Patient-Specific Articulated Digital Twins from a Single Full-Body CT Scan

Boyang Zhao; Han Zhang; Mathias Unberath

arxiv: 2607.02156 · v1 · pith:6WJXLDWXnew · submitted 2026-07-02 · 💻 cs.CV

Patient-Specific Articulated Digital Twins from a Single Full-Body CT Scan

Han Zhang , Boyang Zhao , Mathias Unberath This is my paper

Pith reviewed 2026-07-03 15:22 UTC · model grok-4.3

classification 💻 cs.CV

keywords patient-specific modelingarticulated digital twinCT imagingpose retargetingSMPL fittingkinematic scaffoldradiographic simulationanatomical modeling

0 comments

The pith

A single full-body CT scan can generate a patient-specific articulated digital twin that moves while preserving the individual's skeletal geometry.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper demonstrates a method to convert one static CT scan into a model of a patient's body that can be repositioned to different poses. This addresses the limitation that most CT models remain fixed at the scan position, even though real anatomy changes appearance with movement. The process starts by fitting a parametric body model to align a kinematic structure with the scan, then attaches the actual segmented bones and organs to this structure. Once assembled, the twin supports pose changes and the generation of new radiographic images from those poses.

Core claim

The central claim is that fitting the SMPL parametric human body model to a full-body CT scan provides a kinematic scaffold to which segmented bones and organs can be bound, enabling retargeting of pose changes while preserving patient-specific skeletal geometry. This is shown through a proof-of-concept on three subjects, with quantitative measures of fit quality and image similarity in original and new poses.

What carries the argument

The patient-aligned kinematic scaffold obtained by fitting the SMPL parametric model to the CT scan, to which segmented bones and organs are bound via an anatomy-aware rig.

If this is right

The fitted scaffold achieves 15.8 mm chamfer distance and 95.9% skeletal enclosure on the test subjects.
Recomposition at the original acquisition pose produces DRRs with SSIM of 0.872 and PSNR of 18.5 dB.
Across unseen target poses the twin maintains 94.4% skeletal enclosure.
The articulated twin supports rendering of pose-dependent DRRs for synthetic imaging studies.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same single-scan construction could supply varied-pose training examples for machine-learning methods in radiographic image analysis.
If extended, the twins might allow pre-operative rehearsal of patient positioning for procedures that depend on exact body orientation.
Direct validation against real multi-pose CT pairs of the same individuals would test whether the retargeted geometry matches physical reality beyond the current metrics.

Load-bearing premise

The SMPL parametric model can be fitted accurately enough to the CT scan to serve as a kinematic scaffold that keeps the patient's unique skeletal geometry intact during pose retargeting.

What would settle it

Acquire a second full-body CT of one of the test subjects in a substantially different pose and measure whether the retargeted twin's bone positions and rendered DRRs match the new scan within the reported chamfer distance, enclosure, SSIM, and PSNR ranges.

Figures

Figures reproduced from arXiv: 2607.02156 by Boyang Zhao, Han Zhang, Mathias Unberath.

**Figure 1.** Figure 1: Overview. A single full-body CT scan (top left) is converted into a patientspecific articulated digital twin that fuses CT-derived skeletal anatomy with a fitted body envelope. The twin can be articulated into diverse unseen poses (bottom) and rendered as pose-dependent synthetic radiographs (top right). SMPL body model [9] that provides an articulated kinematic scaffold. Given a full-body CT volume, we f… view at source ↗

**Figure 2.** Figure 2: Overview of the pipeline. Starting from this initialization, we perform multi-phase optimization to refine the SMPL shape and pose parameters. The articulated body model is optimized by minimizing (R ∗ , T ∗ , β∗ , θ∗ ) = arg min R,T ,β,θ (λchamferLchamfer + λinclusionLinclusion + λpriorLprior), where Lchamfer enforces geometric consistency between the SMPL body surface and the CT-derived body surface, L… view at source ↗

