Abstract
Background:
Induction therapy has not been proven to be beneficial for patients with clinical T2N0 esophageal adenocarcinoma. Surgery alone is associated with disappointing survival for patients found to have nodal disease on final pathologic examination. The aim of this study was to identify factors that predict pathologic nodal involvement in patients with endoscopic ultrasound (EUS)–proven T2N0 esophageal adenocarcinoma.
Methods:
We retrospectively reviewed patients with EUS-staged T2N0 (uT2N0) esophageal adenocarcinoma treated with surgery alone. Final pathologic staging was compared with clinical staging. Demographic and clinicopathologic variables were evaluated as putative risk factors for nodal metastases. Logistic regression models were used to identify factors associated with nodal involvement. Kaplan-Meier analysis was performed to compare overall and recurrence-free survival between patients with (N+) and without (N−) nodal disease.
Results:
We identified 80 patients with uT2N0 esophageal adenocarcinoma treated with surgery alone. Clinical staging with EUS was inaccurate for 73 patients (91%). Twenty-eight patients (35%) had pathologic N+ disease at resection. Five-year overall survival was 67% for N− patients and 41% for N+ patients (p=0.006). Recurrence-free survival was 65% for N− patients and 32% for N+ patients (p=0.0043). Univariable analysis identified vascular invasion and neural invasion as risk factors for nodal metastasis. Multivariable analysis identified vascular invasion as an independent predictor of pathologic nodal involvement.
Conclusions:
EUS is inaccurate for staging of T2N0 esophageal adenocarcinoma and often fails to identify nodal involvement. Identification of vascular invasion on preoperative biopsy should be explored as a prognostic marker to select patients for induction therapy.
Keywords: Esophageal adenocarcinoma, endoscopic ultrasound, clinical stage, nodal metastasis, vascular invasion
Introduction
The current consensus is that patients with early-stage, T1N0 esophageal adenocarcinoma should undergo resection (either endoscopic or surgical) without induction therapy. Similarly, it is not a topic of debate that locally advanced disease (T3N1–3) should be treated with neoadjuvant therapy followed by surgery (1). Controversy, however, surrounds the management of patients with clinical T2N0 (cT2N0) esophageal adenocarcinoma.
One of the main difficulties in treating patients with T2N0 disease is the inability to obtain reliable clinical staging (2, 3). Although the use of endoscopic ultrasound (EUS) has improved the ability to clinically stage esophageal cancer, accuracy is lower for early-stage tumors (T1–2N0) than for advanced tumors (4, 5). It has been reported that approximately 50% of patients with cT2N0 esophageal cancer are upstaged at resection (3, 6). The finding that nodal involvement is often seen on pathologic examination of cT2N0 patients and the knowledge that nodal metastasis is associated with poor survival in those treated with surgery alone have led many to question whether induction therapy may be beneficial for these patients. Although some authors advocate neoadjuvant therapy followed by surgery for patients with cT2N0 disease (7), previous studies have failed to establish a benefit of a multimodality approach, in terms of survival, for these patients (8, 9). Furthermore, patients with stage I and II disease who were treated with neoadjuvant chemoradiotherapy had higher postoperative mortality than those treated with surgery alone (10).
Personalized treatment through selective use of induction therapy for patients with a higher risk of nodal disease may offer potential survival benefits for patients who are clinically understaged and may prevent unnecessary morbidity for patients who are overstaged. Whether we can preoperatively predict the final pathologic nodal stage for patients with cT2N0 disease, however, remains to be answered. Therefore, the aim of this study was to identify risk factors for pathologic nodal involvement in patients with EUS-staged T2N0 (uT2N0) esophageal adenocarcinoma.
Methods
This study was approved by the Institutional Review Board at Memorial Sloan Kettering Cancer Center. Patients with uT2N0 esophageal adenocarcinoma were identified from a prospectively maintained database and retrospectively reviewed. Patients who underwent neoadjuvant treatment were excluded. Surgical approach was dependent on tumor location and surgeon preference and included open or minimally invasive Ivor Lewis, transhiatal, or three-hole esophagectomy. To assess the accuracy of ultrasound staging, final pathologic staging was compared with clinical staging. The presence of lymph node metastasis was identified from the surgical pathologic report, and patients with nodal disease (N+) were compared with patients without nodal disease (N−). Nodal involvement was analyzed for association with overall survival (OS) and recurrence-free survival (RFS).
