Sergios Gatidis
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Sergios Gatidis has a total of 397 co-authors. Here are the top co-authors by number of shared publications:
- Thomas Küstner
University Hospital Tübingen
23 shared publications - Konstantin Nikolaou
Department of Diagnostic and Interventional Radiology, University Hospital Tübingen, Tübingen, Germany
17 shared publications - Christian la Fougère
Nuklearmedizin und Klinische Molekulare Bildgebung, Universitätsklinikum Tübingen, Tübingen, Germany
11 shared publications Konstantin Nikolaou
Department of Radiology, University Hospital Tuebingen, Tuebingen, Germany
11 shared publications- Jan Hendrik Moltz
Fraunhofer Institute for Digital Medicine MEVIS, Max-von-Laue-Straße 2, 28359 Bremen, Germany
7 shared publications - Ahmed E. Othman
Department of Neuroradiology, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany
6 shared publications
Latest Publications
- 26 September 2026
- Published by Springer Nature
- p. 100-110
CheXanatomy: Anatomy–Aware Vision–Language Modeling for Chest Radiographs
- 26 September 2026
- Published by Springer Nature
- p. 100-110
CheXanatomy: Anatomy–Aware Vision–Language Modeling for Chest Radiographs
- 15 August 2026
- Published by Springer Nature
- p. 17-24
Tumor Tracking as Point Tracking: CoTracker3 in the TrackRAD2025 Challenge
- 15 August 2026
- Published by Springer Nature
- p. 17-24
Tumor Tracking as Point Tracking: CoTracker3 in the TrackRAD2025 Challenge
- 3 August 2026
- Published by Springer Nature in Pediatric Radiology
- Vol. 56 (12) , 2807-2813
Assessing family engagement and acceptance of patient-friendly radiology reports at two academic children’s hospitals
- 3 August 2026
- Published by Springer Nature in Pediatric Radiology
- Vol. 56 (12) , 2807-2813
Assessing family engagement and acceptance of patient-friendly radiology reports at two academic children’s hospitals
- Open Access
- 18 May 2026
- Published by Springer Nature in Scientific Data
- Vol. 13 (1)
A longitudinal whole-body CT dataset with manually annotated tumor lesions
We introduce Longitudinal-CT, a publicly available resource of whole-body computed tomography (CT) studies with exhaustive expert manual annotations of tumor lesions across two timepoints. The dataset comprises 600 CT studies from 300 patients diagnosed with metastatic malignant melanoma, each including a baseline and a follow-up examination acquired during systemic therapy at the University Hospital Tübingen, Germany. In total, it contains 7,182 manually segmented tumor lesions - 4,079 at baseline and 3,103 at follow-up - each labeled with anatomical location, volume, and longitudinal correspondence to capture lesion evolution such as persistence, regression, merging, or new appearance. All CT data are provided in anonymized NIfTI format with corresponding segmentation masks and lesion metadata. Longitudinal-CT establishes a standardized foundation for developing and validating artificial intelligence methods for automated lesion detection, segmentation, and temporal tracking in oncology. As a reference, a baseline deep learning segmentation model trained using nnU-Net v2 demonstrates the dataset’s potential for advancing research in automated oncologic whole-body CT lesion segmentation.
- 18 May 2026
- Published by Springer Nature in Scientific Data
- Vol. 13 (1)
Open AccessA longitudinal whole-body CT dataset with manually annotated tumor lesions
We introduce Longitudinal-CT, a publicly available resource of whole-body computed tomography (CT) studies with exhaustive expert manual annotations of tumor lesions across two timepoints. The dataset comprises 600 CT studies from 300 patients diagnosed with metastatic malignant melanoma, each including a baseline and a follow-up examination acquired during systemic therapy at the University Hospital Tübingen, Germany. In total, it contains 7,182 manually segmented tumor lesions - 4,079 at baseline and 3,103 at follow-up - each labeled with anatomical location, volume, and longitudinal correspondence to capture lesion evolution such as persistence, regression, merging, or new appearance. All CT data are provided in anonymized NIfTI format with corresponding segmentation masks and lesion metadata. Longitudinal-CT establishes a standardized foundation for developing and validating artificial intelligence methods for automated lesion detection, segmentation, and temporal tracking in oncology. As a reference, a baseline deep learning segmentation model trained using nnU-Net v2 demonstrates the dataset’s potential for advancing research in automated oncologic whole-body CT lesion segmentation.
