|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2005.03.6038 on January 17 2006 © 2006 American Society of Clinical Oncology. Analysis of Micrometastatic Disease in Sentinel Lymph Nodes From Resectable Colon Cancer: Results of Cancer and Leukemia Group B Trial 80001From the Brigham and Women's Hospital; Dana-Farber Cancer Institute, Boston, MA; University of Missouri/Ellis Fischel Cancer Center, Columbia; Washington University Medical Center, St Louis, MO; Cancer and Leukemia Group B Statistical Center, Duke University, Durham, NC; Ohio State University, Columbus, OH; State University of New York Upstate Medical University and Veteran's Administration Medical Center, Syracuse, NY; and McGill University, Montreal, Quebec, Canada Address reprint requests to Monica M. Bertagnolli, MD, Brigham and Women's Hospital, 75 Francis St, Boston, MA 02115; e-mail: mbertagnolli{at}partners.org
PURPOSE: To determine whether sentinel lymph node (LN) sampling (SLNS) could reduce the number of nodes required to characterize micrometastatic disease (MMD) in patients with potentially curable colon cancer. PATIENTS AND METHODS: Cancer and Leukemia Group B 80001 was a study to determine whether SLNS could identify a subset of LNs that predicted the status of the nodal basin for resectable colon cancer and, therefore, could be extensively evaluated for the presence of micrometastases. Patients enrolled onto this study underwent SLNS after injection of 1% isosulfan blue, and both sentinel nodes (SNs) and non-SNs obtained during primary tumor resection were sectioned at multiple levels and stained using anti–carcinoembryonic antigen and anticytokeratin antibodies. RESULTS: Using standard histopathology, SNs failed to predict the presence of nodal disease in 13 (54%) of 24 node-positive patients. Immunostains were performed for patients whose LNs were negative by standard histopathology. Depending on the immunohistochemical criteria used to assign LN positivity, SN examination resulted in either an unacceptably high false-positive rate (20%) or a low sensitivity for detection of MMD (40%). CONCLUSION: By examining both SNs and non-SNs, this multi-institutional study showed that SNs did not accurately predict the presence of either conventionally defined nodal metastases or MMD. As a result, SLNS is not a useful technique for the study of MMD in patients with colon cancer.
Sentinel lymph node (LN) sampling (SLNS) identifies a small number of regional LNs that accurately predict the status of all regional nodes in a patient undergoing cancer surgery. SLNS is based on the assumption that lymphatic flow drains sequentially from peripheral to central tissue locations, with limited functional collaterals outside of the dominant vascular supply. Consequently, tracer substances injected at a tumor site must follow the same pathway by which metastatic tumor cells traverse lymphatic channels. If these conditions are met, the first LN encountered, termed the sentinel node (SN), is a reliable indicator of the tumor status of the entire nodal basin. To a high degree, these principles hold true for the integumentary system because SNLS is a clinically validated indicator of nodal status for both breast cancer and melanoma.1,2 Unlike surgery for breast cancer or melanoma, the anatomy of the colon permits wide lymphadenectomy without significantly increasing the difficulty or morbidity of tumor resection. SLNS for colon cancer is nevertheless useful if it is therapeutic or if it improves the accuracy of pathologic staging.3-7 One population for whom staging could be improved includes patients with node-negative (N0) disease by conventional histopathology. Of patients undergoing potentially curative surgery for colon cancer, 35% to 45% will have N0 nodal status.8 Approximately 25% of these stage II patients will experience post-treatment disease recurrence, suggesting that, for one in four patients with N0 disease, current histopathologic staging methods fail to identify those patients destined to manifest tumor progression. The application of better techniques to detect these high-risk patients would identify patients who may benefit from adjuvant chemotherapy. These patients are an important population for evaluating new treatments to target minimal residual disease. One possibility for improving the staging of N0 colon cancer patients involves detection of clinically relevant micrometastatic disease (MMD). By current American Joint Committee on Cancer criteria, LNs are positive if they contain deposits of tumor cells visible by hematoxylin and eosin (H&E) that are at least 0.2 mm (200 µm) in diameter.9,10 Pathologists have long recognized the presence of smaller volume disease within regional LN. The term MMD broadly describes evidence of tumor metastases within regional LN that are not scored as positive by conventional histopathologic criteria. Examples of MMD include small foci of tumor cells visualized by H&E stain or individual tumor cells identified only after application of immunohistochemistry (IHC) for tumor antigens such as carcinoembryonic antigen (CEA) or cytokeratins. For colon cancer, the clinical significance of regional MMD is unknown. Retrospective studies using a variety of disease definitions do not consistently demonstrate a relationship between the presence of MMD and post-treatment tumor recurrence. Prospective studies of MMD are hampered by the large numbers of patients required, the substantial time and expense involved in the methodologies to identify micrometastases, and the lack of a clear definition of this entity. Current data show that adequate staging of colon cancer requires examination of a minimum of 12 nodes per patient.11,12 As a result, a trial to determine the clinical relevance of MMD in 1,200 conventionally staged N0 patients would require IHC examination of at least 14,400 nodes. Clinical studies of MMD would be greatly facilitated if SLNS reduced the number of nodes required for staging to a small fraction of this number. We performed a multicenter trial in patients with resectable colon cancer to determine whether SLNS would identify a small number of LNs that accurately predicted the status of the nodal basin. Our goal was to collect preliminary data that would inform the design of clinical studies of MMD for colon cancer.
