Prognostic evaluation for squamous cell carcinomas of the lower thoracic esophagus treated with three-field lymph node dissection

Abstract

Objective: The efficacy of esophagectomy with three-field lymph node dissection in surgical treatment for patients with squamous cell carcinomas of the lower thoracic esophagus remains controversial. This report documents the outcomes of this surgical procedure for a large series. Methods: From February 1986 to November 1998, 437 patients with squamous cell carcinomas of the thoracic esophagus underwent transthoracic esophagectomy with three-field lymph node dissection. One hundred and sixteen of these had cancer of the lower thoracic esophagus. To avoid the influence of adjuvant therapy on survival, 20 who also received radiation and/or chemotherapy were excluded, leaving 96 patients who were retrospectively analyzed. Results: The operative morbidity, and 30-day and in-hospital mortality rates were 62, 0, and 3%, respectively. The overall 1-, 3-, and 5-year survival rates were 89, 65, and 59%, with a median survival of 76 months. In those with lymph node metastases (66% of cases), the values were 87, 56, and 48%, as compared with 94, 84, and 79%, respectively (P=0.005) for patients without lymph node metastasis. Factors significantly influencing the overall survival rates were patient age (≧65 vs. ≪65), clinical N status (cN1 vs. cN0), clinical M status (cM1 vs. cM0), longitudinal tumor length of resected specimen (≧5 vs. ≪5 cm), pathologic T status (pT3 vs. pT1, 2), pathologic N status (pN1 vs. pN0), lymphatic invasion (positive vs. negative), vascular invasion (positive vs. negative) and intramural metastasis (present vs. absent). Independent prognostic factors for survival determined by multivariate analysis were pathologic T status (P=0.02), pathologic N status (P=0.03), and presence of intramural metastasis (P=0.04). Additional pathologic M1 status, cervical or celiac lymph node metastasis, was without significant influence. Conclusions: Patients with pathologic T3 tumors with both pathologic N1 status and the presence of intramural metastasis in the lower thoracic esophagus had a poor prognosis. Cervical or celiac lymph node metastasis in patients with carcinomas of the lower thoracic esophagus should be distinguished from pathologic M1 status in the UICC-TNM staging system.

Squamous cell carcinoma, Three field lymph node dissection, Prognostic factor

1 Introduction

Carcinoma of the thoracic esophagus has a poor prognosis, despite advances in operative procedures, perioperative care, and multimodal treatment. The reason for the poor survival rate is that most patients present with advanced stage disease, tumor metastasizing to the lymph nodes even in the early phase of wall penetration.

Esophageal surgeons throughout the world have been attempting to improve survival with this disease using innovative techniques for over 80 years [1]. Some surgical groups have adopted a more radical resection for carcinoma of the thoracic esophagus in the hope of improving the outcome by removing the regional nodes [2–5]. However, the majority of Western surgeons have favored a limited surgical intervention, such as transhiatal resection, with a multimodal treatment approach because they consider this cancer to be systemic or incurable and palliative surgery is more appropriate than a curative intervention [6].

Most importantly, a global standard surgical procedure has not been established for carcinoma of the thoracic esophagus because there are various options (e.g. en bloc esophagectomy through right or left thoracotomy, esophagectomy with two-field lymph node dissection or three-field lymph node dissection) even with the more extensive resections. We have advocated esophagectomy with three-field lymph node dissection, the most aggressive surgical procedure, in our treatment of carcinoma of the thoracic esophagus from 1982, believing that systematic dissection of cervical, mediastinal, and abdominal lymph nodes improves survival and leads to potential cure.

Increase of carcinomas, especially adenocarcinomas located in the lower thoracic esophagus has been reported in Western countries in recent years [7–9]. The present report concerns our results with three-field lymph node dissection for patients with squamous cell carcinoma of the lower thoracic esophagus.

2 Patients and methods

2.1 Patients

From February 1986 to November 1998, 437 patients with carcinomas of the thoracic esophagus underwent transthoracic esophagectomy with three-field lymph node dissection at the National Cancer Center Hospital, Tokyo. In our department, three-field lymph node dissection is indicated for all patients who have carcinomas of the thoracic esophagus requiring esophagectomy. Three hundred and ninety-three (90%) patients had squamous cell carcinomas and 116 (27%) had squamous cell carcinomas of the lower thoracic esophagus. Of these latter, 20 who also received radiation and/or chemotherapy were excluded from this study to avoid the influence of additional therapy on survival. The records of 96 (85 male and 11 female) patients were thus retrospectively reviewed.

