Articoli di riviste sul tema "Auc-Roc"
Segui questo link per vedere altri tipi di pubblicazioni sul tema: Auc-Roc.
Cita una fonte nei formati APA, MLA, Chicago, Harvard e in molti altri stili
Vedi i top-50 articoli di riviste per l'attività di ricerca sul tema "Auc-Roc".
Accanto a ogni fonte nell'elenco di riferimenti c'è un pulsante "Aggiungi alla bibliografia". Premilo e genereremo automaticamente la citazione bibliografica dell'opera scelta nello stile citazionale di cui hai bisogno: APA, MLA, Harvard, Chicago, Vancouver ecc.
Puoi anche scaricare il testo completo della pubblicazione scientifica nel formato .pdf e leggere online l'abstract (il sommario) dell'opera se è presente nei metadati.
Vedi gli articoli di riviste di molte aree scientifiche e compila una bibliografia corretta.
1
Hong, Chong Sun, e So Yeon Choi. "ROC curve generalization and AUC". Journal of the Korean Data And Information Science Society 31, n. 4 (31 luglio 2020): 477–88. http://dx.doi.org/10.7465/jkdi.2020.31.4.477.
2
Hong, Chong Sun, e Dae Soon Yang. "ROC curve and AUC for linear growth models". Journal of the Korean Data and Information Science Society 26, n. 6 (30 novembre 2015): 1367–75. http://dx.doi.org/10.7465/jkdi.2015.26.6.1367.
3
Минин, А. С. "Бинаризация вероятностного прогноза методом ROC AUC". ТЕНДЕНЦИИ РАЗВИТИЯ НАУКИ И ОБРАЗОВАНИЯ 104, n. 14 (2023): 87–91. http://dx.doi.org/10.18411/trnio-12-2023-789.
Abstract (sommario):
В работе проведено исследование влияния порога бинаризации вероятностного прогноза классификатора k-ближайших соседей на значение метрики ROC AUC. Путем варьирования порога бинаризации прогнозов и расчета ROC AUC выявлен оптимальный порог, при котором достигается максимальное значение метрики. Актуальность работы обусловлена широким практическим применением вероятностных классификаторов и необходимостью преобразования их непрерывных прогнозов в дискретные классы. Целью исследования является нахождение оптимального значения порога бинаризации для конкретного классификатора и набора данных на основе анализа зависимости метрики качества ROC AUC от величины порога. Полученные результаты могут быть использованы для настройки и оптимизации работы классификаторов, основанных на вероятностных прогнозах.
4
Krupinski, Elizabeth A. "Evaluating AI Clinically—It’s Not Just ROC AUC!" Radiology 298, n. 1 (gennaio 2021): 47–48. http://dx.doi.org/10.1148/radiol.2020203782.
5
Mukhametshin, Rustam F., Olga P. Kovtun e Nadezhda S. Davydova. "Respiratory parameters as a predictor of hospital outcomes in newborns requiring medical evacuation". Russian Journal of Pediatric Surgery, Anesthesia and Intensive Care 12, n. 4 (19 gennaio 2023): 441–52. http://dx.doi.org/10.17816/psaic1292.
Abstract (sommario):
BACKGROUND: Assessment of the clinical condition, prediction of risks and possible outcomes during the transfer of newborns remains an important part of the work of transport teams. Respiratory disorders remain a significant indication for transfer to medical organizations of a higher level of care.
AIM: To study the predictive value of the parameters of respiratory support in newborns requiring medical evacuation for the outcomes of treatment.
MATERIALS AND METHODS: The observational, cohort, retrospective study included data from neonatal to patients on ventilators (286 newborns) in the period from August 1, 2017 to December 31, 2018. Anamnesis parameters, intensive care volume, respiratory support settings, and assessments on scales (KSHONN, NTISS, TRIPS) were evaluated. Analyzed: 24-hours mortality, 7 days mortality, hospital mortality, air leakage syndrome. The assessment and comparison of the predictive value of the parameters in relation to the hospital outcomes was performed.
RESULTS: The AUC ROC of SpO2/FiO2 for predicting 24-hours mortality was 0.984 [0.9661.000], which is significantly higher than the ROC of the saturation oxygenation index (AUC 0.972 [0.9490.995], p = 0.004). The area under the ROC of the 24-hours mortality on the TRIPS scale does not significantly differ from the saturation index of oxygenation (AUC 0.972 [0.9490.995], p = 0.113) and the mean airway pressure (AUC 0.943 [0.8841.000], p = 0.107). When predicting 7-day mortality, the saturation oxygenation index has AUC ROC (0.702 [0.5490.854]) significantly lower than AUC ROC for SpO2/FiO2 (0.762 [0.6380.887], p = 0.001). SpO2/FiO2 predicts total mortality with AUC ROC (0.759 [0.6770.841]).
CONCLUSIONS: The mean airway pressure, saturation oxygenation index and SpO2/FiO2 have a high (AUC 0,9) predictive value for 24-hours mortality, while only SpO2/FiO2 reliably predicts total mortality with AUC ROC 0,7.
6
Muschelli, John. "ROC and AUC with a Binary Predictor: a Potentially Misleading Metric". Journal of Classification 37, n. 3 (23 dicembre 2019): 696–708. http://dx.doi.org/10.1007/s00357-019-09345-1.
7
Khaidarov, A. G., A. I. Soloviev e D. A. Budko. "STUDY OF THE MOST EFFICIENT MODELS AND ATRIBUTION ALGORITHMS USING THE ROC AUC INDICATOR". Современные наукоемкие технологии (Modern High Technologies), n. 7 2022 (2022): 63–68. http://dx.doi.org/10.17513/snt.39234.
8
García de Guadiana-Romualdo, Luis, María Dolores Albaladejo-Otón, Mario Berger, Enrique Jiménez-Santos, Roberto Jiménez-Sánchez, Patricia Esteban-Torrella, Sergio Rebollo-Acebes, Ana Hernando-Holgado, Alejandro Ortín-Freire e Javier Trujillo-Santos. "Prognostic performance of pancreatic stone protein in critically ill patients with sepsis". Biomarkers in Medicine 13, n. 17 (dicembre 2019): 1469–80. http://dx.doi.org/10.2217/bmm-2019-0174.
Abstract (sommario):
Aim: To assess the prognostic value for 28-day mortality of PSP in critically ill patients with sepsis. Material & methods: 122 consecutive patients with sepsis were enrolled in this study. Blood samples were collected on admission and day 2. Results: On admission, the combination of PSP and lactate achieved an area under the receiver operating characteristic (AUC-ROC) of 0.796, similar to sequential organ failure assessment score alone (AUC-ROC: 0.826). On day 2, PSP was the biomarker with the highest performance (AUC-ROC: 0.844), although lower (p = 0.041) than sequential organ failure assessment score (AUC-ROC: 0.923). Conclusion: The combination of PSP and lactate and PSP alone, on day 2, have a good performance for prognosis of 28-day mortality and could help to identify patients who may benefit most from tailored intensive care unit management.
9
Sauka, Kudzai, Gun-Yoo Shin, Dong-Wook Kim e Myung-Mook Han. "Adversarial Robust and Explainable Network Intrusion Detection Systems Based on Deep Learning". Applied Sciences 12, n. 13 (25 giugno 2022): 6451. http://dx.doi.org/10.3390/app12136451.
Abstract (sommario):
The ever-evolving cybersecurity environment has given rise to sophisticated adversaries who constantly explore new ways to attack cyberinfrastructure. Recently, the use of deep learning-based intrusion detection systems has been on the rise. This rise is due to deep neural networks (DNN) complexity and efficiency in making anomaly detection activities more accurate. However, the complexity of these models makes them black-box models, as they lack explainability and interpretability. Not only is the DNN perceived as a black-box model, but recent research evidence has also shown that they are vulnerable to adversarial attacks. This paper developed an adversarial robust and explainable network intrusion detection system based on deep learning by applying adversarial training and implementing explainable AI techniques. In our experiments with the NSL-KDD dataset, the PGD adversarial-trained model was a more robust model than DeepFool adversarial-trained and FGSM adversarial-trained models, with a ROC-AUC of 0.87. The FGSM attack did not affect the PGD adversarial-trained model’s ROC-AUC, while the DeepFool attack caused a minimal 9.20% reduction in PGD adversarial-trained model’s ROC-AUC. PGD attack caused a 15.12% reduction in the DeepFool adversarial-trained model’s ROC-AUC and a 12.79% reduction in FGSM trained model’s ROC-AUC.
10
Amala, R., e Sudesh Pundir. "ROC Curve and AUC for A Left-Truncated Sample from Rayleigh Distribution". American Journal of Mathematical and Management Sciences 34, n. 2 (31 dicembre 2014): 89–116. http://dx.doi.org/10.1080/01966324.2014.969461.
11
Takenouchi, Takashi, Osamu Komori e Shinto Eguchi. "An Extension of the Receiver Operating Characteristic Curve and AUC-Optimal Classification". Neural Computation 24, n. 10 (ottobre 2012): 2789–824. http://dx.doi.org/10.1162/neco_a_00336.
Abstract (sommario):
While most proposed methods for solving classification problems focus on minimization of the classification error rate, we are interested in the receiver operating characteristic (ROC) curve, which provides more information about classification performance than the error rate does. The area under the ROC curve (AUC) is a natural measure for overall assessment of a classifier based on the ROC curve. We discuss a class of concave functions for AUC maximization in which a boosting-type algorithm including RankBoost is considered, and the Bayesian risk consistency and the lower bound of the optimum function are discussed. A procedure derived by maximizing a specific optimum function has high robustness, based on gross error sensitivity. Additionally, we focus on the partial AUC, which is the partial area under the ROC curve. For example, in medical screening, a high true-positive rate to the fixed lower false-positive rate is preferable and thus the partial AUC corresponding to lower false-positive rates is much more important than the remaining AUC. We extend the class of concave optimum functions for partial AUC optimality with the boosting algorithm. We investigated the validity of the proposed method through several experiments with data sets in the UCI repository.
