Diseases of the Esophagus (2014) 27, 770–776 DOI: 10.1111/dote.12154
Original article
Concordance of studies for nodal staging is prognostic for worse survival in esophageal cancer R. Dhupar, A. M. Correa, J. Ajani, S. Betancourt, R. J. Mehran, S. G. Swisher, W. L. Hofstetter Department of Thoracic and Cardiovascular Surgery, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
SUMMARY. Pretreatment clinical staging in esophageal cancer influences prognosis and treatment strategy. Current staging strategies utilize multiple imaging modalities, and often the results are contradictory. No studies have examined the implications of concordance of computed tomography (CT), positron emission tomography (PET), and endoscopic ultrasound (EUS) when used for the evaluation of nodal disease. The objective of this study was to determine if concordance of CT, PET, or EUS for nodal disease predicts worse overall survival. We reviewed 615 esophageal cancer patients with pretreatment CT, PET, and EUS that underwent esophagectomy for survival outcomes based on concordance of studies for nodal disease. Concordant N+ is defined as two or three studies positive for nodal disease; non-concordant N+ is defined as only one positive study. Node-positive disease by any study predicted shorter survival than node-negative disease (42% vs. 73% 5-year survival; P < 0.001). Additionally, non-concordant N+ patients had shorter survival than N− patients (52% vs. 73% 5-year survival; P < 0.001). Concordant N+ patients had shorter survival than non-concordant N+ patients (38- vs. 61-month median survival; P = 0.017). There were no statistically significant differences in survival based on specific combinations of studies. When PET was disregarded, patients with both CT+ and EUS+ had shorter survival than patients with either CT+ or EUS+ (39- vs. 58-month median survival; P = 0.029). Pretreatment CT, PET, or EUS concordance for node-positive disease predicts shorter overall survival in patients that undergo esophagectomy for esophageal cancer. Predicting survival in esophageal cancer should consider the synergistic capabilities of CT, PET, and EUS in evaluating nodal status. KEY WORDS: CT scan, esophageal cancer, EUS, PET scan, survival. ABBREVIATIONS: CT, computed tomography; EUS, endoscopic ultrasound; N+, pretreatment node positive; N−, pretreatment node negative; PET, positron emission tomography
INTRODUCTION Esophageal cancer is an aggressive malignancy that frequently presents at an advanced stage. In the United States, the incidence of esophageal cancer continues to rise, with approximately 17 460 new cases and 15 070 deaths expected in 2012.1 The last two decades have seen the implementation of multimodality therapy to improve cure rates and disease-free survival. However, the prognosis of this cancer remains grim, with low 5-year survivals. Address correspondence to: Dr Wayne L Hofstetter, MD, MD Anderson Cancer Center, Department of Thoracic and Cardiovascular Surgery, 1515 Holcombe Blvd, Unit 445, Houston, TX 77030, USA. Email: [email protected] 770
There are many factors that may be prognostic for overall survival in esophageal cancer. Tumor grade, growth patterns, sex, molecular markers, and stage are among a few that are believed to be important.2–6 The nodal status at the time of diagnosis can be particularly significant for determining treatment strategy, as well as prognosticating survival.7,8 However, the currently utilized methods to evaluate nodal status are limited by their accuracy.9 Computed tomography (CT), fluorodeoxyglucose positron emission tomography (PET), and endoscopic ultrasound (EUS) are frequently used to evaluate lymph nodes before treatment. Because nodal status is a key determinant of treatment and prognosis, understanding the best combination and sequence of studies for accurate clinical staging is critical for
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
Concordant studies in esophageal cancer
these patients. While all three modalities are widely used, there is a lack of consensus regarding their use alone, in combination, or in sequence.10–12 Several studies have attempted to define the sensitivity, specificity, and accuracy of these tests, but no studies have evaluated the prognostic significance of concordance of two or more modalities.10,13–16 In this study, we retrospectively reviewed a prospective database of 615 esophageal cancer patients with pretreatment nodal assessment by CT, PET, and EUS that underwent esophagectomy to determine if concordance of these studies is prognostic for survival.
MATERIALS AND METHODS Approval was granted by the MD Anderson Cancer Center Institutional Review Board for this study. We performed a retrospective review of a prospectively collected database of patients with esophageal adenocarcinoma or squamous cell cancer that underwent esophagectomy between 2002 and 2011. We included patients evaluated by pretreatment CT scan, PET, and EUS done at MD Anderson Cancer Center. We excluded patients with cervical or type III esophagogastric tumors, salvage esophagectomy, emergency esophageal surgery, and redo esophagectomy. For the 615 patients that met these criteria, we analyzed demographics, survival, and findings from CT, PET, and EUS. We did not analyze CT, PET, or EUS performed after the initiation of therapy for cancer. Criteria for interpreting CT, PET, or EUS as node positive were different for each study. For CT scan, the radiologist interpretation was used, with criterion of nodes greater than 1 cm being considered positive. A CT scan with intravenous contrast was not required. For PET, regional abnormal (standardized uptake value [SUV] greater than 2) was considered positive nodal disease, regional normal (SUV less than or equal to 2) as no nodal disease. For EUS, either pathological results from fine needle aspiration (FNA) or the gastroenterologist interpretation of nodal disease was used (which is a standard practice as defined by National Comprehensive Cancer Network guidelines for esophageal cancer staging). For the 615 total patients, FNA data were available for 95% (583/615); 30% (175/583) had an FNA, and 42% of those (73/175) were positive for cancer, which is 12% (73/615) of the entire study group. There was no pathological assessment of nodal disease as 77% of patients underwent neoadjuvant therapy (chemotherapy and/or radiation). Assessment of final pathological nodal disease after neoadjuvant therapy would not necessarily be reflective of pretreatment malignancy status. Because CT, PET, and EUS determine clinical nodal status, which guides treatment, this study focuses on clinical status only.
