Receiver Operating Characteristic Area Research Articles

Early adverse physiological event detection using commercial wearables: challenges and opportunities

Data from commercial off-the-shelf (COTS) wearables leveraged with machine learning algorithms provide an unprecedented potential for the early detection of adverse physiological events. However, several challenges inhibit this potential, including (1) heterogeneity among and within participants that make scaling detection algorithms to a general population less precise, (2) confounders that lead to incorrect assumptions regarding a participant’s healthy state, (3) noise in the data at the sensor level that limits the sensitivity of detection algorithms, and (4) imprecision in self-reported labels that misrepresent the true data values associated with a given physiological event. The goal of this study was two-fold: (1) to characterize the performance of such algorithms in the presence of these challenges and provide insights to researchers on limitations and opportunities, and (2) to subsequently devise algorithms to address each challenge and offer insights on future opportunities for advancement. Our proposed algorithms include techniques that build on determining suitable baselines for each participant to capture important physiological changes and label correction techniques as it pertains to participant-reported identifiers. Our work is validated on potentially one of the largest datasets available, obtained with 8000+ participants and 1.3+ million hours of wearable data captured from Oura smart rings. Leveraging this extensive dataset, we achieve pre-symptomatic detection of COVID-19 with a performance receiver operator characteristic (ROC) area under the curve (AUC) of 0.725 without correction techniques, 0.739 with baseline correction, 0.740 with baseline correction and label correction on the training set, and 0.777 with baseline correction and label correction on both the training and the test set. Using the same respective paradigms, we achieve ROC AUCs of 0.919, 0.938, 0.943 and 0.994 for the detection of self-reported fever, and 0.574, 0.611, 0.601, and 0.635 for detection of self-reported shortness of breath. These techniques offer improvements across almost all metrics and events, including PR AUC, sensitivity at 75% specificity, and precision at 75% recall. The ring allows continuous monitoring for detection of event onset, and we further demonstrate an improvement in the early detection of COVID-19 from an average of 3.5 days to an average of 4.1 days before a reported positive test result.

Improving Clinician Performance in Classifying EEG Patterns on the Ictal-Interictal Injury Continuum Using Interpretable Machine Learning.

In intensive care units (ICUs), critically ill patients are monitored with electroencephalography (EEG) to prevent serious brain injury. EEG monitoring is constrained by clinician availability, and EEG interpretation can be subjective and prone to interobserver variability. Automated deep-learning systems for EEG could reduce human bias and accelerate the diagnostic process. However, existing uninterpretable (black-box) deep-learning models are untrustworthy, difficult to troubleshoot, and lack accountability in real-world applications, leading to a lack of both trust and adoption by clinicians. We developed an interpretable deep-learning system that accurately classifies six patterns of potentially harmful EEG activity - seizure, lateralized periodic discharges (LPDs), generalized periodic discharges (GPDs), lateralized rhythmic delta activity (LRDA), generalized rhythmic delta activity (GRDA), and other patterns - while providing faithful case-based explanations of its predictions. The model was trained on 50,697 total 50-second continuous EEG samples collected from 2711 patients in the ICU between July 2006 and March 2020 at Massachusetts General Hospital. EEG samples were labeled as one of the six EEG patterns by 124 domain experts and trained annotators. To evaluate the model, we asked eight medical professionals with relevant backgrounds to classify 100 EEG samples into the six pattern categories - once with and once without artificial intelligence (AI) assistance - and we assessed the assistive power of this interpretable system by comparing the diagnostic accuracy of the two methods. The model's discriminatory performance was evaluated with area under the receiver-operating characteristic curve (AUROC) and area under the precision-recall curve. The model's interpretability was measured with task-specific neighborhood agreement statistics that interrogated the similarities of samples and features. In a separate analysis, the latent space of the neural network was visualized by using dimension reduction techniques to examine whether the ictal-interictal injury continuum hypothesis, which asserts that seizures and seizure-like patterns of brain activity lie along a spectrum, is supported by data. The performance of all users significantly improved when provided with AI assistance. Mean user diagnostic accuracy improved from 47 to 71% (P<0.04). The model achieved AUROCs of 0.87, 0.93, 0.96, 0.92, 0.93, and 0.80 for the classes seizure, LPD, GPD, LRDA, GRDA, and other patterns, respectively. This performance was significantly higher than that of a corresponding uninterpretable black-box model (with P<0.0001). Videos traversing the ictal-interictal injury manifold from dimension reduction (a two-dimensional representation of the original high-dimensional feature space) give insight into the layout of EEG patterns within the network's latent space and illuminate relationships between EEG patterns that were previously hypothesized but had not yet been shown explicitly. These results indicate that the ictal-interictal injury continuum hypothesis is supported by data. Users showed significant pattern classification accuracy improvement with the assistance of this interpretable deep-learning model. The interpretable design facilitates effective human-AI collaboration; this system may improve diagnosis and patient care in clinical settings. The model may also provide a better understanding of how EEG patterns relate to each other along the ictal-interictal injury continuum. (Funded by the National Science Foundation, National Institutes of Health, and others.).