12
So, Yoon Kyoung, Zero Kim, Taek Yoon Cheong, Myung Jin Chung, Chung-Hwan Baek, Young-Ik Son, Jungirl Seok et al. "Detection of Cancer Recurrence Using Systemic Inflammatory Markers and Machine Learning after Concurrent Chemoradiotherapy for Head and Neck Cancers". Cancers 15, n. 14 (8 luglio 2023): 3540. http://dx.doi.org/10.3390/cancers15143540.
Abstract (sommario):
Pretreatment values of the neutrophil-to-lymphocyte ratio (NLR) and the platelet-to-lymphocyte ratio (PLR) are well-established prognosticators in various cancers, including head and neck cancers. However, there are no studies on whether temporal changes in the NLR and PLR values after treatment are related to the development of recurrence. Therefore, in this study, we aimed to develop a deep neural network (DNN) model to discern cancer recurrence from temporal NLR and PLR values during follow-up after concurrent chemoradiotherapy (CCRT) and to evaluate the model’s performance compared with conventional machine learning (ML) models. Along with conventional ML models such as logistic regression (LR), random forest (RF), and gradient boosting (GB), the DNN model to discern recurrences was trained using a dataset of 778 consecutive patients with primary head and neck cancers who received CCRT. There were 16 input features used, including 12 laboratory values related to the NLR and the PLR. Along with the original training dataset (N = 778), data were augmented to split the training dataset (N = 900). The model performance was measured using ROC-AUC and PR-AUC values. External validation was performed using a dataset of 173 patients from an unrelated external institution. The ROC-AUC and PR-AUC values of the DNN model were 0.828 ± 0.032 and 0.663 ± 0.069, respectively, in the original training dataset, which were higher than the ROC-AUC and PR-AUC values of the LR, RF, and GB models in the original training dataset. With the recursive feature elimination (RFE) algorithm, five input features were selected. The ROC-AUC and PR-AUC values of the DNN-RFE model were higher than those of the original DNN model (0.883 ± 0.027 and 0.778 ± 0.042, respectively). The ROC-AUC and PR-AUC values of the DNN-RFE model trained with a split dataset were 0.889 ± 0.032 and 0.771 ± 0.044, respectively. In the external validation, the ROC-AUC values of the DNN-RFE model trained with the original dataset and the same model trained with the split dataset were 0.710 and 0.784, respectively. The DNN model with feature selection using the RFE algorithm showed the best performance among the ML models to discern a recurrence after CCRT in patients with head and neck cancers. Data augmentation by splitting training data was helpful for model performance. The performance of the DNN-RFE model was also validated with an external dataset.
13
Afrianto, Mochammad Agus, e Meditya Wasesa. "Booking Prediction Models for Peer-to-peer Accommodation Listings using Logistics Regression, Decision Tree, K-Nearest Neighbor, and Random Forest Classifiers". Journal of Information Systems Engineering and Business Intelligence 6, n. 2 (27 ottobre 2020): 123. http://dx.doi.org/10.20473/jisebi.6.2.123-132.
Abstract (sommario):
Background: Literature in the peer-to-peer accommodation has put a substantial focus on accommodation listings' price determinants. Developing prediction models related to the demand for accommodation listings is vital in revenue management because accurate price and demand forecasts will help determine the best revenue management responses.Objective: This study aims to develop prediction models to determine the booking likelihood of accommodation listings.Methods: Using an Airbnb dataset, we developed four machine learning models, namely Logistics Regression, Decision Tree, K-Nearest Neighbor (KNN), and Random Forest Classifiers. We assessed the models using the AUC-ROC score and the model development time by using the ten-fold three-way split and the ten-fold cross-validation procedures.Results: In terms of average AUC-ROC score, the Random Forest Classifiers outperformed other evaluated models. In three-ways split procedure, it had a 15.03% higher AUC-ROC score than Decision Tree, 2.93 % higher than KNN, and 2.38% higher than Logistics Regression. In the cross-validation procedure, it has a 26,99% higher AUC-ROC score than Decision Tree, 4.41 % higher than KNN, and 3.31% higher than Logistics Regression. It should be noted that the Decision Tree model has the lowest AUC-ROC score, but it has the smallest model development time.Conclusion: The performance of random forest models in predicting booking likelihood of accommodation listings is the most superior. The model can be used by peer-to-peer accommodation owners to improve their revenue management responses.
14
Janssens, A. Cecile J. W., e Forike K. Martens. "Reflection on modern methods: Revisiting the area under the ROC Curve". International Journal of Epidemiology 49, n. 4 (22 gennaio 2020): 1397–403. http://dx.doi.org/10.1093/ije/dyz274.
Abstract (sommario):
Abstract The area under the receiver operating characteristic (ROC) curve (AUC) is commonly used for assessing the discriminative ability of prediction models even though the measure is criticized for being clinically irrelevant and lacking an intuitive interpretation. Every tutorial explains how the coordinates of the ROC curve are obtained from the risk distributions of diseased and non-diseased individuals, but it has not become common sense that therewith the ROC plot is just another way of presenting these risk distributions. We show how the ROC curve is an alternative way to present risk distributions of diseased and non-diseased individuals and how the shape of the ROC curve informs about the overlap of the risk distributions. For example, ROC curves are rounded when the prediction model included variables with similar effect on disease risk and have an angle when, for example, one binary risk factor has a stronger effect; and ROC curves are stepped rather than smooth when the sample size or incidence is low, when the prediction model is based on a relatively small set of categorical predictors. This alternative perspective on the ROC plot invalidates most purported limitations of the AUC and attributes others to the underlying risk distributions. AUC is a measure of the discriminative ability of prediction models. The assessment of prediction models should be supplemented with other metrics to assess their clinical utility.
15
Khoshtinat, Saeed, Babak Aminnejad, Yousef Hassanzadeh e Hasan Ahmadi. "Application of GIS-based models of weights of evidence, weighting factor, and statistical index in spatial modeling of groundwater". Journal of Hydroinformatics 21, n. 5 (10 luglio 2019): 745–60. http://dx.doi.org/10.2166/hydro.2019.127.
Abstract (sommario):
Abstract The present research aims at applying three geographic information system (GIS)-based bivariate models, namely, weights of evidence (WOE), weighting factor (WF), and statistical index (SI), for mapping of groundwater potential for sustainable groundwater management. The locations of wells with groundwater yields more than 11 m3/h were selected for modeling. Then, these locations were grouped into two categories with 70% (52 locations) in a training dataset to build the model and 30% (22 locations) in a testing dataset to validate it. Conditioning factors, namely, altitude, slope degree, plan curvature, slope aspect, rainfall, soil, land use, geology, distance from fault, and distance from river were selected. Finally, the three achieved maps were compared using area under receiver operating characteristic (ROC) and area under the ROC curve (AUC). The ROC method result showed that the SI model better fitted the training dataset (AUC = 0.747) followed by WF (AUC = 0.742) and WOE (AUC = 0.737). Results of the testing dataset show that the WOE model (AUC = 0.798) outperforms SI (AUC = 0.795) and WF (AUC = 0.791). According to the WF model, altitude and rainfall had the highest and lowest impacts on groundwater well potential occurrence, respectively. With regard to Friedman test, the difference in performances of these three models was not statistically significant.
16
Mikula, Anthony L., Seth K. Williams e Paul A. Anderson. "The use of intraoperative triggered electromyography to detect misplaced pedicle screws: a systematic review and meta-analysis". Journal of Neurosurgery: Spine 24, n. 4 (aprile 2016): 624–38. http://dx.doi.org/10.3171/2015.6.spine141323.
Abstract (sommario):
OBJECT Insertion of instruments or implants into the spine carries a risk for injury to neural tissue. Triggered electromyography (tEMG) is an intraoperative neuromonitoring technique that involves electrical stimulation of a tool or screw and subsequent measurement of muscle action potentials from myotomes innervated by nerve roots near the stimulated instrument. The authors of this study sought to determine the ability of tEMG to detect misplaced pedicle screws (PSs). METHODS The authors searched the US National Library of Medicine, the Web of Science Core Collection database, and the Cochrane Central Register of Controlled Trials for PS studies. A meta-analysis of these studies was performed on a per-screw basis to determine the ability of tEMG to detect misplaced PSs. Sensitivity, specificity, and receiver operating characteristic (ROC) area under the curve (AUC) were calculated overall and in subgroups. RESULTS Twenty-six studies were included in the systematic review. The authors analyzed 18 studies in which tEMG was used during PS placement in the meta-analysis, representing data from 2932 patients and 15,065 screws. The overall sensitivity of tEMG for detecting misplaced PSs was 0.78, and the specificity was 0.94. The overall ROC AUC was 0.96. A tEMG current threshold of 10–12 mA (ROC AUC 0.99) and a pulse duration of 300 µsec (ROC AUC 0.97) provided the most accurate testing parameters for detecting misplaced screws. Screws most accurately conducted EMG signals (ROC AUC 0.98). CONCLUSIONS Triggered electromyography has very high specificity but only fair sensitivity for detecting malpositioned PSs.
17
Dashina, P., e R. Vishnu Vardhan. "ESTIMATION OF AUC OF BI-GENERALIZED EXPONENTIAL ROC CURVE AND ITS ASYMPTOTIC RESULTS". Advances and Applications in Statistics 79 (3 agosto 2022): 105–19. http://dx.doi.org/10.17654/0972361722062.
18
Zhang, Feng, Song Qiao, Ning Yao, Chunqiao Li, Marie-Christin Weber, Benedict Jefferies, Helmut Friess, Stefan Reischl e Philipp-Alexander Neumann. "Anastomotic Rings and Inflammation Values as Biomarkers for Leakage of Stapled Circular Colorectal Anastomoses". Diagnostics 12, n. 12 (22 novembre 2022): 2902. http://dx.doi.org/10.3390/diagnostics12122902.
