Department of Biomedical Engineering

Background: The assessment of lung function and structure in infants and toddlers is highly limited due to the constraints of existing lung function and imaging techniques. Common lung function methods such as spirometry are not feasible in this age group, and research methods are very sophisticated and not widely available. Chest X-rays lack sensitivity for detecting subtle respiratory changes, and radiation burden of computed tomography limits its utility in this vulnerable age group. Conversely, magnetic resonance imaging (MRI) offers a radiation-free alternative with superior soft tissue contrast, making it a crucial tool in diagnostic medicine. In recent years, MRI has emerged as a non-invasive lung imaging modality and shown excellent sensitivity for the assessment of structural or functional lung impairments. However, MRI of the lung is rarely performed in infants or toddlers as it presents unique challenges compared to imaging older children or adults. The small size of infant lungs requires improved spatial resolution for adequate visualization of anatomical structures, whereas rapid respiratory and cardiac cycles necessitate rapid image acquisition or efficient gating techniques. Hence, there is an urgent need for the development of non-invasive and radiation-free diagnostic techniques for lung imaging in the youngest patients.

Aims and Methods: We aim to develop and validate novel proton-based and contrast-agent free MRI techniques for assessing lung impairment in infants and toddlers with chronic respiratory diseases and age-appropriate settings for lung MRI without general anesthesia. To achieve this, we will build upon our previous expertise in lung imaging using functional MRI in children. The project will focus on developing non-Cartesian steady-state free precession imaging techniques to improve spatial-temporal resolution, addressing the unique physiological conditions in young patients, as well as adaptation of our current techniques for morphological lung imaging. An important part of this project is the translation of this specialized MRI methodology into clinical application for improved scalability and seamless integration info clinical workflows. We will develop the optimal settings for lung MRI during natural sleep or with slight sedation only. Eventually, we will evaluate the new age-adapted lung MRI techniques in healthy infants and toddlers and in patients with different lung diseases including prematurity with and without bronchopulmonary dysplasia, cystic fibrosis and others. 

Significance and Broader Impact: In contrast to conventional radiographic modalities, the proposed MRI techniques will allow a radiation-free and non-invasive functional and structural assessment. This addresses a critical gap in current diagnostic capabilities and has the potential to change current clinical and research approaches. The ability to safely and repeatedly assess lung function and structure in young children could significantly advance our understanding of early lung development and the pathogenesis of various respiratory diseases. This could lead to earlier detection and earlier intervention in pediatric lung diseases, potentially improving long-term outcomes. The integration of these techniques into clinical workflows will improve patient care, guide treatment choices, predict disease progression, and support the development of individualized therapeutic measures. 

In conclusion, this project will advance the field of pediatric lung imaging by providing a safe, effective, and scalable diagnostic tool, ultimately improving our understanding and management of early lung diseases. Developing MRI techniques tailored for infants and toddlers will offer new perspectives on the early stages and advancement of chronic respiratory diseases, ultimately enhancing patient care and outcomes.

The frequency of thrombectomy (i.e. endovascular therapy for acute ischemic stroke) is rising, but complications of thrombectomy are poorly explored. Vessel perforation with consecutive intracranial hemorrhage is a severe complication of thrombectomy. The risk factors of vessel perforation are largely unknown. There is also almost no data to guide endovascular hemostatic therapy or to decide whether thrombectomy should be continued after the event of vessel perforation.

PREVENT is a multicenter, prospective and retrospective registry with the aim to

(1) Identify risk factors of vessel perforations

(2) Explore the pathophysiology of vessel perforations

(3) Develop a classification system

(4) Compare different hemostatic treatment strategies

(5) Compare continuation of thrombectomy after vessel perforation and abortion of thrombectomy

(6) Develop a novel, safety-optimized thrombectomy technique

Anticipated study cohort: 500 patients with vessel perforation during thrombectomy and 500 matched patients without perforation during thrombectomy. 