Ages ranged from 42 to 86 years, with a mean of 61.6 years. Preoperative evaluation was performed for all patients with barium swallow examination, endoscopy with biopsy, computed tomography scans from the neck to the abdomen, ultrasonography of the neck and the upper abdominal compartment, and endoscopic ultrasonography. Distant organ metastasis, except in the cervical or celiac nodes, was not evident in any of the patients on preoperative evaluation. Postoperative staging according to the 1997 UICC-TNM classification [10] was based on histologic examination of the resected specimens, including the dissected lymph nodes.

2.2 Surgical procedure

All 96 patients underwent right fifth intercostal thoracotomy. The bilateral recurrent laryngeal nerves, upper periesophageal, right paratracheal and infra-aortic arch nodes were dissected as the upper mediastinal lymph nodes. The middle mediastinal nodes comprised the subcarinal, bilateral bronchial, and middle periesophageal nodes, and the lower mediastinal nodes comprised the lower periesophageal, posterior mediastinal, and supradiaphragmatic nodes. All were removed.

Each patient was placed in a supine position after closure of the chest. The deep cervical lymph nodes comprised of internal jugular nodes below the level of the cricoid cartilage and supraclavicular nodes, and the cervical paratracheal nodes were removed bilaterally via collar skin incisions in the neck. The abdominal lymph nodes, comprised of the bilateral paracardiac nodes, the perigastric lesser curvature nodes, the left gastric nodes, the common hepatic artery nodes, the splenic artery nodes, and the celiac nodes were also removed.

Gastrointestinal continuity was restored with a stomach in 90 patients through a retrosternal route and a colon interposition in six patients through a subcutaneous route because of previous gastric surgery for peptic ulcers. Anastomoses of all patients were performed at the neck.

2.3 Statistical analysis

The median follow-up period for 57 surviving patients was 51 months, with a range of 4–128 months. Survival time was measured as the time from the date of surgery until death or until most recent follow-up investigation, March 2000. Survival curves were calculated according to the Kaplan–Meier method including all causes of death, and the log-rank statistics were used for comparisons. Multivariate analyses were performed using the Cox proportional hazards model. Forward and backward stepwise procedures were used to determine the combination of factors that were essential in predicting prognosis. The χ² test or Fisher's exact probability test for smaller numbers was used for comparison of proportions. P≪0.05 was considered to be significant. Statistical calculations were conducted with StatView 5.0J (Abacus Concepts Inc., Berkeley, CA).

3 Results

3.1 Short-term outcomes

The mean±SD duration of surgery was 483±85 min and the operative blood loss was 528±367 ml. Postoperative complications are listed in Table 1 . The operative morbidity was 62%, 37 patients having an uncomplicated postoperative course. Postoperative complications occurred in 59 patients. The in-hospital mortality rate was 3% (3 patients). No patient (0%) died of postoperative complication within 30 days of surgery.

Table 1

Postoperative complications

3.2 Survival rates according to the UICC-TNM staging system

The overall 1-, 3-, and 5-year survival rates of all 96 patients were 89, 65, and 59%, respectively. In the 63 (66%) with lymph node metastasis, the 1-, 3-, and 5-year survival rates were 87, 56, and 48%, while those of 33 patients without lymph node metastasis were 94, 84, and 79%, respectively (P=0.005). Numbers of patients who had lymph node metastasis according to the pathologic T status were as follows: none of four pT1-mucosal tumors, 12 (52%) of 23 pT1-submucosal tumors, 11 (69%) of 16 pT2 tumors, and 40 (76%) of 53 pT3 tumors. Survival curves according to the pathologic UICC-TNM stage are shown in Fig. 1 . The 1-, 3-, and 5-year survival rates were as follows: 100, 83, and 71% for 15 patients with pStage I disease, 89, 83, and 83% for 18 patients with pStage IIA disease, 100, 81, and 73% for 18 patients with pStage IIB disease, 79, 42, and 36% for 30 patients with pStage III disease, 72, 48, and 48% for seven patients with pStage IVA disease, and 88, 50, and 25% for eight patients with pStage IVB disease.

Fig. 1

Survival curves for patients with squamous cell carcinomas of the lower thoracic esophagus according to the pathologic UICC-TNM stage.

3.3 Survival rates based on clinicopathologic factors

The overall survival rates of 96 patients with squamous cell carcinoma of the lower thoracic esophagus based on clinicopathologic factors are listed in Table 2 . The significant factors influencing survivals were patient age (≧65 vs. ≪65), clinical N status (cN1 vs. cN0), clinical M status (cM1 vs. cM0), maximum longitudinal tumor length in resected specimen (≧5 vs. ≪5 cm), pathologic T status (pT3 vs. pT1, 2), pathologic N status (pN1 vs. pN0), lymphatic invasion (positive vs. negative), vascular invasion (positive vs. negative), and intramural metastasis (present vs. absent).

