PosterPDF Available

Towards the Automatization of Cranial Implant Design for 3D Printing

Authors:

Abstract

Fast and fully automatic design of 3-D printed patient-specific cranial implant is highly desired in cranioplasty. To this end, various deep learning-based approaches are investigated. To facilitate supervised training, a database containing 200 high-resolution healthy CT skulls acquired in clinical routine is constructed. Due to the unavailability of large number of defected skulls from clinic, artificial defects are introduced to simulate that caused in a real cranial surgery.
Towards the Automatization of Cranial Implant Design for 3D Printing
Jianning Li, Jan Egger
Institute of Computer Graphics and Vision, Graz University of Technology, Graz, Austria
Contact: jianning.li@icg.tugraz.at
COMET K-Project
Clinical Additive Manufacturing for Medical Applications
CAMED FACTS & FIGURES
20 Project Partners from Science and Industry
~6 Mio. € Project Volume
Public Funding from Federal & State
CAMed Office:
Auenbruggerplatz 22
A-8036 Graz
Tel.: +43 316 385 72044
eMail: ulrike.fasching@medunigraz.at
www.studierfenster.at
... In 2020, the Medical Imaging Computing and Computer Assisted Intervention (MICCAI) Society hosted the AutoImplant Challenge to assess the best methods for automatically designing cranial implants [1]. The organizers designed the challenge around a dataset of skull images that contained artificial defects and corresponding implant images that filled in the artificial defects to restore the original unaltered skull. ...
... Ideally, the implant should be at least 50% thinner than the skull. 1 The primary reason that the implant design submissions were not feasible for real cranioplasty cases is due to the applicability of the available training sets. All participating teams used deep learning models to predict the desired implant shape. ...
... The second step is to use the completed skull combined with the defective skull to determine the ideal implant shape. As previous work has shown, deep learning models are highly effective at performing the skull completion step [1,5,6,8,9]. Performing this step alone may save a significant amount of time in the implant design process. ...
Preprint
Full-text available
As part of the 2021 MICCAI AutoImplant Challenge, CT scans from 11 patients who had undergone cranioplasty using artificial implants were collected. Images of the reconstructed defective skulls before cranioplasty for these patients were shared with participating teams. Three teams submitted cranial implant designs. An experienced neurosurgeon evaluated the submissions to judge the feasibility of the implant designs for use in cranioplasty procedures. None of the submitted cranial implant designs were deemed feasible for use in cranioplasty procedures without modifications. While many implants adequately restored the skull shape by covering the defect area, most contained excess material outside of the defect, fit poorly within the defect and were too thick. Future research should move beyond solely restoring the skull shape and focus on designing implants that contain smooth transitions between skull and implant, cover the entire defect, contain no material outside of the defect, have minimal thickness, and are implantable.
... The AutoImplant Challenge [9] is organized in order to tackle the problem of automatic cranial implant design in a data-driven manner, without relying explicitly on geometric shape priors of human skulls [10]. The organizers provide 3D binary images of defective skulls, complete skulls and implants as the datasets, with which the reconstruction of implants can be proceeded either directly from defective skulls, or from the differences between defective and complete skulls. ...
... The AutoImplant Challenge [9] database comprises 200 binary skull datasets (100 for training and 100 for testing) generated from CQ500 dataset [3]. The dimension of these skulls is 512 × 512 × Z, where Z is the number of axial slices. ...
Conference Paper
In this study, we proposed two methods for AutoImplant (https://autoimplant.grand-challenge.org/)-the cranial implant design challenge. The shape of the implant is predicted based on the inputted defective skull. This task can be accomplished either by directly predicting the implant with the defective skull, or indirectly rebuilding the complete skull and then taking the difference between the defective and complete skulls. In our work, a deep learning model is applied to automatically predict the implant. In order to solve the problem that high resolution images can often not be directly inputted to the deep learning model, two proposed methods of resize and patch-based are examined. On the test set, the proposed resize method achieves an average dice similarity score (DSC) of 0.7350 and a Hausdorff distance (HD) of 7.2425 mm, while the proposed patch-based method achieves an average DSC of 0.8887 and a HD of 5.5339 mm.
... The fully automatic deep learning-based algorithm has been selfsupervised during training by injecting artificial defects in healthy skulls [49,50]. An in-depth review of algorithms for an automatic cranial implant design can be found in AutoImplant 2020 summary paper [51] and challenge proceedings [52]. ...
