Marc André Weber

Prof. Dr. Marc André Weber

Diagnostic and Interventional Radiology, Universitätsmedizin Rostock, Rostock, Mecklenburg-Western Pomerania, Germany
orcid

Research Focus & Metrics

Academic Career Length

20Years (2007-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
23 November 2021 in Alzheimer's Research & Therapy
Verified
Journal Article
Open access
PDF file
MDMartin Dyrba
Martin Dyrba
MHMoritz Hanzig
Moritz Hanzig
SASlawek Altenstein
Slawek Altenstein
SBSebastian Bader
Sebastian Bader
TBTommaso Ballarini
Tommaso Ballarini
STStefan J. Teipel
Stefan J. Teipel
Background: Although convolutional neural networks (CNNs) achieve high diagnostic accuracy for detecting Alzheimer’s disease (AD) dementia based on magnetic resonance imaging (MRI) scans, they are not yet applied in clinical routine. One important reason for this is a lack of model comprehensibility. Recently developed visualization methods for deriving CNN relevance maps may help to fill this gap as they allow the visualization of key input image features that drive the decision of the model. We investigated whether models with higher accuracy also rely more on discriminative brain regions predefined by prior knowledge. Methods: We trained a CNN for the detection of AD in N = 663 T1-weighted MRI scans of patients with dementia and amnestic mild cognitive impairment (MCI) and verified the accuracy of the models via cross-validation and in three independent samples including in total N = 1655 cases. We evaluated the association of relevance scores and hippocampus volume to validate the clinical utility of this approach. To improve model comprehensibility, we implemented an interactive visualization of 3D CNN relevance maps, thereby allowing intuitive model inspection. Results: Across the three independent datasets, group separation showed high accuracy for AD dementia versus controls (AUC ≥ 0.91) and moderate accuracy for amnestic MCI versus controls (AUC ≈ 0.74). Relevance maps indicated that hippocampal atrophy was considered the most informative factor for AD detection, with additional contributions from atrophy in other cortical and subcortical regions. Relevance scores within the hippocampus were highly correlated with hippocampal volumes (Pearson’s r ≈ −0.86, p < 0.001). Conclusion: The relevance maps highlighted atrophy in regions that we had hypothesized a priori. This strengthens the comprehensibility of the CNN models, which were trained in a purely data-driven manner based on the scans and diagnosis labels. The high hippocampus relevance scores as well as the high performance achieved in independent samples support the validity of the CNN models in the detection of AD-related MRI abnormalities. The presented data-driven and hypothesis-free CNN modeling approach might provide a useful tool to automatically derive discriminative features for complex diagnostic tasks where clear clinical criteria are still missing, for instance for the differential diagnosis between various types of dementia.

8 July 2025 in Nature Communications
Verified
Journal Article
Open access
PDF file
MZMaximilian Zenk
Maximilian Zenk
Ujjwal BaidUjjwal Baid
Ujjwal Baid
SPSarthak Pati
Sarthak Pati
ALAkis Linardos
Akis Linardos
BEBrandon Edwards
Brandon Edwards
Spyridon BakasSpyridon Bakas
Spyridon Bakas
Computational competitions are the standard for benchmarking medical image analysis algorithms, but they typically use small curated test datasets acquired at a few centers, leaving a gap to the reality of diverse multicentric patient data. To this end, the Federated Tumor Segmentation (FeTS) Challenge represents the paradigm for real-world algorithmic performance evaluation. The FeTS challenge is a competition to benchmark (i) federated learning aggregation algorithms and (ii) state-of-the-art segmentation algorithms, across multiple international sites. Weight aggregation and client selection techniques were compared using a multicentric brain tumor dataset in realistic federated learning simulations, yielding benefits for adaptive weight aggregation, and efficiency gains through client sampling. Quantitative performance evaluation of state-of-the-art segmentation algorithms on data distributed internationally across 32 institutions yielded good generalization on average, albeit the worst-case performance revealed data-specific modes of failure. Similar multi-site setups can help validate the real-world utility of healthcare AI algorithms in the future.

12 October 2017 in Seminars in Musculoskeletal Radiology
Verified
Journal Article
Closed access
PDF file
RLRadhesh Lalam
Radhesh Lalam
JBJohan L. Bloem
Johan L. Bloem
INIris M. Noebauer-Huhmann
Iris M. Noebauer-Huhmann
KWKlaus Wörtler
Klaus Wörtler
ATAlberto Tagliafico
Alberto Tagliafico
MWMarc André Weber
Marc André Weber
Benign bone tumors are rare but are more common than primary malignant bone tumors. The early accurate diagnosis and reliable differentiation of these rare benign tumors and tumor mimickers from the even rarer malignant tumors with subsequent appropriate treatment or watchful waiting is crucial for the clinical outcome. Bone tumors are often a source of diagnostic and therapeutic uncertainty. Thus this European Society of Musculoskeletal Radiology consensus document is intended to help radiologists in their decision making and support discussion among clinicians who deal with patients with suspected or proven bone tumors. Evaluating these tumors starts with a patient history and physical examination. Radiography is the principal imaging modality and often can reliably diagnose a benign bone tumor by providing information about localization, matrix, aggressiveness, size, and (potential) multiplicity. In a significant number of cases, additional imaging is not necessary. Potentially malignant entities recognized by radiography should be referred for magnetic resonance imaging, which also serves as a preoperative local staging modality, with specific technical requirements. Indeterminate tumors, or tumors in which therapy depends on histology results, should be biopsied. For biopsy, we strongly recommend referral to a specialist regional sarcoma treatment center (RSTC), where a multidisciplinary tumor team, including a specialist pathologist, radiologist, and sarcoma surgeon, are involved. Additional staging modalities are entity specific and should be performed according to the recommendations of the RSTC.