**Figure 3.** Figure 3: (a) Body fitting: the fitted body envelope encloses the CT-derived skeletal anatomy, shown alongside the original CT. (b) Synthetic X-ray: DRRs rendered from the raw CT, the recomposed digital twin at the acquisition pose, and the articulated twin under an unseen target pose. (c) Pose manipulation: examples of the patientspecific twin articulated into diverse target poses. volumes, the twin is voxelized i… view at source ↗

read the original abstract

Patient-specific anatomical models provide individualized context for surgical planning, image-guided intervention, and algorithm development. However, most CT-derived models are static: they preserve the body configuration captured at scan time, but cannot represent how the same anatomy would appear after patient repositioning. This limitation is especially important for radiographic imaging, where appearance depends jointly on imaging geometry and patient pose. We present a proof-of-concept for constructing a patient-specific articulated digital twin from a single full-body CT scan. The method fits a parametric human body model (SMPL) to obtain a patient-aligned kinematic scaffold, binds segmented bones and organs to an anatomy-aware rig, and retargets body-pose changes while preserving skeletal geometry. On three full-body CT subjects, the fitted scaffold achieved 15.8 $\pm$ 4.0 mm chamfer distance and 95.9 $\pm$ 1.8% skeletal enclosure. Recomposition at the acquisition pose preserved major radiographic structure, with overall SSIM of 0.872 $\pm$ 0.016 and PSNR of 18.5 $\pm$ 1.4 dB across paired DRRs. Across unseen target poses, the resulting twins enabled articulation while maintaining high skeletal enclosure (94.4 $\pm$ 0.4%). As a feasibility demonstration, we render the articulated twin as pose-dependent DRRs. These results suggest the feasibility of extending static, view-controllable CT simulation toward pose-controllable anatomical twins for future synthetic imaging and positioning studies.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The work gives a practical pipeline for patient-specific poseable anatomical models but the support for geometry preservation after retargeting rests on thin evidence.

read the letter

This paper demonstrates a pipeline for creating pose-controllable digital twins from one CT scan, but the evidence that it preserves patient-specific geometry in new poses is not strong enough yet.

They take a full-body CT, fit an SMPL model to get the skeleton pose and shape, segment the bones and organs from the scan, bind those segments to the SMPL rig, and then show they can change the pose and render new DRRs. On the three test cases the original pose DRRs match the real ones with SSIM of 0.87 and the skeleton stays mostly inside the original volume when posed differently.

The combination is new as an application, and the fact that they include organs and show radiographic output is a plus. The method is described clearly enough in the abstract to understand the steps.

The main issue is the validation for unseen poses. Skeletal enclosure at 94% is reported, but that only says the bones are still roughly in the body outline; it does not check if their relative positions or the organ placements are anatomically correct or if the rendered images would look right. Since there is no ground truth for the target poses, they cannot run the same SSIM or PSNR tests. The initial fit error of 15.8 mm chamfer also leaves room for the binding to drift. Three subjects is a very small sample for claiming feasibility across patients.

This kind of work is for people building tools for surgical simulation or generating training data for pose-aware imaging algorithms. A reader working on medical digital twins would get some ideas from the pipeline even if the results need more testing.

I think it should go to peer review. The idea is practical and the current results are honest about being a proof of concept, so referees can point out what additional checks would make the claim solid.

Referee Report

2 major / 1 minor

Summary. The manuscript presents a proof-of-concept for constructing patient-specific articulated digital twins from a single full-body CT scan. The method fits the SMPL parametric model to obtain a patient-aligned kinematic scaffold, binds segmented bones and organs to an anatomy-aware rig, and retargets body-pose changes while preserving skeletal geometry. On three subjects, it reports chamfer distance of 15.8 ± 4.0 mm and 95.9 ± 1.8% skeletal enclosure for the fit, SSIM of 0.872 ± 0.016 and PSNR of 18.5 ± 1.4 dB for DRRs at the acquisition pose, and 94.4 ± 0.4% skeletal enclosure for unseen target poses, with rendered pose-dependent DRRs as a feasibility demonstration.

