Transthoracic versus Transhiatal Esophagectomy for Esophageal Squamous Cell Carcinoma

Abstract

Background:

Esophageal squamous cell carcinoma (ESCC) is associated with high morbidity and mortality, and esophagectomy remains the mainstay of curative treatment. The optimal surgical technique transthoracic esophagectomy (TTE) versus transhiatal esophagectomy (THE) continues to be debated due to differences in surgical invasiveness and the extent of lymphadenectomy. This study aimed to compare clinical and oncological outcomes of TTE and THE in patients treated for ESCC.

Materials and Methods:

This retrospective study included patients who underwent curative-intent esophagectomy for histologically confirmed ESCC between 2015 and 2023. Patients were categorized according to surgical approach (TTE vs THE), and demographic characteristics, perioperative data, pathological findings, postoperative complications, and survival outcomes were compared between the 2 groups. Complications were classified according to the Clavien–Dindo system. Disease-free survival (DFS) and overall survival (OS) were assessed using Kaplan–Meier analysis.

Results:

A total of 39 patients were analyzed (21 TTE, 18 THE). Baseline demographic features and most perioperative variables were comparable. TTE was associated with a significantly longer operative time and higher lymph node yield, whereas THE showed higher rates of intraoperative transfusion and pleural effusion. Other postoperative complications were similar between groups. The TTE approach demonstrated significantly better DFS (P = .004). No significant difference in OS was observed between the groups (P = .165). No additional clinical or pathological factors were associated with survival outcomes.

Conclusion:

TTE resulted in a greater lymph node yield and significantly improved DFS, indicating a clear oncologic advantage over the transhiatal approach.

INTRODUCTION

Esophageal cancer remains a significant global health problem. According to GLOBOCAN 2020 data, it ranks 7th in incidence with approximately 604,000 new cases and 6th in cancer-related mortality with 544,000 deaths.¹ The disease is 2 to 3 times more common in men than in women, with approximately 70% of cases diagnosed in males.¹

Surgical resection forms the cornerstone of curative treatment in esophageal cancer. Despite advances in multimodal therapeutic approaches, surgery remains the most effective method for achieving long-term survival.^2,3 However, esophagectomy is associated with high morbidity and mortality rates, and optimization of surgical technique and perioperative care directly affects patient prognosis.² In recent years, with the increasing adoption of minimally invasive approaches particularly thoracoscopic and laparoscopic techniques. It has been reported that these methods significantly reduce surgery-related complications and accelerate postoperative recovery.³ Recent studies have demonstrated that minimally invasive esophagectomy (MIE) offers superior short-term morbidity outcomes while providing comparable results to open surgery in terms of lymph node dissection and long-term survival.^4–6

Although various approaches have been described for esophagectomy, there is no consensus on the optimal technique to maximize both operative and oncological outcomes.⁷ Currently, the modern surgical trend favors transthoracic esophageal dissection with either a cervical (3-field) or intrathoracic (Ivor Lewis) anastomosis to achieve superior oncologic outcomes and more extensive nodal clearance. Transhiatal esophagectomy (THE) and transthoracic esophagectomy (TTE) are among the main contemporary surgical alternatives.^8,9 The THE procedure is performed through abdominal and cervical incisions, whereas the TTE approach involves thoracotomy or thoracoscopy, allowing for an intrathoracic or cervical anastomosis and a more extensive lymphadenectomy.¹⁰ It has been suggested that THE may be limited in mediastinal lymph node dissection, potentially leading to inferior long-term survival.¹¹ Conversely, TTE, which enables a radical “en bloc” dissection in some centers, has been reported to provide improved long-term survival compared to THE.^11–13

In this study, we aimed to compare the clinical and oncological outcomes of transthoracic TTE and THE approaches in patients who underwent surgery for esophageal cancer. By analyzing the effects of these 2 techniques on perioperative morbidity, mortality, extent of lymph node dissection, and survival, we sought to contribute to the clinical decision-making process in selecting the most appropriate surgical approach.