Early Detection and Monitoring of Anastomotic Leaks via Naked Eye-Readable, Non-Electronic Macromolecular Network Sensors.

Anastomotic leakage (AL) is the leaking of non-sterile gastrointestinal contents into a patient's abdominal cavity. AL is one of the most dreaded complications following gastrointestinal surgery, with mortality rates reaching up to 27%. The current diagnostic methods for anastomotic leaks are limited in sensitivity and specificity. Since the timing of detection directly impacts patient outcomes, developing new, fast, and simple methods for early leak detection is crucial. Here, a naked eye-readable, electronic-free macromolecular network drain fluid sensor is introduced for continuous monitoring and early detection of AL at the patient's bedside. The sensor array comprises three different macromolecular network sensing elements, each tailored for selectivity toward the three major digestive enzymes found in the drainage fluid during a developing AL. Upon digestion of the macromolecular network structure by the respective digestive enzymes, the sensor produces an optical shift discernible to the naked eye. The diagnostic efficacy and clinical applicability of these sensors are demonstrated using clinical samples from 32 patients, yielding a Receiver Operating Characteristic Area Under the Curve (ROC AUC) of 1.0. This work has the potential to significantly contribute to improved patient outcomes through continuous monitoring and early, low-cost, and reliable AL detection.

Detecting anomalies in blockchain transactions using machine learning classifiers and explainability analysis

As the use of blockchain for digital payments continues to rise, it becomes susceptible to various malicious attacks. Successfully detecting anomalies within blockchain transactions is essential for bolstering trust in digital payments. However, the task of anomaly detection in blockchain transaction data is challenging due to the infrequent occurrence of illicit transactions. Although several studies have been conducted in the field, a limitation persists: the lack of explanations for the model's predictions. This study seeks to overcome this limitation by integrating explainable artificial intelligence (XAI) techniques and anomaly rules into tree-based ensemble classifiers for detecting anomalous Bitcoin transactions. The shapley additive explanation (SHAP) method is employed to measure the contribution of each feature, and it is compatible with ensemble models. Moreover, we present rules for interpreting whether a Bitcoin transaction is anomalous or not. Additionally, we introduce an under-sampling algorithm named XGBCLUS, designed to balance anomalous and non-anomalous transaction data. This algorithm is compared against other commonly used under-sampling and over-sampling techniques. Finally, the outcomes of various tree-based single classifiers are compared with those of stacking and voting ensemble classifiers. Our experimental results demonstrate that: (i) XGBCLUS enhances true positive rate (TPR) and receiver operating characteristic-area under curve (ROC-AUC) scores compared to state-of-the-art under-sampling and over-sampling techniques, and (ii) our proposed ensemble classifiers outperform traditional single tree-based machine learning classifiers in terms of accuracy, TPR, and false positive rate (FPR) scores.

SOFA score performs worse than age for predicting mortality in patients with COVID-19.

The use of the Sequential Organ Failure Assessment (SOFA) score, originally developed to describe disease morbidity, is commonly used to predict in-hospital mortality. During the COVID-19 pandemic, many protocols for crisis standards of care used the SOFA score to select patients to be deprioritized due to a low likelihood of survival. A prior study found that age outperformed the SOFA score for mortality prediction in patients with COVID-19, but was limited to a small cohort of intensive care unit (ICU) patients and did not address whether their findings were unique to patients with COVID-19. Moreover, it is not known how well these measures perform across races. In this retrospective study, we compare the performance of age and SOFA score in predicting in-hospital mortality across two cohorts: a cohort of 2,648 consecutive adult patients diagnosed with COVID-19 who were admitted to a large academic health system in the northeastern United States over a 4-month period in 2020 and a cohort of 75,601 patients admitted to one of 335 ICUs in the eICU database between 2014 and 2015. We used age and the maximum SOFA score as predictor variables in separate univariate logistic regression models for in-hospital mortality and calculated area under the receiver operator characteristic curves (AU-ROCs) and area under precision-recall curves (AU-PRCs) for each predictor in both cohorts. Among the COVID-19 cohort, age (AU-ROC 0.795, 95% CI 0.762, 0.828) had a significantly better discrimination than SOFA score (AU-ROC 0.679, 95% CI 0.638, 0.721) for mortality prediction. Conversely, age (AU-ROC 0.628 95% CI 0.608, 0.628) underperformed compared to SOFA score (AU-ROC 0.735, 95% CI 0.726, 0.745) in non-COVID-19 ICU patients in the eICU database. There was no difference between Black and White COVID-19 patients in performance of either age or SOFA Score. Our findings bring into question the utility of SOFA score-based resource allocation in COVID-19 crisis standards of care.

