Evaluating non-contrast 3D reconstructions for pulmonary segmentectomy: a reliable alternative to contrast-enhanced models in thoracic surgery

Introduction

Thoracoscopic lobectomy is the standard treatment for early-stage non-small cell lung cancer (NSCLC) (1). However, sublobar resections, particularly segmentectomy, are increasingly gaining prominence due to their advantages in preserving pulmonary function, maintaining quality of life, and accommodating patients with comorbidities or smaller tumors. This shift is particularly relevant given the growing proportion of older patients and the need for surgical interventions in cases of second primary cancers. Recent clinical trials, such as JCOG 0802 and CALBG 140503, have demonstrated the oncological equivalence of sublobar resections compared to lobectomy for early-stage NSCLC (2,3). These studies highlight the benefits of segmentectomy in terms of survival and reduced morbidity, albeit with a slightly increased risk of local recurrence. Consequently, this procedure is increasingly recognized as a valuable option for select patients.

Segmentectomy requires precise understanding of the vascular and bronchial anatomy, as anatomical variations are frequent and can pose significant challenges during surgery. To optimize outcomes, the European Society of Thoracic Surgeons (ESTS) 2023 guidelines strongly recommend the use of preoperative three-dimensional (3D) reconstructions (4). These reconstructions help identify anatomical variations, accurately locate tumors, and ensure adequate resection margins. Typically derived from intravenous (contrast-enhanced) (IV) computed tomography (CT) scans, such models have been shown to improve surgical outcomes by reducing operative time, blood loss, and complications. However, the use of contrast agents presents certain drawbacks, including risks of allergic reactions, renal impairment, increased radiation exposure, and higher financial costs. Additionally, logistical challenges, such as the need to repeat scans, can further complicate the preoperative pathway for patients.

To address these limitations, Visible Patient, renowned for the quality, reliability, and proven utility of its IV 3D models (5), has developed a novel 3D reconstruction technique based on non-intravenous (non-contrast imaging) (NIV) CT scans. The objective of this study is to evaluate the anatomical equivalence of these NIV models compared to their IV counterparts and assess their impact on therapeutic decision-making for pulmonary segmentectomy. We present this article in accordance with the SUPER reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-716/rc).

Methods

This mixed-designed study with prospective enrollment and retrospective data analysis was conducted in the thoracic department of Institut Mutualiste Montsouris between December 15, 2023 and June 15, 2024, involving adult patients scheduled for segmentectomy. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Ethical approval was not required according to institutional policy, as the study involved retrospective analysis of anonymized data without any intervention or modification of patient care. Informed consent was obtained from all patients.

Inclusion criteria

Patients scheduled for minimally invasive pulmonary segmentectomy for solid, mixed nodules, or ground-glass opacities (GGOs) peripheral and clinical stage I for NSCLC were included. Eligibility required the feasibility of achieving a safe resection margin greater than or equal to the nodule diameter. Both benign and malignant nodules were permitted, with no restrictions on age or gender.

Preoperative surgical simulation

Each patient underwent two CT scans, one NIV-based and one IV-based. The NIV-based CT scan was performed as part of the initial diagnostic workup for the disease, while the IV-based CT scan was conducted within one month prior to surgery in accordance with best practice guidelines. Only the IV-based model was used for preoperative planning, as the NIV-based model was not made available for visualization until 48 hours after the surgery.

Using 3D reconstruction imaging, we can identify the precise location of the nodules and their respective lung segments. The location of the tumor determined the extent of resection (single segment or multiple segments) to ensure safe oncological margins (margin/tumor ratio >1). We verified the target segmental bronchus, arteries, segmental and intersegmental veins, subsequently planning the surgical approach. The quantification of the pulmonary nodule’s maximum diameter was conducted via the CT lung window.