Table 2

Clinicopathologic characteristics and survival rates of 96 patients with squamous cell carcinomas of the lower thoracic esophagus undergoing three-field lymph node dissection

Independent prognostic factors determined by multivariate analysis (listed in Table 3) were pathologic T status (pT3 vs. pT1, 2, P=0.02), pathologic N status (pN1 vs. pN0, P=0.03), and presence of intramural metastasis (present vs. absent, P=0.04) (Model 1). The 5-year survival rate of patients with pT1, 2 tumors with neither pN1 status nor presence of intramural metastasis was 77%, whereas no patients with pT3 tumors with both pN1 status and presence of intramural metastasis survived more than 25 months after surgery, the median survival interval being 12.5 months (Fig. 2) .

Table 3

Multivariate analysis of prognostic factors for squamous cell carcinoma of the lower thoracic esophagus

Fig. 2

Survival curves for patients with squamous cell carcinomas of the lower thoracic esophagus based on the significant prognostic factors of pathologic T status, pathologic N status, and presence or absence of intramural metastasis.

A second model showed pathologic M1 status (cervical or celiac lymph node metastasis positive) to not affect survival. Survival curves based on these factors are shown in Fig. 3 .

Fig. 3

Survival curves for patients with squamous cell carcinomas of the lower thoracic esophagus based on the prognostic factors of pathologic T status, pathologic N status, and pathologic M status.

The number of patients and 5-year survival rates according to the status of lymph node metastasis are listed in Table 4 . The incidences of mediastinal and perigastric, or middle mediastinal and paracardiac lymph node metastases in patients with pT3 tumors were significantly higher than those in patients with pT1, 2 tumors. Furthermore, the incidence of upper mediastinal lymph node metastasis in patients with pT3 tumors tended to be higher. Frequencies of lymph node metastasis in each subgroup of mediastinal nodes were similar among the 96 patients with squamous cell carcinomas of the lower esophagus.

Table 4

Relationship between the tumor penetration and the status of lymph node metastasis among the patients with squamous cell carcinoma of the lower thoracic esophagus

The incidences of celiac and cervical lymph node metastases did not differ between pT3 and pT1, 2 tumors. Dissecting these nodes systematically, the 5-year survival rate of 10 patients with celiac lymph node metastasis, including three with simultaneous cervical lymph node metastasis, was 35%, and that of eight patients with cervical lymph node metastasis reached 25%.

4 Discussion

Three-field lymph node dissection is an aggressive surgical treatment for patients with carcinoma of the lower thoracic esophagus. However, it offers an approximately 60% 5-year survival rate for patients with squamous cell carcinoma of the lower thoracic esophagus, and provides a 48% 5-year survival rate for patients with lymph node metastasis not receiving adjuvant therapy.

Many surgical groups, including ourselves, have attempted to eradicate carcinoma of the thoracic esophagus by extending the lymphadenectomy field. Three-field lymph node dissection, the most extensive, which adds the removal of cervical nodes to the two-field (mediastinal and abdominal) approach, does in fast give better operative mortality and survival [11]. The incidences of operative complications were also lower in patients after three-field lymph node dissection than in those who underwent two-field dissection. Anastomotic leaks, which were the most frequent cause of postoperative complications, healed spontaneously and no further surgical intervention was needed. The incidence of vocal cord palsy, caused by the thorough dissection of lymph nodes, was low compared with two-field dissection. The meticulous dissection of lymph nodes along the recurrent laryngeal nerves via both right transthoracic and cervical approaches could rather avoid injury.

The presence of intramural metastasis in patients with squamous cell carcinoma of the thoracic esophagus has been reported to preclude long-term survival after surgery on the basis of univariate analyses [12–15]. In our study, the 17% 5-year survival rate for patients with intramural metastasis was poor, and no patient with a pT3 tumor with both pN1 status and the presence of intramural metastasis survived more than 25 months, even after three-field lymph node dissection. Nishimaki and colleagues reported a 13.1% 3-year survival rate after extended lymphadenectomy and a 5.7% 3-year survival rate after less extensive lymphadenectomy for patients with intramural metastasis [15]. We have found that conventional adjuvant radiation and/or chemotherapy were ineffective at improving prognosis [13].