Demographic characteristics (age, sex, race), comorbidities (cardiac disease, pulmonary disease, diabetes), tumor location, size, and pathologic variables (presence of signet ring cells, tumor differentiation, vascular invasion, neural invasion, multicentricity, and number of nodes resected) were evaluated as putative risk factors for nodal metastases.
Descriptive data are summarized as frequencies and percentages for categorical variables and medians and interquartile ranges (IQRs) for continuous variables. Characteristics were compared by nodal status using Fisher’s exact test and the Wilcoxon rank sum test for categorical and continuous variables, respectively. Univariable analysis for association between clinicopathologic factors and pathologic nodal metastasis was performed using logistic regression analysis. Variables with p<0.1 on univariable analyses and tumor length, considered a clinically important risk factor for nodal involvement, were included in multivariable regression analysis. OS and RFS were calculated from the date of surgery until the event of interest (death from any cause for OS and recurrence or death for RFS). Patients still alive were censored on the date of the last follow-up. OS and RFS were estimated using a Kaplan-Meier method and compared between nodal status using log-rank tests. Two-sided p<0.05 was considered significant. All statistical analysis were performed using R version 3.2.4 (www.R-project.org).
Results
We identified 116 patients with uT2N0 disease treated at our center from 1997 to 2016. Thirty-six patients who received induction therapy were excluded, and 80 patients treated with either surgery alone (86%, N=69) or surgery followed by adjuvant chemotherapy (14%, N=11) were included as the study population. The median age of these patients was 64 years (IQR, 57–71 years); 62 (78%) were male. An Ivor Lewis approach was performed in 62 patients (78%), three-hole esophagectomy in 6 (8%), transhiatal esophagectomy in 3 (4%), and other approach in 9 (11%). A minimally invasive approach was used in 18 patients (23%).
Clinical staging with EUS was inaccurate for 73 patients (91%). At resection, 40 patients (50%) were downstaged, 33 (41%) were upstaged, and only 7 (9%) had no change in stage. Overall, 28 patients (35%) were found to have positive nodes at resection: 13 with N1, 8 with N2, and 7 with N3 disease (Table 1). The median number of metastatic nodes in N+ patients was as follows: 1 for N1, 3.5 (IQR, 3.0–4.0) for N2, and 11 (IQR, 10.0–11.0) for N3.
Table 1.
Final pathologic staging of patients with clinical T2N0 esophageal adenocarcinoma (n=80)
| Stage | T1a | T1b | T2 | T3 | Total |
|---|---|---|---|---|---|
| N0 | 9 | 31 | 7 | 5 | 52 |
| N1 (1–2 nodes) | 0 | 5 | 5 | 3 | 13 |
| N2 (3–6 nodes) | 0 | 2 | 2 | 4 | 8 |
| N3 (≥7 nodes) | 0 | 2 | 0 | 5 | 7 |
Furthermore, the number of lymph nodes retrieved was affected by surgical approach. Patients who underwent minimally invasive esophagectomy had higher numbers of nodes harvested (median, 27 [IQR 20–45]) than patients who underwent resection with an open approach (median, 23 [IQR 19–28]). Similarly, more nodes were harvested with Ivor Lewis (median, 24 [IQR, 19–32]) and 3-hole (median, 27 [IQR 22–33]) esophagectomies than with a transhiatal approach (median, 8 [IQR, 7.5–16]).
Demographic characteristics, tumor location and size, and pathologic variables for N+ and N− patients are summarized in Table 2. Eleven of 28 patients with positive nodes on final pathologic evaluation received adjuvant treatment; among these patients, 4 had pN1 disease (with 1 or 2 nodes involved), 4 had pN2 disease, and 3 had >7 metastatic nodes (pN3 disease).
Table 2.
