Fax: (713) 792-4049
The number of lymph nodes with metastasis predicts survival in patients with esophageal or esophagogastric junction adenocarcinoma who receive preoperative chemoradiation
Article first published online: 2 FEB 2006
Copyright © 2006 American Cancer Society
Volume 106, Issue 5, pages 1017–1025, 1 March 2006
How to Cite
Gu, Y., Swisher, S. G., Ajani, J. A., Correa, A. M., Hofstetter, W. L., Liao, Z., Komaki, R. R., Rashid, A., Hamilton, S. R. and Wu, T.-T. (2006), The number of lymph nodes with metastasis predicts survival in patients with esophageal or esophagogastric junction adenocarcinoma who receive preoperative chemoradiation. Cancer, 106: 1017–1025. doi: 10.1002/cncr.21693
- Issue published online: 17 FEB 2006
- Article first published online: 2 FEB 2006
- Manuscript Accepted: 30 SEP 2005
- Manuscript Revised: 25 AUG 2005
- Manuscript Received: 17 JUN 2005
- lymph node;
The survival of patients with locoregional adenocarcinoma of the esophagus or the esophagogastric junction (EGJ) who receive preoperative chemoradiation is reported to be better among patients who achieve a pathologic complete response than among patients who have residual tumor, including lymph node (LN) metastasis. However, the prognostic significance of the number of LNs with residual metastasis remains unclear.
The authors studied 187 consecutive patients who received chemoradiation followed by an esophagectomy. The number of positive LNs and the size of metastatic tumor in each positive LN were examined with regard to overall survival (OS) and recurrence-free survival (RFS).
A pathologic complete response was achieved by 29% of patients. No LN metastasis (posttherapy pathologic negative LN status [ypN0]) was present in 49% of patients who had residual carcinoma, and LN metastasis (ypN1) was present in 51% of patients. The 5-year OS and 2-year RFS rates achieved by patients who had 1 positive LN (34% and 45%, respectively) were similar to the rates achieved by patients in the ypN0 group (38% [P = 0.84] and 50% [P = 0.77], respectively) but were significantly better than the rates achieved by patients who had ≥ 2 positive LNs (6% [P = 0.02] and 18% [P = 0.01], respectively). The size of metastatic tumor in LNs among patients who had 1 positive LN was a prognostic factor (≥ 4 mm vs. < 4 mm; P = 0.04). In multivariate analysis, OS was better in patients who had 1 LN metastasis among patients in the ypN1 group (P = 0.02) independent of their posttherapy pathologic tumor status.
The current results suggested that the number of LNs with metastasis is an independent prognostic factor in patients with residual adenocarcinoma of the esophagus or the EGJ after preoperative chemoradiation. The authors suggest modification of the tumor-lymph node-metastasis (TNM) staging classification (ypTNM) to include the number of positive LNs in the ypN1 category. Cancer 2006. © 2006 American Cancer Society.
The incidence of esophageal adenocarcinomas has increased by 5–10% annually in the U.S. during the past 2 decades.1–5 Adenocarcinoma of the esophagus is now more prevalent than squamous cell carcinoma in the U.S. and western Europe, with most tumors located in the distal esophagus.1–5 The prognosis for patients who have adenocarcinoma of the esophagus and the esophagogastric junction (EGJ) is dismal, with a 5-year survival rate of approximately 20% after curative resection.5, 6 Although the role of preoperative therapy for esophageal carcinoma remains controversial,7–11 it was shown that preoperative chemoradiation in addition to surgery for esophageal carcinoma downstaged esophageal carcinoma and prolonged survival in some trials.7, 12–24 This is particularly true for patients who achieved a pathologic complete response (CR) after therapy.12–19, 24
It has been shown that the prognosis for patients with esophageal carcinoma who undergo curative surgical resection but do not receive neoadjuvant therapy is correlated most strongly with the American Joint Committee on Cancer (AJCC) pathologic tumor-lymph node-metastasis (pTNM) status.25 In addition, the status of lymph node (LN) metastasis (pN1) is an important independent factor in predicting the prognosis for patients who undergo surgical resection.25–27 In addition, it has been found that the number of metastatic LNs, the location of LNs, and extracapsular involvement also influence the survival of patients who undergo surgery alone without neoadjuvant therapy.28–34
In patients who receive preoperative chemoradiation and undergo subsequent surgery, the posttherapy pathologic status and the extent of residual carcinoma according to the posttherapy pTNM (ypTNM) classification system are independent prognostic factors of overall survival (OS).35, 36 However, in this setting, the prognostic significance of the characteristics of LN metastases remains unclear. In the current study, we investigated whether the number of involved LNs in patients who had positive posttherapy LN status (ypN1) and the size of LN metastases affected the survival of patients with adenocarcinoma of esophagus or EGJ who received preoperative chemoradiation.