Significance. If the central claim holds under stronger validation, the work could enable pose-controllable anatomical models from static CTs, with potential value for surgical planning, image-guided interventions, and synthetic radiographic datasets. The quantitative metrics on fitting accuracy and acquisition-pose DRR similarity, combined with the use of an established parametric model (SMPL) for the scaffold, provide initial evidence of technical feasibility on a small cohort.

major comments (2)

[Abstract] Abstract: The claim that retargeting preserves patient-specific skeletal geometry under arbitrary pose changes rests on skeletal enclosure of 94.4 ± 0.4% for unseen poses, but no ground-truth DRR or geometric fidelity metrics (e.g., organ/bone relative positions) are provided for target poses, in contrast to the acquisition-pose results (SSIM/PSNR). Skeletal enclosure alone does not confirm maintenance of radiographic appearance or anatomical relationships.
[Methods] Methods (binding and retargeting description): No derivation details, error analysis, or validation protocol are given for how segmented bones and organs are bound to the SMPL-derived rig or how the anatomy-aware rig enforces constraints during retargeting. This leaves unclear whether the mechanism goes beyond the initial fit (chamfer distance 15.8 mm) to preserve patient-specific geometry.

minor comments (1)

[Abstract] Abstract: The dataset of three full-body CT subjects is not described (source, acquisition parameters, or subject diversity), which would help evaluate the reported standard deviations and generalizability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on this proof-of-concept manuscript. We address each major comment below.

read point-by-point responses

Referee: [Abstract] Abstract: The claim that retargeting preserves patient-specific skeletal geometry under arbitrary pose changes rests on skeletal enclosure of 94.4 ± 0.4% for unseen poses, but no ground-truth DRR or geometric fidelity metrics (e.g., organ/bone relative positions) are provided for target poses, in contrast to the acquisition-pose results (SSIM/PSNR). Skeletal enclosure alone does not confirm maintenance of radiographic appearance or anatomical relationships.

Authors: We agree that skeletal enclosure is a geometric proxy and does not substitute for direct radiographic metrics such as SSIM/PSNR or organ relative positions on target poses. Because the study uses single static CT acquisitions, no ground-truth multi-pose data exist for the same subjects, precluding those comparisons. We will revise the abstract to state the retargeting result more precisely as preservation of skeletal enclosure while explicitly noting the absence of full radiographic validation for unseen poses. revision: partial
Referee: [Methods] Methods (binding and retargeting description): No derivation details, error analysis, or validation protocol are given for how segmented bones and organs are bound to the SMPL-derived rig or how the anatomy-aware rig enforces constraints during retargeting. This leaves unclear whether the mechanism goes beyond the initial fit (chamfer distance 15.8 mm) to preserve patient-specific geometry.

Authors: The manuscript presents the binding and retargeting at a conceptual level. We accept that additional technical detail is warranted. In revision we will expand the Methods section with the mathematical formulation of the binding step, an error-propagation analysis for retargeting, and the explicit validation protocol used to confirm that the rig preserves the initial patient-specific fit beyond the reported Chamfer distance. revision: yes

Circularity Check

0 steps flagged

No circularity: procedural pipeline with independent metrics

full rationale

The paper presents a standard fitting-and-binding pipeline (SMPL fit to CT, bone/organ binding to rig, pose retargeting) whose outputs are evaluated by direct geometric and radiographic metrics (chamfer distance, skeletal enclosure, SSIM/PSNR on DRRs). These quantities are computed from the constructed models rather than being forced by redefinition or self-citation; the central claim is a feasibility demonstration of the pipeline itself, not a derived prediction that collapses to the input data by construction. No load-bearing self-citations, uniqueness theorems, or ansatzes appear in the provided text.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

Ledger extracted from abstract only. The claim rests on the accuracy of SMPL fitting and the validity of rigid binding to preserve geometry.

free parameters (1)