Clinical Characteristics This study was performed by surgeons and pathologists at 12 academic medical centers affiliated with the Cancer and Leukemia Group B (CALGB). Patients were eligible for enrollment if they had a clinical or pathologic diagnosis of invasive colon cancer for which surgery with curative intent was planned. Patients with a previous history of colon surgery other than appendectomy were excluded. Informed consent was obtained before surgery according to an institutional review board–approved study protocol (CALGB Protocol 80001). Patient registration, data collection, and data quality review were managed by the CALGB Statistical Center and by the study chairperson following standard CALGB policies. Statistical analyses were performed by CALGB statisticians. This limited-access protocol involved surgical investigators who were proficient in SNLS for other diseases and who also received specific instruction in methods for SNLS in colon cancer. Benign LNs from surgically resected patients with diverticulosis were used as control tissues and were obtained from Brigham and Women's Hospital (Institutional Review Board Protocol No. 2002-P-000898).
Surgical Procedure
Specimen Processing and Examination
IHC All IHC stains were examined by a single pathologist (M.R.) who was blinded to the results reported by the treating institution and to the status of the node as SN versus non-SN. Nodes were scored as positive for either AE1/AE3 or CEA if discrete positive cytoplasmic or membranous staining was present in either single cells or small cell clusters (Fig 1). The number of positive single cells, the number of positive cell clusters, and the number of cells per positive cell cluster were recorded for each LN.
Examination of SNs and Non-SNs by Conventional Histopathologic Criteria Details of SNLS using conventional histopathologic specimen review have been previously reported.5 Conventional histopathologic review is defined as examination of a single H&E-stained section from each LN. In summary, of 66 colon cancer patients undergoing successful SNLS, 24 (36%) had at least one positive node (SN or non-SN) diagnosed by conventional criteria. Of these 24 patients, 11 (46%) had positive SN, and 13 (54%) had negative SN, yielding a false-negative rate of SN analysis of 54%. Blocks containing at least one SN and one non-SN were obtained from each of the 42 patients who underwent successful SNLS and had negative LN by conventional criteria. Multilevel sectioning was performed on each of the 169 nodes received from these patients (Table 1). Examination of five sections per node, obtained at 75-µm intervals, upstaged only two (1.2%) of 169 nodes examined and did not change the overall nodal status of any of the patients.
Classification of LN MMD Using CEA and Cytokeratin IHC Using IHC to detect CEA and cytokeratin-positive cells, patients deemed node negative by conventional histopathology were reclassified as containing metastatic disease if nodes contained one or more positively staining tumor cells in any section (Fig 1). A total of 55 SNs and 46 non-SNs from the 42 patients with N0 disease were examined (average = 1.9 SN and 1.6 non-SN per patient). Of these 42 patients, 30 had both SNs and non-SNs available, five had only SNs available, and two had only non-SNs available for IHC studies. For five node-negative patients, no additional sections could be obtained. Micrometastases were identified by IHC for either CEA or cytoketatin in 54 (approximately 54%) of 101 nodes. Agreement between CEA and cytokeratin stains on a node-by-node basis was 79% (Table 2). IHC identified tumor metastasis in 70% of the 37 patients originally diagnosed as N0 by conventional histopathology (Fig 2 and Table 3). As a result of IHC, the number of node-positive patients increased from 24 (positive by conventional examination) to 50 (positive by either conventional examination or IHC).
To determine the effect of IHC on the results of SN examination, patients were classified as node positive if either an SN or a non-SN from that patient contained tumor cells as detected by either conventional histology or IHC. By this definition, the specificity of SN examination in assigning nodal status is 100%, whereas the sensitivity is the proportion of node-positive patients for whom the SN was positive. We found 13 node-positive patients for whom conventional histopathology failed to detect disease in the SN. On the basis of IHC results, seven of these patients had SNs containing micrometastases (Fig 2 and Table 4). Non-SNs from these patients were not re-examined because they had previously been determined to contain metastatic disease by standard criteria. Identification of CEA- or cytokeratin-positive cells within these SNs decreased the false-negative rate of SN examination to 12% (Table 5). When accuracy was defined as the proportion of patients for whom SN status predicts overall nodal status, the addition of IHC resulted in an accuracy of 88% for SN examination. The subset of 30 conventional node-negative patients for whom both SNs and non-SNs were available for IHC was separately examined. In this selected group, we found a similar false-negative value for SN examination (14%; three IHC-negative SN patients out of 21 node-positive patients).
Evaluation of False-Positive Rate of MMD Studies Unfortunately, IHC methods can identify positively staining cells in LNs from patients who do not have cancer. We evaluated the false-positive rate of CEA and cytokeratin IHC by examining LNs removed from 10 patients undergoing colectomy for benign disease. These included five patients enrolled onto CALGB Protocol 80001 whose final pathology revealed either diverticulosis or benign adenomas and five patients who underwent colonic resection for diverticulosis. An average of 4.7 nodes were identified per patient (Table 6). Positively staining cells were observed in 6.4% of these nodes, with two nodes containing isolated CEA-positive cells and one node containing cells staining for both CEA and cytokeratin. As a result of these assays, two (20%) of the 10 patients studied would be characterized as node positive.
Reduction of False-Positive Results by Modification of IHC Criteria The issue of false-positive staining of tissues in patients with benign disease is well recognized, and others have proposed IHC criteria to overcome this problem.13,14 Therefore, we reclassified the relevant nodes using a more stringent threshold for assigning positivity. Nodes were designated as containing metastatic disease if they contained at least five individual positively staining cells or a cluster of two or more cohesive positively staining cells9 (Figs 1C to 1F). Using this definition, none of the nodes removed from patients with benign disease were scored as positive. