Impact of thoracic inlet anatomy and reconstruction route selection based on computed tomography three-dimensional reconstruction on cervical anastomotic leakage after esophagectomy

Introduction

Background

Esophageal cancer ranks as the eighth most common cancer globally and represents the sixth leading cause of cancer-related mortality (1). Despite advancements in diagnostic and therapeutic strategies, the overall 5-year survival rate for patients with esophageal cancer remains between 15% and 20% (1). Owing to the often insidious and non-specific nature of its early symptoms, coupled with the limited use of routine gastroscopy, many patients are diagnosed at an advanced stage (2). Consequently, surgery plays a pivotal role in the management of esophageal cancer. For patients with tumors located in the middle or upper thoracic esophagus, the McKeown procedure (three-incision esophagectomy via right thoracotomy, laparotomy, and left neck incision) has been established as a standard surgical approach (3). This technique, utilizing a right thoracic access, effectively circumvents obstruction by the aortic arch, thereby providing optimal exposure for radical resection of mid- to upper-esophageal tumors and extensive mediastinal lymph node dissection (4). A defining characteristic of this procedure is the placement of the alimentary tract anastomosis in the neck. This design ensures that even if the most concerning complication—anastomotic leakage—occurs, the consequence is typically a manageable cervical infection rather than a life-threatening intrathoracic sepsis, substantially improving perioperative outcomes (5,6).

Rationale and knowledge gap

However, alongside its oncological radicality and reduced risk of thoracic complications, the McKeown procedure presents specific challenges, with cervical anastomotic leakage remaining one of its most frequent major complications (7). Currently, the choice of reconstruction route for gastric conduit pull-up to the neck during McKeown esophagectomy [via the orthotopic posterior mediastinal (PM) route vs. the retrosternal (RS) route] remains a subject of debate, a decision potentially critical for anastomotic tension and blood supply (3,8). More importantly, the thoracic inlet, constituting a narrow bony “bottleneck”, must be traversed by the gastric conduit irrespective of the chosen path (9,10). Whether its individualized anatomical narrowness induces mechanical compression on the gastric conduit, thereby acting as an independent risk factor for cervical anastomotic leakage, currently lacks systematic imaging-based investigation.

Objective

This study aimed to: (I) precisely quantify and compare the lengths of PM and RS reconstruction routes using computed tomography (CT) three-dimensional reconstruction; (II) investigate the association between thoracic inlet anatomical parameters and postoperative cervical anastomotic leakage. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1-2559/rc).

Methods

Inclusion and exclusion criteria

Inclusion criteria were: (I) first-time diagnosis of esophageal cancer confirmed by preoperative biopsy; (II) availability of preoperative and postoperative CT imaging data, with CT scans showing no evidence of distant metastasis; (III) performance of McKeown esophagectomy; (IV) complete clinical data.

Exclusion criteria comprised were as follows: (I) presence of distant metastasis; (II) diagnosis of concurrent other malignancies; (III) history of previous abdominal or thoracic surgery; (IV) severe impairment of cardiac, hepatic, cerebral, renal, or other vital organ function.

Patient population and study design

This retrospective study enrolled 100 patients with esophageal cancer who were admitted to Zhujiang Hospital of Southern Medical University, between October 2019 and October 2025. The cohort included 39 patients in the PM group and 61 patients in the RS group. Reconstruction route selection was determined based on preoperative CT imaging, tumor location, prior radiotherapy, and the attending surgeon’s clinical judgment and experience. Propensity score matching (PSM) was performed using a caliper width of 0.05 in a 1:1 ratio to balance potential confounders, resulting in 33 successfully matched pairs. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Zhujiang Hospital, Southern Medical University (approval No. 2025-KY-03-01). Written informed consent for surgery was obtained from all participants.

Route length measurement

Postoperative thin-slice CT DICOM images were imported into Mimics software (version 21.0). An appropriate threshold range (40–70 HU) was applied to create a mask. Three-dimensional models of the cricoid cartilage and the pyloric ring were reconstructed. Using the midpoint of the inferior cricoid border as the proximal landmark and the pyloric ring as the distal landmark, the ‘trace’ function was employed to delineate the central esophageal path layer by layer. The software automatically computed and generated the reconstruction route, the length of which was obtained using the ‘properties’ function (see Figure 1).

Figure 1 3D reconstruction and measurement of the gastric conduit reconstruction route. (A) The ‘thin structure’ enhancement function (red arrow) was applied to clarify the anatomy. The central path of the reconstruction route was then manually traced (blue arrow) from the proximal landmark toward the distal landmark. (B) Key bony and anatomical landmarks, including the sternum and the pyloric ring, are visualized to provide spatial context for the measured path. (C) The software automatically computed and generated the final 3D reconstruction route, with its total length displayed. 3D, three-dimensional.