Article
Imaging modalities such as computed tomography (CT) and magnetic resonance imaging (MRI) are widely used in diagnostics, clinical studies, and treatment planning. Automatic algorithms for image analysis have thus become an invaluable tool in medicine. Examples of this are two- and three-dimensional visualizations, image segmentation, and the registration of all anatomical structure and pathology types. In this context, we introduce Studierfenster ( www.studierfenster.at ): a free, non-commercial open science client-server framework for (bio-)medical image analysis. Studierfenster offers a wide range of capabilities, including the visualization of medical data (CT, MRI, etc.) in two-dimensional (2D) and three-dimensional (3D) space in common web browsers, such as Google Chrome, Mozilla Firefox, Safari, or Microsoft Edge. Other functionalities are the calculation of medical metrics (dice score and Hausdorff distance), manual slice-by-slice outlining of structures in medical images, manual placing of (anatomical) landmarks in medical imaging data, visualization of medical data in virtual reality (VR), and a facial reconstruction and registration of medical data for augmented reality (AR). More sophisticated features include the automatic cranial implant design with a convolutional neural network (CNN), the inpainting of aortic dissections with a generative adversarial network, and a CNN for automatic aortic landmark detection in CT angiography images. A user study with medical and non-medical experts in medical image analysis was performed, to evaluate the usability and the manual functionalities of Studierfenster. When participants were asked about their overall impression of Studierfenster in an ISO standard (ISO-Norm) questionnaire, a mean of 6.3 out of 7.0 possible points were achieved. The evaluation also provided insights into the results achievable with Studierfenster in practice, by comparing these with two ground truth segmentations performed by a physician of the Medical University of Graz in Austria. In this contribution, we presented an online environment for (bio-)medical image analysis. In doing so, we established a client-server-based architecture, which is able to process medical data, especially 3D volumes. Our online environment is not limited to medical applications for humans. Rather, its underlying concept could be interesting for researchers from other fields, in applying the already existing functionalities or future additional implementations of further image processing applications. An example could be the processing of medical acquisitions like CT or MRI from animals [Clinical Pharmacology & Therapeutics, 84(4):448-456, 68], which get more and more common, as veterinary clinics and centers get more and more equipped with such imaging devices. Furthermore, applications in entirely non-medical research in which images/volumes need to be processed are also thinkable, such as those in optical measuring techniques, astronomy, or archaeology.
... A relevant study was conducted by Morais et al, where an encoder-decoder network is used to predict a complete skull from a defective skull [13]. However, the study deals with very coarse skulls of low dimensionality (30 3 , 60 3 and 120 3 ) extracted from MRI data, whereas in practice, the common imaging modality used for head scan acquisition is computed tomography (CT), with a typical resolution of 512 × 512 × Z. [9,10] further extended shape completion to high-resolution volumetric CT skulls by using a patch-based training strategy. In this study, we primarily elaborate on the former formulation, i.e., given a defective skull, we directly predict the shape of the implant, which is a challenging task as the implant has to be congruent with the defective skull in terms of shape, bone thickness and boundaries of the defected region [11]. ...
Chapter
In this study, we present a baseline approach for AutoImplant (https://autoimplant.grand-challenge.org/) – the cranial implant design challenge, which can be formulated as a volumetric shape learning task. In this task, the defective skull, the complete skull and the cranial implant are represented as binary voxel grids. To accomplish this task, the implant can be either reconstructed directly from the defective skull or obtained by taking the difference between a defective skull and a complete skull. In the latter case, a complete skull has to be reconstructed given a defective skull, which defines a volumetric shape completion problem. Our baseline approach for this task is based on the former formulation, i.e., a deep neural network is trained to predict the implants directly from the defective skulls. The approach generates high-quality implants in two steps: First, an encoder-decoder network learns a coarse representation of the implant from downsampled, defective skulls; The coarse implant is only used to generate the bounding box of the defected region in the original high-resolution skull. Second, another encoder-decoder network is trained to generate a fine implant from the bounded area. On the test set, the proposed approach achieves an average dice similarity score (DSC) of 0.8555 and Hausdorff distance (HD) of 5.1825 mm. The codes are available at https://github.com/Jianningli/autoimplant.
... An automatic, lowcost design and manufacturing of cranial implants can bring significant benefits and improvements to the current clinical workflow for cranioplasty [2]. The AutoImplant Challenge [3] is organized in order to tackle the problem of automatic cranial implant design in a data-driven manner, without relying explicitly on geometric shape priors of human skulls [4]. The organizers provide 3D binary images of defective skulls, complete skulls and implants as the datasets, with which the reconstruction of implants can be proceeded either directly from defective skulls, or from the differences between defective and complete skulls. ...
Poster
Full-text available
A cranial defect usually occurs after injury, tumor invasion or infection. The current process of cranial implant design and manufacturing usually involves costly commercial software and highly-trained professional users. An automatic, lowcost design and manufacturing of cranial implants can bring significant benefits and improvements to the current clinical workflow for cranioplasty. The AutoImplant Challenge is organized in order to tackle the problem of automatic cranial implant design in a data-driven manner, without relying explicitly on geometric shape priors of human skulls. The organizers provide 3D binary images of defective skulls, complete skulls and implants as the datasets, with which the reconstruction of implants can be proceeded either directly from defective skulls, or from the differences between defective and complete skulls.