SMPL shape and pose parameters
Fitted to CT scan to align the kinematic scaffold with patient anatomy

axioms (2)

domain assumption SMPL provides a suitable kinematic scaffold for human anatomy
Invoked to obtain patient-aligned rig for retargeting
domain assumption Segmented bones and organs can be bound rigidly to the rig while preserving geometry
Required for articulation without deformation

pith-pipeline@v0.9.1-grok · 5805 in / 1342 out tokens · 38042 ms · 2026-07-03T15:22:00.540987+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

16 extracted references · 14 canonical work pages · 1 internal anchor

[1]

IEEE Transactions on Pattern Analysis and Machine Intelligence (2019)

Cao, Z., Hidalgo Martinez, G., Simon, T., Wei, S., Sheikh, Y.A.: Openpose: Real- time multi-person 2d pose estimation using part affinity fields. IEEE Transactions on Pattern Analysis and Machine Intelligence (2019)

2019
[2]

https://doi.org/10.48550/arXiv.2405.11133

Dahal, L., Ghojoghnejad, M., Ghosh, D., Bhandari, Y., Kim, D., Ho, F.C., Tushar, F.I.,Luoa,S.,Lafata,K.J.,Abadi,E.,Samei,E.,Lo,J.Y.,Segars,W.P.:XCAT-3.0: A Comprehensive Library of Personalized Digital Twins Derived from CT Scans (Sep.). https://doi.org/10.48550/arXiv.2405.11133

work page doi:10.48550/arxiv.2405.11133
[3]

https://doi.org/10.25827/5s8c-n515, https://doi.org/10.25827/5s8c-n515

Edgar, H.J.H., Daneshvari Berry, S., Moes, E., Adolphi, N.L., Bridges, P., Nolte, K.B.: New Mexico Decedent Image Database (2020). https://doi.org/10.25827/5s8c-n515, https://doi.org/10.25827/5s8c-n515

work page doi:10.25827/5s8c-n515 2020
[4]

Nature machine intelligence5, 294 – 308 (2023), https://api.semanticscholar.org/CorpusID:257668838

Gao, C., Killeen, B., Hu, Y., Grupp, R., Taylor, R.H., Armand, M., Unberath, M.: Synthetic data accelerates the development of generalizable learning-based algo- rithms for x-ray image analysis. Nature machine intelligence5, 294 – 308 (2023), https://api.semanticscholar.org/CorpusID:257668838

2023
[5]

arXiv preprint arXiv:2312.06358 (2023)

Gopalakrishnan, V., Dey, N., Golland, P.: Intraoperative 2D/3D image registration via differentiable x-ray rendering. arXiv preprint arXiv:2312.06358 (2023)

work page arXiv 2023
[6]

In: Medical Image Computing and Com- puter Assisted Intervention 2025

Killeen, B.D., Wang, L.J., Iñígo, B., Zhang, H., Armand, M., Taylor, R.H., Os- good, G., Unberath, M.: FluoroSAM: A Language-Promptable Foundation Model for Flexible X-Ray Image Segmentation. In: Medical Image Computing and Com- puter Assisted Intervention 2025. https://doi.org/10.1007/978-3-032-04981-0_24

work page doi:10.1007/978-3-032-04981-0_24 2025
[7]

https://doi.org/10.1007/978-3-031-43996-4_13

Killeen, B.D., Zhang, H., Mangulabnan, J., Armand, M., Taylor, R.H., Osgood, G., Unberath, M.: Pelphix: Surgical Phase Recognition from X-Ray Images in Per- cutaneous Pelvic Fixation. https://doi.org/10.1007/978-3-031-43996-4_13

work page doi:10.1007/978-3-031-43996-4_13
[8]

International Journal of Computer As- sisted Radiology and Surgery (2024)

