Michela Ferrari

Michela Ferrari

Auto-generated profile, based on publicly available information

Michela Ferrari

University of Pavia

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 April 2024 in European Heart Journal Supplements
Verified
Journal Article
Open access
PDF file
MGMaurizia Grasso
Maurizia Grasso
DBDavide Bondavalli
Davide Bondavalli
VVViviana Vilardo
Viviana Vilardo
CCClaudia Cavaliere
Claudia Cavaliere
IGIlaria Gatti
Ilaria Gatti
EAEloisa Arbustini
Eloisa Arbustini
In the ESC 2023 guidelines, cardiomyopathies are conservatively defined as ‘myocardial disorders in which the heart muscle is structurally and functionally abnormal, in the absence of coronary artery disease, hypertension, valvular disease, and congenital heart disease sufficient to cause the observed myocardial abnormality’. They are morpho-functionally classified as hypertrophic, dilated, restrictive, and arrhythmogenic right ventricular cardiomyopathy with the addition of the left ventricular non-dilated cardiomyopathy that describes intermediate phenotypes not fulfilling standard disease definitions despite the presence of myocardial disease on cardiac imaging or tissue analysis. The new ESC guidelines provide ‘a guide to the diagnostic approach to cardiomyopathies, highlight general evaluation and management issues, and signpost the reader to the relevant evidence base for the recommendations’. The recommendations and suggestions included in the document provide the tools to build up pathways tailored to specific cardiomyopathy (phenotype and cause) and define therapeutic indications, including target therapies where possible. The impact is on clinical cardiology, where disease-specific care paths can be assisted by the guidelines, and on genetics, both clinics and testing, where deep phenotyping and participated multi-disciplinary evaluation provide a unique tool for validating the pathogenicity of variants. The role of endomyocardial biopsy remains underexploited and confined to particular forms of restrictive cardiomyopathy, myocarditis, and amyloidosis. New research and development will be needed to cover the gaps between science and clinics. Finally, the opening up to disciplines such as bioinformatics, bioengineering, mathematics, and physics will support clinical cardiologists in the best governance of the novel artificial intelligence–assisted resources.

20 November 2025 in Computer Methods and Programs in Biomedicine
Verified
Journal Article
Open access
MFMichela Ferrari
Michela Ferrari
MUMario Urtis
Mario Urtis
ESEdoardo Spairani
Edoardo Spairani
ATAntonio Tescari
Antonio Tescari
FSFrancesca Sessa
Francesca Sessa
GMGiovanni Magenes
Giovanni Magenes
X-ray Coronary Angiography (XCA) enables visualization of coronary arteries for disease and morphology assessment. Accurate segmentation of major coronary vessels is crucial for automated analysis of their geometric features but presents challenges due to anatomical complexity. This study introduces CoroSAM, an adaptation of the LiteMedSAM Foundation Model, employing parameter-efficient fine-tuning for interactive coronary artery segmentation in XCA images. The proposed approach incorporates Convolutional Adapter layers within the image encoder's TinyViT blocks to enhance domain-specific feature extraction while maintaining computational efficiency. A point-based prompting strategy directly encodes vessel endpoints and branch points as additional input channels. Model evaluation employed 5-fold cross-validation on the ARCADE dataset and zero-shot testing on external datasets (XCAD, DCA1). Performance was compared with state-of-the-art models for user-guided segmentation. CoroSAM demonstrated superior performance on the ARCADE test set (Dice=0.87, Precision=0.86, Recall=0.89) while requiring fewer trainable parameters compared to competitive models. Statistical analysis confirmed significant improvements over alternative Adapter configurations. Zero-shot generalization yielded competitive performance on external datasets (XCAD: Dice=0.82; DCA1: Dice=0.73), demonstrating robust transferability across different image qualities. Integrating specialized Convolutional Adapters and channel-encoded point prompts enables accurate delineation of major coronary vessels with minimal user intervention. CoroSAM's architecture facilitates efficient inference on standard computing hardware without dedicated GPUs, providing a practical tool for clinical applications. This approach establishes an adaptation framework that effectively balances segmentation accuracy with computational efficiency, making it suitable for routine analysis workflows.

