Low dose tomographic fluoroscopy: 4D intervention guidance with running prior

Flach, Barbara; Kuntz, Jan; Brehm, Marcus; Kueres, R.; Bartling, Sönke; Kachelrieß, Marc

doi:10.1118/1.4819826

Cited by 4 publications

(11 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Kuntz et al 17,18 proposed a compressed sensing-based algorithm for the 3D reconstruction of sparse objects, like guidewires and stents, from 16 simultaneous projections. This algorithm was later improved by Flach et al 22,23 to handle patient motion. Although the number of projections per 3D reconstruction (16) is much lower than in CTfluoroscopy, it is still too high for use in 3D fluoroscopy.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Real‐time 3D reconstruction of guidewires and stents using two update X‐ray projections in a rotating imaging setup

Vöth

Koenig²,

Eulig

et al. 2023

Medical Physics

Self Cite

View full text Add to dashboard Cite

BackgroundVascular diseases are often treated minimally invasively. The interventional material (stents, guidewires, etc.) used during such percutaneous interventions are visualized by some form of image guidance. Today, this image guidance is usually provided by 2D X‐ray fluoroscopy, that is, a live 2D image. 3D X‐ray fluoroscopy, that is, a live 3D image, could accelerate existing and enable new interventions. However, existing algorithms for the 3D reconstruction of interventional material require either too many X‐ray projections and therefore dose, or are only capable of reconstructing single, curvilinear structures.PurposeUsing only two new X‐ray projections per 3D reconstruction, we aim to reconstruct more complex arrangements of interventional material than was previously possible.MethodsThis is achieved by improving a previously presented deep learning‐based reconstruction pipeline, which assumes that the X‐ray images are acquired by a continuously rotating biplane system, in two ways: (a) separation of the reconstruction of different object types, (b) motion compensation using spatial transformer networks.ResultsOur pipeline achieves submillimeter accuracy on measured data of a stent and two guidewires inside an anthropomorphic phantom with respiratory motion. In an ablation study, we find that the aforementioned algorithmic changes improve our two figures of merit by 75 % (1.76 mm → 0.44 mm) and 59 % (1.15 mm → 0.47 mm) respectively. A comparison of our measured dose area product (DAP) rate to DAP rates of 2D fluoroscopy indicates a roughly similar dose burden.ConclusionsThis dose efficiency combined with the ability to reconstruct complex arrangements of interventional material makes the presented algorithm a promising candidate to enable 3D fluoroscopy.

show abstract

Section: Introductionmentioning

confidence: 99%

“…proposed a compressed sensing‐based algorithm for the 3D reconstruction of sparse objects, like guidewires and stents, from 16 simultaneous projections. This algorithm was later improved by Flach et al 22,23 . to handle patient motion.…”

Section: Introductionmentioning

confidence: 99%

Real‐time 3D reconstruction of guidewires and stents using two update X‐ray projections in a rotating imaging setup

Vöth

Koenig²,

Eulig

et al. 2023

Medical Physics

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the second category of more general tool reconstruction algorithms, the reconstruction of guidewires and stents from about 16 X-ray projections has been demonstrated. [7][8][9][10] Recently, the reconstruction of guidewires, stents and coils from only four X-ray projections has been demonstrated on synthetic data. 11,12 In this work, we improve this algorithm and demonstrate on measured data of an anthropomorphic phantom, that guidewires and stents can be reconstructed from only three X-ray projections.…”

Section: Introductionmentioning

confidence: 99%

3D reconstruction of stents and guidewires in an anthropomorphic phantom from three x-ray projections