MATERIALS AND METHODS
This study included 187 consecutive patients who had histologically confirmed adenocarcinoma of the esophagus or the EGJ and were treated at The University of Texas M. D. Anderson Cancer Center with preoperative chemoradiation followed by esophagectomy between 1985 and 2003. Patients with clinical Stage II–IVA disease according to the AJCC staging system were eligible for the current study, but patients who had clinical Stage I or Stage IVB disease with systemic metastases were ineligible.25 This analysis was approved by the Institutional Review Board.
Three agents were used for preoperative chemotherapy: fluorouracil, cisplatin, and a taxane. All patients underwent fluoroscopic simulation (before 1999) or computed tomographic simulation (since 1999) for radiation therapy. Macroscopic tumor plus 5-cm proximal and distal margins and with 2-cm lateral margins were encompassed in the treatment field before 3-dimensional conformal radiotherapy (3-D CRT) became available in 1999. When 3-D CRT came into use, the clinical target volume was defined as the macroscopic tumor volume plus a 5-cm margin superior to the highest extension and inferior to the lowest extension of the carcinoma with a 2-cm radial margin. The planning target volume was defined as the clinical target volume plus a 5-mm margin. The total dose of radiation therapy was 45.0 grays (Gy) in 25 fractions or 50.4 Gy in 28 fractions prescribed to cover at least 95% of the planning target volume. Radiation therapy was given with megavoltage (cobalt or 6–18-megavolt photons) equipment with anterior and posterior fields for 20–22 fractions followed by oblique or lateral fields for the remaining fractions to spare the spinal cord. For patients who received treatment before the 3-D CRT era, all fields were treated daily with 3-D CRT. Four to 6 weeks after the completion of preoperative therapy, patients underwent esophagectomy. The surgical approaches used were Ivor–Lewis esophagectomy (abdominal-right thoracic approach) in 111 patients (59.4%), 3-field McKeown esophagectomy (right thoracic abdominal-cervical approach) in 24 patients (12.8%), and transhiatal esophagectomy (abdominal-cervical approach) in 51 patients (27.3%). One patient (0.5%) underwent 2-stage esophagectomy.
Assessment of Residual Carcinoma and LN Status
Residual disease status in patients who received preoperative chemoradiation was determined in esophagogastrectomy specimens. Residual carcinoma pathologic stage (Stage 0, tumor in situ [Tis]-N0M0; Stage I, T1N0M0; Stage IIA, T2N0M0 and T3N0M0; Stage IIB, T1N1M0 and T2N1M0; Stage III, T3N1M0 and T4N0–N1M0; Stage IV, Tis-4N0–N1M1; Stage IVA, Tis-4N0–N1M1a; and Stage IVB, Tis-4N0–N1M1b) was assessed according to the AJCC staging manual25 and the extent of residual carcinoma was evaluated and assigned to 1 of 4 categories: no residual carcinoma, 1–10% residual carcinoma, 11–50% residual carcinoma, and > 50% residual carcinoma, as described previously.35
All LN sections were examined, and the number of LNs with tumor metastasis was recorded. Posttreatment fibrosis without viable tumor cells or mucin pools without viable tumor cells were classified as no tumor in the LNs (ypN0). The percentage of tumor that involved each LN was measured semiquantitatively as the percentage of surface area involved by the metastatic tumor. The size of metastasis in the LN was measured as the greatest dimension of tumor in each positive LN. If multiple tumor cell nests were present in a fibrotic LN, then the size of the metastasis was measured as the greatest dimension among tumor nests when they were in the same × 40 power field. If tumor cells clusters were in a mucin pool, then the size of viable tumor cell nests was considered the size of the metastasis.
Chi-square or Fisher exact tests were used to compare categorical data. OS and recurrence-free survival (RFS) were calculated from the time of surgery to the time of death from any cause, or disease recurrence, or to time of last follow-up, at which point the data were censored. OS and RFS curves were constructed by using the Kaplan–Meier method, and log-rank tests were used to evaluate the statistical significance of differences. The cumulative disease-specific 5-year OS and 2-year RFS rates were calculated by using the Kaplan–Meier method. In the multivariate analysis, independent prognostic factors for survival were determined by using a Cox regression hazards model.