The Visible Patient 3D modeling algorithm is based on convolutional neural networks (CNNs) trained on a large, curated database of annotated CT scans, including 188 manually segmented NIV datasets. Initially validated on IV CT scans, it has since been adapted for NIV acquisitions. Its performance has been externally validated against expert manual annotations, with reported Dice similarity coefficients exceeding 0.9 for lobar and segmental bronchovascular structures in both IV and NIV datasets. The system is integrated into the Visible Patient Suite 1.2 (CE0459 certified under regulation 2017/745) for routine clinical application. All reconstructions were delivered as finalized models by Visible Patient and used without manual modifications by the investigators.

Surgical technique and procedure

Thoracoscopic approach

All procedures were performed by experienced thoracic surgeons under general anesthesia with double-lumen endotracheal intubation. A full thoracoscopic approach, either video-assisted thoracoscopic surgery (VATS) or robot-assisted thoracoscopic surgery (RATS), was employed to maximize functional preservation, as recommended by ESTS guidelines for early stage lung cancer (4).

The fissure-first technique was utilized to provide comprehensive visualization of vascular anatomical variations within the fissure, enhancing surgical precision (6). Preoperative 3D reconstructions based on IV CT scans were made available in the operating room to guide the surgical procedure. Following the preoperative 3D surgical plan, the targeted arteries, veins, and bronchi were transected while preserving the intersegmental veins. To delineate the intersegmental plane (ISP), 2 to 5 mL of indocyanine green (ICG) was injected intravenously, using a near-infrared fluorescence (NIF) camera. Stapling was used to divide the ISP as recommended by the ESTS and to avoid prolonged air leakage (4).

Management of margins and lymph nodes

In cases of confirmed or suspected primary lung malignancy, a systematic lymph node dissection was performed, following International Association for the Study of Lung Cancer (IASLC) guidelines (7,8). A frozen section on the segmental nodes was performed in all cases (except for patients unfit for lobectomy), and if a positive node was identified during the frozen section examination, the segmentectomy was extended to a lobectomy (4).

The ISPs were defined based on the anatomy of the segment and identified using an NIF camera with 2 to 5 mL ICG systemic injection. If the margins between the ISP and the nodule was less than 1 cm, an additional parenchymal resection was performed to achieve an adequate oncologic margin (4). The margin was measured intraoperatively, and if its adequacy was uncertain, frozen-section pathological analysis was used for verification. In case of positive margins at intraoperative pathological examination, the resection was extended to the adjacent parenchyma (up to a complete lobectomy). This approach ensured compliance with oncologic principles while minimizing unnecessary tissue loss.

Comparing IV and NIV model

Forty-eight hours after the surgery, the NIV-based model was made available, and the surgeon evaluated its concordance with the IV-based model and intraoperative findings. Anatomical variations were documented in a comparative table (Table 1), highlighting any discrepancies between the models and the intraoperative anatomical observations. In cases of inconsistency between the two models, a second surgeon, blinded to the patient and surgical details, independently reviewed the NIV-based model to assess its potential impact on surgical indications and procedural feasibility. Concordance was defined as full agreement between IV and NIV models for the identification of major vascular and bronchial structures. Partial overlap or minor discrepancies were not considered equivalent. These findings were also documented in Table 1.

Statistical analysis

Anatomical concordance between NIV and IV 3D models was expressed as a percentage, with 95% confidence intervals (CIs) calculated using the Wilson score method. Minor discrepancies not impacting surgical planning were noted separately. Descriptive statistics summarized patient characteristics and outcomes.

Results

Twenty-four patients were included in this study December 15, 2023 and June 15, 2024. The analysis compared vascular and bronchial anatomical variations between IV and NIV 3D reconstructions, with an overall concordance rate of 83.3% (20/24 cases, 95% CI: 0.62–0.95).

Intraoperative and postoperative data

Thoracoscopic segmentectomies were successfully performed for all individuals, with no need for conversion to thoracotomy, lobectomy, resection of adjacent segments, or additional parenchymal resection due to insufficient margins. All surgical margins were within acceptable limits according to the established surgical parameters.