Thoracic inlet anatomical measurements

Measurements were performed on 1 or 5 mm thin-slice CT scans. On axial images, the following parameters were recorded: thickness of the clavicle (TC), thickness of the sternum (TS), interclavicular distance (ICD), sternum-trachea distance (STD), sternum-vertebra distance (SVD), and the STD/SVD ratio. At the level of the sixth thoracic vertebra, the thoracic anterior-posterior diameter (AD) and horizontal diameter (HD), along with their ratio, were measured (see Figure 2).

Figure 2 Anatomical measurements of the thoracic inlet and chest cavity. (A) AD: distance from sternum to T6 vertebra. (B) TC: thickness at sternoclavicular joint. (C) Thoracic inlet measurements: ICD, TS, STD, and SVD. (D) HD: transverse diameter at T6 level. AD, thoracic anterior-posterior diameter; HD, thoracic horizontal diameter; ICD, interclavicular distance; STD, sternum-trachea distance; SVD, sternum-vertebra distance; TC, thickness of the clavicle; TS, thickness of the sternum.

Data collection

Demographic information, laboratory results, clinical staging according to the 7th edition tumor-node-metastasis (TNM) classification, and treatment details were retrieved from medical records. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Smoking intensity was expressed in pack-years (number of cigarettes smoked daily/20 × duration in years). Alcohol consumption was quantified in kilogram-years (daily alcohol intake in mL/1,000 × duration in years).

Definition of complications

Anastomotic leakage was defined by evidence of conduit rupture or extravasation of contents on upper gastrointestinal contrast study, CT imaging, or clinical findings such as salivary or gastric content discharge from the cervical wound. A routine upper gastrointestinal (GI) series performed on postoperative day 7, interpreted by a staff radiologist, served as the primary objective assessment; contrast extravasation defined leakage. Anastomotic stricture was evaluated by a radiologist primarily via upper gastrointestinal series or endoscopy 3 months postoperatively; CT was used adjunctively if clinically indicated. Chylothorax was diagnosed based on a positive chylous test of chest tube drainage fluid. Recurrent laryngeal nerve injury was defined as postoperative hoarseness confirmed by otorhinolaryngological evaluation. Pneumonia was diagnosed based on radiographic findings on CT and supportive laboratory evidence.

Statistical analysis

PSM was performed using a logistic regression model with a caliper of 0.05 and a 1:1 matching ratio to minimize potential selection bias. Categorical variables were compared using Fisher’s exact test or the χ² test, as appropriate. Continuous variables were compared using the Mann-Whitney U test. Correlations between factors related to thoracic inlet dimensions and patient characteristics were assessed using Spearman’s rank correlation coefficient. Univariate logistic regression was employed to evaluate potential predictors of anastomotic leakage. Variables yielding a P value <0.1 in univariate analysis were included in a subsequent multivariate binary logistic regression model to identify independent risk factors, calculating odds ratios (ORs) and 95% confidence intervals (CIs). The Wald test was used to determine the statistical significance of the ORs. Missing numerical data (e.g., height, weight) were handled using multiple imputation. A two-sided P value <0.05 was considered statistically significant. All analyses were conducted using R software (version 4.4.1).

Results

Baseline characteristics after PSM

Initially, 61 patients in the RS group and 39 patients in the PM group were considered for analysis. Following 1:1 PSM based on clinical stage, age, BMI, and tumor location, 33 matched pairs were generated. Table 1 summarizes the baseline characteristics of the 66 matched patients (33 per group). The two groups were well-balanced regarding clinical and demographic features, including age, sex, BMI, history of hypertension, smoking and alcohol consumption, tumor location, clinical stage and histologic subtype.

Comparison of intraoperative and postoperative outcomes

All 66 patients underwent successful surgery without major intraoperative complications or mortality (Table 2). No statistically significant differences were observed between the groups in the incidence of postoperative complications, including anastomotic leakage, anastomotic stricture, empyema, surgical site infection, chylothorax, or mortality (all P values >0.05).

Reconstruction route length comparison

The mean reconstructed route length was significantly shorter in the RS group (336±39.3 mm) compared to the PM group (356±32.1 mm) (P<0.05). No significant differences were found in the measured thoracic inlet parameters between the two groups (Table 3).