Background: In recent years, the sensitivity of magnetic resonance imaging (MRI) could be dramatically increased and now in principle enables high-resolution quantitative imaging in clinically feasible scan times. In the future, MRI will therefore transform from a mere camera yielding qualitative images to a quantitative instrument that can map tissue composition. However, robust mapping of the human brain tissue structure is currently impaired, because it was observed that the MRI signal also depends on the orientation of the nerve fibers relative to the direction of the main magnetic field of the MRI system. Problematically, this anisotropy was shown to be of similar magnitude as disease-related alterations. Furthermore, it is only poorly understood how tissue composition and anisotropy influence quantitative MRI (qMRI) investigations. These limitations therefore represent major barriers for qMRI becoming a clinically applied tool. Based on the findings of our preliminary work, the proposed project aims on closing these knowledge gaps by combining post mortem in situ, ex situ and in vivo examinations, as well as histological and biochemical assessments in order to elucidate the origin of anisotropy in qMRI and to foster the translation of qMRI into clinics.

Methods: Post mortem in situ subjects, in vivo calibration subjects, in vivo volunteers, patients with aceruloplasminemia and patients with multiple sclerosis will be included in this collaborative study performed at the sites Basel, Switzerland and Innsbruck, Austria. After post mortem in situ MRI, the brains will be extracted and examined ex situ. Different ex situ experiments will be performed to compare qMRI anisotropy indices calculated by actually rotating the brain with estimating the fiber angle using diffusion tensor imaging (DTI) and advanced diffusion modeling. Furthermore, specific experiments are planned to study qMRI anisotropy related to paramagnetic and diamagnetic tissue components. Additionally, histological, biochemical and molecular analyses will be performed to assess tissue composition in specific brain regions. This will allow to correlate qMRI anisotropy with tissue components, such as iron and myelin content, as well as specific lipids and proteins. Linear mixed models and principal component analysis will be applied to identify connections between the multiple parameters. Possible multi-center bias will be detected based on calibration examinations with traveling volunteers. The volunteers and patients will be scanned at two field strengths to detect B₀ dependent susceptibility contributions and disease related effects on qMRI anisotropy. Disease related effects on the anisotropy will be studied in both patient groups for evaluating the use of the anisotropy as clinical biomarker.

Expected results: We expect to observe differences in anisotropy indices between all acquired qMRI parameters and to confirm the validity of estimating qMRI anisotropy using DTI and advanced diffusion modeling. By separating quantitative MRI parameters into paramagnetic and diamagnetic components, we expect to observe increased anisotropy indices related to the diamagnetic tissue components, such as lipids and proteins. In addition, we anticipate that the anisotropy will differ between B₀ field strengths and between patients and healthy volunteers. Importantly, it can further be expected that we will successfully identify the underlying tissue components causing the qMRI anisotropy. By combining results from the post mortem experiments with in vivo observations in patients, we will be able to correlate pathological alterations in qMRI in vivo to the most likely underlying tissue components.

Relevance: Based on our cutting-edge approach that combines post mortem and in vivo examinations, we will be able to thoroughly study the impact of various factors on anisotropic qMRI, including field strengths, different tissue components such as paramagnetic iron and diamagnetic myelin, as well as lipids and proteins. By studying two distinct patient groups, we will be laying a solid foundation for the effective use of the anisotropy index as a simple measure in routine clinical practice. This research holds immense significance for the field of medicine as a whole, as it not only sheds light on the origin of anisotropy in qMRI, but also opens up possibilities for accurate quantitative mapping of tissue composition in patients and elimination of systematic biases in current diagnostic imaging.

Instrumented clinical gait analysis is used routinely to inform decision-making in neuro-orthopaedics. In addition to gait analysis, musculoskeletal modeling may become a powerful and non-invasive tool to guide clinical management and predict treatment outcomes. However, musculoskeletal modeling needs to integrate patient-specific adaptations, and its outputs need to be validated on a larger scale before it may be used in standard clinical practice.

The goal of this project is to develop patient-specific gait simulations by means of an open-source musculoskeletal modeling software. Results will be validated against existing clinical data pre vs post a typical intervention in neuro-orthopaedics.

Personalized musculoskeletal models from 30 children who received botulinum toxin injection will be developed from gait analysis data obtained before the intervention. To predict patient's response, the botulinum toxin effect will be simulated by weakening the model muscle and running a forward dynamic simulation. I will compare the outcome against existent data post-injection and analyze how induced muscle weakness alters the gait of children with cerebral palsy, providing validation for this specific musculoskeletal modeling application and overall confidence in our framework reliability.