Patient characteristics according to nodal status
| Characteristic | N− (n=52) | N+ (n=28) | P |
|---|---|---|---|
| Positive nodes, median (IQR) | 0 | 3.0 (1.0–6.5) | |
| Age, years, median (IQR) | 64.0 (57.0–71.0) | 64.5 (57.8–69.0) | 0.86 |
| Male | 40 (77) | 22 (79) | 0.99 |
| White | 44 (85) | 27 (96) | 0.15 |
| Cardiac disease | 12 (23) | 6 (21) | 0.99 |
| Pulmonary disease | 4 (8) | 2 (7) | 0.99 |
| Diabetes | 10 (19) | 5 (18) | 0.99 |
| Tumor location | 0.85 | ||
| Distal | 20 (38) | 10 (36) | |
| Gastroesophageal junction | 24 (46) | 15 (54) | |
| Proximal stomach | 8 (15) | 3 (11) | |
| Tumor size, cm, median (IQR) | 2.4 (1.6–3.0) | 2.6 (2.0–3.2) | 0.23 |
| Signet ring cell | 4 (8) | 4 (14) | 0.44 |
| Poor differentiation | 18 (36)a | 16 (57) | 0.096 |
| Vascular invasion | 14 (28)a | 20 (74)b | <.001 |
| Neural invasion | 7 (14)a | 10 (38)a | 0.021 |
| Multicentricity | 2 (4) | 2 (7) | 0.62 |
| Total nodes resected, median (IQR) | 24.0 (18.8–29.2) | 22.5 (19.0–32.0) | 0.96 |
Data are no. (%) unless otherwise noted. IQR, interquartile range; N−: no nodal metastasis; N+: presence of nodal metastasis.
Data unavailable for 2 patients
Data unavailable for 1 patient.
PET scan was performed in 72 patients. The median PET standardize uptake value (SUV) was 4 (IQR, 1.8–6.4) in patients without nodal involvement and 6.2 (IQR, 2.6–8.9) in patients with nodal metastasis; comparison of PET-SUV showed no statistically significant difference between the two groups of patients (p=0.24).
Univariable analysis identified vascular invasion and neural invasion as risk factors for nodal metastases, whereas tumor size, tumor location, and presence of signet ring cells were not correlated with nodal involvement. A weak association was found with poor differentiation grade (Supplementary Table 1). In the final multivariable analysis, only vascular invasion was an independent predictor of pathologic nodal involvement (odds ratio [OR], 5.24 [95% confidence interval [CI], 1.68–17.73]; p=0.005) (Table 3).
Table 3.
Multivariable logistic regression for odds of nodal metastasis
| Characteristic | OR (95% CI) | P |
|---|---|---|
| Tumor size, per cm | 1.01 (0.63–1.53) | 0.96 |
| Poor differentiation | ||
| No | 1.0 (reference) | |
| Yes | 1.34 (0.38–4.47) | 0.63 |
| Vascular invasion | ||
| No | 1.0 (reference) | |
| Yes | 5.24 (1.68–17.73) | 0.005 |
| Neural invasion | ||
| No | 1.0 (reference) | |
| Yes | 2.2 (0.58–8.34) | 0.24 |
There were 36 deaths in this study, with a median follow-up among survivors of 5.1 years (IQR, 0.23–16.5). Five-year OS was 67% for N− patients and 41% for N+ patients (p=0.006) (Figure 1). There were 24 recurrences and 14 deaths without recurrence, for a total of 38 RFS events. The majority of patients with recurrence (79%) had distant metastases. Five-year RFS was 65% for N− patients and 32% for N+ patients (p=0.0043) (Figure 2). Survival was not only associated with nodal status but was also strongly driven by the number of metastatic nodes.
Figure 1.
Three-year and 5-year overall survival for patients with (N+) and without (N−) nodal disease.
Figure 2.
Three-year and 5-year recurrence-free survival for patients with (N+) and without (N−) nodal disease.
Discussion
In our cohort of patients with T2N0 disease, ultrasound staging was often unreliable, with an inaccuracy rate of 91%. Nodal disease was identified at resection in 35% of patients with uT2N0 disease, and these patients, who were treated with surgery alone or with surgery and adjuvant chemotherapy, had disappointing survival outcomes. Vascular invasion was identified as an independent risk factor for nodal involvement, and this finding may help guide treatment for patients with cT2N0 esophageal adenocarcinoma.