MATERIALS AND METHODS

This retrospective study was conducted following the approval of the institutional ethics committee. The study included patients who underwent esophagectomy due to esophageal cancer at our center between January 2015 and December 2023. Patients were divided into 2 groups based on the surgical approach: the THE group and the TTE group. Demographic, perioperative, and pathological data were retrieved from hospital electronic medical records and operative reports.

Patients diagnosed with histopathologically confirmed esophageal squamous cell carcinoma (ESCC) who underwent curative-intent esophagectomy were eligible for inclusion. Patients with adenocarcinoma were excluded from the study. Exclusion criteria were as follows: distant metastasis detected preoperatively, concurrent primary malignancy, incomplete clinical or pathological data, history of previous thoracic or upper abdominal surgery and conversion from minimally invasive to open surgery, and insufficient follow-up data. After applying these criteria, a total of 39 patients were included in the final analysis (18 in the THE group and 21 in the TTE group).

All TTEs were performed using a conventional 2-dimensional thoracoscopic system, whereas THEs were performed through an upper midline abdominal incision without thoracotomy. A cervical end-to-end esophagogastric anastomosis was performed in all patients, and the stomach was used as the conduit for reconstruction. No robotic procedures were performed during the study period. All operations were performed by the same experienced upper gastrointestinal surgical team consisting of 3 surgeons who were proficient in both surgical techniques.

The following parameters were recorded and analyzed: demographic variables, including age, sex and comorbidities; operative characteristics, such as type of surgery, operative time, estimated blood loss, and intraoperative transfusion requirement; and postoperative outcomes, including intensive care unit (ICU) stay, total hospital stay, postoperative complications, and mortality. Postoperative complications were graded according to the Clavien–Dindo classification system, while respiratory complications, including pneumonia, were identified based on clinical and radiological findings. Pathological evaluation encompassed tumor histology, resection margin status, tumor length, lymph node yield, number of metastatic lymph nodes, and pathological TNM staging.

Statistical Analysis

Overall survival (OS) and disease-free survival (DFS) were defined as the time from surgery to death and to recurrence or death, respectively. Continuous variables were expressed as mean ± standard deviation or median (interquartile range [IQR]) based on distribution, while categorical variables were presented as frequencies and percentages. Group comparisons were performed using the Mann–Whitney U test for non-normally distributed continuous variables and the independent t test for normally distributed variables. Categorical variables were analyzed using the χ² test, and when assumptions were not met, Fisher’s exact test or Yates’ corrected χ² test was applied. Survival outcomes (OS and DFS) were estimated using the Kaplan–Meier method, and differences between groups were assessed using the log-rank (Mantel–Cox) test. Statistical analyses were performed by a team member with formal training in biostatistics. All statistical analyses were conducted using SPSS version 28.0 (IBM Corp., Armonk, NY, USA), and a P-value < .05 was considered statistically significant.

RESULTS

This study included a total of 39 patients who underwent esophagectomy. The mean age of the cohort was 61.7 ± 9.7 years. Of the patients, 25 (64.1%) were female and 14 (35.9%) were male. Tumor localization was distributed as follows: upper third in 4 patients (10.2%), middle third in 16 patients (41.0%), and lower third in 19 patients (48.7%). The mean operation time was 331.03 ± 59.37 minutes, with an average ICU stay of 6.67 ± 7.95 days and total hospital stay of 21.95 ± 17.28 days. The mean DFS and OS times were 25.31 ± 27.01 and 30.26 ± 30.27 months, respectively. The mean number of retrieved lymph nodes was 12.33 ± 7.17. Of the patients, 21 (53.8%) underwent TTE, while 18 (46.2%) underwent the transhiatal approach. The clinical characteristics of the patients and differences between tumor locations are summarized in Table 1.

Table 1.

