Insigneo Seminar: Reproducibility of medical imaging through open-source

We are pleased to welcome Dr Christoph Kolbitsch, Physikalisch-Technische Bundesanstalt (Berlin), to give a talk on 'Reproducibility of medical imaging through open-source' at 12.00pm on 1 December 2025. This will be a hybrid seminar. 

Abstract

Advances in medical image reconstruction have strongly improved diagnostic accuracy while reducing scan times and improving patient comfort. Iterative optimisation schemes using classical approaches such as compressed sensing or deep learning-based methods ensure high image quality from very little acquired data. Motion-corrected image reconstruction provides artefact-free images without the need of long and exhausting breathholds.

One challenge of image reconstruction is how to ensure a fair comparison between different methods and how to ensure that algorithms are generalisable and applicable to heterogenous clinical data. Open-source ensures full transparency and allows for independent evaluation of methods. Nevertheless, full reproducibility requires more than just making an algorithm open-access.

This talk will focus on quantitative MR and PET-MR image reconstruction techniques and show how open-source approaches can be used to achieve reproducible multi-site imaging.

Biography

Christoph Kolbitsch completed his PhD at King’s College London, UK, in 2012, focusing on Advanced Techniques for Cardiovascular Magnetic Resonance Imaging in Cases of Irregular Motion. After his PhD, he worked as a postdoctoral research associate at King’s College London, specializing in motion correction for PET-MR applications. In 2015, he joined the Physikalisch-Technische Bundesanstalt (PTB) in Berlin, Germany. Since 2017, he has led the “Quantitative MRI” research group at PTB.

Dr. Kolbitsch has contributed to several open-source software packages for medical image reconstruction, including https://github.com/SyneRBI/SIRF and https://github.com/PTB-MR/mrpro. Currently, he serves as the project leader of the EU-funded A4IM initiative, which is focused on developing an open-source MRI scanner (https://www.a4im.ptb.de/home).