|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 25, No 5 (February 10), 2007: pp. 507-512 © 2007 American Society of Clinical Oncology. DOI: 10.1200/JCO.2006.08.0101 American Joint Committee on Cancer Staging System Does Not Accurately Predict Survival in Patients Receiving Multimodality Therapy for Esophageal Adenocarcinoma
From the Thoracic Service, Department of Surgery; Biostatistics Service, Department of Epidemiology and Biostatistics; Pathology Department; Gastrointestinal Oncology Service, Department of Medicine; and the Radiation Oncology Department, Memorial Sloan-Kettering Cancer Center, New York, NY Address reprint requests to Nabil P. Rizk, MD, Department of Surgery, Thoracic Service, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, Ste C883, New York, NY 10021; e-mail: rizkn{at}mskcc.org
Purpose: In patients with adenocarcinoma of the esophagus who receive preoperative chemoradiotherapy (CRT), American Joint Committee on Cancer (AJCC) stage, pathologic complete response (pCR), and estimated treatment response are various means used to stratify patients prognostically after surgery. However, none of these methods has been formally evaluated. The purpose of this study was to establish prognostic pathologic variables after CRT. Patients and Methods: A retrospective review was performed of patients with esophageal adenocarcinoma who received CRT before esophagectomy. Data collected included demographics, CRT details, pathologic findings, and survival. Statistical methods included recursive partitioning and Kaplan-Meier analyses. Results: Two hundred seventy-six patients were appropriate for this analysis. Kaplan-Meier analysis indicates that the current AJCC system poorly distinguishes between stages 0 to IIA (P = .52), IIB to III (P = .87), and IVA to IVB (P = .30). The presence of a pCR conferred improved survival over residual disease (P = .01). Recursive partitioning analysis indicates that involved lymph nodes and metastatic disease are the best predictors of survival and that depth of invasion and degree of treatment response are less predictive. Conclusion: The current AJCC staging system is not a good predictor of survival after CRT. Although patients with a pCR do have improved long-term survival relative to patients with residual disease, this method places too much emphasis on residual depth of invasion and fails to identify patients with residual disease who have good long-term survival. Recursive partitioning analysis more accurately identifies nodal disease and metastatic disease as the most important prognostic variables. Degree of treatment response is less prognostic than nodal involvement.
The benefit of induction chemoradiotherapy (CRT) followed by resection for patients with locally advanced esophageal cancer remains controversial, but this treatment paradigm is increasingly being used.1 One of the many unresolved questions regarding this approach is how to restage patients after completion of their treatment. Accurate restaging of patients after CRT is important because it provides prognostic information, which can help direct subsequent treatment decisions. One frequently used method is simply to classify patients according to whether or not they have experienced a pathologic complete response (pCR). Indeed, many studies show that patients with a pCR have a better long-term prognosis than patients with residual disease.2-6 Another method used to restage patients after CRT is to use the current American Joint Committee on Cancer (AJCC) staging system.7,8 Alternately, some investigators rely on the estimated treatment response as a means to stratify patients into prognostic groups.9,10 Each of these methods has limitations. Grouping patients into pCR versus residual disease ignores potentially diverse subsets of patients in the non-pCR group. However, the AJCC system has no stage for pCR, and it may ascribe too much weight to the depth of the primary tumor in patients in whom this may be difficult to measure after radiation. Lastly, estimation of treatment response is subjective and might not correspond to outcome. In this article, we review our experience in a group of patients who received CRT followed by resection and explore the relationship between pathologic findings and overall survival. Because previously published data support the use of the number of involved lymph nodes for staging noninduction patients,11 we include this variable in our analyses. In addition, because both the literature6,12,13 and a preliminary analysis of our data indicate that adenocarcinoma and squamous cell carcinoma respond differently to CRT, we present data on adenocarcinoma only in this article.
Acquisition of Clinical Data We undertook a retrospective review of all patients undergoing resection for adenocarcinoma of the esophagus at Memorial Sloan-Kettering Cancer Center between January 1996 and June 2005. January 1996 is the time point at which an institutional electronic medical record system was instituted and is, therefore, a date from which highly reliable data can be obtained. Patients without survival information were excluded from this analysis. We also excluded any patient who did not undergo CRT or who underwent intended definitive nonsurgical therapy but was eventually resected (primary CRT). The data collected included patient age, preoperative clinical stage (based on a combination of radiographic and endoscopic studies), type of CRT (type of chemotherapy, amount of radiation, and length of time from the completion of radiation to surgery), depth of tumor invasion, estimated treatment effect, and the number of all malignant and benign lymph nodes. Overall survival, as calculated from the time of operation, was obtained from the electronic records at Memorial Sloan-Kettering Cancer Center and confirmed by the Social Security Death Index. January 2006 was the censoring date for survival.