Killeen, B.D., Zhang, H., Wang, L.J., Liu, Z., Kleinbeck, C., Rosen, M., Taylor, R.H., Osgood, G., Unberath, M.: Stand in surgeon’s shoes: virtual reality cross- training to enhance teamwork in surgery. International Journal of Computer As- sisted Radiology and Surgery (2024). https://doi.org/10.1007/s11548-024-03138-7

work page doi:10.1007/s11548-024-03138-7 2024
[9]

In: Seminal Graphics Papers: Pushing the Boundaries, Volume 2 (Aug.), https://dl.acm.org/doi/10.1145/3596711.3596800

Loper, M., Mahmood, N., Romero, J., Pons-Moll, G., Black, M.J.: SMPL: A Skinned Multi-Person Linear Model. In: Seminal Graphics Papers: Pushing the Boundaries, Volume 2 (Aug.), https://dl.acm.org/doi/10.1145/3596711.3596800

work page doi:10.1145/3596711.3596800
[10]

AMASS: Archive of Motion Capture as Surface Shapes

Mahmood, N., Ghorbani, N., Troje, N.F., Pons-Moll, G., Black, M.J.: AMASS: Archive of Motion Capture as Surface Shapes (Apr 2019). https://doi.org/10.48550/arXiv.1904.03278

work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1904.03278 2019
[11]

https://doi.org/10.1038/s41746-025-01575-5 10 H

Mekki, Y.M., Luijten, G., Hagert, E., Belkhair, S., Varghese, C., Qadir, J., So- laiman, B., Bilal, M., Dhanda, J., Egger, J., Deng, J., Khanduja, V., Frangi, A.F., Zughaier, S.M., Stotland, M.A.: Digital twins for the era of personalized surgery 8(1), 283 (May 2025). https://doi.org/10.1038/s41746-025-01575-5 10 H. Zhang et al

work page doi:10.1038/s41746-025-01575-5 2025
[12]

Journal of Clinical Imaging Sci- ence13, 34

Midtgaard, M., Pedersen, M.R.V., Christensen, N.L., McKnight, K.L., Jensen, J.: Patient positioning during the radiographic procedure affects the radiological signs of acetabular retroversion - A systematic review. Journal of Clinical Imaging Sci- ence13, 34. https://doi.org/10.25259/JCIS_82_2023

work page doi:10.25259/jcis_82_2023
[13]

Jour- nal of Clinical Medicine14(11) (2025)

Prządka, M., Pająk, W., Kleinrok, J., Pec, J., Michno, K., Karpiński, R., Baj, J.: Advances in 3d printing applications for personalized ortho- pedic surgery: From anatomical modeling to patient-specific implants. Jour- nal of Clinical Medicine14(11) (2025). https://doi.org/10.3390/jcm14113989, https://www.mdpi.com/2077-0383/14/11/3989

work page doi:10.3390/jcm14113989 2025
[14]

https://doi.org/10.48550/arXiv.2303.04923

Shetty, K., Birkhold, A., Jaganathan, S., Strobel, N., Egger, B., Kowarschik, M., Maier, A.: BOSS: Bones, Organs and Skin Shape Model (Mar 2023). https://doi.org/10.48550/arXiv.2303.04923

work page doi:10.48550/arxiv.2303.04923 2023
[15]

In: Medical Image Computing and Computer Assisted Intervention

Unberath, M., Zaech, J.N., Lee, S.C., Bier, B., Fotouhi, J., Armand, M., Navab, N.: DeepDRR – A Catalyst for Machine Learning in Fluoroscopy-Guided Procedures. In: Medical Image Computing and Computer Assisted Intervention. Cham (2018). https://doi.org/10.1007/978-3-030-00937-3_12

work page doi:10.1007/978-3-030-00937-3_12 2018
[16]

Ra- diology: Artificial Intelligence5(5) (2023)

Wasserthal, J., Breit, H.C., Meyer, M.T., Pradella, M., Hinck, D., Sauter, A.W., Heye, T., Boll, D.T., Cyriac, J., Yang, S., Bach, M., Segeroth, M.: Totalseg- mentator: Robust segmentation of 104 anatomic structures in ct images. Ra- diology: Artificial Intelligence5(5) (2023). https://doi.org/10.1148/ryai.230024, http://dx.doi.org/10.1148/ryai.230024

work page doi:10.1148/ryai.230024 2023

[1] [1]