771
In all operations, regional abdominal and thoracic lymph node dissection was performed, with the exception of those performed in a transhiatal fashion (approximately 11%). We defined concordant node positive (N+) to mean that the results of at least two modalities (CT, PET, or EUS) were interpreted as node-positive disease. We defined non-concordant N+ to mean that the results of only one modality were interpreted as node-positive disease. For statistical analysis, Kaplan–Meier survival curves were constructed based on staging modality and node positivity. Values were considered statistically significant if P < 0.05. We report 5-year survival instead of median survival only if groups did not reach 50% mortality.
RESULTS Survival of patients with node positive disease based on study modality Patient and tumor characteristics are summarized in Table 1. Of the 615 patients with esophageal cancer that underwent staging with CT, PET, and EUS, 61% (377/615) had pretreatment node-positive (N+) disease. The median age was 62 (range 23–83), 14% (87/615) were women, and 92% (565/615) had adenocarcinoma. There was no preoperative chemotherapy or radiation in 23% (139/615) of patients. The distribution of tumors was 6% (38/615) upper/middle esophagus, 29% (175/615) lower esophagus, and 65% (401/615) gastroesophageal junction. The 5-year survival for N+ patients was 42% versus 73% for N− patients (P < 0.001; Fig. 1a). Because all patients had CT, PET, and EUS, we analyzed the percentage of studies that were positive in N+ patients: 58.6% (221/ 377) of patients were CT+, 39.8% (150/377) PET+, and 83.8% (316/377) EUS+ (Fig. 2a). In order to determine if there was a difference in survival based on which modality was positive, we created Kaplan–Meier survival curves for each modality. There was a statistically significant
Table 1
Characteristics of esophageal cancer patients and tumors
Variable
All patients % (n)
n Median age (years) % female % Caucasian % adenocarcinoma % squamous cell cancer % upper/middle esophagus % lower esophagus % gastroesophageal junction % neoadjuvant therapy cN− cN+
100 (615) 62 (na) 14 (87) 91 (562) 92 (565) 8 (50) 6 (38) 29 (175) 65 (401) 77 (475) 39 (238) 61 (377)
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
Diseases of the Esophagus
P < 0.001
1.0
Cumulative Survival Probability
a
b Cumulative Survival Probability
772
0.8 N– 0.6
0.4 N+ 0.2
0.0
1.0
0.8
0.4 CT N+ 0.2
0.0 0
Time (months)
0.8
PET N–
0.4 PET N+ 0.2
0.0
d Cumulative Survival Probability
P < 0.001
0.6
12 24 36 48 60 72 84 96 108 120 132 144
Time (months)
1.0
Cumulative Survival Probability
CT N–
0.6
0 12 24 36 48 60 72 84 96 108 120 132 144
c
P < 0.001
P < 0.001
1.0
0.8 EUS N– 0.6
0.4 EUS N+ 0.2
0.0 0
12 24 36 48 60 72 84 96 108 120 132 144
Time (months)
0
12 24 36 48 60 72 84 96 108 120 132 144
Time (months)
Fig. 1 Survival based on modality of finding N+ disease. (a) Kaplan–Meier survival curve of all N+ versus all N− esophageal cancer patients; (b) CT+ versus CT− patients; (c) PET+ versus PET− patients; (d) EUS+ versus EUS− patients.
difference in all groups (P < 0.001). Patients that were CT+ had a 5-year survival of 40% versus 64% if CT−; PET+ had a 5-year survival of 34% versus 61% if PET−; EUS+ had a 5-year survival of 39% versus 69% if EUS− (Fig. 1b–d). Patient survival if concordant N+ Overall, 35% (213/615) of the entire population and 57% (213/377) of the N+ population was concordant N+ (Fig. 2b). In order to determine if there was a difference in survival based on which studies were concordant, we created Kaplan–Meier survival curves of all combinations of concordant studies (Fig. 3). There was no statistically significant differ-
ence in survival based on which studies were concordant (P = 0.999), with median survival for CT+PET+EUS– of 38.8 months; CT+PET–EUS+ of 31.9 months; CT−PET+EUS+ of 32.3 months; and CT+PET+EUS+ of 39 months. Table 2 summarizes 5-year and median survival for all groups based on study results for nodal disease. Patient survival if non-concordant N+ Overall, 27% (164/615) of the entire population and 44% (164/377) of the N+ population was nonconcordant N+ (Fig. 2c). In order to determine if there was a difference in survival based on which studies were non-concordant N+, we created
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
Concordant studies in esophageal cancer
a
773
All Patients 400 350
84%
300 59%
250 200
40%
150 100 50 0
(377)
(221/377)
cN+
cT+
b
(150/377) PET+
(316/377) EUS+
Concordant N+ 400 350 300 250 200 150
26%
100 50 0
(377) cN+
c
4%
(97/377)
(15/377)
23%
4%
(87/377)
(14/377)
CT+PET+EUS+ CT+PET+EUS– CT+PET–EUS+ CT–PET+EUS+
Non-Concordant N+ 400 350 300 250 200 31%
150 6%
100 50 0
6%
(377)
(24/377)
(118/377)
cN+
CT–PET+EUS–
CT–PET–EUS+
(22/377) CT+PET–EUS–
Fig. 2 Bar graphs of patients with N+ disease and modality of recognition. (a) Recognition by any modality; (b) concordant N+, percentages represent fraction of N+ population (n = 377); (c) non-concordant N+, percentages represent fraction of N+ population (n = 377).