Abstract (sommario):
Reliable markers to predict or diagnose anastomotic leakage (AL) of stapled circular anastomoses following colorectal resections are an important clinical need. Here, we aim to quantitatively investigate the morphology of anastomotic rings as an early available prognostic marker for AL and compare them to established inflammatory markers. We perform a prospective single-center cohort study, including patients undergoing stapled circular anastomosis between August 2020 and August 2021. The predictive value of the anastomotic ring configuration and the neutrophil-to-lymphocyte ratio (NLR) regarding anastomotic leakage is examined by ROC analyses and compared to the C-reactive protein (CRP) as an established marker. We included 204 patients, of which 19 suffered from anastomotic leakage (LEAK group), while in 185 patients the anastomoses healed well (HEAL group). The minimal height of the anastomotic rings as a binary classifier had a good ROC-AUC of 0.81 but was inferior to the NLR at postoperative day (POD) 5, with an excellent ROC-AUC of 0.93. Still, it was superior to the NLR at POD 3 (0.74) and the CRP at POD 3 (ROC-AUC 0.54) and 5 (ROC-AUC 0.70). The minimal height of the anastomotic rings as indicator for technically insufficient anastomoses is a good predictor of AL, while postoperatively the NLR was superior to the CRP in prediction of AL.
19
Foley, Christina S., Edwina C. Moore, Mira Milas, Eren Berber, Joyce Shin e Allan E. Siperstein. "RECEIVER OPERATING CHARACTERISTIC ANALYSIS OF INTRAOPERATIVE PARATHYROID HORMONE MONITORING TO DETERMINE OPTIMUM SENSITIVITY AND SPECIFICITY: ANALYSIS OF 896 CASES". Endocrine Practice 25, n. 11 (novembre 2019): 1117–26. http://dx.doi.org/10.4158/ep-2019-0191.
Abstract (sommario):
Objective: While intraoperative parathyroid hormone (IOPTH) monitoring with a ≥50% drop commonly guides the extent of exploration for primary hyperparathyroidism (pHPT), receiver operating characteristic (ROC) analysis has not been performed to determine whether other criteria yield better sensitivity and specificity. The aim of this study was to identify the optimum percent change of IOPTH following removal of the abnormal parathyroid pathology, in order to predict biochemical cure. Secondary aims were to identify patient subgroups with increased area under the ROC curve (AUC) and the need for moderated criteria. Methods: A retrospective review was performed on patients undergoing primary parathyroid surgery for sporadic pHPT between 1999 and 2010 at a tertiary center for endocrine surgery. Eight hundred and ninety-six patients with primary hyperparathyroidism were included. Multigland disease (MGD) was defined as the intraoperative detection of more than 1 enlarged hypercellular gland or persistent disease after single gland excision. ROC analysis was used to determine the value with the best performance at predicting MGD, following bilateral exploration. Results: MGD was diagnosed in 174 patients (19.4%). ROC analysis demonstrated an AUC of 0.69. An IOPTH drop of 72% was the point of optimal discrimination with a sensitivity of 55% and specificity of 76% for predicting MGD. Subgroup analysis by preoperative calcium, preoperative PTH, localization studies, or pre- and post-excision IOPTH, did not identify any factors associated with an improved AUC. Conclusion: To our knowledge, this is the first study to use ROC analysis in a large patient cohort. An IOPTH drop of 72% was found to have optimal discriminating ability. We failed to identify a subset of patients for whom there was substantial improvement in the AUC, sensitivity, or specificity. Abbreviations: AUC = area under the ROC curve; BE = bilateral neck exploration; FE = focal parathyroid exploration; IOPTH = intraoperative parathyroid hormone; MGD = multigland disease; MIBI = Tc99m-sestamibi I-123 subtraction single-photon emission computed tomography/computed tomography; pHPT = primary hyperparathyroidism; ROC = receiver operating characteristic; SGD = single gland disease; US = surgeon-performed neck ultrasound
20
Hakim, Arif Rahman, Windu Gata, Alda Zevana Putri Widodo, Oky Kurniawan e Arief Rama Syarif. "Analisis Perbandingan Algoritma Machine Learning Terhadap Sentimen Analis Pemindahan Ibu Kota Negara". Jurnal JTIK (Jurnal Teknologi Informasi dan Komunikasi) 7, n. 2 (1 aprile 2023): 179–85. http://dx.doi.org/10.35870/jtik.v7i2.701.
Abstract (sommario):
Bangsa indonesia diramaikan dengan berita tentang pemindahan Ibu Kota Negara (IKN). Rencana pemerintah memindahkan IKN beralaskan pada visi misi Indonesia tahun 2045 yaitu Indonesia maju. Twitter menjadi salah satu alat komunikasi microblogging yang digunakan untuk menyampaikan opini. Berbagai algoritma telah digunakan untuk menganalisa sentimen terhadap suatu opini seperti Support Vector Machine, Naive Bayes dan Random Forest. Penelitian ini bertujuan untuk membandingkan kinerja tiga algoritma klasifikasi yaitu Support Vector Machine, Naïve Bayes dan Random Fores. Hasil akurasi tertinggi menggunakan algoritma Support Vector Machine dan di tambah feature Synthetic Minority Oversampling Technique Method (SMOTE) sebesar 82.82%, Presisi 79.34%, Recall 88.75%, 87,78% dan ROC AUC 82.82%. Akurasi Naive bayes sebesar 81.18%, Presisi 84.89%, Recall 75.86%, 80.13% dan ROC AUC 81.18%. dan akurasi Random Forest sebebar 79.55 Presisi 84.48%, Recall 72.39%, 77.97% dan ROC AUC 77.55%.
21
Castagno, Simone, Mark Birch, Mihaela van der Schaar e Andrew McCaskie. "A PRECISION HEALTH APPROACH FOR OSTEOARTHRITIS: PREDICTION OF RAPID KNEE OSTEOARTHRITIS PROGRESSION USING AUTOMATED MACHINE LEARNING". Orthopaedic Proceedings 105-B, SUPP_16 (17 novembre 2023): 23. http://dx.doi.org/10.1302/1358-992x.2023.16.023.
Abstract (sommario):
AbstractIntroductionPrecision health aims to develop personalised and proactive strategies for predicting, preventing, and treating complex diseases such as osteoarthritis (OA), a degenerative joint disease affecting over 300 million people worldwide. Due to OA heterogeneity, which makes developing effective treatments challenging, identifying patients at risk for accelerated disease progression is essential for efficient clinical trial design and new treatment target discovery and development.ObjectivesThis study aims to create a trustworthy and interpretable precision health tool that predicts rapid knee OA progression based on baseline patient characteristics using an advanced automated machine learning (autoML) framework, “Autoprognosis 2.0”.MethodsAll available 2-year follow-up periods of 600 patients from the FNIH OA Biomarker Consortium were analysed using “Autoprognosis 2.0” in two separate approaches, with distinct definitions of clinical outcomes: multi-class predictions (categorising patients into non-progressors, pain-only progressors, radiographic-only progressors, and both pain and radiographic progressors) and binary predictions (categorising patients into non-progressors and progressors). Models were developed using a training set of 1352 instances and all available variables (including clinical, X-ray, MRI, and biochemical features), and validated through both stratified 10-fold cross-validation and hold-out validation on a testing set of 339 instances. Model performance was assessed using multiple evaluation metrics, such as AUC-ROC, AUC-PRC, F1-score, precision, and recall. Additionally, interpretability analyses were carried out to identify important predictors of rapid disease progression.ResultsOur final models yielded high accuracy scores for both multi-class predictions (AUC-ROC: 0.858, 95% CI: 0.856–0.860; AUC-PRC: 0.675, 95% CI: 0.671–0.679; F1-score: 0.560, 95% CI: 0.554–0.566) and binary predictions (AUC-ROC: 0.717, 95% CI: 0.712–0.722; AUC-PRC: 0.620, 95% CI: 0.616–0.624; F1-score: 0.676, 95% CI: 0.673–0679). Important predictors of rapid disease progression included the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) scores and MRI features. Our models were further successfully validated using a hold-out dataset, which was previously omitted from model development and training (AUC-ROC: 0.877 for multi-class predictions; AUC-ROC: 0.746 for binary predictions). Additionally, accurate ML models were developed for predicting OA progression in a subgroup of patients aged 65 or younger (AUC-ROC: 0.862, 95% CI: 0.861–0.863 for multi-class predictions; AUC-ROC: 0.736, 95% CI: 0.734–0.738 for binary predictions).ConclusionsThis study presents a reliable and interpretable precision health tool for predicting rapid knee OA progression using “Autoprognosis 2.0”. Our models provide accurate predictions and offer insights into important predictors of rapid disease progression. Furthermore, the transparency and interpretability of our methods may facilitate their acceptance by clinicians and patients, enabling effective utilisation in clinical practice. Future work should focus on refining these models by increasing the sample size, integrating additional features, and using independent datasets for external validation.Declaration of Interest(b) declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported:I declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research project.
22
Owen, Julian A., Matthew B. Fortes, Saeed Ur Rahman, Mahdi Jibani, Neil P. Walsh e Samuel J. Oliver. "Hydration Marker Diagnostic Accuracy to Identify Mild Intracellular and Extracellular Dehydration". International Journal of Sport Nutrition and Exercise Metabolism 29, n. 6 (1 novembre 2019): 604–11. http://dx.doi.org/10.1123/ijsnem.2019-0022.
