PIPPI creates a platform for researchers coming from different backgrounds - from MR physicists to clinical experts to Machine Learning experts - but with a strong common interest in perinatal imaging. This facilitates the exchange of new ideas to answer the specific challenges in this area and provides fruitful ground for new collaborations and new impulses across fields.Dr Jana Hutter, workshop organiser, Research Fellow in Healthcare Technologies at School of Biomedical Engineering & Imaging Sciences
03 November 2020
School of Biomedical Engineering & Imaging Sciences researchers organise successful workshop on perinatal and paediatric image analysis
The fifth annual Perinatal, Preterm and Paediatric Image Analysis workshop (PIPPI), founded and co-organised by staff from the School of Biomedical Engineering & Imaging Sciences was held online on 8 October 2020.
In addition to Dr Andrew Melbourne and Dr Jana Hutter from the School, the workshop’s organising committee included Roxane Licandro from Medical University of Vienna, Esra Abaci Turk from Boston Children's Hospital and Jordina Torrents from Hewlett Packard.
As in the previous years, the workshop formed an integral part of the prestigious International Conference on Medical Image Computing & Computer Assisted Intervention (MICCAI) and it explored innovative topics such as placental MRI analysis, fetal MRI of the nervous system, automated detection of congenital heart disease and spontaneous preterm birth prediction using convolutional neural networks.
This year's event was a resounding success, the online format contributed to our record attendance, and PIPPI had two amazing speakers with a keynote by Professor Polina Golland from the Massachusetts Institute of Technology (MIT), entitled ‘From Pixels to Clinical Insight: Placental MRI Analysis’ and an invited talk from Dr Michael Aertsen from Katholieke Universiteit Leuven focusing on fetal MRI of the central nervous system. The quality of the contributions from both speakers and the audience was an exceptionally high level this year.
The best paper award was received by Junshen Xu from MIT, for his work on 3D fetal pose estimation with adaptive variance and conditional generative adversarial network. The deep learning method proposed in his research improves accuracy in fetal pose estimation, tackling the dominant challenge posed by fetal motion to reliable performance and diagnostic quality of fetal MRI. As in previous years, Junshen’s best paper prize was a collection of Pippi Longstocking short stories.
Research from the School of Biomedical Engineering & Imaging Sciences was strongly represented in the schedule of the day. Dr Alena Uus, Dr Veronika Zimmer, Irina Grigorescu and Hassna Irzan gave oral presentations about neonatal brain MRI segmentation, motion correction in placental MRI datasets and brain volume differences in extremely preterm adolescents, while Julie Nihouarn and Ramya Thanikasalam explored in vivo tractography in the neonatal brain and the effects of preterm birth on intelligence in the poster sessions.
Dr Uus said: “My work introduced a novel pipeline for motion correction in quantitative 4D T2* volumes and 3D multimodal T2* and T2 placental and fetal MRI datasets based on the deformable slice-to-volume registration (DSVR) method. Motion correction of 4D T2* volumes allows the analysis of long time series. For example, accurate quantification changes of T2* values during the administration of medical oxygen help with characterisation of anomalies of placenta function.”
The online format contributed to a record attendance, with participants joining from 11 countries and 25 universities, attesting the demand for a specialist workshop focused on perinatal and paediatric imaging.
Dr Roxane Licandro from the organising committee said: “As a result of holding the workshop online, scientists and students from underrepresented groups (for example, clinicians who do not usually attend MICCAI) had the possibility to participate without the requirement of travelling and being onsite. In comparison to previous years, we observed an increase of participants and submissions and thus, we also aim to organise future editions of PIPPI in a hybrid setup.”
Dr Andrew Melbourne said: “Five years ago, there was no dedicated MICCAI workshop for early-life researchers. It has been fantastic to watch the community grow and develop over the past years and we look forward to start planning our sixth event in Strasbourg next year.”