Patient and Tumor Characteristics

Variables
Age, median-IQR	61.7 ± 9.7
Sex
Male n (%)	14 (35.9%)
Female n (%)	25 (64.1%)
Tumor localization (n, %)
Upper (n, %)	4(10.2%)
	0.318
Middle (n, %)	16 (41%)
Lower (n, %)	19 (48.7%)
Operation time (minutes, mean ± SD)	331.03 ± 59.37
ICU stay (days, mean ± SD)	6.67 ± 7.95
Hospital stay (days, mean ± SD)	21.95 ± 17.28
Disease-free survival (months, mean ± SD)	25.31 ± 27.005
Overall survival (months, mean ± SD)	30.26 ± 30.27
Number of harvested lymph nodes (mean ± SD)	12.33 ± 7.17
Operation type
Transhiatal (n, %)	18 (46.2%)
Transthoracic (n, %)	21 (53.8%)

When patients were compared according to surgical approach, 21 (53.8%) underwent transthoracic and 18 (46.2%) underwent THE. There was no significant difference in mean age between the groups (56.7 ± 8.6 vs 67.6 ± 7.4 years, P = .599). Similarly, gender distribution and smoking status did not differ significantly between groups (P = .718 and P = .688, respectively). However, a statistically significant difference was found in T stage distribution (P = .026), with early-stage tumors (T0) being more frequent in the transthoracic group (52.4%) and T2 disease more common in the transhiatal group (27.8%). The median operative time was significantly longer in the transthoracic group than in the transhiatal group (330 [IQR: 95] vs 290 [IQR: 45] minutes, P = .017). No significant difference was observed in median intraoperative blood loss between the transthoracic and transhiatal groups (300 [IQR: 425] vs 420 [IQR: 375] mL, P = .269). The median ICU stay was 4 (IQR: 8) days in the transthoracic group and 3 (IQR: 5) days in the transhiatal group (P = .945). No significant difference was observed in median total hospital stay between the groups (12 [IQR: 16] vs 26 [IQR: 20] days, P = .223). Regarding perioperative parameters, intraoperative blood transfusion was significantly more frequent in the transhiatal group (55.6% vs 14.3%, P = .006). Postoperative transfusion, however, did not reach statistical significance (P = .075). Pleural effusion was also significantly more common in patients undergoing transhiatal surgery (50% vs 19%, P = .041). No significant difference was observed in postoperative pneumonia rates between the transhiatal and transthoracic approaches (27.8% vs 23.8%, P = 1.000). Regarding pathological outcomes, the mean resection margin was similar between groups (61.0 ± 22.2 mm vs 53.2 ± 28.9 mm, P = .346). The mean number of retrieved lymph nodes was significantly higher in the transthoracic group compared to the transhiatal group (16.7 ± 6.8 vs 7.3 ± 3.3, P < .001). No significant differences were observed in anastomotic leak rate, wound infection, ileus, or anastomotic stricture between the groups (all P > .05) (Table 2).

Table 2.

Comparison of Clinicopathological Features Between Transthoracic and Transhiatal Esophagectomy Patients