Although EUS remains the gold standard for staging esophageal cancer, its accuracy for T2N0 disease is poor. One of the main limitations of EUS staging is its low sensitivity to detect nodal disease, and in our cohort, most of the understaged patients were misstaged due to occult nodal involvement. Similarly, Crabtree et al. (11) analyzed a smaller cohort of patients and found that 39% of cT2N0 patients had unanticipated pathologic nodal disease. Stiles et al. (12) reported that the prevalence of previously unsuspected nodal disease in cT2 patients was 55%. The addition of fine-needle aspiration has been shown to improve the sensitivity and specificity of EUS for the diagnosis of nodal metastases in patients with esophageal cancer (13, 14). However, suspicious nodes are often peritumoral and would require transtumoral biopsy for sampling, with obvious concerns for specimen contamination. Unfortunately, as most of the staging EUS in our cohort were performed at outside institutions, we couldn’t obtain information regarding the systematic use of fine-needle aspiration with EUS. Theoretically, radiologic imaging may help to accurately stage these patients. Nevertheless, PET has also been shown to have poor sensitivity for the diagnosis of nodal disease (15, 16), and although 90% of patients underwent staging PET in our study, nodal disease was missed in 35% of cases. Moreover, in our study, we found no difference in SUV between patients with and without nodal involvement.
As nodal status is one of the most important long-term prognostic factors for patients with esophageal adenocarcinoma (17, 18), the identification of clinical or histopathologic predictors of nodal involvement is imperative. Numerous studies have shown that, currently, the four most reliable predictors of nodal metastases in patients with superficial lesions are tumor size >2 cm, lymphovascular invasion, high-grade disease (poor differentiation), and invasion depth >500 microns or deeper into the submucosa. However, data on T2 tumors are lacking. A previous study showed that more poorly differentiated tumors were associated with pathologic upstaging in patients with cT2N0 disease (3). Concordantly, Rice et al. (2) reported that overstaged patients were more likely to have poorly differentiated tumors. Recently, Samson et al. (19) found that lymphovascular invasion (OR, 6.0 [95% CI, 2.9–12.5]) and poorly differentiated tumor grade (OR, 9.4 [95% CI, 1.8–48.4]) were independently associated with an increased risk of pathologic upstaging. We found that poor differentiation, neural invasion, and vascular invasion were associated with nodal involvement, but only vascular invasion was an independent risk factor for nodal disease in multivariable analysis (OR, 5.24 [95% CI, 1.68–17.73]).
An important challenge will be to obtain information on vascular invasion from the preoperative endoscopic biopsy. In fact, the frequency with which vascular invasion is reported from endoscopic biopsy is highly variable between institutions, as its detection depends on biopsy size and specific staining techniques. This issue was addressed by Pouw et al. (20), who affirmed that diagnostic endoscopic resection should be included as a final step in the workup of these patients, as “diagnostic endoscopic resection provides a relatively large tissue specimen that allows accurate histologic staging of the infiltration depth as well as other prognostic factors such as tumor differentiation grade and lymphatic and vascular involvement.” The larger samples afforded by endoscopic resection, compared with those harvested by cold forceps biopsy, facilitate investigation of these pathologic features for their prognostic value. Further supporting the use of endoscopic resection, it alone may have been curative in the cohort of downstaged patients with intramucosal adenocarcinoma (9 patients in our study).
Gaur et al. (21) showed that endoscopic tumor length >2 cm was a predictor of nodal metastasis (OR, 7 [95% CI, 2.7–18.1]; p<0.001). In the present study, we evaluated pathologic tumor length for risk of nodal involvement; unfortunately, this variable was not a predictor of nodal metastasis.
