World J Surg (2008) 32:255–261 DOI 10.1007/s00268-007-9213-1
Long-term Results after Dissection of Positive Thoracic Lymph Nodes in Patients with Esophageal Squamous Cell Carcinoma Hideaki Shimada Æ Shin-ichi Okazumi Æ Hisahiro Matsubara Æ Tooru Shiratori Æ Yasunori Akutsu Æ Yoshihiro Nabeya Æ Tooru Tanizawa Æ Kazuyuki Matsushita Æ Hideaki Hayashi Æ Kaichi Isono Æ Takenori Ochiai
Published online: 7 December 2007 Socie´te´ Internationale de Chirurgie 2007
Abstract Background Although thoracic lymph node metastasis in patients with thoracic esophageal squamous cell carcinoma (SCC) has been reported to be a negative risk factor for long-term survival, only a few studies have evaluated the clinicopathologic difference between the impact of metastasis to the paraesophageal lymph nodes and to the nonparaesophageal lymph nodes. The purpose of this study was to evaluate surgical outcome after the clearance of metastatic thoracic lymph nodes. Methods Retrospectively reviewed were 164 consecutive patients with thoracic esophageal SCC who had not had preoperative treatment and underwent surgery from 1980 to 2005 and were found to have thoracic lymph node metastases. Of these patients, 83 underwent surgery from 1980 to 1994 and 81 from 1995 to 2005. Univariate and multivariate analyses were performed to evaluate the impact of nonparaesophageal lymph node metastasis on survival. Results Univariate analysis revealed that T3/T4 tumors and the presence of nonparaesophageal node metastases H. Shimada (&) S. Okazumi H. Matsubara T. Shiratori Y. Akutsu Y. Nabeya K. Matsushita K. Isono T. Ochiai Department of Frontier Surgery, Chiba University, Graduate School of Medicine, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan e-mail:
[email protected] T. Tanizawa Department of Pathology, Chiba University, Graduate School of Medicine, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan H. Hayashi Center for Frontier Medical Engineering, Chiba University, Graduate School of Medicine, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan
were associated with only a 20% overall five-year survival rate. The overall five-year survival for the most recent period was significantly better than for the former period (42% vs. 13%, p \ 0.01). Based on a multivariate analysis of prognostic impact of each nonparaesophageal node, the presence of metastatic subcarinal and/or posterior mediastinal nodes was an independent risk factor for reduced survival. Conclusion Surgical outcome for patients with thoracic esophageal cancer and metastatic thoracic lymph nodes has improved during the last 25 years. Although postoperative chemotherapy might improve survival, the presence of T3/T4 tumors and/or metastatic nonparaesophageal nodes were unfavorable factors for survival.
When primary thoracic esophageal carcinoma spreads to nonparaesophageal lymph nodes, in most instances this is regarded as a distant metastasis and classified as stage IV [1]. In Japan, three-field lymph node dissection, i.e., neck, mediastinal, and abdominal lymphadenectomy, has been introduced to improve the long-term survival of such patients [2–4]. Nevertheless, despite improved surgical techniques and adjuvant therapy [5–8], early recurrence is still frequently observed in patients with positive nodes [9, 10]. Because only a few reports have analyzed the prognostic factors for patients with thoracic node metastases, including paraesophageal and nonparaesophageal nodes, it is still unclear which patients are expected to have a poor prognosis, even after the clearance of metastatic thoracic nodes. Therefore, we retrospectively analyzed a consecutive series of 164 patients with curatively resected primary thoracic esophageal squamous cell carcinoma (SCC) and metastases of the thoracic lymph nodes to
123
256
evaluate the impact of metastasis to nonparaesophageal nodes on the long-term survival of this type of patient.
Patients and methods Patients Between January 1980 and the end of 2005, 623 patients were surgically treated for primary thoracic esophageal SCC at the Department of Surgery, Chiba University Hospital (Chiba, Japan). Patients were classified pathologically according to the pTNM/UICC classification [1]. A total of 417 patients underwent curative resection without any preoperative therapy. Of the 417 patients, 233 (56%) were pathologically diagnosed with lymph node metastases, of which 164 had thoracic node metastases. We focused on these 164 patients. Metastasis in 97 (59%) of these 164 patients involved nonparaesophageal nodes. After surgery, all patients underwent clinical examinations and imaging studies on a regular basis either until death or to the end of 2006.