Abstract (sommario):
Identifying mild dehydration (≤2% of body mass) is important to prevent the negative effects of more severe dehydration on human health and performance. It is unknown whether a single hydration marker can identify both mild intracellular dehydration (ID) and extracellular dehydration (ED) with adequate diagnostic accuracy (≥0.7 receiver-operating characteristic–area under the curve [ROC-AUC]). Thus, in 15 young healthy men, the authors determined the diagnostic accuracy of 15 hydration markers after three randomized 48-hr trials; euhydration (water 36 ml·kg−1·day−1), ID caused by exercise and 48 hr of fluid restriction (water 2 ml·kg−1·day−1), and ED caused by a 4-hr diuretic-induced diuresis begun at 44 hr (Furosemide 0.65 mg/kg). Body mass was maintained on euhydration, and dehydration was mild on ID and ED (1.9% [0.5%] and 2.0% [0.3%] of body mass, respectively). Urine color, urine specific gravity, plasma osmolality, saliva flow rate, saliva osmolality, heart rate variability, and dry mouth identified ID (ROC-AUC; range 0.70–0.99), and postural heart rate change identified ED (ROC-AUC 0.82). Thirst 0–9 scale (ROC-AUC 0.97 and 0.78 for ID and ED) and urine osmolality (ROC-AUC 0.99 and 0.81 for ID and ED) identified both dehydration types. However, only the thirst 0–9 scale had a common dehydration threshold (≥4; sensitivity and specificity of 100%; 87% and 71%, 87% for ID and ED). In conclusion, using a common dehydration threshold ≥4, the thirst 0–9 scale identified mild intracellular and ED with adequate diagnostic accuracy. In young healthy adults’, thirst 0–9 scale is a valid and practical dehydration screening tool.
23
Muhammad Mufti Sofyanoor, Yunita Widyastuti, Juni Kurniawaty e Djayanti Sari. "Validity of Acute Physiology and Chronic Health Evaluation (APACHE) IV for the Prediction of Prolonged Intensive Care Unit (ICU) Length of Stay in Dr. Sardjito General Hospital in the COVID Era". Journal of Anesthesiology and Clinical Research 4, n. 2 (19 maggio 2023): 426–33. http://dx.doi.org/10.37275/jacr.v4i2.302.
Abstract (sommario):
Introduction: APACHE IV was a good predictor of ICU length of stay in the USA and some countries outside the USA but poor in others. It is important to develop a scoring system for the Indonesian population, especially in this scope, Dr. Sardjito General Hospital. To develop such a scoring system, it is reasonable to study the validity of APACHE IV in ICU Dr. Sardjito General Hospital for predicting prolonged length of stay.
Methods: A retrospective cohort observational study using data from January 1st, 2020, to December 31st, 2020, taken from the ICU of Dr. Sardjito General Hospital. The data are the patient's observed ICU LOS and data required in calculating APACHE IV score and ICU LOS prediction. Discrimination is calculated using the area under (AUC) the receiver operating characteristic curve (ROC) and calibration by the Hosmer-Lemeshow test.
Results: Samples were 329 patients. APACHE IV ICU length of stay prediction showed moderate discriminatory ability (AUC-ROC: 0.74) and poor calibration (p <0.001) to predict prolonged ICU stay. The APACHE IV score has a strong discriminatory ability (AUC-ROC: 0.83). Using the DeLong method, the AUC from ROC APACHE IV score was greater than the AUC from ROC predicted length of stay in APACHE IV ICU (p <0.001). APACHE IV predicted ICU length of stay overestimated observed ICU length of stay.
Conclusion: APACHE IV ICU length of stay prediction has moderate discrimination and poor calibration to predict prolonged ICU stay. The APACHE IV score has better discrimination than the APACHE IV ICU length of stay prediction in predicting prolonged ICU stay.
24
Mohammadi, Mehrdad, Barbara L. McFarlin, Michelle Villegas-Downs, Aiguo Han, Douglas G. Simpson e William D. O'Brien. "Quantitative ultrasound for preterm birth risk prediction—Part 1: Statistical evaluation". Journal of the Acoustical Society of America 153, n. 3_supplement (1 marzo 2023): A351. http://dx.doi.org/10.1121/10.0019123.
Abstract (sommario):
Hypothesis: Predicting the spontaneous preterm birth (sPTB) risk level is enhanced when using both historical clinical (HC) data and quantitative ultrasound (QUS) data compared to using only HC data. HC data defined herein includebirth history prior to that of the current pregnancy as well as, from the current pregnancy, a clinical cervical length assessment, and physical examination data. Study population included 248 full-term births (FTBs) and 26 sPTBs. Advanced statistical analyses were performed for supervised classification containing 53 scaled candidate features (48 QUS, 5 HC) using nested fivefold cross-validation of L1-penalized linear logistic regression with 1000 repetitions to identify potential predictors. Statistical models for HC data alone and HC + QUS data were compared with likelihood-ratio test, cross-validated receiver operating characteristic (ROC) area under the curve (AUC), sensitivity, and specificity. To assess performance, the ROC-AUC was estimated with 10-fold cross-validation logistic regression and 1000 repetitions. Averaged ROC curves plus AUCs were computed using threshold averaging. AUC confidence intervals and test statistics to test the two ROC curves’ differences were constructed via DeLong method. Combined HC and QUS data identified women at sPTB risk with better AUC (0.68; 95% CI, 0.57–0.78) than those of HC data alone (0.53; 95% CI, 0.40–0.66). [Work supported by NIHR01HD089935.]
25
Berrar, D. "An Empirical Evaluation of Ranking Measures With Respect to Robustness to Noise". Journal of Artificial Intelligence Research 49 (17 febbraio 2014): 241–67. http://dx.doi.org/10.1613/jair.4136.
Abstract (sommario):
Ranking measures play an important role in model evaluation and selection. Using both synthetic and real-world data sets, we investigate how different types and levels of noise affect the area under the ROC curve (AUC), the area under the ROC convex hull, the scored AUC, the Kolmogorov-Smirnov statistic, and the H-measure. In our experiments, the AUC was, overall, the most robust among these measures, thereby reinvigorating it as a reliable metric despite its well-known deficiencies. This paper also introduces a novel ranking measure, which is remarkably robust to noise yet conceptually simple.
26
Narasimhan, Harikrishna, e Shivani Agarwal. "Support Vector Algorithms for Optimizing the Partial Area under the ROC Curve". Neural Computation 29, n. 7 (luglio 2017): 1919–63. http://dx.doi.org/10.1162/neco_a_00972.
Abstract (sommario):
The area under the ROC curve (AUC) is a widely used performance measure in machine learning. Increasingly, however, in several applications, ranging from ranking to biometric screening to medicine, performance is measured not in terms of the full area under the ROC curve but in terms of the partial area under the ROC curve between two false-positive rates. In this letter, we develop support vector algorithms for directly optimizing the partial AUC between any two false-positive rates. Our methods are based on minimizing a suitable proxy or surrogate objective for the partial AUC error. In the case of the full AUC, one can readily construct and optimize convex surrogates by expressing the performance measure as a summation of pairwise terms. The partial AUC, on the other hand, does not admit such a simple decomposable structure, making it more challenging to design and optimize (tight) convex surrogates for this measure. Our approach builds on the structural SVM framework of Joachims ( 2005 ) to design convex surrogates for partial AUC and solves the resulting optimization problem using a cutting plane solver. Unlike the full AUC, where the combinatorial optimization needed in each iteration of the cutting plane solver can be decomposed and solved efficiently, the corresponding problem for the partial AUC is harder to decompose. One of our main contributions is a polynomial time algorithm for solving the combinatorial optimization problem associated with partial AUC. We also develop an approach for optimizing a tighter nonconvex hinge loss–based surrogate for the partial AUC using difference-of-convex programming. Our experiments on a variety of real-world and benchmark tasks confirm the efficacy of the proposed methods.
27
Luque-Fernandez, Miguel Angel, Daniel Redondo-Sánchez e Camille Maringe. "cvauroc: Command to compute cross-validated area under the curve for ROC analysis after predictive modeling for binary outcomes". Stata Journal: Promoting communications on statistics and Stata 19, n. 3 (settembre 2019): 615–25. http://dx.doi.org/10.1177/1536867x19874237.
Abstract (sommario):
Receiver operating characteristic (ROC) analysis is used for comparing predictive models in both model selection and model evaluation. ROC analysis is often applied in clinical medicine and social science to assess the tradeoff between model sensitivity and specificity. After fitting a binary logistic or probit regression model with a set of independent variables, the predictive performance of this set of variables can be assessed by the area under the curve (AUC) from an ROC curve. An important aspect of predictive modeling (regardless of model type) is the ability of a model to generalize to new cases. Evaluating the predictive performance (AUC) of a set of independent variables using all cases from the original analysis sample often results in an overly optimistic estimate of predictive performance. One can use K-fold cross-validation to generate a more realistic estimate of predictive performance in situations with a small number of observations. AUC is estimated iteratively for k samples (the “test” samples) that are independent of the sample used to predict the dependent variable (the “training” sample). cvauroc implements k-fold cross-validation for the AUC for a binary outcome after fitting a logit or probit regression model, averaging the AUCs corresponding to each fold, and bootstrapping the cross-validated AUC to obtain statistical inference and 95% confidence intervals. Furthermore, cvauroc optionally provides the cross-validated fitted probabilities for the dependent variable or outcome, contained in a new variable named _fit; the sensitivity and specificity for each of the levels of the predicted outcome, contained in two new variables named _sen and _spe; and the plot of the mean cross-validated AUC and k-fold ROC curves.
28
Arcaro, Marina, Chiara Fenoglio, Maria Serpente, Andrea Arighi, Giorgio G. Fumagalli, Luca Sacchi, Stefano Floro et al. "A Novel Automated Chemiluminescence Method for Detecting Cerebrospinal Fluid Amyloid-Beta 1-42 and 1-40, Total Tau and Phosphorylated-Tau: Implications for Improving Diagnostic Performance in Alzheimer’s Disease". Biomedicines 10, n. 10 (21 ottobre 2022): 2667. http://dx.doi.org/10.3390/biomedicines10102667.