Clinical Stage
TNM Classification
Estimation of Treatment Effect and pCR
Statistical Analysis
Clinical Data During the study period, 856 esophagectomies were performed, and 276 were appropriate for further analysis. Four hundred eighty patients were excluded because they did not receive CRT. Other reasons for exclusion included the following: squamous cell carcinoma (n = 93), underwent primary CRT (n = 12), inadequate pathologic data (n = 3), and inadequate survival data (n = 3). Two hundred twenty-four patients (81.2%) had clinical evidence of regional disease before therapy. Two hundred eight patients (75.0%) received 50.4 Gy of radiation in daily fractions of 1.8 Gy over 28 treatments in a manner described in a previous publication.17 Similarly, most patients (n = 266, 95.3%) were administered various concurrent cisplatin-based chemotherapy regimens, following a protocol previously described17 (Table 1).
AJCC Staging System According to the AJCC staging system, after treatment, 52 patients had a pCR and thus did not have a defined stage, 32 had stage I disease, 85 had stage IIA, 32 had stage IIB, 41 had stage III, 12 had stage IVA, and 22 had stage IVB (Table 2). A Kaplan-Meier survival analysis shows that the AJCC staging system does not discriminate well between patients who are stages pCR to IIa (P = .52), patients who are stages IIB to III (P = .87), and patients who are stages IVA to IVB (P = .30; Fig 1).
Outcome in Patients With a pCR to Induction Therapy Overall, 52 patients (18.8%) had evidence of a pCR at the time of surgery (Table 2). The survival of pCR patients was significantly better than in patients who had residual disease. pCR patients had a 3-year survival rate of 70.4% compared with 41.8% in patients with residual disease (P = .01; Fig 2). Of note, the likelihood of a pCR was not influenced by the clinical stage of disease (P = .23) or by the EUS assessment of depth of invasion (P = .54; data not shown).
Recursive Partitioning Analysis of TNM Analysis of survival using recursive partitioning methods and TNM category as variables shows that the initial stratification results from the presence of metastatic disease (M category; Fig 3). The second level of stratification then distinguishes node-negative from node-positive patients. Within the subset of N0M0 patients, T0-1 patients did best (n = 84; 3-year survival rate, 70.5%), followed by T2-4 patients (n = 91; 3-year survival rate, 51.6%). Among the patients who were N1M0, depth of invasion (T0-1 v T2-4) again separated patients (3-year survival rate, 51.9% and 28.9%, respectively).
Recursive Partitioning Analysis of TNM and Number of Malignant Lymph Nodes A recursive partitioning analysis that includes TNM category as well as the number of malignant lymph nodes as variables shows that involvement of more than one lymph node is the initial discriminator of survival (Fig 4). The next stratification occurs based on the presence of metastatic disease. Further analysis identifies patients who had T0-1N(0+-1+)M0 tumors as having the best prognosis [only eight of 92 patients had (1+) node, whereas all other patients were N(0+)], followed by patients with T2-4N(0+)M0 tumors, followed by patients with T2-4N(1+)M0 tumors. The best prognostic group (including the 52 patients with a pCR) had a 3-year survival rate of 70.3% (n = 92), the next best prognostic group (n = 87) had a 3-year survival rate of 52.4%, and the third best prognostic group had a 3-year survival rate of 32.2% (n = 20). Patients with more than two involved lymph nodes and without metastatic disease (n = 53), regardless of the depth of invasion, did poorly, with an estimated 3-year survival rate of 26.3%. This result was similar, however, to patients with T2-4N(1+)M0 tumors (P = .35). Further reinforcing the importance of the presence of any involved lymph nodes in determining survival, an analysis that compared the survival of patients relative to the number of involved lymph nodes showed a sharp decrease in survival when any lymph node was involved (Appendix Fig A1, online only).
