinical MRI (cMRI) or physio-metabolic MRI data. A total of 215 patients (first center: 166 participants + 16 participants for independent internal testing of the algorithms versus second site: 33 participants for independent external testing) were enrolled using two different physio-metabolic MRI protocols. The algorithms trained with physio-metabolic data demonstrated the best classification performance in independent internal testing: precision, 91.7%; accuracy, 87.5%; area under the receiver operating curve (AUROC), 0.979. In external testing, traditional machine learning models trained with cMRI data exhibited the best IDH classification results: precision, 84.9%; accuracy, 81.8%; and AUROC, 0.879. The poor performance for the physio-metabolic MRI approach appears to be explainable by site-dependent differences in data acquisition methodologies. The physio-metabolic MRI approach potentially supports reliable classification of IDH gene status in the presurgical stage of glioma patients. However, non-standardized protocols limit the level of evidence and underlie the need for a reproducible framework of data acquisition techniques.

Utility of demographic and clinical signs as diagnostic predictors for leptospiral uveitis: A retrospective study.

Leptospirosis is a waterborne zoonotic disease prevalent in tropical regions, causing significant morbidity and mortality. It can involve any organ in its primary stage, and uveitis is its late complication. While advanced laboratory diagnosis is available only in tertiary care centers globally, a cost-effective bedside assessment of clinical signs and their scoring could offer a provisional diagnosis. To analyze the diagnostic potential of demographic and clinical signs in a large cohort of serologically confirmed leptospiral uveitis patients. In this retrospective study, demographic and clinical parameters of 876 seropositive leptospiral uveitis patients and 1042 nonleptospiral uveitis controls were studied. Multivariable logistic regression analysis with bootstrap confidence interval (CI) characterized the diagnostic predictors. The performance of the model was evaluated using the area under the receiver operating curve (AUROC). Presence of nongranulomatous uveitis (odds ratio [OR] = 6.9), hypopyon (OR = 4.6), vitreous infiltration with membranous opacities (OR = 4.3), bilateral involvement (OR = 4), panuveitis (OR = 3.3), vasculitis (OR = 1.9), disc hyperemia (OR = 1.6), absence of retinochoroiditis (OR = 15), and absence of cystoid macular edema (OR = 8.9) emerged as predictive parameters. The AUROC value was 0.86 with 95% CI of 0.846-0.874. At a cut-off score of 40, the sensitivity and specificity were 79.5 and 78.4, respectively. The study demonstrates that ocular signs can serve as diagnostic predictors for leptospiral uveitis, enabling primary care ophthalmologists to make bedside diagnosis. This can be further confirmed by laboratory methods available at tertiary care centers.

Applying Machine Learning to Blood Count Data Predicts Sepsis with ICU Admission.

Timely diagnosis is crucial for sepsis treatment. Current machine learning (ML) models suffer from high complexity and limited applicability. We therefore created an ML model using only complete blood count (CBC) diagnostics. We collected non-intensive care unit (non-ICU) data from a German tertiary care centre (January 2014 to December 2021). Using patient age, sex, and CBC parameters (haemoglobin, platelets, mean corpuscular volume, white and red blood cells), we trained a boosted random forest, which predicts sepsis with ICU admission. Two external validations were conducted using data from another German tertiary care centre and the Medical Information Mart for Intensive Care IV database (MIMIC-IV). Using the subset of laboratory orders also including procalcitonin (PCT), an analogous model was trained with PCT as an additional feature. After exclusion, 1 381 358 laboratory requests (2016 from sepsis cases) were available. The CBC model shows an area under the receiver operating characteristic (AUROC) of 0.872 (95% CI, 0.857-0.887). External validations show AUROCs of 0.805 (95% CI, 0.787-0.824) for University Medicine Greifswald and 0.845 (95% CI, 0.837-0.852) for MIMIC-IV. The model including PCT revealed a significantly higher AUROC (0.857; 95% CI, 0.836-0.877) than PCT alone (0.790; 95% CI, 0.759-0.821; P < 0.001). Our results demonstrate that routine CBC results could significantly improve diagnosis of sepsis when combined with ML. The CBC model can facilitate early sepsis prediction in non-ICU patients with high robustness in external validations. Its implementation in clinical decision support systems has strong potential to provide an essential time advantage and increase patient safety.