Statistical analyses were performed using SPSS software (version 184.108.40.206 for Windows; SPSS Inc., Chicago, IL). Kaplan–Meier survival curves were drawn with GraphPad Prism (version 4 for Windows; GraphPad Software, San Diego, CA). We used a 2-sided significance level of 0.05 and a power of 0.90 for all statistical analyses.
Characteristics of Patients and Pathologic Findings
The mean patient age was 59.7 years (range, 32–79 yrs), and the mean follow-up was 52.7 months (range, 17.7–161.7 mos). There were 174 men and 13 women. The vast majority of tumors (98%) were located in the distal esophagus or EGJ, and only 2% were located in the middle or upper esophagus. Pretreatment clinical stage was determined by barium-swallow esophagogram, computed tomographic scan, endoscopic ultrasonography, and positron emission tomography. The pretreatment clinical stage was Stage II in 54% of patients, Stage III in 39% of patients, and Stage IVA with celiac LN involvement in 6% of patients. A pathologic CR was achieved by 54 of 187 patients (29%). The remaining 133 patients had residual disease either at the primary site (65 patients), in the LNs (6 patients), or both (62 patients).
The clinicopathologic characteristics of 133 patients who had residual disease are summarized in Table 1. Among 133 patients who had residual disease, 65 patients (49%) had no LN metastasis (ypN0), and 68 patients (51%) had LN metastasis (ypN1). There were no significant differences in age, gender, tumor location, operative procedure, positive margins, or chemoradiation sequence between N0 patients and N1 patients. Pathologic stage was significantly higher and extent of residual carcinoma was significantly greater in patients who had LN involvement (P < 0.001 and P = 0.001, respectively).
|Characteristic||No. of patients (%)|
|ypN0 status (n = 65 patients)||ypN1 status|
|One positive LN (n = 21 patients)||Two or more positive LNs (n = 47 patients)|
|Male||61 (93.8)||21 (100.0)||45 (95.7)|
|Female||4 (6.2)||0 (0.0)||2 (4.3)|
|Age in yrs (mean ± SD)||59.6 ± 10.0||61.8 ± 7.94||59.7 ± 9.9|
|Cervical/upper/middle||1 (1.5)||0 (0.0)||1 (2.2)|
|Lower/EGJ||64 (98.5)||21 (100)||46 (97.8)|
|T0||0 (0.0)a||5 (23.8)ab||1 (2.1)b|
|T1||21 (32.3)a||3 (14.3)ab||6 (12.8)b|
|T2||16 (24.6)a||3 (14.3)ab||3 (6.4)b|
|T3||28 (43)a||10 (47.6)ab||37 (78.7)b|
|Type of esophagogastrectomy|
|Transthoracic (Ivor– Lewis)||34 (52.3)||13 (61.9)||32 (68.1)|
|Transhiatal||22 (33.8)||4 (19.0)||9 (19.1)|
|Total (three-field technique)||9 (13.8)||4 (19.0)||5 (o13.2)|
|Two-stage procedure||0 (0.0)||0 (0.0)||1 (2.1)|
|Chemoradiation||35 (53.8)||10 (47.6)||31 (66.0)|
|Chemotherapy chemoradiation||30 (46.2)||11 (52.4)||16 (43.0)|
|Positive margin, radial, proximal, distal|
|No||60 (92.3)||19 (90.5)||40 (85.1)|
|Yes||5 (7.7)||2 (9.5)||7 (14.9)|
|No. of dissected LNs (mean ± SD)||14.2 ± 7.1||17.8 ± 12.1||18.1 ± 8.8|
|Extent of residual carcinoma|
|0%||0 (0.0)c||5 (23.8)cd||1 (8.8)d|
|1–50%||51 (78.5)c||10 (47.6)cd||25 (51.5)d|
|> 50%||14 (21.5)c||6 (28.6)cd||21 (44.7)d|
In 68 patients who had positive LN status, 21 patients (31%) had 1 positive LN, 19 patients (28%) had 2 positive LNs, 15 patients (22%) had 3 or 4 positive LNs, and 13 patients (19%) had ≥ 5 positive LNs. There were no significant differences with regard to age, gender, tumor location, surgical procedure, positive margins, number of LNs dissected, or chemoradiation sequence between the group with 1 positive LN and the group with ≥ 2 positive LNs. Pathologic stage was significantly higher and the extent of residual carcinoma was significantly greater in patients who had ≥ 2 positive LNs than in patients who had 1 positive LN (P = 0.009 and P = 0.02, respectively) (Table 1). The extent of residual carcinoma (P < 0.001) and pathologic stage (P = 0.002) in the group with 1 positive LN differed significantly compared with the group with ypN0 status, because there were 5 patients who had LN involvement in 1 LN but no residual primary tumor in the esophagus (Table 1).