Postoperative adjuvant chemotherapy Over the years both postoperative management and postoperative adjuvant therapy for esophageal cancer have gradually improved. Postoperative adjuvant chemotherapy was administered to all eligible patients according to the protocols of the Japan Esophageal Oncology Group, which conducted three consecutive randomized controlled trials after 1980 [5–7]. A total of 41 patients received postoperative chemotherapy that consisted of one of the following three regimens: (1) two courses of combination chemotherapy consisting of cisplatin (50 mg/m2) and vindesine (3 mg/m2) were given to seven patients; (2) two courses of chemotherapy consisting of cisplatin (50 mg/m2) was given to eight patients; and (3) two courses of combination chemotherapy consisting of cisplatin (70 mg/m2/day) on day 1 and 5-fluorouracil (700 mg/m2/day) for 4 days were given to 26 patients.
Surgical procedure Thoracic lymph node dissection was performed according to standard procedures that have been described previously [4]. This procedure, called a ‘‘D2 lymphadenectomy’’ according to the Japanese classification of esophageal carcinoma [11], involves dissecting the nonparaesophageal nodes for all tumors. Although for lower-third tumors dissection of upper thoracic nonparaesophageal nodes is
123
World J Surg (2008) 32:255–261
not a routine part of the standard D2 lymphadenectomy, nevertheless, we performed the same dissection for these tumors as we did for upper-third tumors. For patients under 75 years of age, cervical lymph nodes were dissected for all tumors except lower-third T1 tumors. Distant abdominal nodes, including the celiac nodes and the common hepatic artery nodes, were dissected for only lower-third tumors.
Statistical analysis Because postoperative care and treatment modalities for recurrent disease have gradually improved during our 25-year retrospective study, the consecutive series of 164 patients was divided into two groups according to the time of surgery: 1980–1994 and 1995–2005. The results of surgical treatment and several other variables regarding prognosis were compared between these two time periods to make the results of this retrospective study as meaningful as possible. Fisher’s exact probability test was used to determine the significance of any group differences. Overall survival is represented by survival curves that include all-cause mortality. Cause-specific survival curves included only esophageal carcinoma-related death. Outcome was evaluated at the end of 2006. Survival probabilities were calculated using the Kaplan-Meier product limit method, and survival differences between the two groups were tested using the log-rank test. The influence of clinicopathologic variables on survival was individually assessed by Cox’s proportional hazards model. All statistical analyses were carried out using StatView 5.0 for Windows (SAS Institute Inc., Cary, NC); p values were considered to be statistically significant at the 5% level.
Results Clinicopathologic features of patients with thoracic node metastasis Of the 164 patients with metastases of the thoracic lymph nodes, 140 were males (85%) and 24 females (15%), with a mean age of 63 years (range = 27–82 years). For this retrospective review, 83 patients had surgery in the period from 1980 through 1994 and 81 from 1995 through 2005. In 22 (12%) patients the tumors were located in the upper third of the thoracic esophagus, in 97 (59%) in the middle third, and in 45 (27%) in the lower third. The mean size of the tumors was 50 mm (range = 10–140 mm). The mean number of totally dissected lymph nodes and metastatic lymph nodes per patient was 57 (range = 11–162) and 6 (range = 1–58), respectively. A total of 110 patients
World J Surg (2008) 32:255–261 Fig. 1 Distribution and positive rates of lymph nodes among 164 patients with thoracic lymph node metastases. Recurrent nerve, tracheobronchial, and main bronchus nodes were bilaterally dissected. Nonparaesophageal nodes are underlined
257 Non-paraesophageal node Main bronchial node (9.1%) Tracheobronchial node (5.5%) Recurrent nerve node (46%)
Upper paraesophageal node (20%)
Subcarinal node (25%) Posterior mediastinal node (8.5%)
Lower paraesophageal node (24%) Middle paraesophageal node (30%)
received three-field lymphadenectomy. Overall hospital morbidity and hospital mortality rates were 57% and 3.7%, respectively.
diseases. The most common sites of recurrence were lymph nodes (57 patients, 60%) followed by organ metastasis (26 patients, 27%), disseminated disease (7 patients, 7%), and local recurrence (6 patients, 6%).