Kaplan–Meier survival curves of all combinations of non-concordant N+ studies (Fig. 4). There was no statistically significant difference in survival based on which studies were non-concordant N+ (P = 0.325), with median survival in CT+PET−EUS− of 86.2 months; CT−PET+EUS− of 45.3 months; and CT−PET−EUS+ of 58.4 months (Table 2). However, as expected, there was a statistically significant survival difference between the non-concordant N+ versus N− patients (5-year survival 52% vs. 73%; P < 0.001; curve not shown).
Patient survival if concordant N+ versus non-concordant N+ In order to determine if there was a difference in survival among patients with concordant N+ versus non-concordant N+ studies, we created Kaplan– Meier survival curves (Fig. 5). There was a significant difference in survival between the two groups (P = 0.017), with median survival for the concordant N+ group of 38.2 months versus 60.6 months in the non-concordant N+ group (Table 2). There was no
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
774
Diseases of the Esophagus
Fig. 3 Survival of concordant N+ patients. Kaplan–Meier survival curve of N+ patients based on CT+PET+EUS+, CT+PET+EUS−, CT+PET−EUS+, or CT−PET+EUS+.
significant difference in survival between patients with all studies positive (CT+PET+EUS+) versus patients with only one or two positive studies (P = 0.998; curve not shown). Because both CT and EUS rely on anatomic findings rather than metabolic activity, we compared survival for CT/EUS concordant N+ and CT/EUS non-concordant N+ (disregarding the result of the PET; curve not shown). There was a significant difference in survival between the two groups (P = 0.029), with median survival for the CT/EUS concordant N+ Table 2 Summary of study interpretations for nodal disease, number of patients in each group, median survivals, and 5-year survivals
Study All studies Concordant N+ CT+PET+EUS+ CT+PET+EUS− CT+PET−EUS+ CT−PET+EUS+ Non-concordant N+ CT+PET−EUS− CT−PET+EUS− CT−PET−EUS+ CT and EUS only Concordant N+ Non-concordant N+
n
Median survival (months)
5-year survival (%)
213 97 15 87 14 164 22 24 118
38.2* 39.0 38.8 31.9 32.3 60.6* 86.2 45.3 58.4
36 33 33 40 17 52 75 48 47
184 169
39.0** 58.4**
36 48
*Statistically significant difference with P = 0.017; **statistically significant difference with P = 0.029. CT, computed tomography; EUS, endoscopic ultrasound; PET, positron emission tomography.
group of 39.0 months versus 58.4 months for the CT/EUS non-concordant N+ group (Table 2). There was no significant difference in survival between CT+EUS− and CT−EUS+ patients (P = 0.466; curve not shown).
DISCUSSION In this study, we performed a retrospective analysis of 615 esophageal cancer patients that underwent esophagectomy to determine if concordance of CT, PET, and EUS in pretreatment nodal staging is prognostic for survival. The key findings of this paper are (i) there is shorter survival if any modality is positive for nodal disease; (ii) the patient group with concordant N+ studies had shorter survival than the group with non-concordant N+ studies; and (iii) the patient group that was CT/EUS concordant N+ had shorter survival than CT/EUS non-concordant N+. While previous studies have compared or attempted to determine the prognostic implication of any single positive study, none have examined the implications of concordance of these studies. Our finding that there is a shorter survival if any modality is positive is not surprising (Fig. 1). Although there are different diagnostic capabilities ascribed to CT, PET, and EUS, generally the patient is treated as N+ if any study is positive. This is a validation of clinical practice. Of note, in our population, 44% (164/377) of the N+ patients were positive
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
Concordant studies in esophageal cancer
775
Fig. 4 Survival of non-concordant N+ patients. Kaplan–Meier survival curve of patients with N+ disease based on CT+PET−EUS−, CT−PET+EUS−, or CT−PET−EUS+.
in one study only (non-concordant N+), 72% (118/ 164) of which were EUS+ only (Fig. 2c). Although there was no statistically significant difference in survival among patients with only one positive study (non-concordant N+ group), the number of N+ patients that were CT+ only or PET+ only was small
Fig. 5 Kaplan–Meier survival curve of patients with concordant N+ versus non-concordant N+ patients.