Comparison of Fibroscan, Shear Wave Elastography, and Shear Wave Dispersion Measurements in Evaluating Fibrosis and Necroinflammation in Patients Who Underwent Liver Biopsy.

Our aim was to predict these stages of hepatic fibrosis and necroinflammation using measurements from two-dimensional shear wave elastography (2D-SWE), transient elastography (Fibroscan, TE), and shear wave dispersion (SWD). In this prospectively designed study, chronic liver patients with nonspecific etiology whose biopsy was performed for up to 1 week were included. Two-dimensional SWE, SWD, and TE measurements were performed. The METAVIR and F-ISHAK classification was used for histopathological evaluation. Two-dimensional SWE and TE were considered significant for detecting hepatic fibrosis. In distinguishing ≥F2, for 2D-SWE, area under the receiver operating characteristics (AUROC) was 0.86 (confidence interval [CI], 0.75-0.96) for the cutoff value of 8.05 kPa ( P = 0.003); for TE, AUROC was 0.79 (CI, 0.65-0.94) for the cutoff value of 10.4 kPa ( P < 0.001). No significance was found for TE in distinguishing ≥F3 ( P = 0.132). However, for 2D-SWE, a cutoff value of 10.45 kPa ( P < 0.001), with AUROC = 0.87 (CI, 0.78-0.97) was determined for ≥F3. Shear wave dispersion was able to determine the presence of necroinflammation ( P = 0.016) and a cutoff value of 15.25 (meter/second)/kiloHertz ([m/s]/kHz) ( P = 0.006) and AUROC of 0.71 (CI, 0.57-0.85) were calculated for distinguishing ≥A2. In addition, a cutoff value of 17.25 (m/s)/kHz ( P = 0.023) and AUROC = 0.72 (CI, 0.51-0.93) were found to detect severe necroinflammation. The cutoff value for SWD was 15.25 (m/s)/kHz ( P = 0.013) for detecting ≥A2 in the reversible stage of fibrosis (F0, F1, and F2), and AUROC = 0.72 (CI, 0.56-0.88). Two-dimensional SWE and TE measurements were significant in detecting the irreversible stage and the stage that should be treated in hepatic fibrosis noninvasively. Shear wave dispersion measurements were significant in detecting necroinflammation noninvasively.

Serologic Detection of Hepatocellular Carcinoma: Application of Machine Learning and Implications for Diagnostic Models.

The gender, age, lens culinaris agglutinin-reactive fraction of alphafetoprotein, alphafetoprotein, des-gamma-carboxyprothrombin (GALAD) score is a biomarker-based statistical model for the serologic diagnosis of hepatocellular carcinoma (HCC) that has been developed and validated using the case-control approach with a view to early detection. Performance has, however, been suboptimal in the first prospective studies which better reflect the real-world situation. In this article, we report the application of machine learning to a large, prospectively accrued, HCC surveillance data set. Models were built on a cohort of 3,473 patients with chronic liver disease within a rigorous surveillance program between 1998 and 2014, during which 459 patients with HCC were detected. Two random forest (RF) models were trained. The first RF model uses the same variables as the original GALAD model (GALAD-RF); the second is based on routinely available clinical and laboratory features (RF-practical). For comparison, we evaluated a logistic regression GALAD model trained on this longitudinal prospective data set (termed GALAD-Ogaki). Models were evaluated using a repetitive cross-validation approach with the metrics averaged over 100 independent runs. As judged by area under the receiver operator curve (AUROC) and F1 score, the GALAD RF model significantly outperformed the original GALAD model. The RF-practical model also outperformed the original GALAD model in terms of both AUROC and F1 score, and both models outperformed the individual biomarkers. An online web application that implemented the GALAD-RF and RF-practical models is presented. RF-based models improve on the diagnostic performance of the original GALAD model in the setting of a standard HCC surveillance program. Further prospective validation studies are warranted using these models and could be expanded to offer prediction of risk of HCC development over defined periods of time.