Risk factors for anastomotic leakage

To identify independent risk factors associated with anastomotic leakage, univariate analysis was performed incorporating anatomical parameters, patient demographics, and surgical factors. Variables with P<0.1 were entered into a multivariate binary logistic regression model. Analysis of the entire 100-patient cohort (Table 4) revealed that in the multivariate model, STD exhibited a significant inverse association with leakage risk (β =−0.106, P=0.01; OR =0.903; 95% CI: 0.827–0.979), indicating an approximately 9.7% reduction in leakage odds per 1-mm increase in STD. Conversely, HD showed a significant positive association (β =0.034, P=0.04; OR =1.031; 95% CI: 1.002–1.063), suggesting an approximately 3.1% increase in leakage odds per 1-mm increase in HD.

Clinical prediction model

Based on the univariate and multivariate logistic regression analyses, a clinical prediction model for anastomotic leakage following esophagectomy was developed. The model incorporated statistically significant independent predictors, ultimately retaining two key anatomical variables.

The model was constructed using logistic regression, defined by the equation: logit(P) = β₀ + β₁ × STD + β₂ × HD, where P represents the predicted probability of anastomotic leakage. Parameter estimates were: intercept β₀ =−6.813 (P=0.046), coefficient for STD β₁ =−0.106 (P=0.01), and coefficient for HD β₂ =0.034 (P=0.04). The individual probability of leakage can be calculated as: P = 1/{1 + exp[−(−6.813 − 0.106 × STD + 0.034 × HD)]}.

Model performance evaluation demonstrated good discriminatory ability, with an area under the curve (AUC) of 0.712 (Figure 3), indicating a satisfactory capacity to distinguish between patients with and without leakage. The optimal cutoff probability, determined via receiver operating characteristic (ROC) analysis, was 0.381, yielding a sensitivity of 61.8% and a specificity of 77.3%. Internal validation using the bootstrap method (1,000 resamples) confirmed model stability, with a calibrated AUC of 0.712.

Figure 3 ROC curve for the clinical prediction model of anastomotic leakage. The AUC was 0.712. AUC, area under the curve; ROC, receiver operating characteristic.

Discussion

Key findings

This study, utilizing CT-based three-dimensional reconstruction, demonstrated that the RS reconstruction route is significantly shorter than the PM route in patients undergoing McKeown esophagectomy. Furthermore, multivariate analysis identified the STD and HD as independent risk factors for cervical anastomotic leakage following RS reconstruction.

Strengths and limitations

The strengths of this study include the application of objective and reproducible CT three-dimensional reconstruction technology for precise measurement, and the comprehensive analysis of the relationship between individualized thoracic inlet anatomy and postoperative complications. Several limitations warrant consideration. Its single-center, retrospective design and modest sample size may limit the generalizability of the findings and the statistical power to detect differences in leakage rates between routes. External validation in larger, multi-institutional cohorts is necessary to confirm the model’s robustness and clinical utility.

Comparison with similar research

Despite refinements in modern surgical technique leading to a gradual decline in incidence, anastomotic leakage remains a relatively frequent complication after esophagectomy compared to other gastrointestinal procedures. Reported rates range from 8% to 35% for cervical leaks and 4% to 44% for intrathoracic leaks (11). The etiology of anastomotic leakage is multifactorial, involving patient-related factors, anastomotic location and technique, reconstruction route, and gastric conduit characteristics (12). However, adequate blood supply and minimal anastomotic tension are universally recognized as paramount determinants (13). While anastomotic tension is influenced by both gastric conduit length and the chosen reconstruction path, and while several studies have provided guidance on assessing vascular viability (14), consensus on optimal route measurement is lacking (3,8,15). This study sought to establish a reproducible method for quantifying reconstruction route length and to elucidate the influence of thoracic inlet anatomy on leakage risk.

Our three-dimensional reconstruction methodology, simulating the esophageal path on patient-specific models, minimized potential confounding from intraoperative positional changes and organ shift. The finding of a shorter RS route aligns with the work of Yasuda et al. [2021] (15), who reported a mean RS distance of 34.7±2.37 cm compared to 36.3±2.27 cm for the PM route—a difference of approximately 2 cm—measured intraoperatively using a non-shrinkable 18 F gastric tube. Previous studies, often reliant on cadaveric measurements or intraoperative assessments, are susceptible to variation due to patient positioning, organ state, measurement tools, and landmark selection. Our CT-based approach mitigates these influences, potentially providing a more accurate representation of the in vivo path length. Furthermore, preoperative route measurement can inform the intraoperative creation of the gastric conduit, potentially avoiding excessive tension from an unduly short conduit or compromised vascularity and kinking from an excessively long one.

