Showing papers by "Gemma Piella published in 2022"

Machine Learning for Clinical Decision-Making: Challenges and Opportunities in Cardiovascular Imaging

Abstract: The use of machine learning (ML) approaches to target clinical problems is called to revolutionize clinical decision-making in cardiology. The success of these tools is dependent on the understanding of the intrinsic processes being used during the conventional pathway by which clinicians make decisions. In a parallelism with this pathway, ML can have an impact at four levels: for data acquisition, predominantly by extracting standardized, high-quality information with the smallest possible learning curve; for feature extraction, by discharging healthcare practitioners from performing tedious measurements on raw data; for interpretation, by digesting complex, heterogeneous data in order to augment the understanding of the patient status; and for decision support, by leveraging the previous steps to predict clinical outcomes, response to treatment or to recommend a specific intervention. This paper discusses the state-of-the-art, as well as the current clinical status and challenges associated with the two later tasks of interpretation and decision support, together with the challenges related to the learning process, the auditability/traceability, the system infrastructure and the integration within clinical processes in cardiovascular imaging.

Regulatory network-based model to simulate the biochemical regulation of chondrocytes in healthy and osteoarthritic environments

Abstract: In osteoarthritis (OA), chondrocyte metabolism dysregulation increases relative catabolic activity, which leads to cartilage degradation. To enable the semiquantitative interpretation of the intricate mechanisms of OA progression, we propose a network-based model at the chondrocyte level that incorporates the complex ways in which inflammatory factors affect structural protein and protease expression and nociceptive signals. Understanding such interactions will leverage the identification of new potential therapeutic targets that could improve current pharmacological treatments. Our computational model arises from a combination of knowledge-based and data-driven approaches that includes in-depth analyses of evidence reported in the specialized literature and targeted network enrichment. We achieved a mechanistic network of molecular interactions that represent both biosynthetic, inflammatory and degradative chondrocyte activity. The network is calibrated against experimental data through a genetic algorithm, and 81% of the responses tested have a normalized root squared error lower than 0.15. The model captures chondrocyte-reported behaviors with 95% accuracy, and it correctly predicts the main outcomes of OA treatment based on blood-derived biologics. The proposed methodology allows us to model an optimal regulatory network that controls chondrocyte metabolism based on measurable soluble molecules. Further research should target the incorporation of mechanical signals.

Prediction of Lung Nodule Progression with an Uncertainty-Aware Hierarchical Probabilistic Network

Abstract: Predicting whether a lung nodule will grow, remain stable or regress over time, especially early in its follow-up, would help doctors prescribe personalized treatments and better surgical planning. However, the multifactorial nature of lung tumour progression hampers the identification of growth patterns. In this work, we propose a deep hierarchical generative and probabilistic network that, given an initial image of the nodule, predicts whether it will grow, quantifies its future size and provides its expected semantic appearance at a future time. Unlike previous solutions, our approach also estimates the uncertainty in the predictions from the intrinsic noise in medical images and the inter-observer variability in the annotations. The evaluation of this method on an independent test set reported a future tumour growth size mean absolute error of 1.74 mm, a nodule segmentation Dice’s coefficient of 78% and a tumour growth accuracy of 84% on predictions made up to 24 months ahead. Due to the lack of similar methods for providing future lung tumour growth predictions, along with their associated uncertainty, we adapted equivalent deterministic and alternative generative networks (i.e., probabilistic U-Net, Bayesian test dropout and Pix2Pix). Our method outperformed all these methods, corroborating the adequacy of our approach.

Machine learning-based phenotyping and risk assessment of hypertrophic cardiomyopathy - linking ECGs, morphology and genotypes

Abstract: Type of funding sources: None. Integrating clinical data to distinguish hypertrophic cardiomyopathy (HCM) phenotypes is relevant in clinical practice. Machine learning (ML) can help - deep learning (DL) networks can automate detection and segmentation of 12-lead electrocardiograms (ECGs), whereas unsupervised learning can group patients to compare ECG, imaging and genetic characteristics. The aim is to automate ECG morphology analysis from all 12 ECG leads and multiple beats, and relate this to HCM genotypes and imaging phenotypes. The single-center cohort included phenotype- and genotype-positive (G+) HCM patients (n = 104) and their phenotype-negative relatives (n = 50, 42% G+). All patients had a digital 12-lead ECG, echocardiography, and a magnetic resonance (CMR) study performed. The workflow is shown in Fig 1. A U-Net DL network was used for ECG delineation (P, QRS, T onsets/offsets) for all cardiac cycles. Three heartbeats were selected for each patient based on their morphology, with the aim of capturing beat-to-beat variability. An unsupervised representation learning algorithm was used to fuse ECG data and assess inter-patient similarities. Patients were clustered based on similarities of ECG biomarkers, and compared with regards to genotypes, family history of sudden cardiac death (SCD), history of ventricular arrhythmias/syncope/aborted SCD, implanted defibrillators (ICD), left ventricular (LV) obstruction, maximal wall thickness, late gadolinium enhancement (LGE), and HCM risk-SCD score. ML based on ECG biomarkers provided a good separation of HCM patients and relatives (Fig 1A), also showing G- and patients with variants of uncertain significance grouping together (Fig 1B). Clustering resulted in 6 ECG phenogroups (C1-6). C1 and 2 were related to less comorbidities, cardiac remodeling, and HCM risk score, capturing the majority of G- patients. C3 and 4 were related to LV obstruction – where C4 consisted of symptomatic patients with high ICD implantation and event rates, high LGE, and impaired systolic function. C5 captured patients with high comorbidities, extremely remodeled hearts, but no obstruction, whereas C6 patients with positive family history and high arrhythmic events (Fig 1C, Table 1). The average ECG morphology is shown side-by-side for C1 and C5 in Fig 1D – negative T waves, increased R/S wave amplitudes, left axis deviation (LAD) and ST elevation can be recognized as macro-biomarkers in C5 (yellow arrows). ML can automate the analysis of complex clinical data, simultaneously taking into account the morphology of all ECG components in all 12-leads, throughout multiple beats, compare it with clinical and imaging data, and identify clinically sensible phenogroups as validated by structural and functional findings, as well as with genotypes and clinical information. Automated and comprehensive cardiac data analysis has diagnostic and research potential to help screen populations and phenotype disease etiologies. Abstract Figure 1: analysis pipeline Abstract Table 1: clinical variables