Value of serum iron and urine neutrophil gelatinase-associated lipocalin in predicting the mortality of critically ill patients with sepsis.

We aimed to investigate the association of iron metabolism-related parameters with 60-day mortality in critically ill patients with sepsis. Serum or urine concentrations of iron metabolism-related parameters on intensive care unit admission were measured in a prospective cohort of 133 eligible patients with sepsis according to the Sepsis-3 criteria, and these values were compared between survivors and nonsurvivors, categorized according to their 60-day survival status. Cox regression analyses were performed to examine the association between iron parameters and 60-day mortality. Kaplan-Meier methods were used to illustrate the differences in survival between different iron parameters. Of the 133 patients included in the study, 61 (45.8%) had died by day 60. After adjusting for confounding variables, higher concentrations of serum iron (cut-off 9.5 μmol/mL) and higher concentrations of urine neutrophil gelatinase-associated lipocalin (uNGAL; cut-off 169.3 ng/mL) were associated with a significantly greater risk of death in the Cox regression analysis. These two biomarkers combined with Sequential Organ Failure Assessment (SOFA) scores increased the area under the receiver operating characteristic (AUROC) curve to 0.85. These findings suggest that higher concentrations of serum iron and uNGAL are each associated with higher 60-day mortality, and they add significant accuracy to this prediction in combination with SOFA. Abbreviations: uNGAL: urine neutrophil gelatinase-associated lipocalin; ICU: intensive care unit; SOFA: Sequential Organ Failure Assessment; APACHE II: the Acute Physiology and Chronic Health Evaluation II; ELISA: enzyme-linked immunosorbent assay; HR: hazard ratio; CIs: confidence intervals; WBC: white blood cell; TBIL: total bilirubin.

Modification of qSOFA Increases Mortality Prediction Accuracy in Sepsis Patients

The Quick Sequential Organ Failure Assessment (qSOFA) is a simple parameter, however, its sensitivity as a mortality predictor is low. This study aimed to improve the predictor performance of qSOFA for in-hospital mortality. This study was a retrospective single-centered cohort using medical record data. This study included 150 patients aged 18-80 years old, who experienced sepsis and received ICU care between September 2021 and August 2022. qSOFA and modified ROX index (mROX) were calculated based on the most severe condition recorded in the emergency department (ED). Each variable's area under the receiver operating characteristic (AUROC) curve, sensitivity, and specificity were compared to predict in-hospital mortality. qSOFA scores ≥2 and mROX values ≤3.20 were independent factors that increased the risk of in-hospital mortality (OR 3.69 and 21.50; p 0.004 and 0.005, respectively). The combination of qSOFA scores ≥2 and mROX value ≤3.20 as in-hospital mortality predictors resulted in AUROC 0.791 with a sensitivity of 71.7% and specificity of 75.7%. The AUROC, sensitivity, and specificity of the combination of qSOFA and mROX were higher than qSOFA (0.766, 70.8%, 70.3%) or mROX (0.760, 68.1%, 67.6%) alone. In conclusion, the combination of qSOFA scores ≥2 and mROX values ≤3.20 increase the sensitivity and specificity for predicting in-hospital mortality in sepsis patients.