The finding that our patients with uT2N0 disease with nodal metastases at resection had poor survival outcomes (5-year OS, 41%; 5-year RFS, 32%) establishes a rationale for further investigating whether these patients would benefit from induction therapy. Although previous reports have shown no survival benefits associated with neoadjuvant therapy in patients with cT2N0 disease (8–10), none of those studies implemented selection criteria for induction therapy. Consequently, neoadjuvant therapy was offered to a large group of patients who would go on to be downstaged after surgery—systemic therapy would logically confer no survival benefit for these patients. Thus, we hypothesize that accurate risk stratification based on pathologic features of cT2N0 tumors may help select patients for induction therapy. Another reasonable option would be to offer adjuvant therapy to patients who are upstaged after surgery. However, morbidity and debilitation after esophagectomy usually preclude patients from being candidates for adjuvant therapy or prevent beginning it in a timely fashion.
This study has several limitations. First, it was a retrospective analysis. Second, a relatively small number of patients were analyzed. Although the small cohort size is a major limitation, owing to the uncommon presentation of cT2N0 esophageal adenocarcinoma, this study still represents one of the largest series of patients with cT2N0 esophageal adenocarcinoma treated by surgery with or without adjuvant chemotherapy at a single institution. Furthermore, all patients included in the analysis were staged with EUS. Additionally, our study period spans nearly 20 years, and this can potentially affect the reliability of EUS. In fact, during this time, EUS has become a key tool for esophageal cancer staging. Its adoption in the staging process has improved both physicians’ expertise and the quality of the study over time. Last, we did not consider clinician variability as a factor that might affect the accuracy of EUS.
Conclusions
Clinical staging was inaccurate for 91% of patients with uT2N0 esophageal adenocarcinoma in our series, and the presence of nodal involvement was undetected in 35% of patients. Survival outcomes for patients with nodal disease treated with surgery alone were discouraging. We support the use of endoscopic resection as a complement to EUS to improve the accuracy of staging in these patients. Endoscopic resection not only provides information about tumor characteristics that are important for gauging the risk of nodal involvement but also can be used to improve the staging accuracy of EUS. Moreover, endoscopic resection may be curative for patients found to have intramucosal cancer. Finally, patients with uT2N0 disease with vascular invasion on preoperative biopsy may benefit from preoperative chemoradiation, as they have an increased risk of harboring occult nodal metastasis.
Supplementary Material
Acknowledgments:
A.B. is supported by a Surgeon Development award from the Esophageal Cancer Education Foundation (ECEF).This work was supported, in part, by NIH/NCI Cancer Center Support Grant P30 CA008748.
Abbreviations:
- EUS
endoscopic ultrasound
- CI
confidence interval
- cT2N0
clinical T2N0
- uT2N0
ultrasound-staged T2N0
- IQR
interquartile range
- N−
no nodal metastasis
- N+
presence of nodal metastasis
- OR
odds ratio
- OS
overall survival
- RFS
recurrence-free survival
- SUV
standardized uptake value
Footnotes
IRB Approval: This study was reviewed and approved by the Institutional Review Board of the Memorial Sloan Kettering Cancer Center with the following IRB number: 16–1631.
COI Statement: All authors have no conflicts of interest.
Disclosure: The authors have no conflicts of interest.
References
- 1.Ajani JA, Barthel JS, Bentrem DJ, et al. Esophageal and esophagogastric junction cancers. J Natl Compr Canc Netw 2011;9(8):830–887. [DOI] [PubMed] [Google Scholar]
- 2.Rice TW, Mason DP, Murthy SC, et al. T2N0M0 esophageal cancer. J Thorac Cardiovasc Surg 2007;133(2):317–324. [DOI] [PubMed] [Google Scholar]