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.

Ultrasound-based jawbone surface quality evaluation after alveolar ridge preservation.

Bone readiness for implant placement is typically evaluated by bone quality/density on 2-dimensional radiographs and cone beam computed tomography at an arbitrary time between 3 and 6months after tooth extraction and alveolar ridge preservation (ARP). The aim of this study is to investigate if high-frequency ultrasound (US) can classify bone readiness in humans, using micro-CT as a reference standard to obtain bone mineral density (BMD) and bone volume fraction (BVTV) of healed sockets receiving ARP in humans. A total of 27 bone cores were harvested during the implant surgery from 24 patients who received prior extraction with ARP. US images were taken immediately before the implant surgery at a site co-registered with the tissue biopsy collection location, made possible with a specially designed guide, and then classified into 3 tiers using B-mode image criteria (1) favorable, (2) questionable, and (3) unfavorable.Bone mineral density (hydroxyapatite) and BVTV were obtained from micro-CT as the gold standard. Hydroxyapatite and BVTV were evaluated within the projected US slice plane and thresholded to favorable (>2200mg/cm3; >0.45mm3/mm3), questionable (1500-2200mg/cm3; 0.4-0.45mm3/mm3), and unfavorable (<1500mg/cm3; <0.4mm3/mm3). The present US B-mode classification inversely scales with BMD. Regression analysis showed a significant relation between US classification and BMD as well as BVTV. T-test analysis demonstrated a significant correlation between US reader scores and the gold standard. When comparing Tier 1 with the combination of Tier 2 and 3, US achieved a significant group differentiation relative to mean BMD (p=0.004, true positive 66.7%, false positive 0%, true negative 100%, false negative 33.3%, specificity 100%, sensitivity 66.7%, receiver operating characteristics area under the curve 0.86). Similar results were found between US-derived tiers and BVTV. Preliminary data suggest US could classify jawbone surface quality that correlates with BMD/BVTV and serve as the basis for future development of US-based socket healing evaluation after ARP.

Evaluation of Serial Procalcitonin Levels for the Optimization of Antibiotic Use in Non-Critically Ill COVID-19 Patients.

Procalcitonin (PCT) has been used as a biomarker to guide antibiotic therapy in various patient populations. However, its role in optimizing antibiotic use in COVID-19 patients has not been well studied to date. Thus, we aimed to evaluate the use of serial PCT monitoring as an antimicrobial stewardship tool for COVID-19 patients. This retrospective study included 240 COVID-19 patients who were admitted to a tertiary medical institution in Saudi Arabia between January 2020 and February 2022. Patients who received empiric antibiotic therapy for community-acquired pneumonia (CAP) and had serial procalcitonin levels were included. The patients were divided into two groups: the normal procalcitonin arm (PCT level < 0.5 ng/mL) and the elevated PCT arm (PCT level > 0.5 ng/mL). The primary and secondary outcomes were the effect of PCT monitoring on the duration of antibiotic exposure and the length of hospital stay, respectively. To measure the accuracy of PCT, the receiver-operating characteristic area under the curve (ROC-AUC) was determined. Among the included patients, 142 were in the normal procalcitonin arm (median PCT, 0.12 ng/mL), and 78 were in the elevated PCT arm (median PCT, 4.04 ng/mL). The baseline characteristics were similar between the two arms, except for the higher prevalence of kidney disease in the elevated PCT arm. There was no statistically significant difference in the duration of antibiotic exposure between the normal and elevated PCT arms (median duration: 7 days in both arms). However, the length of hospital stay was significantly shorter in the normal PCT arm (median stay, 9 days) than in the elevated PCT arm (median stay, 13 days; p = 0.028). The ROC-AUC value was 0.54 (95% CI: 0.503-0.595). Serial PCT monitoring did not lead to a reduction in the duration of antibiotic exposure in COVID-19 patients. However, it was associated with a shorter hospital stay. These findings suggest that PCT monitoring may be useful for optimizing antibiotic use and improving outcomes in COVID-19 patients. While PCT-guided algorithms have the potential to enable antibiotic stewardship, their role in the context of COVID-19 treatment requires further investigation.

The Lipid Accumulation Product Index (LAP) and the Cardiometabolic Index (CMI) Are Useful for Predicting the Presence and Severity of Metabolic Syndrome in Adult Patients with Obesity.