EUS-guided liver palpation as a screening tool for advanced fibrosis and cirrhosis in patients with suspected metabolic dysfunction—associated steatotic liver disease: a pilot study

Background and AimsEndoscopic liver “palpation” can be performed by indenting the liver surface under endoscopic ultrasound (EUS). Indentation depth is measured with sonographic calipers. We hypothesize that fibrotic livers are more difficult to indent, and indentation can accurately predict liver fibrosis staging. We compared EUS-guided liver palpation to conventional screening modalities in patients with suspected metabolic dysfunction-associated steatotic liver disease (MASLD). MethodsThis was a cross-sectional pilot study. Consecutive patients at three hospitals between 2021-2023 underwent EUS-guided palpation with liver biopsy. Liver palpation was compared to Fibrosis-4 index (FIB-4), AST to Platelet Ratio Index (APRI), NAFLD Fibrosis Score (NFS), and transient elastography in predicting fibrosis staging on histology. Area under the receiver operator characteristic (AUROC) curve analysis was performed. Results73 patients were included. Mean age was 49.1 and 71.2% were female. Mean body mass index was 41.1. Indentation depth was negatively correlated with fibrosis stage (Kruskal-Willis test, p<0.0001). EUS palpation demonstrated c-statistic of 0.79 and 0.95 discriminating advanced fibrosis and cirrhosis respectively. EUS-liver palpation was superior to NFS in predicting advanced fibrosis (p=0.0057) and superior to APRI and NFS predicting cirrhosis (p=0.0099 and 0.045 respectively). EUS palpation was not significantly different versus FIB-4. EUS palpation was superior to transient elastography in predicting cirrhosis (p=0.045). Using optimal cut-offs, indentation measurement ≤3.5mm yielded 100% predictive value ruling in advanced fibrosis, and ≥4.0mm yielded 100% predictive value ruling out cirrhosis. ConclusionsEUS liver palpation is a novel, accurate, and easy-to-use screening tool for advanced fibrosis and cirrhosis in patients with MASLD.

Deep Semisupervised Transfer Learning for Fully Automated Whole-Body Tumor Quantification and Prognosis of Cancer on PET/CT.

Automatic detection and characterization of cancer are important clinical needs to optimize early treatment. We developed a deep, semisupervised transfer learning approach for fully automated, whole-body tumor segmentation and prognosis on PET/CT. Methods: This retrospective study consisted of 611 18F-FDG PET/CT scans of patients with lung cancer, melanoma, lymphoma, head and neck cancer, and breast cancer and 408 prostate-specific membrane antigen (PSMA) PET/CT scans of patients with prostate cancer. The approach had a nnU-net backbone and learned the segmentation task on 18F-FDG and PSMA PET/CT images using limited annotations and radiomics analysis. True-positive rate and Dice similarity coefficient were assessed to evaluate segmentation performance. Prognostic models were developed using imaging measures extracted from predicted segmentations to perform risk stratification of prostate cancer based on follow-up prostate-specific antigen levels, survival estimation of head and neck cancer by the Kaplan-Meier method and Cox regression analysis, and pathologic complete response prediction of breast cancer after neoadjuvant chemotherapy. Overall accuracy and area under the receiver-operating-characteristic (AUC) curve were assessed. Results: Our approach yielded median true-positive rates of 0.75, 0.85, 0.87, and 0.75 and median Dice similarity coefficients of 0.81, 0.76, 0.83, and 0.73 for patients with lung cancer, melanoma, lymphoma, and prostate cancer, respectively, on the tumor segmentation task. The risk model for prostate cancer yielded an overall accuracy of 0.83 and an AUC of 0.86. Patients classified as low- to intermediate- and high-risk had mean follow-up prostate-specific antigen levels of 18.61 and 727.46 ng/mL, respectively (P < 0.05). The risk score for head and neck cancer was significantly associated with overall survival by univariable and multivariable Cox regression analyses (P < 0.05). Predictive models for breast cancer predicted pathologic complete response using only pretherapy imaging measures and both pre- and posttherapy measures with accuracies of 0.72 and 0.84 and AUCs of 0.72 and 0.76, respectively. Conclusion: The proposed approach demonstrated accurate tumor segmentation and prognosis in patients across 6 cancer types on 18F-FDG and PSMA PET/CT scans.