Abstract (sommario):
Recently, a fully automated instrument for the detection of the Cerebrospinal Fluid (CSF) biomarker for Alzheimer’s disease (AD) (low concentration of Amyloid-beta 42 (Aβ42), high concentration of total tau (T-tau) and Phosphorylated-tau (P-tau181)), has been implemented, namely CLEIA. We conducted a comparative analysis between ELISA and CLEIA methods in order to evaluate the analytical precision and the diagnostic performance of the novel CLEIA system on 111 CSF samples. Results confirmed a robust correlation between ELISA and CLEIA methods, with an improvement of the accuracy with the new CLEIA methodology in the detection of the single biomarkers and in their ratio values. For Aβ42 regression analysis with Passing–Bablok showed a Pearson correlation coefficient r = 0.867 (0.8120; 0.907% 95% CI p < 0.0001), T-tau analysis: r = 0.968 (0.954; 0.978% 95% CI p < 0.0001) and P-tau181: r = 0.946 (0.922; 0.962 5% 95% CI p < 0.0001). The overall ROC AUC comparison between ROC in ELISA and ROC in CLEIA confirmed a more accurate ROC AUC with the new automatic method: T-tau AUC ELISA = 0.94 (95% CI 0.89; 0.99 p < 0.0001) vs. AUC CLEIA = 0.95 (95% CI 0.89; 1.00 p < 0.0001), and P-tau181 AUC ELISA = 0.91 (95% CI 0.85; 0.98 p < 0.0001) vs. AUC CLEIA = 0.98 (95% CI 0.95; 1.00 p < 0.0001). The performance of the new CLEIA method in automation is comparable and, for tau and P-tau181, even better, as compared with standard ELISA. Hopefully, in the future, automation could be useful in clinical diagnosis and also in the context of clinical studies.
29
Zhou, Allen S., Anthony A. Prince, Alice Z. Maxfield e Jennifer J. Shin. "Psychological Status as an Effect Modifier of the Association Between Sinonasal Instrument and Imaging Results". Otolaryngology–Head and Neck Surgery 163, n. 5 (26 maggio 2020): 1044–54. http://dx.doi.org/10.1177/0194599820926129.
Abstract (sommario):
Objective: To determine whether psychological status is an effect modifier of the previously observed low discriminatory capacity of Sinonasal Outcome Test-22 (SNOT-22) scores for Lund-Mackay computed tomography (CT) results. Study Design: Observational outcomes study. Setting: Tertiary care center. Subjects and Methods: We assessed patients presenting with chronic sinonasal complaints who underwent CT of the sinuses within 1 month of completing the SNOT-22 instrument. SNOT-22 overall and domain scores were calculated, as were Lund-Mackay CT scores. The discriminatory capacity of SNOT-22 scores for CT results was determined using the receiver-operator characteristic area under the curve (ROC-AUC). Patient-Reported Outcome Measurement Information System (PROMIS) mental health T-scores were assessed, and stratified analyses were used to test for effect modification by psychological status. Results: In stratified analyses, patients with better PROMIS mental health scores had SNOT-22 overall (ROC-AUC 0.96) and nasal domain scores (ROC-AUC 0.97-0.98) that were highly discriminatory for Lund-Mackay scores, while those with worse mental health scores did not (ROC-AUC 0.42-0.55, P < .007). Patients with better SNOT-22 psychological domain scores also had nasal scores that discriminated among CT results significantly better than those with worse psychological domain scores (ROC-AUC 0.65-0.69 and 0.34-0.35, respectively, P < .013). Conclusions: Psychological status is an effect modifier of the relationship between SNOT-22 and Lund-Mackay scores. SNOT-22 scores were discriminatory for Lund-Mackay CT results in patients with better psychological status, while they were nondiscriminatory in those with worse psychological status. When assessing the relationship between subjective and objective measures of chronic rhinosinusitis, accounting for effect modification may have practical utility.
30
Yang, Zixuan. "Prediction of Autism Spectrum Disorder: Comparison and Tuning of Machine Learning Models". Lecture Notes in Education Psychology and Public Media 35, n. 1 (3 gennaio 2024): 1–6. http://dx.doi.org/10.54254/2753-7048/35/20232015.
Abstract (sommario):
The early diagnosis in Autism Spectrum Disorder (ASD) is crucial for timely interventions to address the patients attentional and social challenges. The currently study aims to use machine learning algorithms to accurately predict ASD outcomes. Dataset from a Kaggle competition was used to perform the prediction analysis. Five supervised machine learning algorithms were employed: Logistic Regression (LR), K-Nearest Neighbors (KNN), Support Vector Machine Classifier (SVC), Random Forest (RF), and Decision Trees (DT). The models were fine-tuned using a range of possible hyperparameters and evaluated using ROC AUC scores. The best-performing model, Random Forest, achieved a training ROC AUC of 0.93. The model's performance in predicting the unseen test set resulted in a ROC AUC score of 0.8623. The outcome demonstrates the potentials of machine learning models in early prediction of ASD symptoms, which provides support for autistic individuals to enhance their quality of life and education.
31
Shaikh, Yasmeen, Vasudev Parvati e Sangappa Ramachandra Biradar. "Early disease prediction algorithm for hypertension-based diseases using data aware algorithms". Indonesian Journal of Electrical Engineering and Computer Science 27, n. 2 (1 agosto 2022): 1100. http://dx.doi.org/10.11591/ijeecs.v27.i2.pp1100-1108.
Abstract (sommario):
This paper <span lang="EN-US">implements a data aware early prediction of hypertension-based diseases. Automated data preprocessing method that adopts for both balanced and unbalanced data is the data aware method included in the disease classification algorithm. Proposed data aware data preprocessing method is evaluated on the ensemble learning based classification algorithm for early disease prediction. Data aware preprocessing method adopts isolation forest algorithm for outlier detection as part of the automation. Automated sampling method of applying the sampling corresponding to either balanced or unbalanced data is adopted. Performance evaluation of the proposed data aware algorithm using isolation forest algorithm for anomaly detection is experimented. Python based implementation of the proposed data aware classification algorithm inferred a better area under the curve (AUC) receiver operating characteristics (ROC) curve for isolation forest implementation in data preprocessing automation thus developed. While the individual classifiers multilayer perceptron classifier approached till 0.918 (AUC) in the ROC-AUC curve. The ensemble learning algorithm that included multilayer perceptron classifier, logistic regression classifier, support vector classifier and decision tree algorithm with the isolation forest-based anomaly detection algorithm performed better than the individual machine learning algorithm with 0.922 (AUC) in the ROC-AUC curve.</span>
32
Nahm, Francis Sahngun. "Receiver operating characteristic curve: overview and practical use for clinicians". Korean Journal of Anesthesiology 75, n. 1 (1 febbraio 2022): 25–36. http://dx.doi.org/10.4097/kja.21209.
Abstract (sommario):
Using diagnostic testing to determine the presence or absence of a disease is essential in clinical practice. In many cases, test results are obtained as continuous values and require a process of conversion and interpretation and into a dichotomous form to determine the presence of a disease. The primary method used for this process is the receiver operating characteristic (ROC) curve. The ROC curve is used to assess the overall diagnostic performance of a test and to compare the performance of two or more diagnostic tests. It is also used to select an optimal cut-off value for determining the presence or absence of a disease. Although clinicians who do not have expertise in statistics do not need to understand both the complex mathematical equation and the analytic process of ROC curves, understanding the core concepts of the ROC curve analysis is a prerequisite for the proper use and interpretation of the ROC curve. This review describes the basic concepts for the correct use and interpretation of the ROC curve, including parametric/nonparametric ROC curves, the meaning of the area under the ROC curve (AUC), the partial AUC, methods for selecting the best cut-off value, and the statistical software to use for ROC curve analyses.
33
Marzban, Caren. "The ROC Curve and the Area under It as Performance Measures". Weather and Forecasting 19, n. 6 (1 dicembre 2004): 1106–14. http://dx.doi.org/10.1175/825.1.
Abstract (sommario):
Abstract The receiver operating characteristic (ROC) curve is a two-dimensional measure of classification performance. The area under the ROC curve (AUC) is a scalar measure gauging one facet of performance. In this short article, five idealized models are utilized to relate the shape of the ROC curve, and the area under it, to features of the underlying distribution of forecasts. This allows for an interpretation of the former in terms of the latter. The analysis is pedagogical in that many of the findings are already known in more general (and more realistic) settings; however, the simplicity of the models considered here allows for a clear exposition of the relation. For example, although in general there are many reasons for an asymmetric ROC curve, the models considered here clearly illustrate that an asymmetry in the ROC curve can be attributed to unequal widths of the distributions. Furthermore, it is shown that AUC discriminates well between “good” and “bad” models, but not between good models.
34
Yerlikaya, Gülen, Veronica Falcone, Tina Stopp, Martina Mittlböck, Andrea Tura, Peter Husslein, Wolfgang Eppel e Christian S. Göbl. "To Predict the Requirement of Pharmacotherapy by OGTT Glucose Levels in Women with GDM Classified by the IADPSG Criteria". Journal of Diabetes Research 2018 (2018): 1–6. http://dx.doi.org/10.1155/2018/3243754.
Abstract (sommario):
The aim of this study was to assess the association between OGTT glucose levels and requirement of pharmacotherapy in GDM patients classified by the IADPSG criteria. This study included 203 GDM patients (108 managed with lifestyle modification and 95 requiring pharmacotherapy). Clinical risk factors and OGTT glucose concentrations at 0 (G0), 60 (G60), and 120 min (G120) were collected. OGTT glucose levels were significantly associated with the later requirement of pharmacotherapy (ROC-AUC: 71.1, 95% CI: 63.8–78.3). Also, the combination of clinical risk factors (age, BMI, parity, and pharmacotherapy in previous gestation) showed an acceptable predictive accuracy (ROC-AUC: 72.1, 95% CI: 65.0–79.2), which was further improved when glycemic parameters were added (ROC-AUC: 77.5, 95% CI: 71.5–83.9). Random forest analysis revealed the highest variable importance for G0, G60, and age. OGTT glucose measures in addition to clinical risk factors showed promising properties for risk stratification in GDM patients classified by the recently established IADPSG criteria.
35
Wei, Chih Chiang. "Receiver Operating Characteristic for Diagnosis of Wine Quality by Bayesian Network Classifiers". Advanced Materials Research 591-593 (novembre 2012): 1168–73. http://dx.doi.org/10.4028/www.scientific.net/amr.591-593.1168.