IEEE Transactions on Pattern Analysis and Machine Intelligence (2019)

Cao, Z., Hidalgo Martinez, G., Simon, T., Wei, S., Sheikh, Y.A.: Openpose: Real- time multi-person 2d pose estimation using part affinity fields. IEEE Transactions on Pattern Analysis and Machine Intelligence (2019)

2019

[2] [2]

https://doi.org/10.48550/arXiv.2405.11133

Dahal, L., Ghojoghnejad, M., Ghosh, D., Bhandari, Y., Kim, D., Ho, F.C., Tushar, F.I.,Luoa,S.,Lafata,K.J.,Abadi,E.,Samei,E.,Lo,J.Y.,Segars,W.P.:XCAT-3.0: A Comprehensive Library of Personalized Digital Twins Derived from CT Scans (Sep.). https://doi.org/10.48550/arXiv.2405.11133

work page doi:10.48550/arxiv.2405.11133

[3] [3]

https://doi.org/10.25827/5s8c-n515, https://doi.org/10.25827/5s8c-n515

Edgar, H.J.H., Daneshvari Berry, S., Moes, E., Adolphi, N.L., Bridges, P., Nolte, K.B.: New Mexico Decedent Image Database (2020). https://doi.org/10.25827/5s8c-n515, https://doi.org/10.25827/5s8c-n515

work page doi:10.25827/5s8c-n515 2020

[4] [4]

Nature machine intelligence5, 294 – 308 (2023), https://api.semanticscholar.org/CorpusID:257668838

Gao, C., Killeen, B., Hu, Y., Grupp, R., Taylor, R.H., Armand, M., Unberath, M.: Synthetic data accelerates the development of generalizable learning-based algo- rithms for x-ray image analysis. Nature machine intelligence5, 294 – 308 (2023), https://api.semanticscholar.org/CorpusID:257668838

2023

[5] [5]

arXiv preprint arXiv:2312.06358 (2023)

Gopalakrishnan, V., Dey, N., Golland, P.: Intraoperative 2D/3D image registration via differentiable x-ray rendering. arXiv preprint arXiv:2312.06358 (2023)

work page arXiv 2023

[6] [6]

In: Medical Image Computing and Com- puter Assisted Intervention 2025

Killeen, B.D., Wang, L.J., Iñígo, B., Zhang, H., Armand, M., Taylor, R.H., Os- good, G., Unberath, M.: FluoroSAM: A Language-Promptable Foundation Model for Flexible X-Ray Image Segmentation. In: Medical Image Computing and Com- puter Assisted Intervention 2025. https://doi.org/10.1007/978-3-032-04981-0_24

work page doi:10.1007/978-3-032-04981-0_24 2025

[7] [7]

https://doi.org/10.1007/978-3-031-43996-4_13

Killeen, B.D., Zhang, H., Mangulabnan, J., Armand, M., Taylor, R.H., Osgood, G., Unberath, M.: Pelphix: Surgical Phase Recognition from X-Ray Images in Per- cutaneous Pelvic Fixation. https://doi.org/10.1007/978-3-031-43996-4_13

work page doi:10.1007/978-3-031-43996-4_13

[8] [8]

International Journal of Computer As- sisted Radiology and Surgery (2024)

Killeen, B.D., Zhang, H., Wang, L.J., Liu, Z., Kleinbeck, C., Rosen, M., Taylor, R.H., Osgood, G., Unberath, M.: Stand in surgeon’s shoes: virtual reality cross- training to enhance teamwork in surgery. International Journal of Computer As- sisted Radiology and Surgery (2024). https://doi.org/10.1007/s11548-024-03138-7

work page doi:10.1007/s11548-024-03138-7 2024

[9] [9]