Variables	Transthoracic (n = 21)	Transhiatal (n = 18)	P-Value
Age (mean ± SD)	56.67 (±8.64)	67.61 (±7.39)	.599
Gender
Male (n, %)	7 (33.3%)	7 (38.9%)	.718
Female (n, %)	14 (66.7%)	11 (61.1%)
T stage
T0	11 (52.4)	5 (27.8%)	.026
T1	2 (9.5%)	0 (0%)
T2	0 (0%)	5 (27.8%)
T3	8 (38.1%)	8 (44.4%)
Smoking
No (n, %)	13 (61.9%)	10 (55.6%)	.688
Yes (n, %)	8 (38.1%)	8 (44.4%)
Preoperative cardiac comorbidity, n (%)
No (n, %)	18 (38.1%)	13 (72.2%)	.432
Yes (n, %)	3 (61.9%)	5 (27.8%)
Preoperative pulmonary comorbidity, n (%)
No (n, %)	20 (95.2%)	13 (72.2%)	.077
Yes (n, %)	1 (4.8%)	5 (27.8%)
Tumor stage
Stage 0	9 (42.9%)	5 (27.8%)	.11
Stage 1	0 (0%)	1 (2.6%)
Stage 2	2 (9.5%)	9 (50%)
Stage 3	5 (23.8%)	0 (0%)
Stage 4	5 (23.8%)	3 (16.7%)
Tumor localization
Upper third 1/3	1 (4.8%)	3 (16.7%)	.584
Middle third 1/3	9 (42.9%)	7 (38.9%)
Lower third 1/3	11 (52.4%)	8 (44.4%)
Tumor grade
Grade 1 (n, %)	2 (9.5%)	1 (5.6%)	1.000
Grade 2 (n, %)	18 (85.7%)	16 (88.7%)
Grade 3 (n, %)	1 (4.8%)	1 (5.6%)
Intraoperative transfusion
No (n, %)	18 (85.7%)	8 (44.4%)	.006
Yes (n, %)	3 (14.3%)	10 (55.6%)
Postoperative transfusion
No (n, %)	13 (61.9%)	6 (33.3%)	.075
Yes (n, %)	8 (38.1%)	12 (66.7%)
Clavien dindo
Grade 1	2 (9.5%)	0 (0%)	.370
Grade 2	9 (42.9%)	7 (38.9%)
Grade 3	3 (14.3%)	6 (33.3%)
Grade 4	2 (9.5%)	0 (0%)
Grade 5	5 (23.8%)	5 (27.8%)
Wound infection
No (n, %)	19 (90.5%)	13 (72.2%)	.215
Yes (n, %)	2 (9.5%)	5 (27.8%)
Evisceration
No (n, %)	20 (95.2%)	16 (88.9%)	.586
Yes (n, %)	1 (4.8%)	2 (11.1%)
Anastomotic leak
No (n, %)	16 (76.2%)	15 (83.3%)	.702
Yes (n, %)	5 (23.8%)	3 (16.7%)
Ileus
No (n, %)	21 (100%)	16 (88.9%)	.206
Yes (n, %)	0 (0%)	2 (11.1%)
Anastomotic stricture
No (n, %)	19 (90.5%)	16 (88.9%)	1.000
Yes (n, %)	2 (9.5%)	2 (11.1%)
Tracheoesophageal fistula
No (n, %)	17 (81%)	16 (88.9%)	.667
Yes (n, %)	4 (19%)	2 (11.1%)
Pleural effusion
No (n, %)	17 (81%)	9 (50%)	.041
Yes (n, %)	4 (19%)	9 (50%)
Pneumonia
No (n, %)	16 (76.2%)	13 (72.2%)	1.000
Yes (n, %)	5 (23.8%)	5 (27.8%)
Number of harvested lymph nodes (mean ± SD)	16.67 ± 6.76	7.28 ± 3.32	.000
Resection margin (mm, mean ± SD)	61.05 ± 22.15	53.22 ± 28.94	.346
Operation time (minutes) (median [IQR])	330 (95)	290 (45)	.017
Blood loss (mL) (median [IQR])	300 (425)	420 (375)	.269
ICU stay (days) (median [IQR])	4 (8)	3 (5)	.945
Hospital stay (days) (median [IQR])	12 (16)	26 (20)	.223

In summary, the transthoracic approach was associated with longer operation time and a higher lymph node yield, while the transhiatal approach demonstrated higher rates of pleural effusion and intraoperative transfusion. Other perioperative and postoperative outcomes were similar between the 2 surgical techniques.

