Simulation‐driven training of vision transformers enables metal artifact reduction of highly truncated CBCT scans
Medical Physics · 2024
DOI: 10.1002/mp.16919 ↗Kurz gefasst
Fachartikel in der Zeitschrift Medical Physics, 2024. Verfasst von Fuxin Fan und 8 weiteren Autor:innen. Beteiligte Einrichtung: Pattern Recognition Lab, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany. Hauptthemen laut MEDLINE: Cone-Beam Computed Tomography, Artifacts, Metals, Image Processing, Computer-Assisted. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 31 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Siemens Healthineers, Friedrich-Alexander-Universität Erlangen-Nürnberg.
Abstract (Originalsprache)
Background
Due to the high attenuation of metals, severe artifacts occur in cone beam computed tomography (CBCT). The metal segmentation in CBCT projections usually serves as a prerequisite for metal artifact reduction (MAR) algorithms.
Purpose
The occurrence of truncation caused by the limited detector size leads to the incomplete acquisition of metal masks from the threshold‐based method in CBCT volume. Therefore, segmenting metal directly in CBCT projections is pursued in this work.
Methods
Since the generation of high quality clinical training data is a constant challenge, this study proposes to generate simulated digital radiographs (data I) based on real CT data combined with self‐designed computer aided design (CAD) implants. In addition to the simulated projections generated from 3D volumes, 2D x‐ray images combined with projections of implants serve as the complementary data set (data II) to improve the network performance. In this work, SwinConvUNet consisting of shift window (Swin) vision transformers (ViTs) with patch merging as encoder is proposed for metal segmentation.
Results
The model's performance is evaluated on accurately labeled test datasets obtained from cadaver scans as well as the unlabeled clinical projections. When trained on the data I only, the convolutional neural network (CNN) encoder‐based networks UNet and TransUNet achieve only limited performance on the cadaver test data, with an average dice score of 0.821 and 0.850. After using both data II and data I during training, the average dice scores for the two models increase to 0.906 and 0.919, respectively. By replacing the CNN encoder with Swin transformer, the proposed SwinConvUNet reaches an average dice score of 0.933 for cadaver projections when only trained on the data I. Furthermore, SwinConvUNet has the largest average dice score of 0.953 for cadaver projections when trained on the combined data set.
Conclusions
Our experiments quantitatively demonstrate the effectiveness of the combination of the projections simulated under two pathways for network training. Besides, the proposed SwinConvUNet trained on the simulated projections performs state‐of‐the‐art, robust metal segmentation as demonstrated on experiments on cadaver and clinical data sets. With the accurate segmentations from the proposed model, MAR can be conducted even for highly truncated CBCT scans.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Pattern Recognition Lab, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany.
- Siemens Healthcare GmbH, Forchheim, Germany.
Förderung
- Siemens Healthineers
- Friedrich-Alexander-Universität Erlangen-Nürnberg
Zitieren
APA
Fan, F., Ritschl, L., Beister, M., Biniazan, R., Wagner, F., Kreher, B., Gottschalk, T. M., Kappler, S., & Maier, A. (2024). Simulation‐driven training of vision transformers enables metal artifact reduction of highly truncated CBCT scans. Medical Physics. https://doi.org/10.1002/mp.16919
Vancouver
Fan F, Ritschl L, Beister M, Biniazan R, Wagner F, Kreher B, et al. Simulation‐driven training of vision transformers enables metal artifact reduction of highly truncated CBCT scans. Medical Physics. 2024. doi:10.1002/mp.16919
BibTeX
@article{Fan2024,
author = {Fan, Fuxin and Ritschl, Ludwig and Beister, Marcel and Biniazan, Ramyar and Wagner, Fabian and Kreher, Björn and Gottschalk, Tristan M and Kappler, Steffen and Maier, Andreas},
title = {Simulation‐driven training of vision transformers enables metal artifact reduction of highly truncated CBCT scans},
journal = {Medical Physics},
year = {2024},
doi = {10.1002/mp.16919}
}
EUVIMED