Vöth

König²,

Eulig

et al. 2022

7th International Conference on Image Formation in X-Ray Computed Tomography

View full text Add to dashboard Cite

Today, the subcutaneous, minimally invasive procedures performed in interventional radiology are usually guided by 2D X-ray fluoroscopy. In 2D X-ray fluoroscopy a series of 2D X-ray images is displayed. For challenging procedures however, 3D X-ray fluoroscopy would be advantageous. In 3D X-ray fluoroscopy, a series of 3D images, which is reconstructed from a series of 2D X-ray images, is displayed. Because the number of images used for guiding an intervention is very high, little dose can be spent per 3D reconstruction of a 3D fluoroscopy. To save dose and to minimize motion artifacts, a reconstruction algorithm that requires very few X-ray projections is desirable. Earlier work showed that guidewires, stents and coils, which are commonly used in interventions, can be reconstructed using only four synthetic X-ray projections. The reconstruction from two or three X-ray projections was only studied briefly. In this work, we improve the method by using a more suitable neural network architecture and by using a multi-channel backprojection instead of a single-channel backprojection. We then apply the improved method to more realistic data measured in an anthropomorphic phantom. The results show that the method produces 3D reconstructions of stents and guidewires with submillimeter accuracy using only three measured X-ray projections.

show abstract

“…Thus a volume-to-volume (3D-3D) registration is applicable. The drawback of this technique is that a reconstruction of sparse data leads to severe image artifacts degrading image quality and thus the registration result (Flach et al 2013b). Alternatively one can compare the data in rawdata domain by applying a volume-to-rawdata (3D-2D) registration to exlude the influence of reconstruction artifacts from the registration process.…”

Section: Introductionmentioning

confidence: 99%

“…The comparison is done with our recently proposed low dose tomographic fluoroscopy as an example application. In low dose tomographic fluoroscopy temporally resolved 3D volumes are reconstructed showing interventional material like guide wires in relation to the surrounding tissue at the same dose level as todays projective fluoroscopy (Flach et al 2013a. A registration is applied in each time step for adjusting the patient's position.…”

Section: Introductionmentioning

confidence: 99%

Deformable 3D–2D registration for CT and its application to low dose tomographic fluoroscopy

et al. 2014

Self Cite

View full text Add to dashboard Cite

Many applications in medical imaging include image registration for matching of images from the same or different modalities. In the case of full data sampling, the respective reconstructed images are usually of such a good image quality that standard deformable volume-to-volume (3D-3D) registration approaches can be applied. But research in temporal-correlated image reconstruction and dose reductions increases the number of cases where rawdata are available from only few projection angles. Here, deteriorated image quality leads to non-acceptable deformable volume-to-volume registration results. Therefore a registration approach is required that is robust against a decreasing number of projections defining the target position. We propose a deformable volume-to-rawdata (3D-2D) registration method that aims at finding a displacement vector field maximizing the alignment of a CT volume and the acquired rawdata based on the sum of squared differences in rawdata domain. The registration is constrained by a regularization term in accordance with a fluid-based diffusion. Both cost function components, the rawdata fidelity and the regularization term, are optimized in an alternating manner. The matching criterion is optimized by a conjugate gradient descent for nonlinear functions, while the regularization is realized by convolution of the vector fields with Gaussian kernels. We validate the proposed method and compare it to the demons algorithm, a well-known 3D-3D registration method. The comparison is done for a range of 4-60 target projections using datasets from low dose tomographic fluoroscopy as an application example. The results show a high correlation to the ground truth target position without introducing artifacts even in the case of very few projections. In particular the matching in the rawdata domain is improved compared to the 3D-3D registration for the investigated range. The proposed volume-to-rawdata registration increases the robustness regarding sparse rawdata and provides more stable results than volume-to-volume approaches. By applying the proposed registration approach to low dose tomographic fluoroscopy it is possible to improve the temporal resolution and thus to increase the robustness of low dose tomographic fluoroscopy.

show abstract

Low dose tomographic fluoroscopy: 4D intervention guidance with running prior

Abstract: The running prior improves the robustness of low dose 3D+T intervention guidance toward intended or unintended patient motion.

Cited by 4 publications

References 46 publications

Real‐time 3D reconstruction of guidewires and stents using two update X‐ray projections in a rotating imaging setup

Real‐time 3D reconstruction of guidewires and stents using two update X‐ray projections in a rotating imaging setup

3D reconstruction of stents and guidewires in an anthropomorphic phantom from three x-ray projections

Deformable 3D–2D registration for CT and its application to low dose tomographic fluoroscopy

Contact Info

Product

Resources

About