Validity of mortality risk prediction scores in critically ill patients with secondary pulmonary embolism.

Pulmonary embolism (PE) is a feared complication in the ICU, significantly impacting morbidity and mortality of the patients affected. Herein, we assess the use of the Acute Physiology and Chronic Health Evaluation-IV (APACHE-IV) and PE-specific risk scores to predict mortality among intensive care unit (ICU) patients who developed secondary PE. This retrospective cohort study used information from 208 United States critical care units recorded in the eICU Collaborative Research Database during 2014 and 2015. We calculated APACHE-IV, Pulmonary Embolism Severity Index (PESI), simplified PESI (sPESI), and ICU-sPESI scores and compared their predicting performance using the area under the receiver operating characteristic (AUROC) curve. Of 812 patients included in our study, 150 died (mortality, 18.5% [95% CI, 15.8%-21.1%]). Compared to survivors, non-survivors had higher APACHE-IV (86 vs 52, P<0.001), PESI (170 vs 129, P<0.001), sPESI (2 vs 2, P<0.001), and ICU-sPESI (4 vs 2, P<0.001) scores. AUROCs were 0.790 (APACHE-IV); 0.737 (PESI); 0.726 (ICU-sPESI); and 0.620 (sPESI). APACHE-IV performed significantly better than all 3 PE-specific mortality scores (APACHE-IV vs PESI, P=0.041; APACHE-IV vs sPESI, P=0.001; and APACHE-IV vs ICU-sPESI, P=0.021). Both the PESI and ICU-sPESI outperformed the sPESI (PESI vs sPESI, P=0.001; ICU-sPESI vs sPESI, P<0.001). APACHE-IV score was found to be the best instrument for predicting mortality risk, but PESI and ICU-sPESI scores may be used when APACHE-IV is unavailable. sPESI AUROC suggests absence of sufficient discriminative value to be used as a predictor of mortality in patients with secondary PE.

Employing Bayesian analysis to establish a cut-off point and assess stigma prevalence in substance use disorder: a comprehensive study of the Chinese version of the Substance Use Stigma Mechanism Scale.

In China, individuals with substance use disorders (SUD) face severe stigma, but reliable stigma assessment tool is lacking. Therefore, this study aimed to validate the Chinese version of the Substance Use Stigma Mechanism Scale (SU-SMS-C) and set its cut-off point. We recruited 1005 individuals with SUDs from Chinese rehabilitation centers. These participants completed a battery of questionnaires that included the SU-SMS-C, The Multidimensional Scale of Perceived Social Support (MSPSS), Center for Epidemiologic Studies Depression Scale (CES-D), General Self-Efficacy Scale (GSES), and Perceived Devaluation and Discrimination (PDD). Confirmatory factor analysis was used to assess the construct validity of the scale. Additionally, the Naive Bayes classifier was used to establish the cut-off point for the SU-SMS-C. We additionally explored the correlation between patient demographic characteristics and stigma. A confirmatory factor analysis was utilized, revealing a second-order five-factor model. Based on the Naive Bayes classifier, the area under the receiver operating characteristic (AUCROC) of 0.746, the cut-off point for the SU-SMS-C was established at 44.5. The prevalence of stigma observed in the study population was 49.05%. Significant disparities were observed in the distribution of stigma across genders, with males experiencing more pronounced stigma than females. Moreover, patients consuming different primary substances reported diverse levels of stigma. Notably, those primarily using heroin endured a higher degree of stigma than users of other substances. The study is the first to identify a cut-off point for the SU-SMS-C by Naive Bayes classifier, bridging a major gap in stigma measurement research. SU-SMS-C may help treat and manage SUDs by reducing stigma.

Automated curation of large-scale cancer histopathology image datasets using deep learning.