Estimated Treatment Response Two hundred sixty-seven patients (96.7%) had adequate data with which to attribute a treatment response. The median treatment response was 90% (mean, 72%). The mean treatment response was higher in node-negative patients (n = 185) than in node-positive patients (n = 82; 76.4% v 62.1%, respectively; P < .001). The median interval of time from completion of radiation to surgery was 49 days (mean, 58 days). There was no correlation between treatment response and the interval between radiation completion and surgery (P = .37; data not shown). Excluding patients with metastatic disease, median survival increased incrementally with increasing treatment response, with a steady increase in survival seen after a more than 50% treatment response (Appendix Fig A2A, online only). When the treatment response was stratified by lymph node involvement, there remained a trend of improved survival with increasing treatment response, with a similar cutoff noted at approximately 50% treatment response. However, even more striking than this trend is the fact that the average survival in patients who are node negative, regardless of estimated treatment response, is better than the survival in their counterparts who are node positive (Appendix Figs A2B and A2C, online only). This underscores the importance of lymph node involvement in determining outcome, even in the face of poor treatment response. In summary, the most important prognostic factors after CRT were node and metastasis status, whereas depth of invasion had less of an impact on survival. Although using pCR identifies a subset of patients who have a better prognosis, using pCR alone fails to recognize a large number of patients with residual disease, namely T1N(0+) patients, who have a similar prognosis. Furthermore, in patients with residual T2-4 disease, nodal status is an important determinant of outcome; in patients with one or more involved lymph node, survival decreases significantly compared with node-negative patients. Finally, although the estimated treatment response can predict survival, this gauge is less predictive than nodal status.
Although establishing the benefits of preoperative CRT for locally and regionally advanced esophageal cancer remains elusive,1,7,18-20 one frequent finding in the many studies that have investigated this treatment strategy is that various subsets of patients seem to benefit significantly from CRT. Several studies show that patients who experience a pCR to CRT have a better prognosis than patients with residual disease.2-6,10 However, other studies do not support this conclusion,12,21-24 and some studies have associated the likelihood of a pCR with pretreatment clinical stage, thereby raising a question of selection bias.25,26 Similarly, some studies have shown that patients with a major response to therapy have improved outcomes compared with patients who do not have a major response,9,10,27 whereas others have used pretherapy clinical staging to correlate downstaging with improved outcome.28 Some studies have also noted that the presence of residual involved lymph nodes seems to predict a particularly poor outcome.29,30 What emerges from these varied results is that CRT before surgery creates subsets of patients whose prognosis is not well described by the current AJCC staging system. Likewise, crudely classifying patients as having either a pCR or residual disease is overly simplistic.
In this analysis, we found that, in patients with adenocarcinoma, the principal determinants of outcome after CRT and surgery are the involvement of lymph nodes and the presence of metastatic disease. Patients who have T0-1 lesions with one or no involved lymph nodes have the best prognosis, followed by patients with T2-4 lesions who are node negative. The small incremental difference in survival between patients who are N(1+) and N( Several other studies evaluating the impact of CRT arrived at similar conclusions. In a retrospective review of 47 patients with adenocarcinoma, Dunne et al27 found that the only predictor of survival after CRT and surgery was lymph node status. Suntharalingam et al26 similarly showed that patients (n = 45) who became node negative after CRT had significantly improved survival compared with patients who had persistent involved lymph nodes. In a group of patients with various tumor histologies (three adenocarcinomas and 15 squamous cell carcinomas), Alexander et al29 also demonstrate that residual nodal status is the most important predictor of outcome. In patients with adenocarcinoma (n = 187), Gu et al30 showed that the number of involved lymph nodes is an additional predictor of outcome, with two or more nodes predicting a worse outcome. Finally, Rice et al28 retrospectively evaluated groups of patients with various tumor histologies who were downstaged from clinical N1 disease to pathologic N0 disease (n = 37) and compared them to a matched group of patients with persistent N1 disease (n = 32). They were able to show a significant difference in survival between the two groups. In most of these studies, the number of patients evaluated was small; three of the five studies were not stratified by histology; and only two studies included multivariable analyses that included other pathologic variables. There are several advantages of our study compared with other published series. This series consists of a large number of patients (n = 276) with uniform histology, with consistent clinical staging studies, central review by one pathologist, and treatment provided over a short and contemporary time span. These characteristics minimize the risk for confounding influences that might occur over a longer time interval. In conclusion, this study shows that, in patients with adenocarcinoma of the esophagus who are treated with preoperative CRT, current pathologic staging criteria (AJCC) do not adequately stratify prognostic groups. Furthermore, patients currently ascribed as having a pCR have similar outcomes as patients with residual T1 disease, and therefore, these two groups should be combined. Finally, in patients with residual T2-4 disease, prognostic groups should be stratified by nodal status (N0, N1). Although estimating tumor response provides additional information about prognosis, these additional data are outweighed by the importance of the nodal status.