(46/377; Fig. 4). Perhaps there is a difference in the prognostic implication of which single modality is positive; however, it is possible that a difference could not be detected because of a small patient population. Our finding that the concordant N+ group had shorter survival than the non-concordant N+ group is a novel finding (Fig. 5). N+ disease is associated with lower overall survival, both before and after treatment.7,17 However, it is not clear if the diagnosis of pretreatment N+ disease by all modalities is equivalent. Our findings suggest that N+ disease by more than one modality may confer a worse prognosis than patients with N+ disease by only one modality. While studies have looked to determine differences in sensitivity, specificity, and accuracy of each modality, perhaps CT, PET, and EUS evaluate separate characteristics of lymph nodes that work in synergy to differentiate N+ from N−.12,16 Although we did not find a statistically significant difference among the different combinations of positive studies, again, the number of patients in this study may limit the ability to detect subtle differences. Finally, we compared CT with EUS, the two modalities that evaluate anatomic rather than metabolic characteristics of lymph nodes. By disregarding the results of the PET scan, we found a statistically significant difference in survival among patients who were CT/EUS concordant N+ versus those who were not (Fig. 5). There was no difference in survival in CT+EUS− versus CT−EUS+ patients (data not shown). As with our other findings, this also implies
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
776
Diseases of the Esophagus
a potential synergy between CT and EUS – two modalities that are thought to evaluate similar characteristics of lymph nodes. This study has several limitations. This is a single institution study, and it is unknown if performance of these procedures and interpretation of studies are generalizable. Also, while we evaluate overall survival in patients that underwent esophagectomy, this group was not homogenous for neoadjuvant treatment (Table 1). In addition, we did not evaluate for differences in EUS+ studies based on biopsies performed during EUS (12% of EUS+). While pathological findings do not rely on ultrasound characteristics, the decision to perform FNA is based on ultrasound findings. However, we did not determine if nodes positive by ultrasound characteristics alone were differentiated from nodes positive by biopsy performed during EUS. Finally, because some groups were small in number, the ability to detect subtle differences among groups may be hindered. For this study, we did not analyze final pathological status of lymph nodes. Although it can be suggested that comparison of the pretreatment staging should be verified with final pathological analysis, this should be considered with caution. Pathological node status is a function of treatment response, which can be unpredictable. A comparison of the clinical stage with the pathological stage would potentially confuse the issues addressed in this paper, as disagreement could be interpreted as either complete response or false positive staging. Along these same lines, the American Joint Committee on Cancer seventh edition staging system was formulated on entirely pathological criteria. Use of this staging system has not been rigorously validated as a clinical tool for guiding treatment. This is, in part, why we undertook this evaluation – to better understand treatment decisions based on clinical (pretreatment) data. Future studies will focus on our ability to detect the number of lymph nodes seen on staging evaluation and how this correlates to outcome. Treatment strategy is guided by prognostic factors for survival in esophageal cancer. Many believe that nodal status is an important consideration, but we are limited by our interpretation of the studies most commonly employed – CT, PET, and EUS. While several groups have tried to determine which are most sensitive, specific, and accurate, because these studies are not evaluating the same characteristics, it is possible that they should not be evaluated individually. Perhaps there is synergy in these studies that is not yet defined, and we should not try to determine which one is most revealing, but rather how these can be interpreted in conjunction to best guide clinical therapy.
Acknowledgments We would like to recognize the generous contributions from the families of Carl Edwards and Stuart Mason, without which our ongoing clinical research would not be possible. References 1 Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin 2012; 62: 10–29. 2 Ajani J A, Correa A M, Hofstetter W L et al. Clinical parameters model for predicting pathologic complete response following preoperative chemoradiation in patients with esophageal cancer. Ann Oncol 2012; 23: 2638–42. 3 Blanchard P, Quero L, Pacault V, Schlageter M-H, Baruch-Hennequin V, Hennequin C. Prognostic significance of anti-p53 and anti-KRas circulating antibodies in esophageal cancer patients treated with chemoradiotherapy. BMC Cancer 2012; 12: 119. 4 Bohanes P, Yang D, Chhibar R S et al. Influence of sex on the survival of patients with esophageal cancer. J Clin Oncol 2012; 30: 2265–72. 5 Chan D S, Twine C P, Lewis W G. Systematic review and meta-analysis of the influence of HER2 expression and amplification in operable oesophageal cancer. J Gastrointest Surg 2012; 16: 1821–9. 6 Ito E, Ozawa S, Kijima H et al. New invasive patterns as a prognostic factor for superficial esophageal cancer. J Gastroenterol 2012; 47: 1279–89. 7 Killinger W A, Jr, Rice T W, Adelstein D J et al. Stage II esophageal carcinoma: the significance of T and N. J Thorac Cardiovasc Surg 1996; 111: 935–40. 8 Rice T W, Blackstone E H, Goldblum J R et al. Superficial adenocarcinoma of the esophagus. J Thorac Cardiovasc Surg 2001; 122: 1077–90. 9 Crabtree T D, Yacoub W N, Puri V et al. Endoscopic ultrasound for early stage esophageal adenocarcinoma: implications for staging and survival. Ann Thorac Surg 2011; 91: 1509–15; discussion 1515-6. 10 Kato H, Nakajima M. The efficacy of FDG-PET for the management of esophageal cancer: review article. Ann Thorac Cardiovasc Surg 2012; 18: 412–9. 11 Schreurs L M, Janssens A C, Groen H et al. Value of EUS in determining curative resectability in reference to CT and FDGPET: the optimal sequence in preoperative staging of esophageal cancer? Ann Surg Oncol 2011 May 6. 12 Tangoku A, Yamamoto Y, Furukita Y, Goto M, Morimoto M. The new era of staging as a key for an appropriate treatment for esophageal cancer. Ann Thorac Cardiovasc Surg 2012; 18: 190–9. 13 Choi J, Kim S G, Kim J S, Jung H C, Song I S. Comparison of endoscopic ultrasonography (EUS), positron emission tomography (PET), and computed tomography (CT) in the preoperative locoregional staging of resectable esophageal cancer. Surg Endosc 2010; 24: 1380–6. 14 Kato H, Kuwano H, Nakajima M et al. Comparison between positron emission tomography and computed tomography in the use of the assessment of esophageal carcinoma. Cancer 2002; 94: 921–8. 15 Kato H, Nakajima M, Sohda M et al. The clinical application of (18)F-fluorodeoxyglucose positron emission tomography to predict survival in patients with operable esophageal cancer. Cancer 2009; 115: 3196–203. 16 van Vliet E P M, Heijenbrok-Kal M H, Hunink M G M, Kuipers E J, Siersema P D. Staging investigations for oesophageal cancer: a meta-analysis. Br J Cancer 2008; 98: 547–57. 17 Hayashi Y, Xiao L, Suzuki A et al. A nomogram associated with high probability of malignant nodes in the surgical specimen after trimodality therapy of patients with oesophageal cancer. Eur J Cancer 2012; 48: 3396–404.