11 December 2020 in Seminars in Musculoskeletal Radiology
Verified
Journal Article
Closed access
PDF file
DDDanoob Dalili
Danoob Dalili
AIAmanda Isaac
Amanda Isaac
Alberto BazzocchiAlberto Bazzocchi
Alberto Bazzocchi
GÅGunnar Åström
Gunnar Åström
JBJonas Bergh
Jonas Bergh
RMRamy Mansour
Ramy Mansour
Musculoskeletal (MSK) image-guided oncologic intervention is an established field within radiology. Numerous studies have described its clinical benefits, safety, cost effectiveness, patient satisfaction, and improved quality of life, thereby establishing image-guided oncologic intervention as a preferred pathway in treating patients presenting with specific benign MSK tumors. But there is a paradigm shift on the horizon because these techniques may also support established pillars (surgery, systemic treatment, radiotherapy) in the treatment of malignant MSK tumors. Unlike benign tumors, where they are used as primary therapy lines with curative intent, such interventions can be selected for malignant tumors as adjuvant treatment in painful or unstable bone or soft tissue lesions or as more palliative therapy strategies. Using examples from our clinical practices, we elaborate on the benefits of applying a multidisciplinary approach (traditionally involving MSK radiologists, oncologists, orthopaedic surgeons, microbiologists, pathologists, physiotherapists, and pain management experts), ideally within a sarcoma treatment center to deliver a patient-specific therapy plan and illustrate methods to assess the benefits of this model of care.In this article, we review the current repertoire of ablation techniques, demonstrate why such procedures offer value-based alternatives to conventional treatments of specific tumors, and reflect on future directions. Additionally, we review the advantages and limitations of each technique and offer guidance to improve outcomes.

30 January 2019 in Seminars in Musculoskeletal Radiology
Verified
Journal Article
Closed access
PDF file
OPOlympia Papakonstantinou
Olympia Papakonstantinou
VNVioleta Vasilevska Nikodinovska
Violeta Vasilevska Nikodinovska
FVFilip M. Vanhoenacker
Filip M. Vanhoenacker
MWMarc André Weber
Marc André Weber
Ewing's sarcoma (ES) is a rare, highly malignant anaplastic stem cell tumor. Histologically, the tumor consists of uniform densely packed small monomorphic cells with round nuclei. The typical appearance at hematoxylin and eosin (H&E) staining is small blue round cells without any matrix formation. On conventional radiography, ES typically presents as a permeative lesion in the diaphysis of a long bone in a child. A large soft tissue component is another characteristic feature, best depicted by magnetic resonance imaging.Primary osseous lymphomas are most commonly highly malignant B-cell lymphomas. At H&E histologic staining, the tumor stroma consists of diffuse round-cell infiltrates that resembles the appearance of ES. Although there is no typical imaging appearance of an osseous lymphoma, it should be considered in an adult presenting with a Lodwick grade II or III lesion in the metaphysis or diaphysis of a large long bone, the pelvis, or the vertebral column. Histologic confirmation is mandatory.

16 July 2020 in The International Journal of Cardiovascular Imaging
Verified
Journal Article
Open access
PDF file
BBBenjamin Böttcher
Benjamin Böttcher
EBEbba Beller
Ebba Beller
ABAnke Busse
Anke Busse
DCDaniel Cantré
Daniel Cantré
SYSeyrani Yücel
Seyrani Yücel
FMFelix G. Meinel
Felix G. Meinel
To investigate the performance of a deep learning-based algorithm for fully automated quantification of left ventricular (LV) volumes and function in cardiac MRI. We retrospectively analysed MR examinations of 50 patients (74% men, median age 57 years). The most common indications were known or suspected ischemic heart disease, cardiomyopathies or myocarditis. Fully automated analysis of LV volumes and function was performed using a deep learning-based algorithm. The analysis was subsequently corrected by a senior cardiovascular radiologist. Manual volumetric analysis was performed by two radiology trainees. Volumetric results were compared using Bland–Altman statistics and intra-class correlation coefficient. The frequency of clinically relevant differences was analysed using re-classification rates. The fully automated volumetric analysis was completed in a median of 8 s. With expert review and corrections, the analysis required a median of 110 s. Median time required for manual analysis was 3.5 min for a cardiovascular imaging fellow and 9 min for a radiology resident (p < 0.0001 for all comparisons). The correlation between fully automated results and expert-corrected results was very strong with intra-class correlation coefficients of 0.998 for end-diastolic volume, 0.997 for end-systolic volume, 0.899 for stroke volume, 0.972 for ejection fraction and 0.991 for myocardial mass (all p < 0.001). Clinically meaningful differences between fully automated and expert corrected results occurred in 18% of cases, comparable to the rate between the two manual readers (20%). Deep learning-based fully automated analysis of LV volumes and function is feasible, time-efficient and highly accurate. Clinically relevant corrections are required in a minority of cases.