Background: The concomitant occurrence of obesity and metabolic syndrome (MetS) causes a significant worsening of a patient's clinical condition. Indexes that employ anthropometric measurements alone or associated with blood parameters have been investigated for their ability to identify MetS. This study aimed to evaluate the diagnostic accuracy of three of these indexes, the body adiposity index (BAI), the lipid accumulation product index (LAP), and the cardiometabolic index (CMI), in a cohort of 1912 adult subjects with obesity. Methods and Results: MetS was found in 62.3% of the enrolled subjects, with a higher prevalence in males (72.5%) than females (60.9%). Receiver operating characteristic (ROC) analysis was used to define which index performed better. The BAI was found to be the lowest-performing index, with an ROC area of 0.50, a sensitivity of 30.31%, a specificity of 74.48%, and a likelihood ratio of 1.19. On the contrary, the LAP and the CMI showed a comparable ROC area of 0.82. The LAP had a sensitivity of 63.06%, a specificity of 86.55%, and a likelihood ratio of 4.69, while the CMI had a sensitivity of 67.59%, specificity of 81.55%, and a likelihood ratio of 3.66. The analysis was also performed in the group divided into males and females, with overlapping results. Conclusions: The LAP and the CMI performed better than the BAI in detecting MetS both in the general population with obesity and in the male/female subgroups. In the future, it will be important to validate these useful diagnostic tools in order to employ them in clinical practices.

Increased [18F]Fluorodeoxyglucose Uptake in the Left Pallidum in Military Veterans with Blast-Related Mild Traumatic Brain Injury: Potential as an Imaging Biomarker and Mediation with Executive Dysfunction and Cognitive Impairment.

Blast-related mild traumatic brain injury (blast-mTBI) can result in a spectrum of persistent symptoms leading to substantial functional impairment and reduced quality of life. Clinical evaluation and discernment from other conditions common to military service can be challenging and subject to patient recall bias and the limitations of available assessment measures. The need for objective biomarkers to facilitate accurate diagnosis, not just for symptom management and rehabilitation but for prognostication and disability compensation purposes is clear. Toward this end, we compared regional brain [18F]fluorodeoxyglucose-positron emission tomography ([18F]FDG-PET) intensity-scaled uptake measurements and motor, neuropsychological, and behavioral assessments in 79 combat Veterans with retrospectively recalled blast-mTBI with 41 control participants having no lifetime history of TBI. Using an agnostic and unbiased approach, we found significantly increased left pallidum [18F]FDG-uptake in Veterans with blast-mTBI versus control participants, p < 0.0001; q = 3.29 × 10-9 [Cohen's d, 1.38, 95% confidence interval (0.96, 1.79)]. The degree of left pallidum [18F]FDG-uptake correlated with the number of self-reported blast-mTBIs, r2 = 0.22; p < 0.0001. Greater [18F]FDG-uptake in the left pallidum provided excellent discrimination between Veterans with blast-mTBI and controls, with a receiver operator characteristic area under the curve of 0.859 (p < 0.0001) and likelihood ratio of 21.19 (threshold:SUVR ≥ 0.895). Deficits in executive function assessed using the Behavior Rating Inventory of Executive Function-Adult Global Executive Composite T-score were identified in Veterans with blast-mTBI compared with controls, p < 0.0001. Regression-based mediation analyses determined that in Veterans with blast-mTBI, increased [18F]FDG-uptake in the left pallidum-mediated executive function impairments, adjusted causal mediation estimate p = 0.021; total effect estimate, p = 0.039. Measures of working and prospective memory (Auditory Consonant Trigrams test and Memory for Intentions Test, respectively) were negatively correlated with left pallidum [18F]FDG-uptake, p < 0.0001, with mTBI as a covariate. Increased left pallidum [18F]FDG-uptake in Veterans with blast-mTBI compared with controls did not covary with dominant handedness or with motor activity assessed using the Unified Parkinson's Disease Rating Scale. Localized increased [18F]FDG-uptake in the left pallidum may reflect a compensatory response to functional deficits following blast-mTBI. Limited imaging resolution does not allow us to distinguish subregions of the pallidum; however, the significant correlation of our data with behavioral but not motor outcomes suggests involvement of the ventral pallidum, which is known to regulate motivation, behavior, and emotions through basal ganglia-thalamo-cortical circuits. Increased [18F]FDG-uptake in the left pallidum in blast-mTBI versus control participants was consistently identified using two different PET scanners, supporting the generalizability of this finding. Although confirmation of our results by single-subject-to-cohort analyses will be required before clinical deployment, this study provides proof of concept that [18F]FDG-PET bears promise as a readily available noninvasive biomarker for blast-mTBI. Further, our findings support a causative relationship between executive dysfunction and increased [18F]FDG-uptake in the left pallidum.