In: Seminal Graphics Papers: Pushing the Boundaries, Volume 2 (Aug.), https://dl.acm.org/doi/10.1145/3596711.3596800

Loper, M., Mahmood, N., Romero, J., Pons-Moll, G., Black, M.J.: SMPL: A Skinned Multi-Person Linear Model. In: Seminal Graphics Papers: Pushing the Boundaries, Volume 2 (Aug.), https://dl.acm.org/doi/10.1145/3596711.3596800

work page doi:10.1145/3596711.3596800

[10] [10]

AMASS: Archive of Motion Capture as Surface Shapes

Mahmood, N., Ghorbani, N., Troje, N.F., Pons-Moll, G., Black, M.J.: AMASS: Archive of Motion Capture as Surface Shapes (Apr 2019). https://doi.org/10.48550/arXiv.1904.03278

work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1904.03278 2019

[11] [11]

https://doi.org/10.1038/s41746-025-01575-5 10 H

Mekki, Y.M., Luijten, G., Hagert, E., Belkhair, S., Varghese, C., Qadir, J., So- laiman, B., Bilal, M., Dhanda, J., Egger, J., Deng, J., Khanduja, V., Frangi, A.F., Zughaier, S.M., Stotland, M.A.: Digital twins for the era of personalized surgery 8(1), 283 (May 2025). https://doi.org/10.1038/s41746-025-01575-5 10 H. Zhang et al

work page doi:10.1038/s41746-025-01575-5 2025

[12] [12]

Journal of Clinical Imaging Sci- ence13, 34

Midtgaard, M., Pedersen, M.R.V., Christensen, N.L., McKnight, K.L., Jensen, J.: Patient positioning during the radiographic procedure affects the radiological signs of acetabular retroversion - A systematic review. Journal of Clinical Imaging Sci- ence13, 34. https://doi.org/10.25259/JCIS_82_2023

work page doi:10.25259/jcis_82_2023

[13] [13]

Jour- nal of Clinical Medicine14(11) (2025)

Prządka, M., Pająk, W., Kleinrok, J., Pec, J., Michno, K., Karpiński, R., Baj, J.: Advances in 3d printing applications for personalized ortho- pedic surgery: From anatomical modeling to patient-specific implants. Jour- nal of Clinical Medicine14(11) (2025). https://doi.org/10.3390/jcm14113989, https://www.mdpi.com/2077-0383/14/11/3989

work page doi:10.3390/jcm14113989 2025

[14] [14]

https://doi.org/10.48550/arXiv.2303.04923

Shetty, K., Birkhold, A., Jaganathan, S., Strobel, N., Egger, B., Kowarschik, M., Maier, A.: BOSS: Bones, Organs and Skin Shape Model (Mar 2023). https://doi.org/10.48550/arXiv.2303.04923

work page doi:10.48550/arxiv.2303.04923 2023

[15] [15]

In: Medical Image Computing and Computer Assisted Intervention

Unberath, M., Zaech, J.N., Lee, S.C., Bier, B., Fotouhi, J., Armand, M., Navab, N.: DeepDRR – A Catalyst for Machine Learning in Fluoroscopy-Guided Procedures. In: Medical Image Computing and Computer Assisted Intervention. Cham (2018). https://doi.org/10.1007/978-3-030-00937-3_12

work page doi:10.1007/978-3-030-00937-3_12 2018

[16] [16]

Ra- diology: Artificial Intelligence5(5) (2023)

Wasserthal, J., Breit, H.C., Meyer, M.T., Pradella, M., Hinck, D., Sauter, A.W., Heye, T., Boll, D.T., Cyriac, J., Yang, S., Bach, M., Segeroth, M.: Totalseg- mentator: Robust segmentation of 104 anatomic structures in ct images. Ra- diology: Artificial Intelligence5(5) (2023). https://doi.org/10.1148/ryai.230024, http://dx.doi.org/10.1148/ryai.230024

work page doi:10.1148/ryai.230024 2023