Anatomical variations identified in both models

In 20 cases (83.3%), the IV and NIV models showed equivalent identification of major vascular anatomical variations:

• Arterial variations: mediastinal lingular artery A4+5 (n=2) (Figure 1), right common trunk A2+6 (n=2) (Figure 2), and left common trunks A2+3 (n=1) and A3+4+5 (n=2) (Figure 1).
• Venous variations: mixed lingular venous drainage patterns V3+4 and V4+5 (n=1), left common veins V2+4 (n=1), V2+6 (n=1), V3+4 (n=1), and right mediastinal recurrent V2 (n=1).
• Segmental variations: shared identification of S6* anatomical variants A6*, V6*, and B6* (n=1).

Figure 1 3D reconstructions of the left pulmonary arterial and bronchial trees: mediastinal lingular artery and common trunk A3+4+5 (n=2): similar IV (A) and NIV (B) models. 3D, three-dimensional; IV, intravenous (contrast-enhanced); NIV, non-intravenous (non-contrast imaging).

Figure 2 3D reconstructions of the right pulmonary arterial and bronchial trees: common trunk A2+A6: similar IV (A) and NIV (B) models. 3D, three-dimensional; IV, intravenous (contrast-enhanced); NIV, non-intravenous (non-contrast imaging).

Anatomical variations between the IV and NIV models

Minor differences were noted in 4 cases (16.7%), primarily involving small vascular structures:

• Arterial variation: an additional small segmental artery of the basal pyramid was visualized only in the IV model for two patients (n=2), and only in the NIV model for on patients (n=1) (Figure 3).
• Venous variation: a segmental left lower vein draining into the lingular vein was identified exclusively in the IV model for one patient (n=1).

Figure 3 3D reconstructions of the left pulmonary arterial and bronchial trees: anatomical variation: small additional segmental artery of the basal pyramid (arrow) visible on NIV models (A), but not on the IV models (B). 3D, three-dimensional; IV, intravenous (contrast-enhanced); NIV, non-intravenous (non-contrast imaging).

No anatomical differences affected the surgical strategy, and no major anatomical discrepancies were identified between the models.

Discussion

Lung cancer, particularly NSCLC, remains the leading cause of cancer-related deaths worldwide (9). The prognosis for patients with early-stage NSCLC has improved due to advancements in imaging technology, minimally invasive surgical techniques, and personalized treatment approaches. Historically, lobectomy with systematic lymph node dissection has been the standard treatment for early-stage NSCLC (1). However, recent studies, such as JCOG0802 and CALGB 140503, have demonstrated that minimally invasive segmentectomy can provide oncological outcomes comparable to lobectomy for tumors smaller than 2 cm while better preserving lung function (2,3). Segmentectomy has gained acceptance due to these trials, which highlighted similar disease-free and overall survival rates for segmentectomy compared to lobectomy, alongside superior lung function preservation. However, segmentectomy is technically more challenging than lobectomy (4). It requires precise preoperative planning, including the localization of nodules, identification of anatomical variations in segmental arteries, veins, and bronchi, and accurate delineation of the ISP. Advanced imaging modalities like 3D reconstruction and NIF imaging are crucial for this meticulous planning (4).

3D reconstruction in pulmonary segmentectomy: enhanced precision and surgical outcomes

3D reconstruction converts conventional two-dimensional (2D) CT images into detailed, patient-specific 3D models. This advanced imaging technology provides precise visualization of pulmonary anatomy, enabling accurate tumor localization, clarification of segment boundaries, identification of anatomical variations, and planning of resection margins. It supports step-by-step preparation for minimally invasive pulmonary segmentectomy. Studies have demonstrated that 3D reconstruction improves surgical outcomes compared to reliance on 2D imaging alone. Reported benefits include shorter operative time, reduced intraoperative bleeding, improved lymph node sampling, fewer postoperative complications, and decreased duration of chest drainage and hospital stay. These findings highlight its role in enhancing surgical efficiency and promoting faster postoperative recovery. Traditional 2D-CT imaging lacks the anatomical detail required for precise segmentectomy planning. In contrast, 3D reconstruction allows surgeons to anticipate complex anatomical variations, such as mediastinal lingular arteries in left upper lobe segmentectomies. It also provides a clear understanding of spatial relationships between the pulmonary nodule, segmental vessels, and bronchi, reducing the risk of incomplete resection margins (10-13).