Although all authors completed the disclosure declaration, the following author or immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment: N/A Leadership: N/A Consultant: N/A Stock: N/A Honoraria: N/A Research Funds: Bruce D. Minsky, Pfizer Testimony: N/A Other: N/A
Conception and design: Nabil P. Rizk Provision of study materials or patients: Manjit S. Bains, Bernard Park, Raja Flores, David H. Ilson, Bruce D. Minsky, Valerie W. Rusch Collection and assembly of data: Nabil P. Rizk Data analysis and interpretation: Nabil P. Rizk, Ennapadam Venkatraman Manuscript writing: Nabil P. Rizk, Ennapadam Venkatraman, Laura Tang, Valerie W. Rusch Final approval of manuscript: Nabil P. Rizk, Ennapadam Venkatraman, Manjit S. Bains, Bernard Park, Raja Flores, Laura Tang, David H. Ilson, Bruce D. Minsky, Valerie W. Rusch
Presented at the 41st Annual Meeting of the American Society of Clinical Oncology, May 13-17, 2005, Orlando, FL. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Geh JI, Crellin AM, Glynne-Jones R, et al: Preoperative (neoadjuvant) chemoradiotherapy in oesophageal cancer. Br J Surg 88:338-356, 2001[CrossRef][Medline] 2. DeVita F, DiMartino N, Orditura M, et al: Preoperative chemoradiotherapy for squamous cell carcinoma and adenocarcinoma of the esophagus. Chest 122:1302-1308, 2002[CrossRef][Medline] 3. Bates BA, Detterbeck FC, Bernard SA, et al: Concurrent radiation therapy and chemotherapy followed by esophagectomy for localized esophageal carcinoma. J Clin Oncol 14:156-163, 1996[Abstract] 4. Stahl M, Wilke H, Fink U, et al: Combined preoperative chemotherapy and radiotherapy in patients with locally advanced esophageal cancer: Interim analysis of a phase II trial. J Clin Oncol 14:829-837, 1996 5. Urba SG, Orringer MB, Turrisi A, et al: Randomized trial of preoperative chemoradiation versus surgery alone in patients with locoregional esophageal carcinoma. J Clin Oncol 19:305-313, 2001 6. Seydel HG, Leichman L, Byhardt R, et al: Preoperative radiation and chemotherapy for localized squamous cell carcinoma of the esophagus: A RTOG study. Int J Radiat Oncol Biol Phys 14:33-35, 1988[Medline] 7. Walsh TN, Noonan N, Hollywood D, et al: A comparison of multimodal therapy and surgery for esophageal adenocarcinoma. N Engl J Med 335:462-467, 1996 8. Green FL, Page DL, Fleming ID, et al (eds): AJCC Cancer Staging Manual (ed 6). New York, NY, Springer-Verlag, 2002 9. Schneider PM, Baldus SE, Mestzger R, et al: Histomorphologic tumor regression and lymph node metastases determine prognosis following neoadjuvant radiochemotherapy for esophageal cancer: Implications for response classification. Ann Surg 242:684-692, 2005[CrossRef][Medline] 10. Chirieac LR, Swisher SG, Ajani JA, et al: Post-therapy pathologic stage predicts survival in patients with esophageal carcinoma receiving preoperative chemoradiation. Cancer 103:1347-1355, 2005[CrossRef][Medline] 11. Korst RJ, Rusch VW, Venkatraman E, et al: Proposed revision of the staging classification for esophageal cancer. J Thorac Cardiovasc Surg 115:660-670, 1998 12. Keller SM, Ryan LM, Coia LR, et al: High dose chemoradiotherapy followed by esophagectomy for adenocarcinoma of the esophagus and gastroesophageal junction. Cancer 83:1908-1916, 1998[CrossRef][Medline] 13. Burmeister BH, Smithers BM, Fitzgerald L, et al: Randomized phase III trial of preoperative chemoradiation followed by surgery (CR-S) versus surgery alone (S) for localized resectable cancer of the esophagus. Proc Am Soc Clin Oncol 21:130a, 2002 (abstr 518) 14. Breiman L, Friedman JH, Olshen RA, et al: Classification and Regression Trees. New York, NY, Chapman & Hall, 1984 15. Lausen B, Schumaker M: Maximally selected rank statistics. Biometrics 48:73-85, 1992[CrossRef] 