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
Copyright of Diseases of the Esophagus is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Original article
Concordance of studies for nodal staging is prognostic for worse survival in esophageal cancer R. Dhupar, A. M. Correa, J. Ajani, S. Betancourt, R. J. Mehran, S. G. Swisher, W. L. Hofstetter Department of Thoracic and Cardiovascular Surgery, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
SUMMARY. Pretreatment clinical staging in esophageal cancer influences prognosis and treatment strategy. Current staging strategies utilize multiple imaging modalities, and often the results are contradictory. No studies have examined the implications of concordance of computed tomography (CT), positron emission tomography (PET), and endoscopic ultrasound (EUS) when used for the evaluation of nodal disease. The objective of this study was to determine if concordance of CT, PET, or EUS for nodal disease predicts worse overall survival. We reviewed 615 esophageal cancer patients with pretreatment CT, PET, and EUS that underwent esophagectomy for survival outcomes based on concordance of studies for nodal disease. Concordant N+ is defined as two or three studies positive for nodal disease; non-concordant N+ is defined as only one positive study. Node-positive disease by any study predicted shorter survival than node-negative disease (42% vs. 73% 5-year survival; P < 0.001). Additionally, non-concordant N+ patients had shorter survival than N− patients (52% vs. 73% 5-year survival; P < 0.001). Concordant N+ patients had shorter survival than non-concordant N+ patients (38- vs. 61-month median survival; P = 0.017). There were no statistically significant differences in survival based on specific combinations of studies. When PET was disregarded, patients with both CT+ and EUS+ had shorter survival than patients with either CT+ or EUS+ (39- vs. 58-month median survival; P = 0.029). Pretreatment CT, PET, or EUS concordance for node-positive disease predicts shorter overall survival in patients that undergo esophagectomy for esophageal cancer. Predicting survival in esophageal cancer should consider the synergistic capabilities of CT, PET, and EUS in evaluating nodal status. KEY WORDS: CT scan, esophageal cancer, EUS, PET scan, survival. ABBREVIATIONS: CT, computed tomography; EUS, endoscopic ultrasound; N+, pretreatment node positive; N−, pretreatment node negative; PET, positron emission tomography
INTRODUCTION Esophageal cancer is an aggressive malignancy that frequently presents at an advanced stage. In the United States, the incidence of esophageal cancer continues to rise, with approximately 17 460 new cases and 15 070 deaths expected in 2012.1 The last two decades have seen the implementation of multimodality therapy to improve cure rates and disease-free survival. However, the prognosis of this cancer remains grim, with low 5-year survivals. Address correspondence to: Dr Wayne L Hofstetter, MD, MD Anderson Cancer Center, Department of Thoracic and Cardiovascular Surgery, 1515 Holcombe Blvd, Unit 445, Houston, TX 77030, USA. Email: [email protected] 770
There are many factors that may be prognostic for overall survival in esophageal cancer. Tumor grade, growth patterns, sex, molecular markers, and stage are among a few that are believed to be important.2–6 The nodal status at the time of diagnosis can be particularly significant for determining treatment strategy, as well as prognosticating survival.7,8 However, the currently utilized methods to evaluate nodal status are limited by their accuracy.9 Computed tomography (CT), fluorodeoxyglucose positron emission tomography (PET), and endoscopic ultrasound (EUS) are frequently used to evaluate lymph nodes before treatment. Because nodal status is a key determinant of treatment and prognosis, understanding the best combination and sequence of studies for accurate clinical staging is critical for
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
Concordant studies in esophageal cancer
these patients. While all three modalities are widely used, there is a lack of consensus regarding their use alone, in combination, or in sequence.10–12 Several studies have attempted to define the sensitivity, specificity, and accuracy of these tests, but no studies have evaluated the prognostic significance of concordance of two or more modalities.10,13–16 In this study, we retrospectively reviewed a prospective database of 615 esophageal cancer patients with pretreatment nodal assessment by CT, PET, and EUS that underwent esophagectomy to determine if concordance of these studies is prognostic for survival.
MATERIALS AND METHODS Approval was granted by the MD Anderson Cancer Center Institutional Review Board for this study. We performed a retrospective review of a prospectively collected database of patients with esophageal adenocarcinoma or squamous cell cancer that underwent esophagectomy between 2002 and 2011. We included patients evaluated by pretreatment CT scan, PET, and EUS done at MD Anderson Cancer Center. We excluded patients with cervical or type III esophagogastric tumors, salvage esophagectomy, emergency esophageal surgery, and redo esophagectomy. For the 615 patients that met these criteria, we analyzed demographics, survival, and findings from CT, PET, and EUS. We did not analyze CT, PET, or EUS performed after the initiation of therapy for cancer. Criteria for interpreting CT, PET, or EUS as node positive were different for each study. For CT scan, the radiologist interpretation was used, with criterion of nodes greater than 1 cm being considered positive. A CT scan with intravenous contrast was not required. For PET, regional abnormal (standardized uptake value [SUV] greater than 2) was considered positive nodal disease, regional normal (SUV less than or equal to 2) as no nodal disease. For EUS, either pathological results from fine needle aspiration (FNA) or the gastroenterologist interpretation of nodal disease was used (which is a standard practice as defined by National Comprehensive Cancer Network guidelines for esophageal cancer staging). For the 615 total patients, FNA data were available for 95% (583/615); 30% (175/583) had an FNA, and 42% of those (73/175) were positive for cancer, which is 12% (73/615) of the entire study group. There was no pathological assessment of nodal disease as 77% of patients underwent neoadjuvant therapy (chemotherapy and/or radiation). Assessment of final pathological nodal disease after neoadjuvant therapy would not necessarily be reflective of pretreatment malignancy status. Because CT, PET, and EUS determine clinical nodal status, which guides treatment, this study focuses on clinical status only.