NIF imaging with ICG

While 3D reconstruction provides valuable insights, integrating 3D and NIF technology further enhances surgical precision (14,15). Accurate identification of the ISP is crucial for segmentectomy, ensures that an oncological safe distance of tumor from resection margins is achieved without unnecessary resection of adjacent lung tissue pertinent to other segments. NIF imaging, combined with intravenous ICG injection, provides a clear demarcation of segmental boundaries. Compared to the inflation-deflation (ID) method, NIF imaging offers faster and more precise visualization (90–95%) (14-16), especially in patients with chronic obstructive pulmonary disease (COPD) or emphysema. It reduces operation time, complications, and air leak rates.

Applications beyond surgical planning

3D reconstruction has transformative applications beyond preoperative planning. For instance, they are invaluable educational tools, enhancing trainee understanding of thoracic anatomy and surgical techniques. Virtual reality (VR) and augmented reality (AR) platforms incorporating 3D models provide immersive training experiences, further bridging the gap between theoretical knowledge and practical expertise (17-20).

NIV vs. IV model

Most of the existing evidence supporting 3D reconstructions is based on IV models. These models offer superior vascular clarity, making them the gold standard for preoperative planning. However, IV-based reconstructions carry inherent limitations, including risks of allergic reactions, renal impairment, increased radiation exposure, and higher financial costs. Additionally, they often involve logistical complexities, such as scheduling additional scans, which can delay the preoperative process and complicate the patient care pathway. In Institut Mutualiste Montsouris, the use of NIV models meant that reconstructions were available earlier, since they were generated directly from the diagnostic CT scan without requiring a second examination.

This study evaluated the anatomical equivalence of NIV 3D models generated by Visible Patient compared to their IV counterparts. The findings confirm the high reliability of NIV reconstructions, which achieved an 83.3% concordance rate with IV-based models. This demonstrates their viability for preoperative planning, particularly for patients contraindicated for contrast agents. Minor anatomical differences were observed in 16.7% of cases, primarily involving small intraparenchymal vascular structures. These variations, described by the operating surgeon as likely small vessels not visible intraoperatively, had no impact on surgical decision-making. Although none of the anatomical differences identified in this study impacted surgical strategy, it is important to acknowledge their potential clinical implications. Missed identification of small segmental vessels could theoretically increase the risk of unexpected intraoperative bleeding, compromise the accuracy of ISP division, or contribute to postoperative complications such as prolonged air leaks. These risks underline the need for surgeons to remain vigilant, to carefully correlate preoperative imaging with intraoperative findings, and to anticipate anatomical variations even when using advanced 3D reconstruction tools. While our experience did not reveal any adverse outcomes, these considerations reinforce that NIV models should complement, but not replace, surgical expertise and intraoperative judgment. While IV models remain the reference for optimal anatomical precision, particularly at the subsegmental level, our findings suggest that NIV models can provide sufficient accuracy for most standard anatomical segmentectomies, such as lingular, superior (S6), or basilar resections. Their use may be most relevant in patients with contraindications to contrast agents, where a non-contrast alternative is preferable. In contrast, IV-based reconstructions are still recommended for highly complex or subsegmental procedures, where maximum vascular clarity is required.

Notably, this study highlights the ability of NIV models to reliably identify significant anatomical variants, such as the mediastinal lingular artery, a well-documented challenge in left upper lobe segmentectomy. This reinforces their potential utility in complex surgical cases.

Other companies have also developed validated IV-based 3D reconstruction models that are widely used in clinical practice, recognized for their high anatomical accuracy. Some of these providers are reportedly exploring NIV 3D models; however, no published results are currently available to validate their performance or reliability. Future publications could provide valuable insights into the methodologies and accuracy of competing NIV models, broadening their acceptance and standardization.

