A Vertebral Segmentation Dataset with Fracture                     Grading

Löffler, Maximilian T.; Sekuboyina, Anjany; Jacob, Alina; Grau, Anna-Lena; Scharr, Andreas; Husseini, Malek El; Kallweit, Mareike; Zimmer, Claus; Baum, Thomas; Kirschke, Jan S.

doi:10.1148/ryai.2020190138

Cited by 94 publications

(72 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, we implemented a deep-learning driven approach to opportunistic osteoporosis screening. Manual interaction was limited to labeling of vertebrae that should be automatically segmented ( 27 , 28 ) and to the inspection of automatic segmentation masks for quality assurance and exclusion of severely degenerated vertebrae. The latter also reflects a routine necessity for DXA scans—the up-to-date reference standard for bone densitometry ( 26 ).…”

Section: Discussionmentioning

confidence: 99%

“…First, vertebrae were manually labelled and, then, automatically segmented using an in-house developed, fully convolutional neural network ( 27 , 28 ) ( Figures 1 and 2A ). In case of postoperative scans that depicted vertebrae with pedicle screws the segmentation algorithm excluded any voxels above a threshold of 1400 HU from the segmentation process by default, thus delineating the screw contours and excluding them from vertebral segmentation, which then only covered bone ( Figure 2 ).…”

Section: Methodsmentioning

confidence: 99%

“…The feasibility and validity of opportunistic BMD screening in existing CT scans has been extensively shown ( 26 ). Here, we present a new screening approach for decreased BMD that involves automatic spine segmentation by a fully convolutional neural network and requires little user interaction ( 27 , 28 ).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Opportunistic Osteoporosis Screening Reveals Low Bone Density in Patients With Screw Loosening After Lumbar Semi-Rigid Instrumentation: A Case-Control Study

et al. 2021

Self Cite

View full text Add to dashboard Cite

ObjectiveDecreased bone mineral density (BMD) impairs screw purchase in trabecular bone and can cause screw loosening following spinal instrumentation. Existing computed tomography (CT) scans could be used for opportunistic osteoporosis screening for decreased BMD. Purpose of this case-control study was to investigate the association of opportunistically assessed BMD with the outcome after spinal surgery with semi-rigid instrumentation for lumbar degenerative instability.MethodsWe reviewed consecutive patients that had primary surgery with semi-rigid instrumentation in our hospital. Patients that showed screw loosening in follow-up imaging qualified as cases. Patients that did not show screw loosening or—if no follow-up imaging was available (n = 8)—reported benefit from surgery ≥ 6 months after primary surgery qualified as controls. Matching criteria were sex, age, and surgical construct. Opportunistic BMD screening was performed at L1 to L4 in perioperative CT scans by automatic spine segmentation and using asynchronous calibration. Processing steps of this deep learning-driven approach can be reproduced using the freely available online-tool Anduin (https://anduin.bonescreen.de). Area under the curve (AUC) was calculated for BMD as a predictor of screw loosening.ResultsForty-six elderly patients (69.9 ± 9.1 years)—23 cases and 23 controls—were included. The majority of surgeries involved three spinal motion segments (n = 34). Twenty patients had low bone mass and 13 had osteoporotic BMD. Cases had significantly lower mean BMD (86.5 ± 29.5 mg/cm³) compared to controls (118.2 ± 32.9 mg/cm³, p = 0.001), i.e. patients with screw loosening showed reduced BMD. Screw loosening was best predicted by a BMD < 81.8 mg/cm³ (sensitivity = 91.3%, specificity = 56.5%, AUC = 0.769, p = 0.002).ConclusionPrevalence of osteoporosis or low bone mass (BMD ≤ 120 mg/cm³) was relatively high in this group of elderly patients undergoing spinal surgery. Screw loosening was associated with BMD close to the threshold for osteoporosis (< 80 mg/cm³). Opportunistic BMD screening is feasible using the presented approach and can guide the surgeon to take measures to prevent screw loosening and to increase favorable outcomes.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Opportunistic Osteoporosis Screening Reveals Low Bone Density in Patients With Screw Loosening After Lumbar Semi-Rigid Instrumentation: A Case-Control Study

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…To capture the bone contour accurately, tetrahedral element (C3D4 in Abaqus material library) was used for FE meshing. The nonhomogeneous and non-isotropic material behavior of the bone was captured by considering image intensity (Hounsfield unit (HU))-based material mapping relation [ 18 , 35 , 36 , 37 , 38 , 39 ]. Table 2 shows the HU-density-elasticity material mapping relations used in the current study [ 40 , 41 , 42 ].…”