Kaplan–Meier survival analysis demonstrated a statistically significant difference in DFS between patients who underwent thoracoscopic esophagectomy and those who underwent THE (Log-rank [Mantel–Cox] χ² = 8.077, P = .004) (Figure 1). The mean DFS time was 65.8 ± 6.7 months (95% confidence interval [CI]: 52.6–79.0) in the thoracoscopic esophagectomy group and 34.5 ± 7.6 months (95% CI: 19.6–49.4) in the transhiatal group. The median DFS time could not be reached in the thoracoscopic group due to ongoing censoring, whereas it was 22.0 ± 2.5 months (95% CI: 17.1–26.9) in the transhiatal group (Figure 1). At 36 months, the estimated DFS rate was approximately 83% for patients in the thoracoscopic esophagectomy group and about 35% for those in the THE group. This indicates a markedly better medium-term oncological outcome following thoracoscopic esophagectomy (Figure 1). Kaplan–Meier analysis demonstrated no statistically significant difference in OS between patients undergoing thoracoscopic and THE (Log-rank [Mantel–Cox] χ² = 1.926, P = .165) (Figure 2). The mean OS time was 51.8 ± 7.5 months (95% CI: 37.2–66.4) in the thoracoscopic esophagectomy group and 41.2 ± 9.8 months (95% CI: 21.9–60.4) in the transhiatal group. The median survival could not be estimated for the thoracoscopic cohort due to censoring, whereas it was 19.0 ± 14.3 months (95% CI: 0–47.1) for the transhiatal cohort (Figure 2). At 36 months, the estimated OS rate was approximately 71% in the thoracoscopic esophagectomy group and about 55% in the THE group. However, no statistically significant difference in OS was observed between the 2 groups.

Figure 1.

Kaplan–Meier disease-free survival curves comparing transthoracic and transhiatal esophagectomy.

Figure 2.

Kaplan–Meier overall survival curves comparing transthoracic and transhiatal esophagectomy.

DISCUSSION

In this study, the transthoracic approach was associated with a significantly higher lymph node yield, which may partly explain the improved DFS observed in this group. Despite the more extensive mediastinal dissection, perioperative morbidity did not increase, and pleural complications were even less frequent. These findings suggest that the enhanced oncologic clearance achievable with the transthoracic approach may translate into better long-term outcomes. This aligns with the results of Hayami et al,¹⁴ who similarly reported improved survival and superior perioperative parameters including shorter operative time, reduced blood loss, and higher lymph node yield with the transthoracic technique. Consistent with these observations, previous studies have also demonstrated higher long-term survival and increased R0 resection rates with TTE.¹⁵ Furthermore, recent evidence has highlighted the prognostic importance of nodal status in ESCC. Sun et al reported that adjuvant chemotherapy may improve survival in selected patients with incidentally discovered lymph node–positive disease after esophagectomy, underscoring the critical role of adequate lymph node assessment and oncologic clearance in determining long-term outcomes.¹⁶

Although OS did not differ significantly between the groups in our study, DFS was significantly improved with the transthoracic approach, suggesting a more effective oncologic clearance. These observations are consistent with the findings of Hayami et al,¹⁴ who, in a large population-based cohort of esophageal cancer patients, demonstrated superior OS and a higher lymph node yield with the transthoracic technique. Supporting the oncologic advantage of the transthoracic approach, the number of dissected lymph nodes was also higher in our study, consistent with the results of Mertens et al,¹⁷ who demonstrated a significantly greater lymph node yield with the transthoracic approach in a nationwide propensity score–matched analysis.

In our study, pleural effusion occurred significantly more often in the transhiatal group, while rates of other postoperative complications including wound infection, evisceration, anastomotic leak, ileus, anastomotic stricture, tracheoesophageal fistula, and pneumonia were comparable between the 2 approaches. The distribution of complication severity based on the Clavien–Dindo classification also showed no meaningful difference. These findings align with previous reports indicating similar postoperative morbidity profiles for TTE and THE, although studies evaluating thoracoscopic techniques have noted higher lymph node retrieval and comparable or even lower complication rates.^18,19

In a study comparing THE and TTE for distal esophageal cancer, Higuchi et al reported that anastomotic leakage was significantly more frequent in patients with preoperative cardiac comorbidities undergoing the transhiatal approach.²⁰ In our cohort, however, cardiac comorbidities were similarly distributed between the 2 groups, and the impact of cardiac comorbidity on anastomotic leakage could not be evaluated. Notably, the overall rate of anastomotic leakage did not differ significantly between the transthoracic and transhiatal techniques.