Additionally, several lung cancer screening programs are under evaluation in France, aiming for nationwide implementation in individuals at risk of NSCLC. These programs are based on low-dose computed tomography (LDCT) scans, which could, in the near future, serve as a foundation for preoperative 3D reconstructions (20). LDCT-based 3D modeling could streamline preoperative planning by eliminating the need for IV CT scans, reducing costs, and minimizing patient risks associated with contrast agents, while also facilitating the patient’s preoperative pathway. However, it is important to note that the present study was conducted on diagnostic-quality CT scans, and the performance of the NIV algorithm on LDCT datasets remains to be formally evaluated. Given the reduced spatial resolution of LDCT compared with standard diagnostic scans, small-caliber vessels or subtle bronchial branches may be more challenging to reconstruct accurately. This could potentially reduce sensitivity for detecting minor anatomical variations. Therefore, while the integration of NIV models into screening programs is promising, further validation specifically on LDCT acquisitions will be necessary before their routine implementation in this setting.

Despite the growing reliance on 3D reconstructions, their application—whether IV or NIV—requires caution. Surgeons must remain vigilant for potential misinterpretations, particularly in cases involving complex anatomical variations. Advanced planning technologies continue to evolve, but they should complement, not replace, surgical expertise and intraoperative assessments. The main limitation of this work is the relatively small sample size (n=24), inherent to its pilot and single-center design. Nevertheless, the concordance proportion of 83.3% remained statistically robust, with a 95% CI ranging from 0.62 to 0.95. These results suggest that, despite the limited cohort, the observed agreement is unlikely to be due to chance alone. Larger multicenter studies are warranted to validate these findings and confirm their generalizability. Future research should aim to validate NIV models across larger patient cohorts and multiple centers. Comparative studies of methodologies from different providers will also be essential to refine and improve the reliability of NIV reconstructions. Although IV-based models currently remain the gold standard, NIV models address critical limitations, including contrast-related risks and logistical inefficiencies. This study serves as a key step forward in demonstrating the viability of NIV reconstructions for thoracic surgical planning, paving the way for broader adoption in routine clinical practice.

