1,802 research outputs found
Integrating radiomic and 3D autoencoder-based features for Non-Small Cell Lung Cancer survival analysis
Get PDFBackground and objectives: The aim of this study is to develop a radiomic and deep learning-based signature for survival analysis of patients with Non-Small Cell Lung Cancer. Methods: Four-hundred twenty-two patients from “Lung1” dataset were included in the study. A 3D convolutional autoencoder (AE) was built and features from the latent space extracted for further analysis. Radiomic features were derived from the 3D volume of the tumor region using PyRadiomics. Both radiomic and AE-based features underwent feature selection, by removing: i) highly correlated and ii) constant features. The selected variables were then used to derive both mono-domain (radiomics, AE and clinic) and multi-domain signatures fitting a Cox Proportional Hazard model with LASSO penalization and evaluated considering the concordance (C)-index as performance metric. Results: Both mono-domain and multi-domain signatures could significantly differentiate high risk from low risk patients. Among the mono-domain signatures, the highest hazard ratio (HR) in the test set was obtained using radiomics (HR = 1.5428) followed by the AE-based signature (HR = 1.5012) and the clinical signature (HR = 1.4770). The best overall performance was achieved by combining all three signatures, resulting in the highest HR (HR = 1.7383), while the combination of AE-based and clinical signatures yielded the highest C-index (C-index = 0.6309). Conclusions: These preliminary results show that combining information carried by AE, radiomic and clinical domain shows potential for improving the prediction of overall survival in NSCLC patients
Assessing the Performance of MRI-Radiomic Prognostic Signatures in Head and Neck Cancer Patients: A Comparative Analysis
Get PDFAn ECG-Based Model for Left Ventricular Hypertrophy Detection: A Machine Learning Approach
Get PDFGoal: Despite the high incidence of left ventricular hypertrophy (LVH), clinical LVH-electrocardiography (ECG) criteria remain unsatisfactory due to low sensitivity. We propose an automatic LVH detection method based on ECG-extracted features and machine learning. Methods: ECG features were automatically extracted from two publicly available databases: PTB-XL with 2181 LVH and 9001 controls, and Georgia with 1012 LVH and 1387 controls. After preprocessing and feature extraction, the most relevant features from PTB-XL were selected to train three models: logistic regression, random forest (RF), and support vector machine (SVM). These classifiers, trained with selected features and a reduced set of five features, were evaluated on the Georgia database and compared with clinical LVH-ECG criteria. Results: RF and SVM models showed accuracies above 90% and increased sensitivity to above 86%, compared to clinical criteria achieving 38% at maximum. Conclusions: Automatic ECG-based LVH detection using machine learning outperforms conventional diagnostic criteria, benefiting clinical practice
Automatic Right Ventricular Hypertrophic Detection Integrating Electrocardiography-based QRS Biomarkers with Machine Learning
Get PDFSpectral Analysis of Blood Pressure Variability in Atrial Fibrillation: the Effect of Tilting
No full textOn the reliability of frequency components in systolic arterial pressure in patients with atrial fibrillation
No full text
Analysis of Surface Atrial Signal Using Spectral Method for Time Series with Missing Data
No full textA Statistical Atrioventricular Node Model Accounting for Pathway Switching During Atrial Fibrillation
No full textObjective: The atrioventricular (AV) node plays a central role in atrial fibrillation (AF) as it influences the conduction of impulses from the atria into the ventricles. In the present paper, the statistical dual pathway AV node model, previously introduced by us, is modified so that it accounts for atrial impulse pathway switching even if the preceding impulse did not cause a ventricular activation. Methods: The proposed change in model structure implies that the number of model parameters subjected to maximum likelihood estimation is reduced from five to four. The model is evaluated using the data acquired in the RATe control in Atrial Fibrillation (RATAF) study, involving 24- h ECG recordings from 60 patients with permanent AF. Results: When fitting the models to the RATAF database, similar results were obtained for both the present and the previous model, with a median fit of 86%. The results show that the parameter estimates characterizing refractory period prolongation exhibit considerably lower variation when using the present model, a finding that may be ascribed to fewer model parameters. Conclusion: The new model maintains the capability to model RR intervals, while providing more reliable parameters estimates. Significance: The model parameters are expected to convey novel clinical information, and may be useful for predicting the effect of rate control drugs
Giambattista Giuliani: dagli aurei trecentisti al vivente linguaggio della Toscana.
Get PDFIn questo volume sono raccolti gli interventi del convegno“Di scritto e di parlato”. Antiche e nuove diamesie organizzato dal Dipartimento di studi letterari filologici e linguistici, dal Dottorato in scienze del patrimonio letterario, artistico e ambientale e dal Calcif, dell’Università degli Studi di Milano e che si è tenuto a Milano il 6 novembre 2015. Il volume è articolato in due parti, corrispondenti alle due sezioni tematiche del convegno: nella prima si raccolgono gli interventi di interesse soprattutto storico-linguistico, mentre nella seconda confluiscono quelli di argomento teorico e glottodidattico e quelli che studiano il rapporto tra scritto e parlato in sincronia. I contributi dei relatori si sono intrecciati in una narrazione che voleva tracciare in prospettiva storica le manifestazioni della voce dalle origini sino all’Ottocento (nei cantari, con Beatrice Barbiellini Amidei; nel Giuliani con Valentina Petrini; nel De Amicis e in altri letterati, linguisti e lessicografi del secondo Ottocento con Matteo Grassano; nella manualistica per le scuole reggimentali con Michela Dota), completandosi con indagini di tipo diverso, sincronico, teorico e glottodidattico (con il tema della simulazione del parlato e dell’enunciazione di Enrico Testa; con l’analisi linguistica dei corpora di apprendenti l’italiano di Elisa Corino e Carla Marello e Franca Bosc; con lo scavo documentario su scritture di semicolti a metà del Novecento di Elisabetta Banfi; con la ricostruzione del pensiero di Spitzer di Diego Stefanelli; con indagini sui media, tradizionali e nuovi, di Ilaria Bonomi ed Elisabetta Mauroni e di Massimo Prada)
- …
