Publications by authors named "Matteo Bregonzio"
Advancing healthcare through data: the BETTER project's vision for distributed analytics.
Front Med (Lausanne)
October 2024
Introduction: Data-driven medicine is essential for enhancing the accessibility and quality of the healthcare system. The availability of data plays a crucial role in achieving this goal.
Methods: We propose implementing a robust data infrastructure of FAIRification and data fusion for clinical, genomic, and imaging data.
Article Synopsis
- Broadband Coherent anti-Stokes Raman (BCARS) microscopy is a fast imaging technique that captures full Raman spectra of biological samples, but the results can be distorted by a non-resonant background (NRB) signal.
- Traditionally, NRB was removed with complex numerical algorithms, but recent advancements in deep learning have made it possible to automate and speed up this process.
- The paper reviews existing deep-learning models for NRB removal and introduces two new architectures, finding that CNN + GRU and VECTOR networks offer the best accuracy, while GAN excels in identifying true positive peaks and is suitable for real-time processing.
Article Synopsis
- Leukemia is a group of bone marrow tumors characterized by rapid cell growth, and current diagnosis methods rely on slow visual inspection of blood samples to identify subtypes.
- The study presents a novel approach using Raman hyperspectral imaging to analyze bone marrow samples from 19 patients with different leukemia subtypes, capturing over 1.3 million cell spectra.
- The automated analysis revealed distinct cellular components and created high-quality images at the single-cell level, demonstrating Raman imaging's potential in enhancing leukemia research despite some challenges with the clustering process.