Artificial intelligence (AI) has numerous applications in pathology, supporting diagnosis and prognostication in cancer. However, most AI models are trained on highly selected data, typically one tissue slide per patient. In reality, especially for large surgical resection specimens, dozens of slides can be available for each patient. Manually sorting and labelling whole-slide images (WSIs) is a very time-consuming process, hindering the direct application of AI on the collected tissue samples from large cohorts. In this study we addressed this issue by developing a deep-learning (DL)-based method for automatic curation of large pathology datasets with several slides per patient. We collected multiple large multicentric datasets of colorectal cancer histopathological slides from the United Kingdom (FOXTROT, N = 21,384 slides; CR07, N = 7985 slides) and Germany (DACHS, N = 3606 slides). These datasets contained multiple types of tissue slides, including bowel resection specimens, endoscopic biopsies, lymph node resections, immunohistochemistry-stained slides, and tissue microarrays. We developed, trained, and tested a deep convolutional neural network model to predict the type of slide from the slide overview (thumbnail) image. The primary statistical endpoint was the macro-averaged area under the receiver operating curve (AUROCs) for detection of the type of slide. In the primary dataset (FOXTROT), with an AUROC of 0.995 [95% confidence interval [CI]: 0.994-0.996] the algorithm achieved a high classification performance and was able to accurately predict the type of slide from the thumbnail image alone. In the two external test cohorts (CR07, DACHS) AUROCs of 0.982 [95% CI: 0.979-0.985] and 0.875 [95% CI: 0.864-0.887] were observed, which indicates the generalizability of the trained model on unseen datasets. With a confidence threshold of 0.95, the model reached an accuracy of 94.6% (7331 classified cases) in CR07 and 85.1% (2752 classified cases) for the DACHS cohort. Our findings show that using the low-resolution thumbnail image is sufficient to accurately classify the type of slide in digital pathology. This can support researchers to make the vast resource of existing pathology archives accessible to modern AI models with only minimal manual annotations.

Prediction of unplanned cesarean section using measurable maternal and fetal characteristics, Ethiopia, a retrospective cohort study

BackgroundWhen a pregnant woman experiences unusual circumstances during a vaginal delivery, an unplanned cesarean section may be necessary to save her life. It requires knowledge and quick assessment of the risky situation to decide to perform an unplanned cesarean section, which only occurs in specific obstetric situations. This study aimed to develop and validate a risk prediction model for unplanned cesarean sections among laboring women in Ethiopia.MethodA retrospective follow-up study was conducted. The data were extracted using a structured checklist. Analysis was done using STATA version 14 and R version 4.2.2 software. Logistic regression was fitted to determine predictors of unplanned cesarean sections. Significant variables were then used to develop a risk prediction model. Performance was assessed using Area Under the Receiver Operating Curve (AUROC) and calibration plot. Internal validation was performed using the bootstrap technique. The clinical benefit of the model was assessed using decision curve analysis.ResultA total of 1,000 laboring women participated in this study; 28.5% were delivered by unplanned cesarean section. Parity, amniotic fluid status, gestational age, prolonged labor, the onset of labor, amount of amniotic fluid, previous mode of delivery, and abruption remained in the reduced multivariable logistic regression and were used to develop a prediction risk score with a total score of 9. The AUROC was 0.82. The optimal cut-off point for risk categorization as low and high was 6, with a sensitivity (85.2%), specificity (90.1%), and accuracy (73.9%). After internal validation, the optimism coefficient was 0.0089. The model was found to have clinical benefits.ConclusionTo objectively measure the risk of an unplanned Caesarean section, a risk score model based on measurable maternal and fetal attributes has been developed. The score is simple, easy to use, and repeatable in clinical practice.

BlastAssist: a deep learning pipeline to measure interpretable features of human embryos.

BlastAssist: a deep learning pipeline to measure interpretable features of human embryos.