Research
Research Areas
Research Line
Medical imaging technologies for early and non-invasive disease diagnosis
Medical imaging is essential for non-invasive clinical decision making. But, it still lacks comprehensive and quantitative disease characterization (required for precision medicine), is not integrated with non-imaging patient data and is expensive. The cost of medical imaging, such as magnetic resonance imaging (MRI), computed tomography (CT), ultrasound (US) and positron emission tomography (PET) are associated with operational costs (scan/staff time), the need for highly trained staff and the equipment itself. To overcome these challenges, in iHEALTH we aim to develop innovative methods that integrate medical imaging-physics/engineering, non-imaging patient data, and artificial intelligence (AI), from data acquisition to clinical decision-making, and evaluate these in hospitals across Chile.
Research Line
Inverse problems for medical image reconstruction and reduction of scan times
Medical imaging requires an analytical model of the acquisition and solving inverse problems to reconstruct images from the raw-data (measured in a different domain, e.g. the frequency domain), which are then used for analysis and interpretation. These inverse problems are ill-posed and require sufficient data to generate good quality images, limiting the ability of medical images to capture the full spectrum of disease-related processes due to the need for long scan times or high radiation doses. In iHEALTH we aim to develop innovative methods that incorporate physical models of the acquisition and organ of interest into the reconstruction process, as well as exploiting synergies across medical imaging modalities, the medical imaging pipeline and/or across systems with different specifications to improve medical imaging efficiency (faster scans and/or lower radiation dose), efficacy (capturing wider-spectrum of disease related-process) and accessibility (solutions for regular and lower-cost systems).
Research Line
Physics-informed artificial intelligence methods for medical imaging acquisition, processing and diagnosis
Artificial intelligence (AI) based methods have been proposed to automate several aspects of the medical imaging pipeline, including medical imaging acquisition, processing and diagnosis. Yet, most current solutions are based on generic data-driven models that require large training sets, which are usually not available in medical imaging. Moreover, most current solutions are developed for specific populations and technologies and mostly lack interpretability. In iHEALTH we aim to develop innovative methods that incorporate physical models of the acquisition and organ of interest into physics-informed AI solutions to exploit both data-driven and physics-driven components for medical imaging acquisition, processing and diagnosis.
Research Line
Reliable and explainable artificial intelligence methods integrating medical images and other clinical patient data
Artificial intelligence (AI) based methods have been proposed to automate several aspects of the medical imaging pipeline. Most of these methods focus on the analysis of the images once acquired and reconstructed, and are based on “black-box” models, with their success depending on large amounts of high-quality images. Yet, access to medical images is limited (even more so for labeled data), non-standardised and with disparate quality, making most methods not generalizable to different scenarios (e.g. between hospitals, patient-cohorts). Moreover, most current solutions are developed for specific populations/technologies and mostly lack interpretability. In iHEALTH we aim to develop explainable AI (XAI) methods that integrate medical imaging and non-imaging (e.g. EHR, sensors) data to provide accurate, reliable and interpretable AI-assisted clinical analysis, diagnosis and reporting.
Research Line
Biomarkers based on imaging and physiological sensors for early diagnosis and characterization of cardiovascular, liver and cancer diseases
Cardiovascular (CVD), non-alcoholic fatty liver (NAFLD) and oncology diseases are the leading cause of morbidity and mortality in the Western world. Clinical standards for diagnosing most of these conditions remain invasive (e.g. x-ray angiography, biopsy). Advances in medical imaging and Artificial Intelligence (AI)-based analysis have been proposed for non-invasive assessment, but current methods have limited sensitivity for early-stage disease-related processes and largely lack reproducibility and generalizability. The latter is also due to the nonexistence of advanced medical imaging biobanks in countries like Chile. In iHEALTH we aim to develop innovative biomarkers based on imaging and physiological sensors to enable early diagnosis and characterization of cardiovascular, liver and cancer diseases.
Research Projects
Research project
Fully Quantitative Magnetic Resonance Imaging
Medical imaging requires sequential acquisition of various (mostly qualitative) anatomical/functional images and subjective expert interpretation to correlate imaging findings with a diagnosis. Quantitative medical imaging is a major research focus in magnetic resonance imaging. However, current solutions are still limited by their accuracy, the restricted number of quantifiable parameters and by lack of reproducibility and standardization. Moreover, image quality can be affected by system imperfections, motion and other confounding factors, hindering analysis and interpretation. In this project we aim to develop accurate and easy to use/interpret (“one-click” exam) MRI able to provide comprehensive quantitative tissue characterization (e.g. fibrosis, inflammation) from a single and fast scan.
Research project
Multiparametric MR for the assessment of acute cardiac toxicity induced by thoracic radiotherapy
Today, cancer ranks as one of the leading causes of death. Despite the large number of novel available therapies, radiotherapy remains as the most effective non-surgical method to cure cancer patients. In fact, approximately 50% of all cancer patients receive some type of radiotherapy and among these, 60% receive radiotherapy-treatment with a curative intent. However, as occurs with any other oncological therapy, radiotherapy treated patients may experience toxicity side effects that range from moderate to severe. Among these, cardiotoxicity represents a significant threat for premature death. Current methods evaluate cardiotoxic damage based on volumetric changes in the Left Ventricle Ejection Fraction (LVEF). Indeed, a 10% drop in LVEF is commonly used as an indicator of cardiotoxicity. More recently, several novel techniques have been developed that significantly improve specificity and sensitivity of heart’s volumetric changes and early detection of cardiotoxicity even in asymptomatic patients. In this project we will use novel multi-parametric MRI to characterize the myocardial tissue and strain analysis to develop early biomarkers of cardiotoxicity in a cohort of patients that will receive radiotherapy.
Research project
Automatic Report Generation from Medical Images - Database and Methods
The task of Medical Image Report Generation can be of great support for physicians. From a computational perspective, the Medical Image Report Generation task can be described as follows: given as input one or more medical images of a patient, a text report is output that is as similar as possible to one generated by a radiologist. From a machine learning point of view, creating a system that performs such a task would require learning a generative model from instances of reports written by radiologists. This project aims at creating the technology to automatically generate reports from medical images.