Abstract (sommario):
This paper is dedicated to demonstrate the use of the receiver operating characteristic (ROC) and the area under the ROC curve (AUC) for diagnosing forecast skill. Several local search heuristic algorithms to discover which one performs better for learning a certain Bayesian networks (BN). Five heuristic search algorithms, including K2, Hill Climbing, Repeated Hill Climber, LAGD Hill Climbing, and TAN, were empirically evaluated and compared. This study tests BN models in a real-world case, the Vinho Verde wine taste preferences. An average AUC of 0.746 and 0.727 respectively in red wine and white wine were obtained by TAN algorithm. The results show that the use of TAN can effectively improve the AUC measures for predicting quality grade.
36
Yang, Jingyan, Christine L. Sardo Molmenti, Joaquin Cagliani, Harish Datta, Elliot Grodstein, Rehana Rasul, Horacio Rilo, Lewis W. Teperman e Ernesto P. Molmenti. "Time-Effect of Donor and Recipient Characteristics on Graft Survival after Kidney Transplantation". International Journal of Angiology 28, n. 04 (1 novembre 2019): 249–54. http://dx.doi.org/10.1055/s-0039-1700500.
Abstract (sommario):
AbstractThe kidney allocation system (KAS) is based on quality-based “longevity matching” strategies that provide only a momentary snapshot of expected outcomes at the time of transplantation. The purpose of our study was to define on a continuous timeline the relative and mutual interactions of donor and recipient characteristics on graft survival.Total 39,108 subjects who underwent kidney transplant between October 25, 1999 and January 1, 2007 were identified in the United Network for Organ Sharing dataset. Our primary outcome was graft survival. Time-dependent receiver operating characteristic (ROC) curves and area under time-dependent ROC curve (AUC) were used to compare the predictive ability of the two allocation systems.During the first year after transplantation, both donor and recipient models showed identical relevance. From the first to the sixth years, although the two ROC curves were nearly identical, the donor model outweighed the recipient model. Both models intersected again at the sixth year. From that time onward, the ROC curve for recipient characteristics model predominated over the ROC curve for donor characteristics model. The predictive value of the recipient model (AUC = 0.752) was greater than that of the donor model (AUC = 0.673)We hope that this model will provide additional guidance and risk stratification to further optimize organ allocation based on the dynamic interaction of both donor and recipient characteristics over time.
37
Fonseca, Isabel, José Carlos Oliveira, Manuela Almeida, Madalena Cruz, Anabela Malho, La Salete Martins, Leonídio Dias et al. "Neutrophil Gelatinase-Associated Lipocalin in Kidney Transplantation Is an Early Marker of Graft Dysfunction and Is Associated with One-Year Renal Function". Journal of Transplantation 2013 (2013): 1–12. http://dx.doi.org/10.1155/2013/650123.
Abstract (sommario):
Urinary neutrophil gelatinase-associated lipocalin (uNGAL) has been suggested as potential early marker of delayed graft function (DGF) following kidney transplantation (KTx). We conducted a prospective study in 40 consecutive KTx recipients to evaluate serial changes of uNGAL within the first week after KTx and assess its performance in predicting DGF (dialysis requirement during initial posttransplant week) and graft function throughout first year. Urine samples were collected on post-KTx days 0, 1, 2, 4, and 7. Linear mixed and multivariable regression models, receiver-operating characteristic (ROC), and areas under ROC curves were used. At all-time points, mean uNGAL levels were significantly higher in patients developing DGF (n=18). Shortly after KTx (3–6 h), uNGAL values were higher in DGF recipients (on average +242 ng/mL, considering mean dialysis time of 4.1 years) and rose further in following days, contrasting with prompt function recipients. Day-1 uNGAL levels accurately predicted DGF (AUC-ROC = 0.93), with a performance higher than serum creatinine (AUC-ROC = 0.76), and similar to cystatin C (AUC-ROC = 0.95). Multivariable analyses revealed that uNGAL levels at days 4 and 7 were strongly associated with one-year serum creatinine. Urinary NGAL is an early marker of graft injury and is independently associated with dialysis requirement within one week after KTx and one-year graft function.
38
Gray, Nicola, Nathan Lawler, Annie Zeng, Monique Ryan, Sze Bong, Berin Boughton, Maider Bizkarguenaga et al. "Diagnostic Potential of the Plasma Lipidome in Infectious Disease: Application to Acute SARS-CoV-2 Infection". Metabolites 11, n. 7 (20 luglio 2021): 467. http://dx.doi.org/10.3390/metabo11070467.
Abstract (sommario):
Improved methods are required for investigating the systemic metabolic effects of SARS-CoV-2 infection and patient stratification for precision treatment. We aimed to develop an effective method using lipid profiles for discriminating between SARS-CoV-2 infection, healthy controls, and non-SARS-CoV-2 respiratory infections. Targeted liquid chromatography–mass spectrometry lipid profiling was performed on discovery (20 SARS-CoV-2-positive; 37 healthy controls; 22 COVID-19 symptoms but SARS-CoV-2negative) and validation (312 SARS-CoV-2-positive; 100 healthy controls) cohorts. Orthogonal projection to latent structure-discriminant analysis (OPLS-DA) and Kruskal–Wallis tests were applied to establish discriminant lipids, significance, and effect size, followed by logistic regression to evaluate classification performance. OPLS-DA reported separation of SARS-CoV-2 infection from healthy controls in the discovery cohort, with an area under the curve (AUC) of 1.000. A refined panel of discriminant features consisted of six lipids from different subclasses (PE, PC, LPC, HCER, CER, and DCER). Logistic regression in the discovery cohort returned a training ROC AUC of 1.000 (sensitivity = 1.000, specificity = 1.000) and a test ROC AUC of 1.000. The validation cohort produced a training ROC AUC of 0.977 (sensitivity = 0.855, specificity = 0.948) and a test ROC AUC of 0.978 (sensitivity = 0.948, specificity = 0.922). The lipid panel was also able to differentiate SARS-CoV-2-positive individuals from SARS-CoV-2-negative individuals with COVID-19-like symptoms (specificity = 0.818). Lipid profiling and multivariate modelling revealed a signature offering mechanistic insights into SARS-CoV-2, with strong predictive power, and the potential to facilitate effective diagnosis and clinical management.
39
Liu, Jiazhou, Shihang Pan, Liang Dong, Guangyu Wu, Jiayi Wang, Yan Wang, Hongyang Qian et al. "The Diagnostic Value of PI-RADS v2.1 in Patients with a History of Transurethral Resection of the Prostate (TURP)". Current Oncology 29, n. 9 (5 settembre 2022): 6373–82. http://dx.doi.org/10.3390/curroncol29090502.
Abstract (sommario):
To explore the diagnostic value of the Prostate Imaging–Reporting and Data System version 2.1 (PI-RADS v2.1) for clinically significant prostate cancer (CSPCa) in patients with a history of transurethral resection of the prostate (TURP), we conducted a retrospective study of 102 patients who underwent systematic prostate biopsies with TURP history. ROC analyses and logistic regression analyses were performed to demonstrate the diagnostic value of PI-RADS v2.1 and other clinical characteristics, including PSA and free/total PSA (F/T PSA). Of 102 patients, 43 were diagnosed with CSPCa. In ROC analysis, PSA, F/T PSA, and PI-RADS v2.1 demonstrated significant diagnostic value in detecting CSPCa in our cohort (AUC 0.710 (95%CI 0.608–0.812), AUC 0.768 (95%CI 0.676–0.860), AUC 0.777 (95%CI 0.688–0.867), respectively). Further, PI-RADS v2.1 scores of the peripheral and transitional zones were analyzed separately. In ROC analysis, PI-RADS v2.1 remained valuable in identifying peripheral-zone CSPCa (AUC 0.780 (95%CI 0.665–0.854; p < 0.001)) while having limited capability in distinguishing transitional zone lesions (AUC 0.533 (95%CI 0.410–0.557; p = 0.594)). PSA and F/T PSA retain significant diagnostic value for CSPCa in patients with TURP history. PI-RADS v2.1 is reliable for detecting peripheral-zone CSPCa but has limited diagnostic value when assessing transitional zone lesions.
40
Ikegami, Takako, Hiroki Nishikawa, Masahiro Goto, Masahiro Matsui, Akira Asai, Kosuke Ushiro, Takeshi Ogura et al. "The Relationship between the SARC-F Score and the Controlling Nutritional Status Score in Gastrointestinal Diseases". Journal of Clinical Medicine 11, n. 3 (24 gennaio 2022): 582. http://dx.doi.org/10.3390/jcm11030582.
Abstract (sommario):
We sought to examine the relationship between the SARC-F score and the Controlling Nutritional Status (CONUT) score in patients with gastrointestinal diseases (GDs, n = 735, median age = 71 years, and 188 advanced cancer cases). The SARC-F score ≥ 4 (highly suspicious of sarcopenia) was found in 93 cases (12.7%). Mild malnutritional condition was seen in 310 cases (42.2%), moderate in 127 (17.3%) and severe in 27 (3.7%). The median SARC-F scores in categories of normal, mild, moderate and severe malnutritional condition were 0, 0, 1 and 1 (overall p < 0.0001). The percentage of SARC-F score ≥ 4 in categories of normal, mild, moderate and severe malnutritional condition were 4.4%, 12.9%, 26.8% and 25.9% (overall p < 0.0001). The SARC-F score was an independent factor for both the CONUT score ≥ 2 (mild, moderate or severe malnutrition) and ≥5 (moderate or severe malnutrition). In the receiver operating characteristic (ROC) curve analysis for the CONUT score ≥ 2, C reactive protein (CRP) had the highest area under the ROC (AUC = 0.70), followed by the SARC-F score (AUC = 0.60). In the ROC analysis for the CONUT score ≥ 5, CRP had the highest AUC (AUC = 0.79), followed by the SARC-F score (AUC = 0.63). In conclusion, the SARC-F score in patients with GDs can reflect malnutritional status.