Metastasis Size in LNs
The total number of LNs dissected in 186 patients was 2969, and the average number was 16 LNs per patient (range, 1–54 LNs per patient). The number of LNs dissected in patients who had LN metastasis (18.0 ± 9.3 LNs) was significantly higher compared with the number in patients without LN metastasis (14.2 ± 7.1 LNs; P = 0.01). There was no significant difference noted in the number of dissected LNs between patients with ypN0 status and patients with 1 positive LN or between the group with 1 positive LN and the group with ≥ 2 positive LNs (Table 1).
In total, there were 211 positive LNs. The size of metastasis in the LNs measured from 0.5 mm to 18 mm in greatest dimension, and the tumor occupied 5–100% of the LNs.
Prognostic Value of LN Status
Patients who achieved a pathologic CR to preoperative chemoradiation had a better prognosis than patients who had residual disease (Fig. 1). The 5-year OS and 2-year RFS rates were 56.2% and 70.1%, respectively, in patients who achieved a pathologic CR and 24.6% and 38.9%, respectively, in patients who had residual tumor (P < 0.001 and P = 0.005, respectively). In patients who had residual tumor, the presence of LN metastasis (ypN1) was associated with decreased survival (Fig. 1). The mean OS was 42.7 months. and the mean RFS was 28.8 months in ypN0 patients, which was significantly better compared with the OS and RFS in noted ypN1 patients (28.7 mos and 22.1 mos, respectively; P = 0.003 and P = 0.04, respectively). The 5-year OS and 2-year RFS rates in patients who had residual tumor were significantly better in the ypN0 group compared with the ypN1 group (38% vs. 15%, respectively [P = 0.003] and 50% vs. 28%, respectively [P = 0.04]).
In patients who had positive LN status (ypN1), the number of LNs involved was associated with OS and RFS (Fig. 2). The mean OS was 47.1 months (5-yr OS rate of 34%), and the mean RFS was 29.6 months (2-year RFS rate of 45%) in patients who had 1 positive LN, compared with 21.2 months (5-year OS rate of 6%) and 14.5 months (2-year RFS rate of 18%), respectively, for patients who had ≥ 2 positive LNs (P = 0.02 and P = 0.01, respectively). The 5-year OS and 2-year RFS rates remained significantly better in patients who had 1 positive LN than in patients who had ≥ 2 positive LNs when 6 patients with no residual tumor in the esophagus (ypT0) were excluded from the analysis (P = 0.02 and P = 0.048, respectively). There was no significant difference in OS noted among the patients with 2 positive LNs, 3 or 4 positive LNs, and ≥ 5 positive LNs (P = 0.35). The OS and RFS were similar among patients who had 1 positive LN and patients in the ypN0 group (P = 0.84 and P = 0.77, respectively).
Univariate Cox regression analysis of OS among patients with LNs metastasis (the ypN1 group) demonstrated that age, the extent of residual disease, and the number of positive LNs were prognostic factors (Table 2). In the multivariate analysis, we observed that the OS was better in patients who had 1 positive LN than in patients who had ≥ 2 positive LNs (P = 0.02; P = 0.04 when the 6 patients with no residual tumor in the esophagus were excluded from analysis), and OS was independent of ypT status (Table 3).
|Characteristic||No. of patients||HR||95% CI||P value|
|Type of esophagectomy||0.99|
|Transthoracic (Ivor–Lewis) esophagectomy (reference)||45||1.00|
|Total (three-field technique) esophagectomy||9||1.11||0.49–2.53||0.80|
|At least one positive margin|
|Extent of residual tumor in esophagus||0.05|
|Lower one-third (reference)||23||1.00|
|No. of positive lymph nodes|
|Variable||HR||95% CI||P value|
|No. of positive lymph nodes|
Prognostic Significance of the Size of Metastatic LNs
The size of metastatic tumor in involved LNs ranged from 0.5 mm to 16 mm in 21 patients who had only 1 positive LN. The percentage of tumor that involved these LNs ranged from 1% to 90%, and no influence was observed on OS. In this group of patients, the OS rate was significantly better when the metastatic tumor measured < 4 mm than when the metastatic tumor measured ≥ 4 mm (P = 0.04) (Fig. 3).
