We study new imaging techniques in CT and MRI for quantitative imaging of the spine. We develop automated evaluation algorithms including artificial intelligence, deep learning and biomechanic modeling to predict disease progression and get new pathophysiologic insights. Currently we focus on osteoporosis, back pain, degenerative spine diseases and inflammatory disorders like multiple sclerosis. We work with in-house datasets as well as with data from GNC and SHIP studies
We develop AI-based methods to automatically analyze and quantify MR images of the spine. We currently work on methods to translate MRI to CT to use existing labeling and segmentation methods. Our goal is to analyze large datasets from GNC to better understand sources of back pain.
To better understand back pain and surgical outcome, we combine quantitative parameters from different imaging modalities (X-rays, MRI, CT) in one biomechanical model, based on multibody simulations (MBS) on a large scale and finite element models (FEM/FCM) on a small scale.
We investigate different techniques to automatically assess spinal lesion load and its clinical relevance for the individual patient.
We closely collaborate with the group of Prof. Mühlau and Prof. Cohen-Adad to further increase functionaliy and robustness of the spinal cord toolbox.
Anduin is a freely
available web-based application able to fully automatically
segment the spine in clinical CT scans. Our spin-off "Bonescreen"
has further developed the segmentation code (contact
them directly for scientific collaborations) and is
dedicated to establish a fully automated opportunistic
osteoporosis diagnosis in clinical routine.
Try out Bonescreen!
Driving forward an interest for fully automated segmentation of the spine in the research community, we publicly released voxel-level-annotated CT data of more than 300 patients and organised the VerSe (Large Scale Vertebrae Segmentation) challenge series MICCAI 2019 and 2020. See the resulting MedIA paper (or ArXiv here)
Together with Bonescreen, we validate and develop automated techniques for opportunistic BMD measurements in any CT dataset. We also implemented automatic fracture detection and prediction algorithms using artificial intelligence. Currently, we focus on the spine.
Somatosensory evoked potentials and magnetic resonance imaging of the central nervous system in early multiple sclerosis. Wuschek A, J Neurol. 2022 Oct 7.
of a Patient-Specific Musculoskeletal Model for Lumbar
Load Estimation Generated by an Automated Pipeline From
Whole Body CT. Lerchl T, Front Bioeng Biotechnol.
2022 Jul 11;10:862804.
mean value analysis - a voxel-based analysis of the
quantitative MR biomarker water T2 in the
presence of fatty infiltration in skeletal muscle tissue
of patients with neuromuscular diseases. Schlaeger S,
NMR Biomed. 2022 Dec;35(12):e4805.
Finite Element Modeling of the Whole Lumbar Spine Using
Clinical Routine Multi-Detector Computed Tomography (MDCT)
Data-A Pilot Study. Rayudu NM, Biomedicines. 2022 Jun
imaging of the musculoskeletal system - spine]. Schlaeger
S, Kirschke JS. Radiologie (Heidelb). 2022
Volumetric BMD Threshold Values for the Prediction of
Incident Vertebral Fractures Using Opportunistic QCT: A
Case-Control Study. Dieckmeyer M, Front Endocrinol
(Lausanne). 2022 May 20;13:882163.
Opportunistic Osteoporosis Screening in Routine Computed
Tomography of the Spine: Comparison With Dedicated
Quantitative CT. Sollmann N, J Bone Miner Res. 2022
sclerosis lesions and atrophy in the spinal cord:
Distribution across vertebral levels and correlation with
disability. Bussas M, Neuroimage Clin.
diagnostic volumetric bone mineral density thresholds for
osteoporosis and osteopenia at the cervicothoracic spine
in correlation to the lumbar spine. Rühling S, Eur
Radiol. 2022 Sep;32(9):6207-6214.
Inference of the 3D Standing Spine Posture from 2D
Radiographs. Bayat A, Tomography. 2022 Feb
computed tomography vertebral segmentation dataset with
anatomical variations and multi-vendor scanner data. Liebl
H, Sci Data. 2021 Oct 28;8(1):284.
detection of the contrast phase in MDCT by an artificial
neural network improves the accuracy of opportunistic bone
mineral density measurements. Rühling S, Eur Radiol.
and reporting of osteoporotic vertebral fractures in
patients with long-term hospital records based on routine
clinical CT imaging. Löffler MT, Kallweit M,
Niederreiter E, Baum T, Makowski MR, Zimmer C, Kirschke
JS.Osteoporos Int. 2022 Mar;33(3):685-694.
proton density fat fraction (PDFF) of the vertebral bone
marrow differentiates between patients with and without
osteoporotic vertebral fractures. Gassert FT, et al.
Osteoporos Int. 2022 Feb;33(2):487-496.
VerSe: A Vertebrae labelling and segmentation benchmark for multi-detector CT images. Sekuboyina A, et al. Med Image Anal. 2021 Oct;73:102166.
of incident vertebral fractures in routine MDCT:
Comparison of global texture features, 3D finite element
parameters and volumetric BMD. Dieckmeyer M, Eur J
Radiol. 2021 Aug;141:109827.
Vertebral Segmentation Dataset with Fracture Grading. Löffler
MT, Radiol Artif Intell. 2020 Jul 29;2(4):e190138.
opportunistic osteoporosis screening in routine CT:
improved prediction of patients with prevalent vertebral
fractures compared to DXA. Löffler MT, Eur Radiol.
Osteoporosis Screening Reveals Low Bone Density in
Patients With Screw Loosening After Lumbar Semi-Rigid
Instrumentation: A Case-Control Study. Löffler MT,
Front Endocrinol (Lausanne). 2021 Jan 11;11:552719.
Technische Universität München, Ismaninger Str. 22, 81675 München