- 3.Crabtree TD, Kosinski AS, Puri V, et al. Evaluation of the reliability of clinical staging of T2 N0 esophageal cancer: a review of the Society of Thoracic Surgeons database. Ann Thorac Surg 2013;96(2):382–390. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kutup A, Link BC, Schurr PG, et al. Quality control of endoscopic ultrasound in preoperative staging of esophageal cancer. Endoscopy 2007;39(8):715–719. [DOI] [PubMed] [Google Scholar]
- 5.Pech O, Günter E, Dusemund F, et al. Accuracy of endoscopic ultrasound in preoperative staging of esophageal cancer: results from a referral center for early esophageal cancer. Endoscopy 2010;42(6):456–461. [DOI] [PubMed] [Google Scholar]
- 6.Dolan JP, Kaur T, Diggs BS, et al. Significant understaging is seen in clinically staged T2N0 esophageal cancer patients undergoing esophagectomy. Dis Esophagus 2016;29(4):320–5. [DOI] [PubMed] [Google Scholar]
- 7.Zhang JQ, Hooker CM, Brock MV, et al. Neoadjuvant chemoradiation therapy is beneficial for clinical stage T2 N0 esophageal cancer patients due to inaccurate preoperative staging. Ann Thorac Surg 2012;93(2):429–435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Speicher PJ, Ganapathi AM, Englum BR, et al. Induction therapy does not improve survival for clinical stage T2N0 esophageal cancer. J Thorac Oncol 2014;9(8):1195–1201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Markar SR, Gronnier C, Pasquer A, et al. Role of neoadjuvant treatment in clinical T2N0M0 oesophageal cancer: results from a retrospective multi-center European study. Eur J Cancer 2016;56:59–68. [DOI] [PubMed] [Google Scholar]
- 10.Mariette C, Dahan L, Mornex F, et al. Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer: final analysis of randomized controlled phase III trial FFCD 9901. J Clin Oncol 2014;32(23):2416–2242. [DOI] [PubMed] [Google Scholar]
- 11.Crabtree TD, Yacoub WN, Puri V, et al. Endoscopic ultrasound for early stage esophageal adenocarcinoma: implications for staging and survival. Ann Thorac Surg 2011;91(5):1509–1515. [DOI] [PubMed] [Google Scholar]
- 12.Stiles BM, Mirza F, Coppolino A, et al. Clinical T2-T3N0M0 esophageal cancer: the risk of node positive disease. Ann Thorac Surg 2011;92(2):491–496. [DOI] [PubMed] [Google Scholar]
- 13.Vazquez-Sequeiros E Nodal staging: number or site of nodes? How to improve accuracy? Is FNA always necessary? Junctional tumors--what’s N and what’s M? Endoscopy 2006;38:S4–8. [DOI] [PubMed] [Google Scholar]
- 14.Marsman WA, Brink MA, Bergman JJ, et al. Potential impact of EUS-FNA staging of proximal lymph nodes in patients with distal esophageal carcinoma. Endoscopy 2006;38:825–9. [DOI] [PubMed] [Google Scholar]
- 15.Little SG, Rice TW, Bybel B, et al. Is FDG-PET indicated for superficial esophageal cancer? Eur J Cardiothorac Surg 2007;31(5):791–796. [DOI] [PubMed] [Google Scholar]
- 16.Sandha GS, Severin D, Postema E, et al. Is positron emission tomography useful in locoregional staging of esophageal cancer? Results of a multidisciplinary initiative comparing CT, positron emission tomography, and EUS. Gastrointest Endosc 2008;67(3):402–409. [DOI] [PubMed] [Google Scholar]
- 17.Cabau M, Luc G, Terrebonne E, et al. Lymph node invasion might have more prognostic impact than R status in advanced esophageal adenocarcinoma. Am J Surg 2013;205(6):711–717. [DOI] [PubMed] [Google Scholar]
- 18.Bhamidipati CM, Stukenborg GJ, Thomas CJ, et al. Pathologic lymph node ratio is a predictor of survival in esophageal cancer. Ann Thorac Surg 2012;94(5):1643–1651. [DOI] [PubMed] [Google Scholar]
- 19.Samson P, Puri V, Robinson C, Lockhart C, et al. Clinical T2N0 Esophageal Cancer: Identifying Pretreatment Characteristics Associated with Pathologic Upstaging and the Potential Role for Induction Therapy. Ann Thorac Surg 2016;101(6):2102–2111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pouw RE, Heldoorn N, Alvarez Herrero L, et al. Do we still need EUS in the workup of patients with early esophageal neoplasia? A retrospective analysis of 131 cases. Gastrointest Endosc 2011;73(4):662–8. [DOI] [PubMed] [Google Scholar]
- 21.Gaur P, Sepesi B, Hofstetter WL et al. A clinical nomogram predicting pathologic lymph node involvement in esophageal cancer patients. Ann Surg 2010; 252:611–617. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.