41
do Nascimento, Carla Ferreira, Hellen Geremias dos Santos, André Filipe de Moraes Batista, Alejandra Andrea Roman Lay, Yeda Aparecida Oliveira Duarte e Alexandre Dias Porto Chiavegatto Filho. "Cause-specific mortality prediction in older residents of São Paulo, Brazil: a machine learning approach". Age and Ageing 50, n. 5 (3 maggio 2021): 1692–98. http://dx.doi.org/10.1093/ageing/afab067.
Abstract (sommario):
Abstract Background Populational ageing has been increasing in a remarkable rate in developing countries. In this scenario, preventive strategies could help to decrease the burden of higher demands for healthcare services. Machine learning algorithms have been increasingly applied for identifying priority candidates for preventive actions, presenting a better predictive performance than traditional parsimonious models. Methods Data were collected from the Health, Well Being and Aging (SABE) Study, a representative sample of older residents of São Paulo, Brazil. Machine learning algorithms were applied to predict death by diseases of respiratory system (DRS), diseases of circulatory system (DCS), neoplasms and other specific causes within 5 years, using socioeconomic, demographic and health features. The algorithms were trained in a random sample of 70% of subjects, and then tested in the other 30% unseen data. Results The outcome with highest predictive performance was death by DRS (AUC−ROC = 0.89), followed by the other specific causes (AUC−ROC = 0.87), DCS (AUC−ROC = 0.67) and neoplasms (AUC−ROC = 0.52). Among only the 25% of individuals with the highest predicted risk of mortality from DRS were included 100% of the actual cases. The machine learning algorithms with the highest predictive performance were light gradient boosted machine and extreme gradient boosting. Conclusion The algorithms had a high predictive performance for DRS, but lower for DCS and neoplasms. Mortality prediction with machine learning can improve clinical decisions especially regarding targeted preventive measures for older individuals.
42
Robles-Arana, Yolanda, e Martín Padilla-Lay. "Análisis psicométrico del Cuestionario de Autorreporte (SRQ) como indicador de depresión y ansiedad en usuarios de establecimientos de salud de Lima". Revista de Neuro-Psiquiatría 86, n. 3 (3 ottobre 2023): 161–70. http://dx.doi.org/10.20453/rnp.v86i3-1.4971.
Abstract (sommario):
Los trastornos mentales muestran alta prevalencia en personas que buscan atención por problemas físicos en establecimientos de salud y que no son apropiadamente identificados y tratados. El Cuestionario de Autorreporte (SRQ) ha sido diseñado para detectar posibles casos y tiene un amplio uso en el campo de la atención primaria. Objetivo: Establecer la capacidad predictiva de ítems del SRQ-27 en la identificación de trastornos depresivos y deansiedad en usuarios de establecimientos de salud de Lima. Método: El estudio utilizó datos de la muestra de 10 885personas del Estudio Epidemiológico de Salud Mental en establecimientos de salud de Lima Metropolitana 2015. Se analizaron las respuestas al SRQ-27 en relación a la presencia de ansiedad y depresión, establecida por la Mini Entrevista Neuropsiquiátrica, utilizando procedimientos estadísticos descriptivos, pruebas de asociación y análisis de la AUC-ROC de cada ítem. Resultados: Dos ítems referidos a “sentirse triste” y “llorar” fueron los de mayor poder predictivo para la presencia de depresión (AUC ROC 0,766 y 0,719) y el ítem referido a “sentirse nervioso o tenso” fue el de mayor poder para ansiedad (AUC ROC 0,695). Otros nueve ítems, comunes para ambos trastornos, mostraron también ser eficaces (AUC ROC >0,60) en su identificación. Conclusiones: Se identificó tres ítems del SRQ-27 con mayor poder predictivo para detectar depresión y ansiedad en general, en tanto que ítems con mayor poder predictivo son comunes para ambas condiciones.
43
de la Rosa, Ezequiel, Diana M. Sima, Jan S. Kirschke, Bjoern Menze e David Robben. "Detecting CTP truncation artifacts in acute stroke imaging from the arterial input and the vascular output functions". PLOS ONE 18, n. 3 (30 marzo 2023): e0283610. http://dx.doi.org/10.1371/journal.pone.0283610.
Abstract (sommario):
Background Current guidelines for CT perfusion (CTP) in acute stroke suggest acquiring scans with a minimal duration of 60-70 s. But even then, CTP analysis can be affected by truncation artifacts. Conversely, shorter acquisitions are still widely used in clinical practice and may, sometimes, be sufficient to reliably estimate lesion volumes. We aim to devise an automatic method that detects scans affected by truncation artifacts. Methods Shorter scan durations are simulated from the ISLES’18 dataset by consecutively removing the last CTP time-point until reaching a 10 s duration. For each truncated series, perfusion lesion volumes are quantified and used to label the series as unreliable if the lesion volumes considerably deviate from the original untruncated ones. Afterwards, nine features from the arterial input function (AIF) and the vascular output function (VOF) are derived and used to fit machine-learning models with the goal of detecting unreliably truncated scans. Methods are compared against a baseline classifier solely based on the scan duration, which is the current clinical standard. The ROC-AUC, precision-recall AUC and the F1-score are measured in a 5-fold cross-validation setting. Results The best performing classifier obtained an ROC-AUC of 0.982, precision-recall AUC of 0.985 and F1-score of 0.938. The most important feature was the AIFcoverage, measured as the time difference between the scan duration and the AIF peak. When using the AIFcoverage to build a single feature classifier, an ROC-AUC of 0.981, precision-recall AUC of 0.984 and F1-score of 0.932 were obtained. In comparison, the baseline classifier obtained an ROC-AUC of 0.954, precision-recall AUC of 0.958 and F1-Score of 0.875. Conclusions Machine learning models fed with AIF and VOF features accurately detected unreliable stroke lesion measurements due to insufficient acquisition duration. The AIFcoverage was the most predictive feature of truncation and identified unreliable short scans almost as good as machine learning. We conclude that AIF/VOF based classifiers are more accurate than the scans’ duration for detecting truncation. These methods could be transferred to perfusion analysis software in order to increase the interpretability of CTP outputs.
44
Emamekhoo, Hamid, Jennifer L. Schehr, Rory M. Bade, Xiao X. Wei, Rana R. McKay, Toni K. Choueiri e Joshua Michael Lang. "Clinical correlation of circulating tumor cell (CTC) PD-L1 and HLA I expression in metastatic renal cell carcinoma (mRCC) using exclusion-based sample preparation technology." Journal of Clinical Oncology 38, n. 6_suppl (20 febbraio 2020): 721. http://dx.doi.org/10.1200/jco.2020.38.6_suppl.721.
Abstract (sommario):
721 Background: Despite therapeutic advancement in Vascular endothelial growth factor receptor (VEGF-R) tyrosine kinase inhibitors (TKIs) and immune checkpoint inhibitors (ICIs) for mRCC treatment, there is currently no reliable predictive biomarkers of response or resistance. Single site biopsies provide limited information given the heterogenous nature of mRCC. Liquid biopsies may overcome these limitations; however, prior CTC capture platforms lacked sufficient sensitivity and specificity to achieve clinically useful detection rates. Methods: Given the prevalence of VEGF dependency, and its reliance on hypoxia, we aimed to increase sensitivity and specificity of capturing and identifying mRCC CTCs using carbonic anhydrase IX (CA IX) and CA XII. In addition, traditional markers for cell capture and identification with epithelial cellular adhesion molecule (EpCAM) and cytokeratin (CK) were included. Exclusion-based Sample Preparation technology was used to maximize cell yield. CD45/34/66b positive blood cells were excluded to ensure high specificity in evaluation of PD-L1 and HLA I expression on CTCs. Results: In a preliminary cohort of 21 mRCC pts (treatment: TKI=12, ICI=5, TKI+ICI=2, baseline=2), we identified heterogeneous populations of CTCs with differential expression of CA XII and CK. We detected CK+ CTCs in 20/21 pts (mean= 5/mL; range 0-53), CAXII+ CTCs in 21/21 pts (mean= 1/mL; range 1-9), and CK+/CAXII+ CTCs in 19/21 pts (mean=7/mL; range 0-102). In pts with multiple CTC samples on treatment, there was a high correlation between the number of CK+ CTCs and treatment response (ROC AUC 0.88). PD-L1 expression in CAXII+ CTCs correlated with response to ICI (ROC AUC 0.77) and TKI (ROC AUC 0.73). HLA I expression in CAXII+ CTCs correlated with response to TKI (ROC AUC 0.73) better than ICI (ROC AUC 0.59). Conclusions: Assessment of CTC heterogeneity may provide valuable molecular insights and diversify tools for early detection of therapeutic response and resistance that may guide treatment decision making. This assay is being tested in ongoing Phase II clinical trials.
45
Guerrero-Bermúdez, Camila Andrea, Simón Villa-Pérez, Ariel Antonio Arteta-Cueto, Juan Camilo Pérez-Cadavid e Fabián Jaimes-Barragán. "Evaluation of the performance of three non-invasive scores for the diagnosis of advanced fibrosis in a population with non-alcoholic fatty liver disease". Hepatología 5, n. 2 (2 maggio 2024): 137–47. http://dx.doi.org/10.59093/27112330.113.