An automatic pipeline for atlas-based fetal and neonatal brain segmentation and analysis

Abstract: BACKGROUND AND OBJECTIVE The automatic segmentation of perinatal brain structures in magnetic resonance imaging (MRI) is of utmost importance for the study of brain growth and related complications. While different methods exist for adult and pediatric MRI data, there is a lack for automatic tools for the analysis of perinatal imaging. METHODS In this work, a new pipeline for fetal and neonatal segmentation has been developed. We also report the creation of two new fetal atlases, and their use within the pipeline for atlas-based segmentation, based on novel registration methods. The pipeline is also able to extract cortical and pial surfaces and compute features, such as curvature, local gyrification index, sulcal depth, and thickness. RESULTS Results show that the introduction of the new templates together with our segmentation strategy leads to accurate results when compared to expert annotations, as well as better performances when compared to a reference pipeline (developing Human Connectome Project (dHCP)), for both early and late-onset fetal brains. CONCLUSIONS These findings show the potential of the presented atlases and the whole pipeline for application in both fetal, neonatal, and longitudinal studies, which could lead to dramatic improvements in the understanding of perinatal brain development.

Transferring 3D facial expressions from adults to children

Abstract: Diagnosis of craniofacial conditions is shifting towards pre- and peri-natal stages, since early assessment has shown to be crucial for the effective treatment of functional and developmental aspects of children. 3D Morphable Models are a valuable tool for such evaluation. However, limited data availability on 3D newborn geometry, and highly variable imaging environments, challenge the construction of 3D baby face models. Our hypothesis is that constructing a bi-linear baby face model that allows identity and expression decoupling, enables to improve craniofacial and brain function assessments. Thus, given that adult and infants facial expression configurations are very similar and that 3D facial expressions in babies are difficult to be scanned in a controlled manner, we propose transferring the facial expressions from the available FaceWarehouse (FW) database to baby scans, to construct a baby-specific bi-linear expression model. First, we defined a spatial mapping between the BabyFM and the FW. Then, we propose an automatic neutralization to remove the expressions from the facial scans. Finally, we apply expression transfer to obtain a complete data tensor. We test the performance and generalization of the resulting bi-linear model with a test set. Results show that the obtained model allow us to successfully and realistically manipulate facial expressions of babies while keeping them decoupled from identity variations.

Unsupervised Machine Learning Exploration of Morphological and Haemodynamic Indices to Predict Thrombus Formation in the Left Atrial Appendage

Abstract: Atrial Fibrillation (AF) is the most common cardiac arrhythmia, and it is associated with an increased risk of embolic stroke. It is known that AF-related thrombus formation occurs predominantly in the left atrial appendage (LAA). However, it is still unknown the structural and functional characteristics of the left atria (LA) that promote low velocities and stagnated blood flow, thus a high risk of thrombogenesis. In this work, we investigated morphological and in-silico haemodynamic indices of the LA and LAA with unsupervised machine learning (ML) techniques, to identify the most relevant features that could subsequently be used to generate thrombus prediction models. A fully automatic pipeline was implemented to extract multiple morphological parameters from a 3D mesh of a LA. Morphological parameters were then combined with particle flow parameters from in-silico fluid simulations. Unsupervised multiple kernel learning (MKL) was used for dimensionality reduction, resulting in a latent space positioning patients based on feature similarity. Clustering applied to the MKL output space estimated clusters with different proportion of thrombus cases. The cluster with the highest risk of thrombus formation was characterised by high values of LAA height, tortuosity and ostium perimeter, as well as total number of flow particles in the LAA and low angle between the LAA and the left superior pulmonary vein, proving the usefulness of unsupervised ML techniques to extract knowledge from the data, and early identify AF patients at higher risk of thrombus formation.