Section: Methodsmentioning

confidence: 99%

“…Table 2 shows the HU-density-elasticity material mapping relations used in the current study [ 40 , 41 , 42 ]. The cortical bone was simplified and assumed as denser trabecular bone and same material mapping relations are used for both the regions [ 28 , 35 ].…”

Section: Methodsmentioning

confidence: 99%

Prediction of Incidental Osteoporotic Fractures at Vertebral-Specific Level Using 3D Non-Linear Finite Element Parameters Derived from Routine Abdominal MDCT

et al. 2021

Self Cite

View full text Add to dashboard Cite

To investigate whether finite element (FE) analysis of the spine in routine thoracic/abdominal multi-detector computed tomography (MDCT) can predict incidental osteoporotic fractures at vertebral-specific level; Baseline routine thoracic/abdominal MDCT scans of 16 subjects (8(m), mean age: 66.1 ± 8.2 years and 8(f), mean age: 64.3 ± 9.5 years) who sustained incidental osteoporotic vertebral fractures as confirmed in follow-up MDCTs were included in the current study. Thoracic and lumbar vertebrae (T5-L5) were automatically segmented, and bone mineral density (BMD), finite element (FE)-based failure-load, and failure-displacement were determined. These values of individual vertebrae were normalized globally (g), by dividing the absolute value with the average of L1-3 and locally by dividing the absolute value with the average of T5-12 and L1-5 for thoracic and lumbar vertebrae, respectively. Mean-BMD of L1-3 was determined as reference. Receiver operating characteristics (ROC) and area under the curve (AUC) were calculated for different normalized FE (Kload, Kdisplacement,K(load)g, and K(displacement)g) and BMD (KBMD, and K(BMD)g) ratio parameter combinations for identifying incidental fractures. Kload, K(load)g, KBMD, and K(BMD)g showed significantly higher discriminative power compared to standard mean BMD of L1-3 (BMDStandard) (AUC = 0.67 for Kload; 0.64 for K(load)g; 0.64 for KBMD; 0.61 for K(BMD)g vs. 0.54 for BMDStandard). The combination of Kload, Kdisplacement, and KBMD increased the AUC further up to 0.77 (p < 0.001). The combination of FE with BMD measurements derived from routine thoracic/abdominal MDCT allowed an improved prediction of incidental fractures at vertebral-specific level.

show abstract

Efficient spine segmentation network based on multi‐scale feature extraction and multi‐dimensional spatial attention

Xu,

Wang,

et al. 2024

Int J Imaging Syst Tech

View full text Add to dashboard Cite

In spine imaging, efficient automatic segmentation is crucial for clinical decision‐making, yet current models increase accuracy at the expense of elevated parameter counts and computational complexity, complicating integration with contemporary medical devices. Addressing identified challenges, this research introduces LE‐NeXt, a spine segmentation framework utilizing multi‐dimensional spatial attention and multi‐scale feature extraction, optimizing the architecture via convolution and MLP. It integrates lightweight convolutions and attention mechanisms within an encoder‐decoder model, enhancing stage‐specific feature extraction while ensuring efficiency. Experimental analyses on VerSe and SpineWeb datasets demonstrate that LE‐NeXt outperforms the lightweight U‐NeXt, enhancing IoU accuracy from 87.7 to 89.8 on VerSe, and exceeds the performance of established networks such as U‐Net and its variants. Significantly, on SpineWeb, LE‐NeXt not only surpasses Trans U‐Net in accuracy but also achieves a considerable reduction in both parameter count and computational complexity. These results emphasize LE‐NeXt's effectiveness in improving segmentation precision efficiently, optimally balancing computational efficiency and accuracy.

show abstract

A Vertebral Segmentation Dataset with Fracture Grading

Abstract: Published under a CC BY 4.0 license. Supplemental material is available for this article .

Cited by 94 publications

References 28 publications

Opportunistic Osteoporosis Screening Reveals Low Bone Density in Patients With Screw Loosening After Lumbar Semi-Rigid Instrumentation: A Case-Control Study

Opportunistic Osteoporosis Screening Reveals Low Bone Density in Patients With Screw Loosening After Lumbar Semi-Rigid Instrumentation: A Case-Control Study

Prediction of Incidental Osteoporotic Fractures at Vertebral-Specific Level Using 3D Non-Linear Finite Element Parameters Derived from Routine Abdominal MDCT

Efficient spine segmentation network based on multi‐scale feature extraction and multi‐dimensional spatial attention

Contact Info

Product

Resources

About