Distribution and positive rates of thoracic lymph nodes The positive rates of lymph nodes at each station are shown in Figure 1. Among the paraesophageal nodes, the positive rate was highest for the middle paraesophageal nodes. Furthermore, the positive rate for middle thoracic nodes with upper tumors was higher than that with lower tumors (32% vs. 18%, p = 0.16). Among nonparaesophageal nodes, the positive rate was highest for recurrent nerve nodes followed by the subcarinal nodes. The positive rate for recurrent nerve nodes in patients with upper tumors was higher than that in those with lower tumors (59% vs. 44%, p = 0.25). On the other hand, the positive rate for subcarinal nodes with upper tumors was lower than that with lower tumors (14% vs. 27%, p = 0.29).
Clinicopathologic features and survival according to the time period of surgery Less intramural metastasis, hospital mortality, and hospital morbidity and more three-field lymphadenectomies were observed in the later period (Table 1). Overall hospital morbidity rate in the later period was significantly lower than that in the former period (28% vs. 84%, p\0.01). The overall five-year survival rate of the entire group of 164 patients was 31%. Kaplan-Meier survival curves were compared between the two groups (Fig. 2). Both overall survival and cause-specific survival in the later period (1995–2005) were significantly better than in the former period (1980–1994). A total of 115 patients (70%) had died by the end of 2006; 96 died of the primary disease and 19 of other
Univariate and multivariate analyses for prognostic variables Using univariate analysis, 7 of the 16 variables provided a significant estimate of the prognosis for both overall survival and cause-specific survival (Table 2). The following were all identified as significant predictors of poor survival: earlier time period of surgery, large tumor, T3/T4 tumor, three or more positive nodes, presence of positive nonparaesophageal nodes, presence of positive abdominal nodes, and hospital morbidity. Survival difference was not significant between two-field and three-field lymphadenectomy. Both overall survival and cause-specific survival in patients with nonparaesophageal node metastasis were significantly worse than in those without nonparaesophageal node metastasis (Fig. 3). Multivariate analysis using Cox’s proportional hazard model was performed to reevaluate the impact of all 16 variables examined by univariate analysis (Table 3). Earlier time period of surgery and/or T3/T4 tumor were identified as independent risk factors for reduced survival. Although the difference was statistically borderline, metastatic nonparaesophageal and abdominal nodes and/or absence of postoperative chemotherapy seemed to be independent risk factors for worse survival. Overall fiveyear survival of the subgroup with T3/T4 tumors positive for nonparaesophageal nodes was only 12%. Multivariate analysis using Cox’s proportional hazard model was also done to evaluate the impact of each nonparaesophageal node station (Table 4). Among nonparaesophageal nodes, metastatic subcarinal nodes and posterior mediastinal nodes were independent risk factors
123
258 Table 1 Comparison of clinicopathologic features in 164 patients with thoracic node metastasis according to time period of surgery
a
Three-field lymphadenectomy
World J Surg (2008) 32:255–261 p valueb
Variables
Total No. patients
Former period (1980–1994) (n = 83)
Later period (1995–2005) (n = 81)
Male/Female
140/24
73/10
67/14
0.38
Age: \65 / ‡65 years
83/81
37/46
46/35
0.12
Tumor location: upper/lower
22/142
7/76
15/66
0.07
Tumor size: \50 mm / ‡50 mm
82/82
42/41
40/41
[0.99
Tumor depth: T1T2/T3T4
52/112
24/59
28/53
0.50
Number positive thoracic nodes: 1 or 2 / ‡3
100/64
47/36
53/28
0.27
Paraesophageal node metastasis (+) Nonparaesophageal metastasis (+)
97 113
51 59
46 54
0.87 0.61
Cervical node metastasis (+)
48
24
24
[0.99
Abdominal node metastasis (+)
80
45
35
0.16 0.10
Venous invasion (+)
124
58
66
Intramural metastasis (+)
37
25
12
3FLD / 2FLDa
110/54
45/38
65/16
Postoperative chemotherapy
41
27
14
0.53
Hospital morbidity
93
70
23
\0.01
Hospital mortality
6
6
0
0.02 \0.01
b
(A) Over-all Survival 100
100
P