Article
Full-text available
Deep learning belongs to the field of artificial intelligence, where machines perform tasks that typically require some kind of human intelligence. Deep learning tries to achieve this by drawing inspiration from the learning of a human brain. Similar to the basic structure of a brain, which consists of (billions of) neurons and connections between them, a deep learning algorithm consists of an artificial neural network, which resembles the biological brain structure. Mimicking the learning process of humans with their senses, deep learning networks are fed with (sensory) data, like texts, images, videos or sounds. These networks outperform the state-of-the-art methods in different tasks and, because of this, the whole field saw an exponential growth during the last years. This growth resulted in way over 10,000 publications per year in the last years. For example, the search engine PubMed alone, which covers only a sub-set of all publications in the medical field, provides already over 11,000 results in Q3 2020 for the search term ‘deep learning’, and around 90% of these results are from the last three years. Consequently, a complete overview over the field of deep learning is already impossible to obtain and, in the near future, it will potentially become difficult to obtain an overview over a subfield. However, there are several review articles about deep learning, which are focused on specific scientific fields or applications, for example deep learning advances in computer vision or in specific tasks like object detection. With these surveys as a foundation, the aim of this contribution is to provide a first high-level, categorized meta-survey of selected reviews on deep learning across different scientific disciplines and outline the research impact that they already have during a short period of time. The categories (computer vision, language processing, medical informatics and additional works) have been chosen according to the underlying data sources (image, language, medical, mixed). In addition, we review the common architectures, methods, pros, cons, evaluations, challenges and future directions for every sub-category.
Article
Full-text available
Patient-specific craniofacial implants are used to repair skull bone defects after trauma or surgery. Currently, cranial implants are designed and produced by third-party suppliers, which is usually time-consuming and expensive. Recent advances in additive manufacturing made the in-hospital or in-operation-room fabrication of personalized implants feasible. However, the implants are still manufactured by external companies. To facilitate an optimized workflow, fast and automatic implant manufacturing is highly desirable. Data-driven approaches, such as deep learning, show currently great potential towards automatic implant design. However, a considerable amount of data is needed to train such algorithms, which is, especially in the medical domain, often a bottleneck. Therefore, we present CT-imaging data of the craniofacial complex from 24 patients, in which we injected various artificial cranial defects, resulting in 240 data pairs and 240 corresponding implants. Based on this work, automatic implant design and manufacturing processes can be trained. Additionally, the data of this work build a solid base for researchers to work on automatic cranial implant designs.
Chapter
Cranial implant design can be thought as a 3D shape completion task predicting the missing part of a defective cranium, which is a time-consuming task in traditional methods. This paper proposes a deep convolutional neural network (CNN) based method which predicts the implant from a binary voxel image of a defective skull. Three networks with the same structure are trained for inpainting sagittal, coronal, and horizontal slices of the defective skull, respectively. After skull size regularization and slice extraction, inpainting results from one or more axes are used to synthesize the final binary implant voxel image. Cross-validation shows that the proposed method has a good performance in the cranial implant design task in terms of both Dice similarity coefficient (DSC) and Hausdorff distance (HD).
Chapter
Cranioplasty is a surgical operation on the repairing of cranial defects caused by the previous operation, ischemic, or hemorrhagic disease, or even after the removal of cranial tumors. It can be performed by filling the defective area with a range of materials. Interactive and semi-automatic computer-aided design tools for cranial implant design are time-consuming and costly. In this paper, we proposed a deep learning method for automatic cranial implant generation. The proposed method mainly included two steps. First, a variational auto-encoder model was trained to learn the latent distribution of complete skulls. Then, the encoder part of the pre-trained VAE together with an encoder-decoder network was trained to generate the complete skull. We design an anatomical regularization term to drive the predicted skull to be more anatomically plausible compared with the ground truth skull. We evaluated the performance of our method using the skull data from the AutoImplant Challenge. The results show that the proposed framework performs well on the 100 test cases while has poor performance on the 10 test cases.
Book
The AutoImplant Cranial Implant Design Challenge (AutoImplant 2020: https:// autoimplant.grand-challenge.org/) was initialized jointly by the Graz University of Technology (TU Graz) and the Medical University of Graz (MedUni Graz), Austria, through an interdisciplinary project “Clinical Additive Manufacturing for Medical Applications” (CAMed: https://www.medunigraz.at/camed/) between the two institutions. The project aims to provide more affordable, faster, and patient-friendly solutions to the design and manufacturing of medical implants, including cranial implants, which is needed in order to repair a defective skull from a brain tumor surgery or trauma.
ResearchGate has not been able to resolve any references for this publication.