Notably, despite the shorter length of the RS route, we observed no significant difference in the incidence of anastomotic leakage between the two groups. Potential explanations include: (I) limited statistical power due to the relatively small sample size post-matching; and (II) a genuine absence of difference, possibly attributable to surgeons’ deliberate construction of longer gastric conduits and meticulous attention to ensuring a tension-free anastomosis, thereby negating the theoretical advantage of the shorter route. This may suggest that, provided adequate conduit length is achieved, the absolute reconstruction distance might have less impact on leakage rates than previously assumed. We would like to clarify that in both reconstruction routes, the gastric conduit was created routinely, and there was no intentional design of a longer conduit specifically in the PM group. By “longer conduit”, we refer to a moderate elongation to avoid anastomotic tension, rather than excessive lengthening. Previous studies have demonstrated that the length and vascular pedicle ratio of the right gastroepiploic artery play a primary role in determining blood supply, and that differences in conduit length within a reasonable range do not significantly affect vascular perfusion (13).

Furthermore, even if anastomotic leakage occurs, the RS route is less likely to lead to severe complications compared to the PM route. Given that the leakage rates were comparable between the two groups, the RS route may offer a distinct advantage in terms of postoperative safety (3,16).

Explanations of findings

Through systematic univariate and multivariate analysis of 100 patients (69 in the RS group and 31 in the PM group), this study successfully identified independent risk factors for anastomotic leakage and established a predictive model with reasonable performance. The multivariate model highlighted STD and HD as key anatomical determinants. STD demonstrated a protective effect (OR =0.903; 95% CI: 0.827–0.979; P=0.01), whereas HD was associated with increased risk (OR =1.031; 95% CI: 1.002–1.063; P=0.04). STD effectively reflects the degree of thoracic inlet narrowing. Previous studies have identified STD as an independent risk factor for anastomotic leakage in the RS route (9,10). In our study, STD remained a predictor of anastomotic leakage even in the PM group. This may be attributed to the proportional relationship between STD and the overall thoracic inlet dimensions. Although STD is anatomically more relevant to the RS route, its reduction may contribute to global crowding at the thoracic inlet, thereby affecting both reconstruction pathways.

A larger HD may reflect a more flattened thoracic morphology, which could predispose the gastric conduit to anterior-posterior compression during its passage through the thoracic inlet (17,18), thereby increasing the risk of anastomotic leakage. An alternative explanation is that a larger HD indicates a larger body habitus, while the length and vascular pedicle ratio of the stomach do not increase proportionally. Consequently, when constructing the gastric conduit, the same length of conduit may become relatively shorter in these patients, leading to increased anastomotic tension and a higher risk of leakage.

This finding introduces a nuanced perspective on route selection. Although the RS route is shorter, its potential benefit might be counteracted in patients with a critically narrow thoracic inlet, where it could exacerbate conduit compression within this bony confine. Thus, the optimal route may not be absolute but should be evaluated in the context of individual patient anatomy.

The composite prediction model, incorporating these independent factors, exhibited good discriminatory power (AUC =0.712). The established cutoff value (0.381) offers a practical tool for stratifying patients into low-risk and high-risk categories, potentially guiding tailored perioperative management strategies.

Implications and actions needed

This study recommends incorporating preoperative CT-based assessment of the thoracic inlet, specifically focusing on parameters like STD and HD, into the standard surgical planning process for esophageal cancer. For patients identified with significant thoracic inlet narrowing (characterized by a small STD and/or a large HD), heightened vigilance is warranted. Surgical strategies may need to be tailored, which could include meticulous conduit preparation to minimize bulk, careful positioning of the conduit through the inlet to avoid compression, or even individualized deliberation on the choice of reconstruction route itself, weighing the shorter distance of the RS path against the potential for increased compression in a narrow space.

Conclusions

This study demonstrates, using CT-based three-dimensional reconstruction, that the RS reconstruction route is shorter than the PM route in McKeown esophagectomy. The STD and the HD were identified as independent risk factors for cervical anastomotic leakage following RS reconstruction. We recommend incorporating preoperative CT-based assessment of the thoracic inlet into the standard surgical planning process for esophageal cancer. For patients identified with significant thoracic inlet narrowing, heightened vigilance and consideration of tailored surgical strategies are warranted to mitigate the risk of anastomotic leakage.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1-2559/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1-2559/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1-2559/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-2025-1-2559/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. The study was approved by the Ethics Committee of Zhujiang Hospital, Southern Medical University (approval No. 2025-KY-03-01). Written informed consent for surgery was obtained from all participants.