Conclusions

This study confirms the reliability of Visible Patient NIV 3D reconstructions as a viable alternative to IV models for preoperative planning in thoracic surgery. NIV models achieved an 83.3% concordance rate with IV reconstructions, effectively identifying major vascular and bronchial variations without impacting surgical decisions. NIV reconstructions are an attractive option, particularly in patients contraindicated for contrast agents. The ability of NIV models to detect critical variants, such as the mediastinal lingular artery, supports their clinical utility. Integration of them into preoperative workflows may streamline patient management and reduce the need for additional IV CT scans, a perspective of particular relevance with the expansion of lung cancer screening programs. Caution remains warranted, as minor omissions are possible, and imaging must complement surgical expertise and intraoperative findings. Larger multicenter studies are required to validate these results and define the role of NIV models across different clinical contexts.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the SUPER reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-716/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-716/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-716/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-716/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Ethical approval was not required according to institutional policy, as the study involved retrospective analysis of anonymized data without any intervention or modification of patient care. Informed consent was obtained from all patients.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. National Comprehensive Cancer Network. National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology. Non-Small Cell Lung Cancer Version 2. 2018. Accessed 23 Feb 2019. Available online: https://www.nccn.org/guidelines/guidelines-detail?category=1&id=1450
2. Saji H, Okada M, Tsuboi M, et al. Segmentectomy versus lobectomy in small-sized peripheral non-small-cell lung cancer (JCOG0802/WJOG4607L): a multicentre, open-label, phase 3, randomised, controlled, non-inferiority trial. Lancet 2022;399:1607-17. [Crossref] [PubMed]
3. Altorki N, Wang X, Damman B, et al. Lobectomy, segmentectomy, or wedge resection for peripheral clinical T1aN0 non-small cell lung cancer: A post hoc analysis of CALGB 140503 (Alliance). J Thorac Cardiovasc Surg 2024;167:338-347.e1. [Crossref] [PubMed]
4. Brunelli A, Decaluwe H, Gonzalez M, et al. European Society of Thoracic Surgeons expert consensus recommendations on technical standards of segmentectomy for primary lung cancer. Eur J Cardiothorac Surg 2023;63:ezad224. [Crossref] [PubMed]
5. Seguin-Givelet A, Grigoroiu M, Brian E, et al. Planning and marking for thoracoscopic anatomical segmentectomies. J Thorac Dis 2018;10:S1187-94. [Crossref] [PubMed]
6. Gossot D, Seguin-Givelet A. Anatomical variations and pitfalls to know during thoracoscopic segmentectomies. J Thorac Dis 2018;10:S1134-44. [Crossref] [PubMed]
7. Rami-Porta R, Wittekind C, Goldstraw P, et al. Complete resection in lung cancer surgery: proposed definition. Lung Cancer 2005;49:25-33. [Crossref] [PubMed]
8. Travis WD, Brambilla E, Noguchi M, et al. International association for the study of lung cancer/american thoracic society/european respiratory society international multidisciplinary classification of lung adenocarcinoma. J Thorac Oncol 2011;6:244-85. [Crossref] [PubMed]
9. Institut National du Cancer. Les chiffres du cancer en France. Épidémiologie des cancers: Les cancers du poumon. 17 September 2024. Available online: https://www.e-cancer.fr/Professionnels-de-sante/Les-chiffres-du-cancer-en-France/Epidemiologie-des-cancers/Les-cancers-les-plus-frequents/Cancer-du-poumon
10. Zhu S, Xu W, Li Z, et al. Branching patterns and variations of the bronchus and blood vessels in the superior segment of the right lower lobe: a three-dimensional computed tomographic bronchography and angiography study. J Thorac Dis 2023;15:6879-88. [Crossref] [PubMed]
11. Chan EG, Landreneau JR, Schuchert MJ, et al. Preoperative (3-dimensional) computed tomography lung reconstruction before anatomic segmentectomy or lobectomy for stage I non-small cell lung cancer. J Thorac Cardiovasc Surg 2015;150:523-8. [Crossref] [PubMed]
12. Lin KH, Huang YJ, Lee SC, et al. Benefit of Three-dimensional Image Simulation in Surgical Resection of Early-stage Lung Cancer. Ann Thorac Surg 2022;114:1029-34. [Crossref] [PubMed]
13. Shimizu K, Nakazawa S, Nagashima T, et al. 3D-CT anatomy for VATS segmentectomy. J Vis Surg 2017;3:88. [Crossref] [PubMed]
14. Bédat B, Triponez F, Sadowski SM, et al. Impact of near-infrared angiography on the quality of anatomical resection during video-assisted thoracic surgery segmentectomy. J Thorac Dis 2018;10:S1229-34. [Crossref] [PubMed]
15. Pardolesi A, Veronesi G, Solli P, et al. Use of indocyanine green to facilitate intersegmental plane identification during robotic anatomic segmentectomy. J Thorac Cardiovasc Surg 2014;148:737-8. [Crossref] [PubMed]
16. Misaki N, Chang SS, Igai H, Tarumi S, Gotoh M, Yokomise H. New clinically applicable method for visualizing adjacent lung segments using an infrared thoracoscopy system. J Thorac Cardiovasc Surg 2010;140:752-6. [Crossref] [PubMed]
17. Sadeghi AH, Maat APWM, Taverne YJHJ, et al. Virtual reality and artificial intelligence for 3-dimensional planning of lung segmentectomies. JTCVS Tech 2021;7:309-21. [Crossref] [PubMed]
18. Jensen K, Ringsted C, Hansen HJ, et al. Simulation-based training for thoracoscopic lobectomy: a randomized controlled trial: virtual-reality versus black-box simulation. Surg Endosc 2014;28:1821-9. [Crossref] [PubMed]
19. Jensen K, Bjerrum F, Hansen HJ, et al. Using virtual reality simulation to assess competence in video-assisted thoracoscopic surgery (VATS) lobectomy. Surg Endosc 2017;31:2520-8. [Crossref] [PubMed]
20. Onorati I, Radu DM, Martinod E. What's new in minimally invasive thoracic surgery? Clinical application of augmented reality and learning opportunities in surgical simulation. Front Surg 2023;10:1254039. [Crossref] [PubMed]