- Open Access
Image Quality Evaluation of Neonatal Brain MRI Using a Deep Learning Reconstruction Algorithm: A Quantitative and Multireader Study Using Variable Denoising Levels at 3 Tesla
Purpose: Neonatal imaging is particularly challenging because newborns have a high likelihood of head motion, which can degrade image quality and complicate interpretation. Improving MRI brain image quality may help reduce diagnostic uncertainty and facilitate the nuanced assessment of early myelinating structures in the neonatal brain. Although deep learning reconstruction algorithms designed to improve MRI image quality have been evaluated in pediatric imaging, they have not been specifically studied in exclusively neonatal populations. We sought to evaluate image quality improvement through the employment of a deep learning reconstruction algorithm in neonatal brain imaging. Methods: 3D T1-weighted brain MRIs were obtained in 15 neonates. A deep-learning reconstruction algorithm was applied to the image sets using low, medium, and high levels of denoising. Three radiologists qualitatively rated image quality (signal-to-noise ratio, presence of artifacts, and overall clarity) on a 4-point scale of eight early myelinating structures. Objective apparent signal-to-noise ratio (aSNR) and apparent contrast-to-noise ratio (aCNR), based on signal intensities of white-and gray-matter, was measured across all three denoising levels. Results: Evaluation by radiologists indicated an overall increase in all image quality categories and increased conspicuity of the early myelinating structures as the level of denoising increased. Objective aSNR and aCNR values also increased progressively with denoising, with significant differences observed for nearly all pairwise comparisons. Conclusion: Our findings suggest that the use of the proposed deep learning reconstruction algorithm improves image quality in 3D T1-weighted neonatal brain MRIs at 3T.
Open AccessImage Quality Evaluation of Neonatal Brain MRI Using a Deep Learning Reconstruction Algorithm: A Quantitative and Multireader Study Using Variable Denoising Levels at 3 Tesla
Purpose: Neonatal imaging is particularly challenging because newborns have a high likelihood of head motion, which can degrade image quality and complicate interpretation. Improving MRI brain image quality may help reduce diagnostic uncertainty and facilitate the nuanced assessment of early myelinating structures in the neonatal brain. Although deep learning reconstruction algorithms designed to improve MRI image quality have been evaluated in pediatric imaging, they have not been specifically studied in exclusively neonatal populations. We sought to evaluate image quality improvement through the employment of a deep learning reconstruction algorithm in neonatal brain imaging. Methods: 3D T1-weighted brain MRIs were obtained in 15 neonates. A deep-learning reconstruction algorithm was applied to the image sets using low, medium, and high levels of denoising. Three radiologists qualitatively rated image quality (signal-to-noise ratio, presence of artifacts, and overall clarity) on a 4-point scale of eight early myelinating structures. Objective apparent signal-to-noise ratio (aSNR) and apparent contrast-to-noise ratio (aCNR), based on signal intensities of white-and gray-matter, was measured across all three denoising levels. Results: Evaluation by radiologists indicated an overall increase in all image quality categories and increased conspicuity of the early myelinating structures as the level of denoising increased. Objective aSNR and aCNR values also increased progressively with denoising, with significant differences observed for nearly all pairwise comparisons. Conclusion: Our findings suggest that the use of the proposed deep learning reconstruction algorithm improves image quality in 3D T1-weighted neonatal brain MRIs at 3T.
- Open Access
- 4 February 2026
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 95 (6) , 3519-3535
Automated Coregistered Segmentation for Volumetric Analysis of Multiparametric Renal MRI
Purpose: This study aims to develop and evaluate a fully automated deep learning‐driven postprocessing pipeline for multiparametric renal MRI, enabling accurate kidney alignment, segmentation, and quantitative feature extraction within a single efficient workflow. Methods: Our method has three main stages. First, a segmentation network delineates renal structures in high‐contrast images. Next, a deep learning‐based pairwise image registration algorithm maps the multiparametric image series to a common target and transfers the predicted annotations between the multiparametric images. Finally, clinically relevant quantitative parameters are extracted through region‐specific assessment of renal structure and function based on the aligned and segmented multiparametric data. We used five‐fold cross‐validation to compare the segmentation outcomes and extracted features with manual analyses in 24 patients with prostate cancer or neuroendocrine tumors and 10 healthy subjects, each undergoing repeated scans. Results: Our automated pipeline achieved high agreement with expert kidney segmentation while delivering significant alignment improvements through registration. Volumetric analysis showed a strong correlation ( r 0.9) with manual results, and feature extraction demonstrated high intraclass correlation coefficients with minimal bias. The complete processing pipeline, encompassing coregistration, segmentation, and feature extraction, required approximately 15 s per scan from raw input to final quantitative output. Conclusion: The study establishes a reliable automated pipeline for renal multiparametric MRI postprocessing. The achieved accuracy and efficiency can support improved diagnosis and treatment planning for patients with kidney disease.