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. Morgan E, Soerjomataram I, Rumgay H, et al. The Global Landscape of Esophageal Squamous Cell Carcinoma and Esophageal Adenocarcinoma Incidence and Mortality in 2020 and Projections to 2040: New Estimates From GLOBOCAN 2020. Gastroenterology 2022;163:649-658.e2. [Crossref] [PubMed]
2. Lagergren J, Smyth E, Cunningham D, et al. Oesophageal cancer. Lancet 2017;390:2383-96. [Crossref] [PubMed]
3. Dat TQ, Thong DQ, Nguyen DT, et al. Comparison of substernal and posterior mediastinal route of reconstruction after minimally invasive esophagectomy for esophageal cancer. Langenbecks Arch Surg 2024;409:27. [Crossref] [PubMed]
4. Watanabe M, Takahashi N, Tamura M, et al. Gastric conduit reconstruction after esophagectomy. Dis Esophagus 2024;37:doae045. [Crossref] [PubMed]
5. Chen C, Ding C, He Y, et al. High cervical anastomosis reduces leakage-related complications after a McKeown esophagectomy. Eur J Cardiothorac Surg 2024;65:ezae050. [Crossref] [PubMed]
6. Yoshida N, Eto K, Matsumoto T, et al. Omental Flap Wrapping Around the Esophagogastric Anastomosis and Association with Anastomotic Leak in Esophagectomy for Esophageal Cancer: A Propensity Score-Matching Analysis. J Am Coll Surg 2023;236:189-97. [Crossref] [PubMed]
7. Sun S, Wang Z, Huang C, et al. Different gastric tubes in esophageal reconstruction during esophagectomy. Esophagus 2023;20:595-604. [Crossref] [PubMed]
8. Booka E, Takeuchi H, Morita Y, et al. What is the best reconstruction procedure after esophagectomy? A meta-analysis comparing posterior mediastinal and retrosternal approaches. Ann Gastroenterol Surg 2023;7:553-64. [Crossref] [PubMed]
9. Sato S, Nakatani E, Higashizono K, et al. Size of the thoracic inlet predicts cervical anastomotic leak after retrosternal reconstruction after esophagectomy for esophageal cancer. Surgery 2020;168:558-66. [Crossref] [PubMed]
10. Mine S, Watanabe M, Okamura A, et al. Superior Thoracic Aperture Size is Significantly Associated with Cervical Anastomotic Leakage After Esophagectomy. World J Surg 2017;41:2598-604. [Crossref] [PubMed]
11. Scheese D, Alwatari Y, Rustom S, et al. Chest vs. neck anastomotic leak post esophagectomy for malignancy: rate, predictors, and outcomes. J Thorac Dis 2023;15:3593-604. [Crossref] [PubMed]
12. Li KK, Wang YJ, Zhang TM, et al. Right gastroepiploic artery length determined anastomotic leakage after minimally invasive esophagectomy for esophageal cancer: a prospective cohort study. Int J Surg 2024;110:2757-64. [Crossref] [PubMed]
13. Kurahashi Y, Hojo Y, Nakamura T, et al. Anastomotic leakage after esophagectomy possibly caused by compression of the gastric conduit behind the sternoclavicular joint: a report of three cases. Surg Case Rep 2021;7:163. [Crossref] [PubMed]
14. Takahashi K, Nishikawa K, Tanishima Y, et al. Anatomical and anastomotic viability indexes for stratifying the risk of anastomotic leakage in esophagectomy with retrosternal reconstruction. Ann Gastroenterol Surg 2023;7:896-903. [Crossref] [PubMed]
15. Yasuda T, Shiraishi O, Kato H, et al. A comparative study of the lengths of different reconstruction routes used after thoracic esophagectomy. Esophagus 2021;18:468-74. [Crossref] [PubMed]
16. Wang Y, Xu J, Qian J, et al. Retrospective Analysis of Risk Factors and Prediction Model of Cervical Anastomotic and Intrathoracic Anastomotic Leakage After Radical Esophagectomy. Cancer Rep (Hoboken) 2025;8:e70286. [Crossref] [PubMed]
17. Okamoto K, Inaki N, Saito H, et al. Analysis of factors associated with operative difficulty in thoracoscopic esophageal cancer surgery in the left-decubitus position: a single-center retrospective study. BMC Surg 2023;23:242. [Crossref] [PubMed]
18. Abe T, Oshikiri T, Goto H, et al. Thoracic cavity-to-cage ratio is a predictor of technical difficulties in minimally invasive esophagectomy. Surgery 2022;172:145-9. [Crossref] [PubMed]