19 February 2025 in European Heart Journal Supplements
Verified
Journal Article
Open access
MUMario Urtis
Mario Urtis
MTMarilena Tagliani
Marilena Tagliani
DBDavide Bondavalli
Davide Bondavalli
CPChiara Paganini
Chiara Paganini
CCClaudia Cavaliere
Claudia Cavaliere
EAEloisa Arbustini
Eloisa Arbustini
The diagnostic work-up of cardiovascular genetic diseases is emerging as a reference model for precision (cause and phenotype) and personalized medicine (tailored individual management). This approach is expanding across all areas of cardiovascular medicine, ranging from monogenic diseases to multifactorial disorders where ‘risk factors’ such as familial hypercholesterolaemia may play a role in the risk profile. In this context, cardiomyopathies provide an ideal reference because they are monogenic yet genetically heterogeneous diseases, much like many other cardiovascular genetic disorders. In this model, the genetic path starts with deep phenotyping of the patient and relatives and progresses with genetic testing including extensive multigene panels, up to whole-exome sequencing and whole-genome sequencing. Although genetic work-ups are increasingly successful, several unresolved challenges and limitations remain. These include the interpretation and reinterpretation of variants, as many pre-American College of Medical Genetics variants previously classified as likely pathogenic or pathogenic are now recognized as variants of uncertain significance or benign/likely benign; pathogenic variants missed with short-read next generation sequencing (NGS) technologies (e.g. deep intronic variants or Copy Number Variations); gene-specific issues such as pseudogenes and pseudo-exons; and differing interpretations of pathogenicity for the same gene defects by commercial pipelines. Despite widespread NGS-based testing, about half of suspected Mendelian conditions still lack a precise molecular diagnosis. New organizational models are needed to integrate emerging knowledge and innovations incorporating both clinical and genetic data into intelligent platforms that may support shared management pathways.


22 June 2025
Verified
Book Chapter
Closed access
PDF file
MFMichela Ferrari
Michela Ferrari
ESEdoardo Spairani
Edoardo Spairani
MUMario Urtis
Mario Urtis
ATAntonio Tescari
Antonio Tescari
MGMaurizia Grasso
Maurizia Grasso
GMGiovanni Magenes
Giovanni Magenes

22 October 2026 in ACM Transactions on Computing for Healthcare
Verified
Journal Article
Open access
FPFrancesco Podda
Francesco Podda
ESEdoardo Spairani
Edoardo Spairani
EBEdoardo Bosco
Edoardo Bosco
MFMichela Ferrari
Michela Ferrari
GMGiovanni Magenes
Giovanni Magenes
Giulia MatroneGiulia Matrone
Giulia Matrone
Transthoracic Echocardiography (TTE) is a primary imaging modality for the evaluation of cardiac anatomy and function. Standardized TTE protocols rely on specific probe maneuvers to acquire predefined anatomical views; however, image acquisition quality is heavily operator-dependent, and particularly novice users often fail to obtain diagnostically adequate images, since optimal probe positioning is inherently a very challenging task. To address this challenge, we propose a novel deep learning-based framework to support less experienced operators in acquiring correct echocardiographic apical views. The proposed approach integrates two core components: a new navigation map that models the relationships between optimal and suboptimal views, linking common acquisition errors to the corresponding corrective probe motion (e.g., tilting, rotation, translation) and a classification system combining a Variational Autoencoder, for semi-supervised latent feature extraction, with a lightweight Multi-Layer Perceptron, for view classification. The framework was first evaluated on two public datasets (CAMUS and HMC-QU) for binary classification (apical 2 and 4 chamber views), achieving average accuracies up to 0.97 while maintaining good generalization across heterogeneous acquisition conditions. The analysis was then extended to a more challenging 11-class classification task, including four standard apical views (2, 3, 4, 5 chambers) and seven suboptimal variants, using a newly collected in-house dataset (PROUD). In this setting, the proposed VAE+MLP achieved an average accuracy of 0.85, representing a promising step toward AI-assisted TTE guidance systems for novice operators.