Narrow interval dual phase 18F-FDG PET/CT: A practical approach for distinguishing tumor recurrence from radiation necrosis in brain metastasis.

Purpose of our research is to demonstrate efficacy of narrow interval dual phase [18F]-fluorodeoxyglucose positron emission tomography/computed tomography (FDG-PET/CT) imaging in distinguishing tumor recurrence (TR) from radiation necrosis (RN) in patients treated for brain metastases. 35 consecutive patients (22 female, 13 male) with various cancer subtypes, lesion size > 1.0 cm3, and suspected recurrence on brain magnetic resonance imaging (MRI) underwent narrow interval dual phase FDG-PET/CT (30 and 90 min after tracer injection). Clinical outcome was determined via sequential MRIs or pathology reports. Maximum standard uptake value (SUVmax) of lesion (L), gray matter (GM), and white matter (WM) was measured on early (1) and delayed (2) imaging. Analyzed variables include % change, late phase, and early phase for L uptake, L/GM uptake, and L/WM uptake. Statistical analysis (P < .01), receiver operator characteristic (ROC) curve and area under curve (AUC) cutoff values were obtained. Change in L/GM ratio of > -2% was 95% sensitive, 91% specific, and 93% accurate (P < .001, AUC = 0.99) in distinguishing TR from RN. Change in SUVmax of lesion alone was the second-best indicator (P < .001, AUC = 0.94) with an ROC cutoff > 30.5% yielding 86% sensitivity, 83% specificity, and 84% accuracy. Other variables (L alone or L/GM ratios in early or late phase, all L/WM ratios) were significantly less accurate. Utilizing narrow interval dual phase FDG-PET/CT in patients with brain metastasis treated with radiation therapy provides a practical approach to distinguish TR from RN. Narrow time interval allows for better patient comfort, greater efficiency of PET/CT scanner, and lower disruption of workflow.

Predicting Myosin Heavy Chain Isoform Through Muscle Fiber Length in Older Adults

Myosin heavy chain (MHC) isoforms directly impact muscle cellular contractile characteristics and are closely associated with other cellular functions (substrate metabolism, calcium handling, etc.). As such, MHC isoform has come to define muscle “fiber type” (MHC I = slow/oxidative and MHC II = fast/glycolytic). Due to the heterogeneity of MHC isoforms expressed within human skeletal muscle tissue, it is essential to determine MHC isoform to allow proper interpretation of tissue level outcomes that may arise from variations within or between fiber types. A priori estimation of fiber type would be a useful tool for ensuring even distribution of fiber type when considering cellular assays in which MHC isoform can only be determined after the fact. Based on the assumption that fibers of various MHC isoform would differently resist mechanical strain, we’ve demonstrated that MHC I fibers can be distinguished from MHC II fibers based on the ratio of fiber to bundle length (degree of stretch) during mechanical dissection. However, our approach used samples from healthy young adults and has not been tested in tissue from older adults, where cellular stiffness is likely elevated. Therefore, our objective is to test our ability to predict MHC isoform in older adults. We hypothesize that muscle fiber strain following dissection will predict MHC isoform in older adults. Methods: Muscle biopsies from the vastus lateralis muscle were obtained from two older adults (one male, one female). A total of 129 muscle fibers were dissected and measured. Fibers were then stored individually and MHC isoform was subsequently identified via sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-Page). Receiver Operating Characteristics (ROC) and confusion matrix were used to assess the predictive capacity for fiber stretch on activation as an index of fiber type. Data: An ROC area under curve (AUC) of 0.634 which indicates that there is a relatively low predictive capacity. Summary: This study aimed to test the feasibility of an approach for predicting muscle fiber type through changes in fiber length upon mechanical dissection. Established in young adults, this study tested samples obtained from older adults to explore the potential effects of age. Preliminary data show that muscle fiber length can predict muscle fiber type in older adults, though the accuracy may be limited when compared with the same approach in younger adults. This lack of accuracy may be attributed to the increase in stiffness that occurs with age, limiting the capacity for strain and thus fiber type variability. A new threshold may need to be determined to increase the accuracy of fiber type prediction. Conclusion: Preliminary data shows that fiber length can provide a prediction for MHC isoform in older adults. NIH National Institute on Aging (1R21AG077125-01A1) and Achievement Rewards for College Scientists (ARCS) Oregon. This is the full abstract presented at the American Physiology Summit 2024 meeting and is only available in HTML format. There are no additional versions or additional content available for this abstract. Physiology was not involved in the peer review process.