771
In all operations, regional abdominal and thoracic lymph node dissection was performed, with the exception of those performed in a transhiatal fashion (approximately 11%). We defined concordant node positive (N+) to mean that the results of at least two modalities (CT, PET, or EUS) were interpreted as node-positive disease. We defined non-concordant N+ to mean that the results of only one modality were interpreted as node-positive disease. For statistical analysis, Kaplan–Meier survival curves were constructed based on staging modality and node positivity. Values were considered statistically significant if P < 0.05. We report 5-year survival instead of median survival only if groups did not reach 50% mortality.
RESULTS Survival of patients with node positive disease based on study modality Patient and tumor characteristics are summarized in Table 1. Of the 615 patients with esophageal cancer that underwent staging with CT, PET, and EUS, 61% (377/615) had pretreatment node-positive (N+) disease. The median age was 62 (range 23–83), 14% (87/615) were women, and 92% (565/615) had adenocarcinoma. There was no preoperative chemotherapy or radiation in 23% (139/615) of patients. The distribution of tumors was 6% (38/615) upper/middle esophagus, 29% (175/615) lower esophagus, and 65% (401/615) gastroesophageal junction. The 5-year survival for N+ patients was 42% versus 73% for N− patients (P < 0.001; Fig. 1a). Because all patients had CT, PET, and EUS, we analyzed the percentage of studies that were positive in N+ patients: 58.6% (221/ 377) of patients were CT+, 39.8% (150/377) PET+, and 83.8% (316/377) EUS+ (Fig. 2a). In order to determine if there was a difference in survival based on which modality was positive, we created Kaplan–Meier survival curves for each modality. There was a statistically significant
Table 1
Characteristics of esophageal cancer patients and tumors
Variable
All patients % (n)
n Median age (years) % female % Caucasian % adenocarcinoma % squamous cell cancer % upper/middle esophagus % lower esophagus % gastroesophageal junction % neoadjuvant therapy cN− cN+
100 (615) 62 (na) 14 (87) 91 (562) 92 (565) 8 (50) 6 (38) 29 (175) 65 (401) 77 (475) 39 (238) 61 (377)
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
Diseases of the Esophagus
P < 0.001
1.0
Cumulative Survival Probability
a
b Cumulative Survival Probability
772
0.8 N– 0.6
0.4 N+ 0.2
0.0
1.0
0.8
0.4 CT N+ 0.2
0.0 0
Time (months)
0.8
PET N–
0.4 PET N+ 0.2
0.0
d Cumulative Survival Probability
P < 0.001
0.6
12 24 36 48 60 72 84 96 108 120 132 144
Time (months)
1.0
Cumulative Survival Probability
CT N–
0.6
0 12 24 36 48 60 72 84 96 108 120 132 144
c
P < 0.001
P < 0.001
1.0
0.8 EUS N– 0.6
0.4 EUS N+ 0.2
0.0 0
12 24 36 48 60 72 84 96 108 120 132 144
Time (months)
0
12 24 36 48 60 72 84 96 108 120 132 144
Time (months)
Fig. 1 Survival based on modality of finding N+ disease. (a) Kaplan–Meier survival curve of all N+ versus all N− esophageal cancer patients; (b) CT+ versus CT− patients; (c) PET+ versus PET− patients; (d) EUS+ versus EUS− patients.
difference in all groups (P < 0.001). Patients that were CT+ had a 5-year survival of 40% versus 64% if CT−; PET+ had a 5-year survival of 34% versus 61% if PET−; EUS+ had a 5-year survival of 39% versus 69% if EUS− (Fig. 1b–d). Patient survival if concordant N+ Overall, 35% (213/615) of the entire population and 57% (213/377) of the N+ population was concordant N+ (Fig. 2b). In order to determine if there was a difference in survival based on which studies were concordant, we created Kaplan–Meier survival curves of all combinations of concordant studies (Fig. 3). There was no statistically significant differ-
ence in survival based on which studies were concordant (P = 0.999), with median survival for CT+PET+EUS– of 38.8 months; CT+PET–EUS+ of 31.9 months; CT−PET+EUS+ of 32.3 months; and CT+PET+EUS+ of 39 months. Table 2 summarizes 5-year and median survival for all groups based on study results for nodal disease. Patient survival if non-concordant N+ Overall, 27% (164/615) of the entire population and 44% (164/377) of the N+ population was nonconcordant N+ (Fig. 2c). In order to determine if there was a difference in survival based on which studies were non-concordant N+, we created
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
Concordant studies in esophageal cancer
a
773
All Patients 400 350
84%
300 59%
250 200
40%
150 100 50 0
(377)
(221/377)
cN+
cT+
b
(150/377) PET+
(316/377) EUS+
Concordant N+ 400 350 300 250 200 150
26%
100 50 0
(377) cN+
c
4%
(97/377)
(15/377)
23%
4%
(87/377)
(14/377)
CT+PET+EUS+ CT+PET+EUS– CT+PET–EUS+ CT–PET+EUS+
Non-Concordant N+ 400 350 300 250 200 31%
150 6%
100 50 0
6%
(377)
(24/377)
(118/377)
cN+
CT–PET+EUS–
CT–PET–EUS+
(22/377) CT+PET–EUS–
Fig. 2 Bar graphs of patients with N+ disease and modality of recognition. (a) Recognition by any modality; (b) concordant N+, percentages represent fraction of N+ population (n = 377); (c) non-concordant N+, percentages represent fraction of N+ population (n = 377).