16. Therneau TM, Atkinson EJ: An Introduction to Recursive Partitioning Using the RPART Routine. Technical Report 61. Rochester, MN, Mayo Clinic, Section of Statistics, 1997 17. Ilson DH, Bains M, Kelsen DP, et al: Phase I trial of escalating-dose irinotecan given weekly with cisplatin and concurrent radiotherapy in locally advanced esophageal cancer. J Clin Oncol 21:2926-2932, 2003 18. Urschel JD, Vasan H: A meta-analysis of randomized controlled trials that compared neoadjuvant chemoradiation and surgery to surgery alone for resectable esophageal cancer. Am J Surg 185:538-543, 2003[CrossRef][Medline] 19. Bosset JF, Gignoux M, Triboulet JP, et al: Chemoradiotherapy followed by surgery compared with surgery alone in squamous-cell cancer of the esophagus. N Engl J Med 337:161-167, 1997 20. Nygard K, Hagen S, Hansen HS, et al: Pre-operative radiotherapy prolongs survival in operable esophageal carcinoma: A randomized multicenter study of pre-operative radiotherapy and chemotherapy—The second Scandinavian trial in esophageal cancer. World J Surg 16:1104-1110, 1992[CrossRef][Medline] 21. Jones DR, Detterbeck FC, Egan TM, et al: Induction chemoradiotherapy followed by esophagectomy in patients with carcinoma of the esophagus. Ann Thorac Surg 64:185-192, 1997 22. Malhaire JP, Labat JP, Lozac P, et al: Preoperative concomitant radiochemotherapy in squamous cell carcinoma of the esophagus: Results of a study of 56 patients. Int J Radiat Oncol Biol Phys 34:429-437, 1996[CrossRef][Medline] 23. Adelstein DJ, Rice TW, Becker M, et al: Use of concurrent chemotherapy, accelerated fractionation radiation, and surgery for patients with esophageal carcinoma. Cancer 80:1011-1020, 1997[CrossRef][Medline] 24. Tamin WZ, Davidson RS, Quinlan RM, et al: Neoadjuvant chemoradiotherapy for esophageal cancer: Is it worthwhile? Arch Surg 133:722-726, 1998 25. Rohatgi P, Swisher SG, Correa AM, et al: Characterization of pathologic complete response after preoperative chemoradiotherapy in carcinoma of the esophagus and outcome after pathologic complete response. Cancer 104:2365-2372, 2005[CrossRef][Medline] 26. Suntharalingam M, Haas ML, Sonett JR, et al: Accurate lymph node assessment prior to trimodality therapy for esophageal carcinoma. Cancer J 7:509-515, 2001[Medline] 27. Dunne B, Reynolds JV, Mulligan E, et al: A pathologic study of tumor regression in oesophageal adenocarcinoma treated with preoperative chemoradiotherapy. J Clin Pathol 54:841-845, 2001 28. Rice TW, Blackstone EH, Adelstein DJ, et al: N1 esophageal carcinoma: The importance of staging and downstaging. J Thorac Cardiovasc Surg 121:454-464, 2001 29. Alexander EP, Lipman T, Harmon J, et al: Aggressive multimodality therapy for stage III esophageal cancer: A phase I/II study. Ann Thorac Surg 69:363-368, 2000 30. Gu Y, Swisher SG, Ajani JA, et al: 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, 2006[CrossRef][Medline] Submitted June 28, 2006; accepted November 17, 2006. Related Correspondence
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2+) does not warrant a distinction between these nodal groupings. Among some noteworthy aspects of these findings is that the number of patients in the best prognostic group is nearly twice the number of patients with a pCR. Furthermore, this analysis highlights the fact the AJCC staging system attributes too much importance to the depth of invasion, as evidenced by the poor discrimination in survival between stages pCR to IIa. Estimation of treatment response did show a correlation with outcome, but this association was less evident if the analysis was stratified by nodal status. Thus, in general, regardless of the estimated treatment response, lymph node status was more predictive of outcome. 