Introducing a machine learning algorithm for delirium prediction-the Supporting SURgery with GEriatric Co-Management and AI project (SURGE-Ahead).

Post-operative delirium (POD) is a common complication in older patients, with an incidence of 14-56%. To implement preventative procedures, it is necessary to identify patients at risk for POD. In the present study, we aimed to develop a machine learning (ML) model for POD prediction in older patients, in close cooperation with the PAWEL (patient safety, cost-effectiveness and quality of life in elective surgery) project. The model was trained on the PAWEL study's dataset of 878 patients (no intervention, age ≥ 70, 209 with POD). Presence of POD was determined by the Confusion Assessment Method and a chart review. We selected 15 features based on domain knowledge, ethical considerations and a recursive feature elimination. A logistic regression and a linear support vector machine (SVM) were trained, and evaluated using receiver operator characteristics (ROC). The selected features were American Society of Anesthesiologists score, multimorbidity, cut-to-suture time, estimated glomerular filtration rate, polypharmacy, use of cardio-pulmonary bypass, the Montreal cognitive assessment subscores 'memory', 'orientation' and 'verbal fluency', pre-existing dementia, clinical frailty scale, age, recent falls, post-operative isolation and pre-operative benzodiazepines. The linear SVM performed best, with an ROC area under the curve of 0.82 [95% CI 0.78-0.85] in the training set, 0.81 [95% CI 0.71-0.88] in the test set and 0.76 [95% CI 0.71-0.79] in a cross-centre validation. We present a clinically useful and explainable ML model for POD prediction. The model will be deployed in the Supporting SURgery with GEriatric Co-Management and AI project.

Femoral Neck Shaft Angle Measurement for Differentiating Femoral Head Stress Fracture from Avascular Necrosis.

The study aimed to evaluate whether the measurement of Femoral Neck Shaft Angle (FNSA) can be helpful in differentiating femoral head Stress Fracture (SF) from Avascular Necrosis (AVN). From September 2019 to April 2022, sixty-four patients [median age 32.0 years, interquartile range (IQR) 23.0-39.0 years] who underwent both hip radiograph and Magnetic Resonance Imaging (MRI) and diagnosed as femoral head SF or AVN were included in our retrospective study. Patients were divided into as having either femoral head SF (n = 34) or AVN (n = 30). The FNSA was measured in anteroposterior hip radiography. Continuous values were compared using the Mann-Whitney U test. The assessment of the predictive value of FNSA for femoral head SF was performed by Receiver Operating Characteristic (ROC) analysis. The FNSA was significantly higher in patients with SF (median 133.5°, IQR 128.0-136.7°) than those with AVN (median 127.5°, IQR 124.0-132.0°) (p = 0.001). In addition, the FNSA was significantly higher in SF femurs (median 134.8°, IQR 129.2-137.4°) than in contralateral normal femurs (median 127.1°, IQR 124.3-132.5°) in patients with unilateral femoral head SF (n = 30) (p &#60; 0.001). In ROC analysis, the sensitivity, specificity, and Area Under the Curve (AUC) for predicting the femoral head SF were 77.3%, 63.3%, and 0.785 (95% confidence interval: 0.666-0.905), respectively, at a cutoff of 130.2°. Increased FNSA was associated with femoral head SF; thus, measurement of FNSA could be helpful for differentiating femoral head SF from AVN.

Prognostic value of creatine kinase (CK)-MB to total-CK ratio in colorectal cancer patients after curative resection.

This study aimed to evaluate the prognostic significance of postoperative Creatine Kinase type M and B (CK-MB) to total Creatine Kinase (CK) ratio (CK-MB/CK) in colorectal cancer (CRC) patients after radical resection. This was a single-center retrospective cohort analysis. Subjects were stage I-III CRC patients hospitalized in Sichuan Cancer Hospital from January 2017 to May 2021. Patients were divided into abnormal group and normal group according to whether the CK-MB/CK ratio was abnormal after surgery. Through a comparative analysis of clinical data, laboratory test results, and prognosis differences between the two groups, we aimed to uncover the potential relationship between abnormal CK-MB > CK results and CRC patients. To gauge the impact of CK-MB/CK on overall survival (OS) and disease-free survival (DFS), we employed the multivariable COX regression and LASSO regression analysis. Additionally, Spearman correlation analysis, logistic regression, and receiver-operating characteristic (ROC) curve analysis were conducted to assess the predictive value of the CK-MB/CK ratio for postoperative liver metastasis. Cox regression analysis revealed that the CK-MB/CK ratio was a stable risk factors for OS (HR = 3.82, p < 0.001) and DFS (HR = 2.31, p < 0.001). To distinguish hepatic metastases after surgery, the ROC area under the curve of CK-MB/CK was 0.697 (p < 0.001), and the optimal cut-off value determined by the Youden index was 0.347. Postoperative abnormal CK-MB/CK ratio predicts worse prognosis in CRC patients after radical resection and serves as a useful biomarker for detecting postoperative liver metastasis.