Kaplan–Meier survival curves of all combinations of non-concordant N+ studies (Fig. 4). There was no statistically significant difference in survival based on which studies were non-concordant N+ (P = 0.325), with median survival in CT+PET−EUS− of 86.2 months; CT−PET+EUS− of 45.3 months; and CT−PET−EUS+ of 58.4 months (Table 2). However, as expected, there was a statistically significant survival difference between the non-concordant N+ versus N− patients (5-year survival 52% vs. 73%; P < 0.001; curve not shown).
Patient survival if concordant N+ versus non-concordant N+ In order to determine if there was a difference in survival among patients with concordant N+ versus non-concordant N+ studies, we created Kaplan– Meier survival curves (Fig. 5). There was a significant difference in survival between the two groups (P = 0.017), with median survival for the concordant N+ group of 38.2 months versus 60.6 months in the non-concordant N+ group (Table 2). There was no
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
774
Diseases of the Esophagus
Fig. 3 Survival of concordant N+ patients. Kaplan–Meier survival curve of N+ patients based on CT+PET+EUS+, CT+PET+EUS−, CT+PET−EUS+, or CT−PET+EUS+.
significant difference in survival between patients with all studies positive (CT+PET+EUS+) versus patients with only one or two positive studies (P = 0.998; curve not shown). Because both CT and EUS rely on anatomic findings rather than metabolic activity, we compared survival for CT/EUS concordant N+ and CT/EUS non-concordant N+ (disregarding the result of the PET; curve not shown). There was a significant difference in survival between the two groups (P = 0.029), with median survival for the CT/EUS concordant N+ Table 2 Summary of study interpretations for nodal disease, number of patients in each group, median survivals, and 5-year survivals
Study All studies Concordant N+ CT+PET+EUS+ CT+PET+EUS− CT+PET−EUS+ CT−PET+EUS+ Non-concordant N+ CT+PET−EUS− CT−PET+EUS− CT−PET−EUS+ CT and EUS only Concordant N+ Non-concordant N+
n
Median survival (months)
5-year survival (%)
213 97 15 87 14 164 22 24 118
38.2* 39.0 38.8 31.9 32.3 60.6* 86.2 45.3 58.4
36 33 33 40 17 52 75 48 47
184 169
39.0** 58.4**
36 48
*Statistically significant difference with P = 0.017; **statistically significant difference with P = 0.029. CT, computed tomography; EUS, endoscopic ultrasound; PET, positron emission tomography.
group of 39.0 months versus 58.4 months for the CT/EUS non-concordant N+ group (Table 2). There was no significant difference in survival between CT+EUS− and CT−EUS+ patients (P = 0.466; curve not shown).
DISCUSSION In this study, we performed a retrospective analysis of 615 esophageal cancer patients that underwent esophagectomy to determine if concordance of CT, PET, and EUS in pretreatment nodal staging is prognostic for survival. The key findings of this paper are (i) there is shorter survival if any modality is positive for nodal disease; (ii) the patient group with concordant N+ studies had shorter survival than the group with non-concordant N+ studies; and (iii) the patient group that was CT/EUS concordant N+ had shorter survival than CT/EUS non-concordant N+. While previous studies have compared or attempted to determine the prognostic implication of any single positive study, none have examined the implications of concordance of these studies. Our finding that there is a shorter survival if any modality is positive is not surprising (Fig. 1). Although there are different diagnostic capabilities ascribed to CT, PET, and EUS, generally the patient is treated as N+ if any study is positive. This is a validation of clinical practice. Of note, in our population, 44% (164/377) of the N+ patients were positive
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
Concordant studies in esophageal cancer
775
Fig. 4 Survival of non-concordant N+ patients. Kaplan–Meier survival curve of patients with N+ disease based on CT+PET−EUS−, CT−PET+EUS−, or CT−PET−EUS+.
in one study only (non-concordant N+), 72% (118/ 164) of which were EUS+ only (Fig. 2c). Although there was no statistically significant difference in survival among patients with only one positive study (non-concordant N+ group), the number of N+ patients that were CT+ only or PET+ only was small
Fig. 5 Kaplan–Meier survival curve of patients with concordant N+ versus non-concordant N+ patients.