LN involvement is an important factor in determining prognosis for patients who undergo curative resection for carcinoma of the esophagus or EGJ.28–42 It has been established that there is a direct correlation between the depth of tumor invasion and the presence of LN metastases. Among patients who have T3 tumors, approximately 80% will have LN involvement.40, 43 Among patients with positive LN status (pN1), the number of involved LNs is a major prognostic factor.28–31, 44, 45 Patients who had limited numbers of metastatic LNs (i.e., from < two to eight LNs in different studies) had a significantly better survival compared with patients who had larger numbers of metastatic LNs in their esophagectomy specimens.27–32, 38 In patients who receive preoperative therapy, the ypTNM classification is applicable, but the prognostic significance of the number of LNs metastases remains unclear.
In the current study, we focused on patients with esophageal adenocarcinoma who had received preoperative chemoradiotherapy followed by surgery. Our study revealed that LN metastasis is an important prognostic factor for OS and RFS in this setting. In addition, the number of LNs with metastasis is important and useful for predicting prognosis. Patients who had a single LN with metastasis had a significantly better OS and RFS than patients who had two or more LNs with metastasis. Multivariate analysis in the current study clearly showed that the involvement of two or more LNs was an independent adverse prognostic determinant in patients with residual carcinoma irrespective of the depth of tumor invasion (ypT category). In addition, the 5-year OS and 2-year RFS rates among patients who had 1 metastatic LN were similar to those among patients who had no LN metastasis. Therefore, patients with a single metastatic LN are considered to have a better prognosis, like patients with no LN metastasis (ypN0). The reasons why patients who had only one positive LN in our study had a better prognosis than patients who had two or more positive LNs after preoperative chemoradiation are not clear. One possible explanation is that the number of LN metastasis is a surrogate marker for the extent of metastatic tumor load that is resistant to chemoradiation.
There has been some controversy regarding the site of LN involvement in patients with esophageal carcinoma who undergo surgery alone. Some authors reported that, if LN involvement was confined to the peritumoral region, then the prognosis was significantly better than if LN involvement was located in more distant LNs.45 In contrast, Nigro reported that the location of involved LNs did not predict the likelihood of recurrence or death in patients with esophageal adenocarcinoma.46 In the curent study, because of its retrospective nature, the regional LNs were not divided into different regions in most patients; therefore, it was not clear whether the location of the involved LNs affected patient outcomes after they received preoperative chemoradiation.
In the current study, we found that patients who had a metastasis size < 4 mm had a better OS than patients who had larger metastases, and this parameter may be used to separate OS further among patients with only 1 positive LN. In breast carcinoma, patients who had smaller LN metastases have demonstrated a significantly better prognosis than patients who had larger LN metastases.47 The size criterion of tumor nests in the LNs has been introduced into the TNM staging system for breast carcinoma. The prognostic significance of the size of tumor metastases in the LNs for patients with esophageal carcinoma has been investigated by two groups among patients who did not receive preoperative treatment.48, 49 Komori et al. found that the area of the largest tumor nest in the LNs was one of the most significant prognostic factor in patients with esophageal carcinoma: Patients whose largest metastasis measured < 4 mm had a prognosis similar to that of patients without LNs metastasis, whereas the 5-year survival rate was extremely poor among patients whose largest metastasis measured > 25 mm2.48 The report by Doi et al. indicated that the size of metastasis in the thickest part of the LN was a meaningful prognostic factor.49 Our results extend the importance of the size of LN metastasis to patients who receive preoperative neoadjuvant chemoradiation.
It has been shown that posttherapy pathologic stage (ypTNM) predicts survival in patients with carcinoma of the esophagus or the EGJ who receive preoperative chemoradiation.35 In the current study, the results suggest that the number of LNs with metastases also is a prognostic factor in patients who have LN metastasis. The OS and RFS rates achieved by patients who had one positive LN were similar to the rates achieved by patients who had no LN metastasis (ypN0) and were significantly better than the rates achieved by patients who had two or more LN metastases. The results of the current study suggest that modification of the ypTNM classification system should include categories of one residual positive LN (ypN1a) and more than one positive LN (ypN1b).
- 22Irinotecan/cisplatin followed by 5-FU/paclitaxel/radiotherapy and surgery in esophageal cancer. Oncology. 2003; 17: 20–22., , , et al.
- 25GreeneFL, PageDL, FlemingID, et al., editors. AJCC cancer staging manual, 6th ed. New York: Springer-Verlag, 2002.