Mean computed tomography value to predict spread through air spaces in clinical N0 lung adenocarcinoma

BackgroundThe aim of this study was to assess the ability of radiologic factors such as mean computed tomography (mCT) value, consolidation/tumor ratio (C/T ratio), solid tumor size, and the maximum standardized uptake (SUVmax) value by F-18 fluorodeoxyglucose positron emission tomography to predict the presence of spread through air spaces (STAS) of lung adenocarcinoma.MethodsA retrospective study was conducted on 118 patients those diagnosed with clinically without lymph node metastasis and having a pathological diagnosis of adenocarcinoma after undergoing surgery. Receiver operating characteristics (ROC) analysis was used to assess the ability to use mCT value, C/T ratio, tumor size, and SUVmax value to predict STAS. Univariate and multiple logistic regression analyses were performed to determine the independent variables for the prediction of STAS.ResultsForty-one lesions (34.7%) were positive for STAS and 77 lesions were negative for STAS. The STAS positive group was strongly associated with a high mCT value, high C/T ratio, large solid tumor size, large tumor size and high SUVmax value. The mCT values were − 324.9 ± 19.3 HU for STAS negative group and − 173.0 ± 26.3 HU for STAS positive group (p < 0.0001). The ROC area under the curve of the mCT value was the highest (0.738), followed by SUVmax value (0.720), C/T ratio (0.665), solid tumor size (0.649). Multiple logistic regression analyses using the preoperatively determined variables revealed that mCT value (p = 0.015) was independent predictive factors of predicting STAS. The maximum sensitivity and specificity were obtained at a cutoff value of − 251.8 HU.ConclusionsThe evaluation of mCT value has a possibility to predict STAS and may potentially contribute to the selection of suitable treatment strategies.

Biomarker Predictors of Clinical Efficacy of the Anti-IgE Biologic Omalizumab in Severe Asthma in Adults: Results of the SoMOSA Study.

Background: The anti-IgE monoclonal antibody omalizumab is widely used for severe asthma. This study aimed to identify biomarkers that predict clinical improvement during 1 year of omalizumab treatment. Methods: One-year open-label Study of Mechanisms of action of Omalizumab in Severe Asthma (SoMOSA) involving 216 patients with severe (Global Initiative for Asthma step 4/5) uncontrolled atopic asthma (at least two severe exacerbations in the previous year) taking high-dose inhaled corticosteroids and long-acting β-agonists with or without maintenance oral corticosteroids. It had two phases: 0-16 weeks, to assess early clinical improvement by Global Evaluation of Therapeutic Effectiveness (GETE); and 16-52 weeks, to assess late responses based on ⩾50% reduction in exacerbations or mOCS dose. All participants provided samples (exhaled breath, blood, sputum, urine) before and after 16 weeks of omalizumab treatment. Measurements and Main Results: A total of 191 patients completed phase 1; 63% had early improvement. Of 173 who completed phase 2, 69% had reduced exacerbations by ⩾50% and 57% (37 of 65) taking mOCSs had reduced their dose by ⩾50%. The primary outcomes 2,3-dinor-11-β-PGF2α, GETE score, and standard clinical biomarkers (blood and sputum eosinophils, exhaled nitric oxide, serum IgE) did not predict either clinical response. Five volatile organic compounds and five plasma lipid biomarkers strongly predicted the ⩾50% reduction in exacerbations (receiver operating characteristic areas under the curve of 0.780 and 0.922, respectively) and early responses (areas under the curve of 0.835 and 0.949, respectively). In an independent cohort, gas chromatography/mass spectrometry biomarkers differentiated between severe and mild asthma. Conclusions: This is the first discovery of omics biomarkers that predict improvement in asthma with biologic agent treatment. Prospective validation and development for clinical use is justified.

