Projects<\/h2>\n\n\n\n
HEARTMAGIC<\/strong><\/a> (funded by the Swiss National Science Foundation<\/a>) – Heart Failure with Preserved Ejection Fraction (HFpEF) affects around half patients suffering from heart failure. It has very poor prognosis and no treatment available. Our hypothesis is that we should be able to identify subtypes of patients for which better therapeutic options can be found. To this end, we will recruit around 600 patients between two hospitals, and acquire genotype, transcriptomics, metabolomic, and MR imaging data using newly-developed sequences. We will also develop new ML algorithms to subtype HFpEF patients from these rich, panoramic data. In collaboration with MR expert Ruud van Heeswijk<\/a> (CHUV) and cardiologists Roger Hullin<\/a> (CHUV) and Philippe Meyer<\/a> (HUG)<\/p>\n\n\n\n Within this project we are developing several methods around graph-based machine learning, including multiplex graphs for multi-omics<\/a> integration and translation, or multiplex graph neural networks for cardiac imaging<\/a>. <\/p>\n\n\n\n <\/p>\n\n\n Advanced Stroke Analytics Platform<\/a> (ASAP)<\/strong> (co-financed by Innosuisse<\/a>) – We are developing ML algorithms for federated learning robust to heterogeneous imaging protocols, in order to improve ischemic tissue volumetry and localisation. In collaboration with SCAN group at Inselspital Bern<\/a> and Siemens Healthcare Advanced Clinical Imaging Technology.<\/p>\n\n\n\n Heart and Brain in Action<\/strong><\/a> (funded by Alliance Campus Rhodanien<\/a>) – We are exploring representations and predictive models for spatiotemporal biomedical imaging data, as a way to translate neuroimaging statistical\/ML modelling techniques to heart imaging, and vice versa. In collaboration with the Statify INRIA team<\/a>.<\/p>\n\n\n\n CHUV-Lundin Brain cancer database<\/strong> (funded by Lundin Family Brain Tumour Research Centre) – We are aiming to build the most comprehensive brain tumor database in the world, comprising clinical data, imaging data, and omics data. With a target size of several thousand patients, and a commitment to open science from the beginning, this ambitious project aims at supporting data-driven research in brain cancer. The project is led at the Lausanne University Hospital and University of Lausanne in close collaboration between experts in oncology<\/a>, radio-oncology<\/a>, oncological biomarkers<\/a>, precision oncology<\/a>, clinical data science<\/a>, with support from the Data Science team in the IT department<\/a>, and external collaborators at the MedGIFT group<\/a>, HES-SO Valais.<\/p>\n\n\n\n Brain Circuitry Therapeutics for Schizophrenia<\/a><\/strong> (funded by the Leenaards Foundation<\/a>) – We are evaluating the effectiveness of transcranial magnetic stimulation to alleviate negative symptoms of schizophrenia. In this precision psychiatry randomized control trial, we are modeling mouse and human motion data, as well as human neuroimaging data (deep phenotyping), to track neural and behaviour correlates of treatment effects. In collaboration with Indrit B\u00e8gue<\/a> (Neuroimaging and translational psychatry<\/a>) and Camilla Bellone<\/a> (Synaptic brain dysfunctions<\/a>)<\/p>\n\n\n\n We enjoy wide-ranging collaborations with several other research and clinical groups groups, where we provide machine learning and data analysis expertise<\/p>\n\n\n\n INRIA<\/a>‘s Statify<\/a> team at Universit\u00e9 Grenoble-Alpes: Advanced Spatiotemporal Statistical Frameworks for brain connectivity<\/em> project, funded by Multidisciplinary Institute in Artificial Intelligence<\/a>.<\/p>\n\n\n\n Geneva University Hospital’s Carrera Lab<\/a>: Manipulating the peri-infarct cortex using Maraviroc to enhance motor skills after stroke<\/a><\/em>, funded by the Swiss National Science Foundation.<\/p>\n\n\n\n Lausanne University Hospital’s Radiodiagnostic and Interventional Radiology service<\/a>: Improving hepatocellular carcinoma screening in Switzerland: new strategies<\/a>, funded by the Swiss National Science Foundation<\/p>\n\n\n\n PACSifier<\/strong><\/a> – batch DICOM query\/retrieve interface enabling easy and targetted mass retrieval of medical images from hospital PACS, using a simple query file format in CSV from standard DICOM elements. Available as docker image or source code.<\/p>\n\n\n\n TML-CTP<\/a><\/strong> – depersonalisation tool. Enhancement around the RSNA MIRC CTP Dicom Anonymizer tool, dockerized and pip-installable, enabling anonymisation (not only coding), customisable date shits, deep-cleaning of potentially dangerous tags containing PHI (e.g. sequences), parallelisation, and deletion of files containing potentially identifiable pixel data.<\/p>\n\n\n\n Lesion disconnectomics<\/strong> – estimate a disconnectome graph without diffusion data. Code<\/a> (Maintained by Veronica Ravano). Paper<\/a>.<\/p>\n\n\n\n Aneurysm detection<\/strong> – Detection-by-segmentation approach. Code<\/a> (Maintained by Tommaso Di Noto). Paper<\/a>.<\/p>\n\n\n\n Flexible Brain Graph Visualizer<\/strong> – Create cool 3D plots of brain graphs. Code<\/a>.<\/p>\n\n\n\n Lausanne University Hospital and University of Lausanne TOF-MRA Aneurysm cohort<\/a> – At N=284, this is the largest open MR angiography dataset for aneurysms.<\/p>\n\n\n\n Lausanne University Hospital and University of Lausanne longitudinal Glioma change detection cohort<\/a> – 1683 T2w difference maps for 183 patients, together with stable\/unstable labels. <\/p>\n\n\n\n![\"\"](\"https:\/\/wp.unil.ch\/tml\/files\/2022\/08\/ASAP_LOGO_FINAL_WBshort.png\"){kind=logo}
<\/figure>\n<\/div>\n\n\nPartnership projects<\/h2>\n\n\n\n
Open source software<\/h2>\n\n\n\n
Open Data<\/h2>\n\n\n\n