(46/377; Fig. 4). Perhaps there is a difference in the prognostic implication of which single modality is positive; however, it is possible that a difference could not be detected because of a small patient population. Our finding that the concordant N+ group had shorter survival than the non-concordant N+ group is a novel finding (Fig. 5). N+ disease is associated with lower overall survival, both before and after treatment.7,17 However, it is not clear if the diagnosis of pretreatment N+ disease by all modalities is equivalent. Our findings suggest that N+ disease by more than one modality may confer a worse prognosis than patients with N+ disease by only one modality. While studies have looked to determine differences in sensitivity, specificity, and accuracy of each modality, perhaps CT, PET, and EUS evaluate separate characteristics of lymph nodes that work in synergy to differentiate N+ from N−.12,16 Although we did not find a statistically significant difference among the different combinations of positive studies, again, the number of patients in this study may limit the ability to detect subtle differences. Finally, we compared CT with EUS, the two modalities that evaluate anatomic rather than metabolic characteristics of lymph nodes. By disregarding the results of the PET scan, we found a statistically significant difference in survival among patients who were CT/EUS concordant N+ versus those who were not (Fig. 5). There was no difference in survival in CT+EUS− versus CT−EUS+ patients (data not shown). As with our other findings, this also implies
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
776
Diseases of the Esophagus
a potential synergy between CT and EUS – two modalities that are thought to evaluate similar characteristics of lymph nodes. This study has several limitations. This is a single institution study, and it is unknown if performance of these procedures and interpretation of studies are generalizable. Also, while we evaluate overall survival in patients that underwent esophagectomy, this group was not homogenous for neoadjuvant treatment (Table 1). In addition, we did not evaluate for differences in EUS+ studies based on biopsies performed during EUS (12% of EUS+). While pathological findings do not rely on ultrasound characteristics, the decision to perform FNA is based on ultrasound findings. However, we did not determine if nodes positive by ultrasound characteristics alone were differentiated from nodes positive by biopsy performed during EUS. Finally, because some groups were small in number, the ability to detect subtle differences among groups may be hindered. For this study, we did not analyze final pathological status of lymph nodes. Although it can be suggested that comparison of the pretreatment staging should be verified with final pathological analysis, this should be considered with caution. Pathological node status is a function of treatment response, which can be unpredictable. A comparison of the clinical stage with the pathological stage would potentially confuse the issues addressed in this paper, as disagreement could be interpreted as either complete response or false positive staging. Along these same lines, the American Joint Committee on Cancer seventh edition staging system was formulated on entirely pathological criteria. Use of this staging system has not been rigorously validated as a clinical tool for guiding treatment. This is, in part, why we undertook this evaluation – to better understand treatment decisions based on clinical (pretreatment) data. Future studies will focus on our ability to detect the number of lymph nodes seen on staging evaluation and how this correlates to outcome. Treatment strategy is guided by prognostic factors for survival in esophageal cancer. Many believe that nodal status is an important consideration, but we are limited by our interpretation of the studies most commonly employed – CT, PET, and EUS. While several groups have tried to determine which are most sensitive, specific, and accurate, because these studies are not evaluating the same characteristics, it is possible that they should not be evaluated individually. Perhaps there is synergy in these studies that is not yet defined, and we should not try to determine which one is most revealing, but rather how these can be interpreted in conjunction to best guide clinical therapy.
Acknowledgments We would like to recognize the generous contributions from the families of Carl Edwards and Stuart Mason, without which our ongoing clinical research would not be possible. References 1 Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin 2012; 62: 10–29. 2 Ajani J A, Correa A M, Hofstetter W L et al. Clinical parameters model for predicting pathologic complete response following preoperative chemoradiation in patients with esophageal cancer. Ann Oncol 2012; 23: 2638–42. 3 Blanchard P, Quero L, Pacault V, Schlageter M-H, Baruch-Hennequin V, Hennequin C. Prognostic significance of anti-p53 and anti-KRas circulating antibodies in esophageal cancer patients treated with chemoradiotherapy. BMC Cancer 2012; 12: 119. 4 Bohanes P, Yang D, Chhibar R S et al. Influence of sex on the survival of patients with esophageal cancer. J Clin Oncol 2012; 30: 2265–72. 5 Chan D S, Twine C P, Lewis W G. Systematic review and meta-analysis of the influence of HER2 expression and amplification in operable oesophageal cancer. J Gastrointest Surg 2012; 16: 1821–9. 6 Ito E, Ozawa S, Kijima H et al. New invasive patterns as a prognostic factor for superficial esophageal cancer. J Gastroenterol 2012; 47: 1279–89. 7 Killinger W A, Jr, Rice T W, Adelstein D J et al. Stage II esophageal carcinoma: the significance of T and N. J Thorac Cardiovasc Surg 1996; 111: 935–40. 8 Rice T W, Blackstone E H, Goldblum J R et al. Superficial adenocarcinoma of the esophagus. J Thorac Cardiovasc Surg 2001; 122: 1077–90. 9 Crabtree T D, Yacoub W N, Puri V et al. Endoscopic ultrasound for early stage esophageal adenocarcinoma: implications for staging and survival. Ann Thorac Surg 2011; 91: 1509–15; discussion 1515-6. 10 Kato H, Nakajima M. The efficacy of FDG-PET for the management of esophageal cancer: review article. Ann Thorac Cardiovasc Surg 2012; 18: 412–9. 11 Schreurs L M, Janssens A C, Groen H et al. Value of EUS in determining curative resectability in reference to CT and FDGPET: the optimal sequence in preoperative staging of esophageal cancer? Ann Surg Oncol 2011 May 6. 12 Tangoku A, Yamamoto Y, Furukita Y, Goto M, Morimoto M. The new era of staging as a key for an appropriate treatment for esophageal cancer. Ann Thorac Cardiovasc Surg 2012; 18: 190–9. 13 Choi J, Kim S G, Kim J S, Jung H C, Song I S. Comparison of endoscopic ultrasonography (EUS), positron emission tomography (PET), and computed tomography (CT) in the preoperative locoregional staging of resectable esophageal cancer. Surg Endosc 2010; 24: 1380–6. 14 Kato H, Kuwano H, Nakajima M et al. Comparison between positron emission tomography and computed tomography in the use of the assessment of esophageal carcinoma. Cancer 2002; 94: 921–8. 15 Kato H, Nakajima M, Sohda M et al. The clinical application of (18)F-fluorodeoxyglucose positron emission tomography to predict survival in patients with operable esophageal cancer. Cancer 2009; 115: 3196–203. 16 van Vliet E P M, Heijenbrok-Kal M H, Hunink M G M, Kuipers E J, Siersema P D. Staging investigations for oesophageal cancer: a meta-analysis. Br J Cancer 2008; 98: 547–57. 17 Hayashi Y, Xiao L, Suzuki A et al. A nomogram associated with high probability of malignant nodes in the surgical specimen after trimodality therapy of patients with oesophageal cancer. Eur J Cancer 2012; 48: 3396–404.
© 2013 Wiley Periodicals, Inc. and the International Society for Diseases of the Esophagus
Copyright of Diseases of the Esophagus is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