- 4 February 2026
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 95 (6) , 3519-3535
Open AccessAutomated Coregistered Segmentation for Volumetric Analysis of Multiparametric Renal MRI
Purpose: This study aims to develop and evaluate a fully automated deep learning‐driven postprocessing pipeline for multiparametric renal MRI, enabling accurate kidney alignment, segmentation, and quantitative feature extraction within a single efficient workflow. Methods: Our method has three main stages. First, a segmentation network delineates renal structures in high‐contrast images. Next, a deep learning‐based pairwise image registration algorithm maps the multiparametric image series to a common target and transfers the predicted annotations between the multiparametric images. Finally, clinically relevant quantitative parameters are extracted through region‐specific assessment of renal structure and function based on the aligned and segmented multiparametric data. We used five‐fold cross‐validation to compare the segmentation outcomes and extracted features with manual analyses in 24 patients with prostate cancer or neuroendocrine tumors and 10 healthy subjects, each undergoing repeated scans. Results: Our automated pipeline achieved high agreement with expert kidney segmentation while delivering significant alignment improvements through registration. Volumetric analysis showed a strong correlation ( r 0.9) with manual results, and feature extraction demonstrated high intraclass correlation coefficients with minimal bias. The complete processing pipeline, encompassing coregistration, segmentation, and feature extraction, required approximately 15 s per scan from raw input to final quantitative output. Conclusion: The study establishes a reliable automated pipeline for renal multiparametric MRI postprocessing. The achieved accuracy and efficiency can support improved diagnosis and treatment planning for patients with kidney disease.
- Open Access
Pipeline Evaluation of a State-of-the-Art AI Algorithm for Detection of Focal Cortical Dysplasia: Insights into Potential Failure Sources
Purpose: MELD Graph is a state-of-the-art artificial intelligence (AI) model for automated detection of focal cortical dysplasia (FCD), but its performance remains limited, highlighting the need to investigate which aspects of the pipeline affect its accuracy. Methods: A retrospective failure-mode analysis of the MELD Graph pipeline was performed in 242 subjects, with model predictions and FreeSurfer segmentations reviewed to classify errors as segmentation-associated or algorithm-related. FCD imaging features salient to humans were quantified, with statistical associations examined for both MELD Graph detection and focal FreeSurfer segmentation failure. Results: MELD Graph demonstrated overall performance similar to previously published non-harmonized results, achieving a sensitivity of 69%, specificity of 44%, and positive predictive value (PPV) of 75%. Focal FreeSurfer segmentation failures were associated with 21% of false negative patients, 25% of false positive clusters in patients, and 16% of false positive clusters in controls. Higher conspicuity on T1-weighted images was associated with MELD Graph detection, whereas greater conspicuity on T2-FLAIR images relative to T1 was associated with detection failure. Bottom-of-sulcus dysplasia (BOSD) and presence of transmantle sign were not associated with detection. Non-BOSD lesions, higher human conspicuity measures, and low T1 image quality were positively associated with focal FreeSurfer segmentation failures. Conclusion: FreeSurfer segmentation failures are a significant potential source of error in the MELD Graph pipeline. FCD imaging features salient to humans and image quality were also associated with variability in the algorithm performance. Robust cortical segmentation and stronger integration of T2-FLAIR imaging features may be beneficial for automated FCD detection tools.
Open AccessPipeline Evaluation of a State-of-the-Art AI Algorithm for Detection of Focal Cortical Dysplasia: Insights into Potential Failure Sources
Purpose: MELD Graph is a state-of-the-art artificial intelligence (AI) model for automated detection of focal cortical dysplasia (FCD), but its performance remains limited, highlighting the need to investigate which aspects of the pipeline affect its accuracy. Methods: A retrospective failure-mode analysis of the MELD Graph pipeline was performed in 242 subjects, with model predictions and FreeSurfer segmentations reviewed to classify errors as segmentation-associated or algorithm-related. FCD imaging features salient to humans were quantified, with statistical associations examined for both MELD Graph detection and focal FreeSurfer segmentation failure. Results: MELD Graph demonstrated overall performance similar to previously published non-harmonized results, achieving a sensitivity of 69%, specificity of 44%, and positive predictive value (PPV) of 75%. Focal FreeSurfer segmentation failures were associated with 21% of false negative patients, 25% of false positive clusters in patients, and 16% of false positive clusters in controls. Higher conspicuity on T1-weighted images was associated with MELD Graph detection, whereas greater conspicuity on T2-FLAIR images relative to T1 was associated with detection failure. Bottom-of-sulcus dysplasia (BOSD) and presence of transmantle sign were not associated with detection. Non-BOSD lesions, higher human conspicuity measures, and low T1 image quality were positively associated with focal FreeSurfer segmentation failures. Conclusion: FreeSurfer segmentation failures are a significant potential source of error in the MELD Graph pipeline. FCD imaging features salient to humans and image quality were also associated with variability in the algorithm performance. Robust cortical segmentation and stronger integration of T2-FLAIR imaging features may be beneficial for automated FCD detection tools.