Optimizing building material selection: A machine learning approach for efficient concrete compressive strength forecasting

Price changes in construction materials have a significant impact on building construction projects. Such price variations occur at random and at varying rates over time. A system that can estimate the magnitude and quantity of the change in material prices with reasonable accuracy is required. The primary goal is to create a machine-learning model that can predict the type of building material chosen based on environmental factors. The compressive strength of concrete is critical in defining its mechanical qualities. Long laboratory testing is needed to determine the compressive strength of concrete. The capacity of powerful machine learning algorithms to forecast concrete compressive strength speeds up these lengthy experimental methods while also lowering expenses. This study provides abilities to precisely anticipate and categorize numerous qualities and traits of distinct materials. The framework includes a broad dataset that details materials, composition, and performance characteristics. Machine learning algorithms such as logistic regression (LR), decision trees (DT), and random forests (RF) train models on the training data. The models are hyper-parameter tweaked and feature developed to achieve the most outstanding performance. The k-fold method is used throughout the training and assessment phase to guarantee robustness and reduce bias. The F1 score and Receiver Operating Characteristic-Area Under Curve (ROC-AUC) curve are two performance measures used to measure how accurate and predictive the trained models are. The study findings provide insights into the qualities of the materials, facilitating improved material selection, quality assurance, and decision-making in the building sector. In the analyses, the best accuracy value was 99.92%, and the precision value was 88.83% using the LR algorithm. As a result, it was determined that the LR algorithm had the least execution 57.826 ms, and is thus the most suitable for use in concrete compressive strength estimation.

Serum Levels of miR-122-5p and miR-125a-5p Predict Hepatotoxicity Occurrence in Patients Undergoing Autologous Hematopoietic Stem Cell Transplantation.

Hepatic complications are an acknowledged cause of mortality and morbidity among patients undergoing hematopoietic stem cell transplantation. In this study, we aimed to evaluate the potential role in the prediction of liver injury of five selected microRNAs (miRNAs)-miR-122-5p, miR-122-3p, miR-15b-5p, miR-99b-5p, and miR-125a-5p-in the setting of autologous hematopoietic stem cell transplantation (ASCT). A total of 66 patients were included in the study: 50 patients (75.8%) with multiple myeloma (MM) and 16 (24.2%) with lymphoma. Blood samples were collected after the administration of the conditioning regimen, on the day of transplant (day 0). The expression levels of selected miRNAs were quantified by reverse transcription-quantitative polymerase chain reaction (RT-qPCR) using the miRCURY LNA miRNA Custom PCR Panels (QIAGEN). In a multivariate logistic regression analysis adjusted for age, sex, and the administered conditioning regimen, two miRNAs, hsa-miR-122-5p (odds ratio, OR 2.10, 95% confidence interval, CI: 1.29-3.42, p = 0.0029) and hsa-miR-125a-5p (OR 0.27, 95% CI: 0.11-0.71, p = 0.0079), were independent for hepatic toxicity occurrence during the 14 days after transplant. Our model in 10-fold cross-validation preserved its diagnostic potential with a receiver operating characteristics area under the curve (ROC AUC) of 0.75, 95% CI: 0.63-0.88 and at optimal cut-off reached 72.0% sensitivity and 74.4% specificity. An elevated serum level of miR-122-5p and decreased level of miR-125a-5p on day 0 are independent risk factors for hepatotoxicity in ASCT recipients, showing promise in accurately predicting post-ASCT complications. Identifying patients susceptible to complications has the potential to reduce procedure costs and optimize the selection of inpatient or outpatient procedures.

Imaging Outcomes of Emergency MR Imaging in Dizziness and Vertigo: A Retrospective Cohort Study.

Patients exhibiting acute dizziness or vertigo often represent a diagnostic challenge, and many undergo neuroimaging for stroke detection. We aimed to demonstrate the imaging outcomes of first-line emergency MR imaging among patients with acute dizziness or vertigo and to determine the clinical risk factors for stroke and other acute pathology. This retrospective study included consecutive patients with acute dizziness or vertigo referred for emergency MR imaging in a tertiary hospital during 5 years. We recorded and analyzed patient characteristics, relevant clinical information, and imaging outcomes. Risk score models were derived to predict which patients were more likely to present with positive MR imaging findings. A total of 1169 patients were included. Acute stroke was found in 17%; other clinically significant pathology, in 8% of patients. In 75% of the patients, emergency MR imaging showed no significant abnormalities. Risk factors for acute stroke included older age, male sex, and a prevalence of cardiovascular risk factors and neurologic signs. Isolated dizziness had no discriminative power on imaging outcomes, and 14% of these patients showed acute stroke. Risk scores had only moderate performance in predicting acute ischemic stroke (receiver operating characteristic area under curve = 0.75) or any significant pathology (receiver operating characteristic area under curve = 0.70). Acute dizziness and vertigo remain challenging even when emergency MR imaging is readily available. One in 4 patients had acute pathology on MR imaging. Predictors for acute pathology (older age, male sex, cardiovascular risk factors, and neurologic signs) may aid in patient selection for MR imaging, optimizing the yield and clinical impact of emergency neuroimaging. Low diagnostic yields of CT and internal acoustic canal MR imaging may offer an opportunity to reduce health care expenditures in the future.