Automatic skull defect restoration and cranial implant generation for cranioplasty

Li, Jianning; Campe, Gord von; Pepe, Antonio; Gsaxner, Christina; Wang, Enpeng; Chen, Xiaojun; Zefferer, Ulrike; Tödtling, Martin; Krall, Marcell; Deutschmann, Hannes; Schäfer, Ute; Schmalstieg, Dieter; Egger, Jan

doi:10.1016/j.media.2021.102171

Cited by 46 publications

(19 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…DISCUSSION In automatic cranial implant design, deep learning-based approaches that rely on a defect-complete or defect-implant pair for training often fail to generalize to large and complex cranial defects in the test set, since the synthetic defects used during training have different distributions to the real defects during evaluation (i.e., domain shift). One popular solution to this problem is resorting to intensive data augmentation: augment the defects [26,40] and/or augment the skull images [32,38]. The former tries to create realistic synthetic defects for training.…”

Section: Task2mentioning

confidence: 99%

“…Automatic cranial implant design is another typical application that uses images as the initial shape representation [24]. Existing deep learning-based methods usually train a deep neural net on hundreds of skull images with either synthetic defects [25][26][27][28] or clinical defects [29], depending on the availability of the clinical images. These approaches are data-and computation-intensive, and most importantly, the quality of their reconstructions for large and complex defects, which are common in cranioplasty, remains inadequate for clinical use [30,31].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Back to the Roots: Reconstructing Large and Complex Cranial Defects using an Image-based Statistical Shape Model

Li¹,

Ellis²,

Pepe³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Designing implants for large and complex cranial defects is a challenging task, even for professional designers. Current efforts on automating the design process focused mainly on convolutional neural networks (CNN), which have produced state-of-the-art results on reconstructing synthetic defects. However, existing CNN-based methods have been difficult to translate to clinical practice in cranioplasty, as their performance on complex and irregular cranial defects remains unsatisfactory. In this paper, a statistical shape model (SSM) built directly on the segmentation masks of the skulls is presented. We evaluate the SSM on several cranial implant design tasks, and the results show that, while the SSM performs suboptimally on synthetic defects compared to CNN-based approaches, it is capable of reconstructing large and complex defects with only minor manual corrections. The quality of the resulting implants is examined and assured by experienced neurosurgeons. In contrast, CNN-based approaches, even with massive data augmentation, fail or produce less-than-satisfactory implants for these cases. Codes are publicly available at https://github.com/Jianningli/ssm

show abstract

Section: Task2mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Back to the Roots: Reconstructing Large and Complex Cranial Defects using an Image-based Statistical Shape Model

Li¹,

Ellis²,

Pepe³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“… The 29 craniotomy skulls together with the corresponding manually designed cranial implants can serve as an evaluation set for automatic cranial implant design algorithms. Researchers can create synthetic cranial defects on the 500 healthy skulls in order to train deep learning algorithms [1] , [5] , [6] , [7] and host challenges [8] . The .stl files included in the MUG500+ dataset are 3D printable and can be used for educational purposes.…”

Section: Value Of the Datamentioning

confidence: 99%

“…Researchers can create synthetic cranial defects on the 500 healthy skulls in order to train deep learning algorithms [1] , [5] , [6] , [7] and host challenges [8] .…”

Section: Value Of the Datamentioning

confidence: 99%

MUG500+: Database of 500 high-resolution healthy human skulls and 29 craniotomy skulls and implants

Krall

Trummer

et al. 2021

Data in Brief

Self Cite

View full text Add to dashboard Cite

“…This is relevant in medical image analysis, where it is difficult to collect large amounts of labelled data for rare diseases. Classical neural networks do not quantify model uncertainty, and Bayesian neural networks, which describe the weight parameters as distributions over the parameters ω i = f (µ i , σ i ) (Kendall and Gall (2017)), can be harder to train, especially in 3D medical imaging, where data can be scarce and requires higher computational costs (Gal and Ghahramani (2016); Li et al (2021)). An alternative approach is to train K models independently.…”

Section: Uncertainty-aware Deep Learningmentioning

confidence: 99%

Automated cross-sectional view selection in CT angiography of aortic dissections with uncertainty awareness and retrospective clinical annotations

Pepe¹,

Egger²,

Codari³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Objective: Surveillance imaging of chronic aortic diseases, such as dissections, relies on obtaining and comparing cross-sectional diameter measurements at predefined aortic landmarks, over time. Due to a lack of robust tools, the orientation of the crosssectional planes is defined manually by highly trained operators. We show how manual annotations routinely collected in a clinic can be efficiently used to ease this task, despite the presence of a non-negligible interoperator variability in the measurements. Impact: Ill-posed but repetitive imaging tasks can be eased or automated by leveraging imperfect, retrospective clinical annotations. Methodology: In this work, we combine convolutional neural networks and uncertainty quantification methods to predict the orientation of such cross-sectional planes. We use clinical data randomly processed by 11 operators for training, and test on a smaller set processed by 3 independent operators to assess interoperator variability. Results: Our analysis shows that manual selection of cross-sectional planes is characterized by 95% limits of agreement (LOA) of 10.6 • and 21.4 • per angle. Our method showed to decrease static error by 3.57 • (40.2%) and 4.11 • (32.8%) against state of the art and LOA by 5.4 • (49.0%) and 16.0 • (74.6%) against manual processing. Conclusion: This suggests that pre-existing annotations can be an inexpensive resource in clinics to ease ill-posed and repetitive tasks like cross-section extraction for surveillance of aortic dissections.

show abstract

Automatic skull defect restoration and cranial implant generation for cranioplasty

Cited by 46 publications

References 27 publications

Back to the Roots: Reconstructing Large and Complex Cranial Defects using an Image-based Statistical Shape Model

Back to the Roots: Reconstructing Large and Complex Cranial Defects using an Image-based Statistical Shape Model

MUG500+: Database of 500 high-resolution healthy human skulls and 29 craniotomy skulls and implants

Automated cross-sectional view selection in CT angiography of aortic dissections with uncertainty awareness and retrospective clinical annotations

Contact Info

Product

Resources

About