- Open Access
Retrospective motion correction in MRI using disentangled embeddings
Physiological motion can affect the diagnostic quality of magnetic resonance imaging (MRI). While various retrospective motion correction methods exist, many struggle to generalize across different motion types and body regions. In particular, machine learning (ML)-based corrections are often tailored to specific applications and datasets. We hypothesize that motion artifacts, though diverse, share underlying patterns that can be disentangled and exploited. To address this, we propose a hierarchical vector-quantized (VQ) variational auto-encoder that learns a disentangled embedding of motion-to-clean image features. A codebook is deployed to capture finite collection of motion patterns at multiple resolutions, enabling coarse-to-fine correction. An auto-regressive model is trained to learn the prior distribution of motion-free images and is used at inference to guide the correction process. Unlike conventional approaches, our method does not require artifact-specific training and can generalize to unseen motion patterns. We demonstrate the approach on simulated whole-body motion artifacts and observe robust correction across varying motion severity. Our results suggest that the model effectively disentangled physical motion of the simulated motion-effective scans, therefore, improving the generalizability of the ML-based MRI motion correction. Our work of disentangling the motion features shed a light on its potential application across anatomical regions and motion types.
Open AccessRetrospective motion correction in MRI using disentangled embeddings
Physiological motion can affect the diagnostic quality of magnetic resonance imaging (MRI). While various retrospective motion correction methods exist, many struggle to generalize across different motion types and body regions. In particular, machine learning (ML)-based corrections are often tailored to specific applications and datasets. We hypothesize that motion artifacts, though diverse, share underlying patterns that can be disentangled and exploited. To address this, we propose a hierarchical vector-quantized (VQ) variational auto-encoder that learns a disentangled embedding of motion-to-clean image features. A codebook is deployed to capture finite collection of motion patterns at multiple resolutions, enabling coarse-to-fine correction. An auto-regressive model is trained to learn the prior distribution of motion-free images and is used at inference to guide the correction process. Unlike conventional approaches, our method does not require artifact-specific training and can generalize to unseen motion patterns. We demonstrate the approach on simulated whole-body motion artifacts and observe robust correction across varying motion severity. Our results suggest that the model effectively disentangled physical motion of the simulated motion-effective scans, therefore, improving the generalizability of the ML-based MRI motion correction. Our work of disentangling the motion features shed a light on its potential application across anatomical regions and motion types.
- Open Access
- 4 November 2025
- Published by Springer Nature in Research Square
Imaging-derived biological age across multiple organs links to mortality and aging-related health outcomes
Aging is a complex, multifactorial process, influencing disease risk and overall health. While chronological age (CA) is widely used in clinical practice, it fails to capture individual aging trajectories. Current approaches to estimate biological age (BA) often focus on single organs or predefined clinical biomarkers, limiting comprehensive assessment. We introduce a novel, purely imaging-driven deep learning framework for organ-specific BA estimation across seven organ systems. Our uncertainty-aware ResNet-based models autonomously learned aging-related features from imaging data in ~70,000 UK Biobank participants, eliminating manual feature selection biases. Training on a healthy cohort, where CA approximates BA, allows learning normative aging patterns. When applied to a broader cohort, deviations from typical aging indicate older or younger BA. Our findings demonstrate the feasibility of BA estimation, even in organs with subtle aging features. While aging is largely heterogeneous across organs, we also identified correlations in aging patterns. We further showed that accelerated aging is predictive of mortality and health outcomes, offering insights for personalized assessments.
- 4 November 2025
- Published by Springer Nature in Research Square
Open AccessImaging-derived biological age across multiple organs links to mortality and aging-related health outcomes
Aging is a complex, multifactorial process, influencing disease risk and overall health. While chronological age (CA) is widely used in clinical practice, it fails to capture individual aging trajectories. Current approaches to estimate biological age (BA) often focus on single organs or predefined clinical biomarkers, limiting comprehensive assessment. We introduce a novel, purely imaging-driven deep learning framework for organ-specific BA estimation across seven organ systems. Our uncertainty-aware ResNet-based models autonomously learned aging-related features from imaging data in ~70,000 UK Biobank participants, eliminating manual feature selection biases. Training on a healthy cohort, where CA approximates BA, allows learning normative aging patterns. When applied to a broader cohort, deviations from typical aging indicate older or younger BA. Our findings demonstrate the feasibility of BA estimation, even in organs with subtle aging features. While aging is largely heterogeneous across organs, we also identified correlations in aging patterns. We further showed that accelerated aging is predictive of mortality and health outcomes, offering insights for personalized assessments.