Despite previous studies suggesting that MIE shortens hospital stay and reduces postoperative morbidity compared with open esophagectomy (OE),²¹ our findings did not demonstrate any significant difference in either ICU stay or total hospital stay between the transthoracic and transhiatal groups. These results indicate that, in our cohort, the choice of surgical approach did not substantially affect postoperative recovery time or the need for intensive care monitoring.

In a comparative study by Takahashi et al, no significant difference in estimated intraoperative blood loss was observed between the transthoracic and transhiatal approaches. However, the operative time was significantly longer in the transthoracic group.²² Similarly, in our cohort, operative duration was significantly longer in patients who underwent TTE, while no significant difference was found in intraoperative blood loss.

With regard to oncological outcomes, although the present study exclusively included patients with ESCC, several previous studies involving esophageal adenocarcinoma have also compared TTE and THE. These studies are discussed separately below because differences in tumor biology and treatment strategies may influence surgical and oncological outcomes.

In another study, Meillat et al²³ reported a 3-year OS rate of 75% after THE for locally advanced gastroesophageal junction adenocarcinoma, whereas the corresponding rate was 55% in our cohort. Differences in tumor histology, biological behavior, patient selection, and adjuvant treatment strategies may account for this discrepancy, particularly given that our study exclusively included patients with ESCC.

In a meta-analysis by Wu et al,²⁴ which included patients with Siewert type II esophagogastric junction adenocarcinoma, the transhiatal approach was associated with reduced intraoperative blood loss, shorter hospital stay, lower incidence of pulmonary complications, and improved 3-year OS. Although their cohort predominantly included adenocarcinoma, the observed trends regarding perioperative outcomes and the relative advantages of the transthoracic approach appear consistent with our findings in squamous cell carcinoma. No significant differences were found between the 2 approaches in operative time, R0 resection rate, number of dissected lymph nodes, perioperative morbidity or mortality, abdominal complications, or anastomotic leakage.

In our study, similar to the findings of Wu et al, intraoperative blood loss, postoperative complications, and length of hospital stay did not differ significantly between the transthoracic and transhiatal groups. However, the transthoracic approach was associated with a significantly longer operative duration and a higher lymph node yield, reflecting a more extensive mediastinal dissection. These findings reinforce the concept that the transthoracic approach may achieve superior oncologic clearance without a concomitant increase in perioperative morbidity.

This study has several limitations that should be acknowledged. First, its retrospective and single-center design introduces the potential for selection bias and limits the generalizability of the findings. Furthermore, our institution represents a relatively low-volume esophageal surgery center serving a regional population, which may partly explain the limited sample size and could have influenced surgical outcomes and perioperative management strategies. In addition, baseline differences between the TTE and THE groups, particularly regarding pathological stage, tumor grade, and tumor location, may have influenced the perioperative and oncological outcomes, including DFS and OS. In addition, several clinically relevant variables, including preoperative functional status, nutritional parameters, and treatment-related toxicities, were not consistently recorded and therefore could not be incorporated into the analysis. The limited sample size precluded meaningful subgroup analyses according to tumor location, pathological stage, and tumor grade, and may also have reduced the statistical power of the study, potentially preventing the detection of subtle differences between groups. Nevertheless, despite these limitations, the present study provides additional real-world evidence and may contribute to the ongoing discussion regarding the optimal surgical approach for ESCC. Larger prospective multicenter studies with more balanced patient populations are warranted to validate these findings.

CONCLUSION

In this study, the transthoracic approach demonstrated a clear oncologic advantage over the transhiatal technique, primarily through a more extensive lymph node dissection. This benefit was reflected in significantly improved DFS, while no significant difference was observed in OS. Importantly, these gains were achieved without an increase in postoperative morbidity. Except for a higher rate of pleural effusion in the transhiatal group, other complication rates and hospital stay were comparable between techniques. Overall, our findings suggest that the more comprehensive lymphadenectomy achievable with TTE may translate into improved oncologic outcomes while maintaining a similar safety and recovery profile. However, these findings should be interpreted in light of the study's limitations and require confirmation in larger prospective multicenter studies.