The lymphatic drainage system of the esophagus, which is well developed in the submucosal layer, forms a complex interconnecting network that extends particularly longitudinally [16]. Lymphoscintigrams of the esophagus detect uptake in the cervical, upper mediastinal, and perigastric nodes [17]. Upward longitudinal lymphatic extension of tumors causes cervical and upper mediastinal lymph node metastasis and downward longitudinal lymphatic extension causes perigastric lymph node metastasis. Lateral lymphatic extension results in periesophageal lymph node metastasis. Deeper penetration of the esophageal wall increases tumor entry into the lymphatics, increasing the potential for longitudinal and lateral lymphatic extension. Lymph node metastases in the upper mediastinal, middle mediastinal, or paracardiac nodes of patients with pT3 tumors were thus significantly more frequent than with pT1, 2 tumors.

Three-field lymph node dissection tends emphasize bilateral cervical lymphadenectomy. However, the importance of this surgical procedure is in the systematic lymph node dissection because the well developed longitudinal lymphatic drainage system of the esophagus cause early lymph node metastasis at great distance. The incidence of lymph node metastasis in the perigastric nodes may thus be higher than in the lower mediastinal nodes, and metastasis in the upper, middle and lower mediastinal nodes was similar among all our patients. Clark and colleagues reported eight (19%) out of 43 patients with adenocarcinomas of the lower thoracic esophagus or cardia after en bloc esophagectomy to suffer recurrence in the upper mediastinal nodes [18]. However, the lack of data on the benefit of three-field lymph node dissection for adenocarcinoma of the lower thoracic or cardia discouraged them from extending the lymphadenectomy field. Our results suggest that their recurrence in the upper mediastinal nodes after en bloc dissection may have arisen in involved nodes that were left because the incidences of lymph node metastasis at other nodal sites in their series were similar to those in our series. By dissecting the upper mediastinal nodes, a 35% 5-year survival rate was achieved for patients with involved nodes.

Carcinomas of the lower thoracic esophagus appear to exhibit a certain incidence of cervical lymph node metastasis independent of the histological type. Altorki and Skinner adopted three-field lymph node dissection to evaluate the patterns of nodal spread of patients with carcinomas of the thoracic esophagus in a prospective study [19]. They documented a 30% (six patients) incidence of preoperatively unsuspected cervical lymph node metastasis in 20 patients with cancers of the lower thoracic esophagus (four of 15 adenocarcinomas and two of five squamous cell carcinomas). Lerut et al. reported a similar rate for cervical nodal involvement. Three patients (17%) had cervical lymph node metastases out of 18 node-positive patients with T3 adenocarcinomas of the esophagogastric junction [20]. They also reported that 6 (35%) out of 17 patients with pT3 adenocarcinomas of the distal esophagus and 4 (20%) out of 20 patients with pT3 adenocarcinomas of the gastroesophageal junction had cervical lymph node metastases [21].

Ultrasonography of the neck is useful for evaluation of cervical lymph node metastasis in patients with carcinomas of the thoracic esophagus [22–24], with reported sensitivity and accuracy of nearly 80 and 90%, respectively. However, it cannot detect micrometastasis in nodes of normal size and shape.

Patients with celiac lymph node metastasis are classified as having pM1a (pStage IVA) disease according to the UICC-TNM staging system, thus being considered incurable. However, Reed and colleagues noted they might have a locally advanced status and therefore be potential candidates for surgical treatment because the celiac lymph nodes can be considered within a regional nodal basin of carcinomas of the lower thoracic esophagus or gastroesophageal junction [25]. A 35% 5-year survival rate of patients with celiac lymph node metastasis, including patients with cervical lymph node metastasis, and a 48% 5-year survival rate of patients with pStage IVA disease were achieved in our series.

Multivariate analysis have showed that the pM1 status, the presence of cervical or celiac lymph node metastasis, of the UICC-TNM staging system is not a significant prognostic factor for survival, the rate for pStage IV case without distant organ metastasis not differing from that for pStage III after three-field lymph node dissection. In contrast, pN1 or pM1 status appeared important. Although a problem exists regarding lymph node metastasis with low incidence, the favorable survival rate obtained by removing involved nodes, suggests at least that cervical or celiac lymph node metastasis should be distinguished from the M categorization.

In conclusion, pathologic T status, pathologic N status, and presence of intramural metastasis are significant prognostic factors for patients with squamous cell carcinomas of the lower thoracic esophagus. Pathologic T3 tumors with both pathologic N1 status and presence of intramural metastasis have a poor prognosis. While pathologic M1 status without distant organ metastasis does not affect survival. Cervical or celiac lymph node metastasis should be distinguished from M1 categorization of the UICC-TNM staging system.