Abstract (sommario):
Introducción. La enfermedad hepática grasa no alcohólica (EHGNA) es la hepatopatía crónica más común en el mundo, y en aproximadamente el 10 % de los casos progresará a cirrosis o a carcinoma hepatocelular. La presencia de fibrosis hepática es el mejor predictor de esta progresión, pero su diagnóstico mediante biopsia hepática es invasivo y con riesgo de complicaciones (alrededor del 2,5 %). Existen puntajes no invasivos que se han desarrollado y validado para estadificar la fibrosis, pero no conocemos su rendimiento en la población colombiana. El objetivo de este estudio fue evaluar el desempeño de los puntajes fibrosis-4 (FIB-4), la relación AST/ALT y el índice AST/plaquetas (APRI) para la detección de fibrosis avanzada en pacientes colombianos con EHGNA. Metodología. Estudio observacional tipo transversal de pacientes con EHGNA, que entre 2008 y 2022 tuvieran disponible el resultado de una biopsia hepática. Se hizo una descripción demográfica básica y se calculó el FIB-4, la relación AST/ALT y el APRI con los laboratorios más recientes previos al procedimiento. Posteriormente se calcularon valores de sensibilidad, especificidad, valores predictivos, razones de verosimilitud y área bajo la curva-característica operativa del receptor (AUC-ROC) para los puntos de corte evaluados previamente en la literatura. Resultados. Se incluyeron 176 pacientes, de los cuales el 14,3 % tenían fibrosis avanzada. El FIB-4 presentó el mejor rendimiento con un valor AUC-ROC de 0,74 para el punto de corte de 1,30 y 2,67. En segundo lugar, estuvo la relación AST/ALT con un valor AUC-ROC de 0,68 con el punto de corte de 0,8, y finalmente el APRI con valor AUC-ROC 0,62 con el punto de corte de 1. Conclusión. En la población analizada los tres puntajes tienen menor rendimiento diagnóstico comparado a los resultados reportados en Europa y Japón. El FIB-4 es el único que alcanza una AUC-ROC con rendimiento razonable, con la limitación que 27,4 % obtuvieron un resultado indeterminado.
46
Vardhan, R. Vishnu, e S. Balaswamy. "Improved Methods for Estimating Areas under the Receiver Operating Characteristic Curves". International Journal of Green Computing 4, n. 2 (luglio 2013): 58–75. http://dx.doi.org/10.4018/jgc.2013070105.
Abstract (sommario):
ROC Curve is the most widely used statistical technique for classifying an individual into one of the two pre-determined groups basing on test result. Area under the curve (AUC) is a measure of accuracy which exhibits the discriminating power of the test with respect to a threshold or cutoff value. In medical diagnosis, this technique has its relevance to study and compare different diagnostic tests. In this paper, a method is proposed to estimate the AUC of Binormal ROC model by taking into account the confidence interval of mean and corresponding variances.
47
Alebić, Miro Šimun, Nataša Stojanović e Marta Žuvić-Butorac. "The IVF Outcome Counseling Based on the Model Combining DHEAS and Age in Patients with Low AMH Prior to the First Cycle of GnRH Antagonist Protocol of Ovarian Stimulation". International Journal of Endocrinology 2013 (2013): 1–7. http://dx.doi.org/10.1155/2013/637919.
Abstract (sommario):
Objective. To investigate the endocrine and/or clinical characteristics of women with low anti-Müllerian hormone (AMH) that could improve the accuracy of IVF outcome prediction based on the female age alone prior to the first GnRH antagonist IVF cycle.Methods. Medical records of 129 patients with low AMH level (<6.5 pmol/L) who underwent their first GnRH antagonist ovarian stimulation protocol for IVF/ICSI were retrospectively analyzed. The main outcome measure was the area under the ROC curve (AUC-ROC) for the models combining age and other potential predictive factors for the clinical pregnancy.Results. Clinical pregnancy rate (CPR) per initiated cycles was 11.6%. For the prediction of clinical pregnancy, DHEAS and age showed AUC-ROC of 0.726 (95%CI 0.641–0.801) and 0.662 (95%CI 0.573–0.743), respectively (). The predictive accuracy of the model combining age and DHEAS (AUC-ROC 0.796; 95%CI 0.716–0.862) was significantly higher compared to that of age alone (). In patients <37.5 years with DHEAS pmol/L, 60% (9/15) of all pregnancies were achieved with CPR of 37.5%.Conclusions. DHEAS appears to be predictive for clinical pregnancy in younger women (<37.5 years) with low AMH after the first GnRH antagonist IVF cycle. Therefore, DHEAS-age model could refine the pretreatment counseling on pregnancy prospects following IVF.
48
Chan, Tommy CY, Yu Meng Wang, Marco Yu e Vishal Jhanji. "Comparison of corneal dynamic parameters and tomographic measurements using Scheimpflug imaging in keratoconus". British Journal of Ophthalmology 102, n. 1 (30 maggio 2017): 42–47. http://dx.doi.org/10.1136/bjophthalmol-2017-310355.
Abstract (sommario):
AimTo compare the diagnostic ability of corneal tomography and corneal dynamic response measurements in normal and keratoconus eyes.MethodsConsecutive patients with grade II–III keratoconus and age-matched normal subjects were recruited. Corneal imaging was performed using Pentacam (Oculus Optikgeräte, Wetzlar, Germany) and Corvis (Oculus Optikgeräte). A beta version of Corvis software was used with three additional parameters: maximal change of arc length, deformation amplitude (DA) ratio 1 and DA ratio 2. Diagnostic ability of both devices to differentiate normal and keratoconus eyes was evaluated using receiver-operating characteristic (ROC) curves. The areas under the ROC curve (AUC) and partial AUC (pAUC) for specificity ≥80% for each parameter of Corvis and final D value of Belin/Ambrosio Enhanced Ectasia Display (BAD) were compared.ResultsForty-two eyes of 42 patients (21 patients with keratoconus and 21 normal subjects) were included. Both groups were age matched (p=0.760). The ROC analysis showed that the final D value of BAD had the highest AUC (0.994) and pAUC (0.194). Maximum inverse radius had the highest AUC (0.954) but a relatively lower pAUC (0.158), while DA ratio 2 had the second highest AUC (0.946) together with the highest pAUC (0.177) among Corvis parameters. There was no significant difference between AUC and pAUC of BAD compared with those of DA ratio 1 (p≥0.162) and DA ratio 2 (p≥0.208).ConclusionsThe results of our study suggest that Corvis measurements have the potential to differentiate keratoconus and normal eyes. The diagnostic ability of novel parameters on Corvis was comparable to Pentacam.
49
Astapovskii, A. A., V. N. Drozdov, E. V. Shikh, N. B. Lazareva e S. Yu Serebrova. "Рrognostic value of proadrenomedullin in patients with COVID-19". Meditsinskiy sovet = Medical Council, n. 14 (12 agosto 2022): 200–205. http://dx.doi.org/10.21518/2079-701x-2022-16-14-200-205.
Abstract (sommario):
Introduction. In the context of a pandemic, it is necessary to quickly and accurately stratify patients into groups based on the severity of their condition and prognostic risk. For these purposes, some available biomarkers, such as proadrenomedullin, can be used. Objective. To determine the prognostic value of regional mean proadrenomedullin (pro-ADM) in comparison with routine clinical and laboratory parameters in patients with a new coronavirus infection COVID-19. Materials and methods. The study included 140 patients who were hospitalized with a diagnosis of community-acquired pneumonia on the background of COVID-19. The level of pro-ADM was determined on the first and third days of hospitalization by ELISA. In accordance with the outcome of the disease, patients were divided into two groups: those discharged with recovery or improvement (n = 110, 78, 6%) and those who died during their stay in the hospital (n = 30, 21, 4%). Results. Pro-ADM had the highest prognostic value as a predictor of adverse outcome on day 1 AUC ROC 0.72 95% CI (0.57–0.84) sensitivity 79.2%, specificity 62.9% and on day 3 AUC ROC 0.98 95% CI (0.86–1.0) sensitivity 100%, specificity 95.6%. ROC analysis results for C-reactive protein AUC ROC 0.55 95% CI (0.41–0.77), sensitivity 73.3%, specificity 45.6%; procalcitonin AUC ROC 0.62 95% CI (0.49–0.73), sensitivity 80%, specificity 48.2%. The relative risk of a poor outcome for a proADM level > 500 pmol/L is 2.3 95% CI (1.23–4.32), and for a proADM level > 700 it is 8.5 95% CI (4.83–14.94) p < 0.001. Conclusions. Compared to C-reactive protein and procalcitonin, regional mean proadrenomedullin has the highest predictive value as a predictor of death in patients with COVID-19.
50
Petraikin, A. V., Zh E. Belaya, A. N. Kiseleva, Z. R. Artyukova, M. G. Belyaev, V. A. Kondratenko, M. E. Pisov et al. "Artificial intelligence for diagnosis of vertebral compression fractures using a morphometric analysis model, based on convolutional neural networks". Problems of Endocrinology 66, n. 5 (25 dicembre 2020): 48–60. http://dx.doi.org/10.14341/probl12605.
Abstract (sommario):
BACKGROUND: Pathological low-energy (LE) vertebral compression fractures (VFs) are common complications of osteoporosis and predictors of subsequent LE fractures. In 84% of cases, VFs are not reported on chest CT (CCT), which calls for the development of an artificial intelligence-based (AI) assistant that would help radiology specialists to improve the diagnosis of osteoporosis complications and prevent new LE fractures.AIMS: To develop an AI model for automated diagnosis of compression fractures of the thoracic spine based on chest CT images.MATERIALS AND METHODS: Between September 2019 and May 2020 the authors performed a retrospective sampling study of ССТ images. The 160 of results were selected and anonymized. The data was labeled by seven readers. Using the morphometric analysis, the investigators received the following metric data: ventral, medial and dorsal dimensions. This was followed by a semiquantitative assessment of VFs degree. The data was used to develop the Comprise-G AI mode based on CNN, which subsequently measured the size of the vertebral bodies and then calculates the compression degree. The model was evaluated with the ROC curve analysis and by calculating sensitivity and specificity values.RESULTS: Formed data consist of 160 patients (a training group - 100 patients; a test group - 60 patients). The total of 2,066 vertebrae was annotated. When detecting Grade 2 and 3 maximum VFs in patients the Comprise-G model demonstrated sensitivity - 90,7%, specificity - 90,7%, AUC ROC - 0.974 on the 5-FOLD cross-validation data of the training dataset; on the test data - sensitivity - 83,2%, specificity - 90,0%, AUC ROC - 0.956; in vertebrae demonstrated sensitivity - 91,5%, specificity - 95,2%, AUC ROC - 0.981 on the cross-validation data; for the test data sensitivity - 79,3%, specificity - 98,7%, AUC ROC - 0.978.CONCLUSIONS: The Comprise-G model demonstrated high diagnostic capabilities in detecting the VFs on CCT images and can be recommended for further validation.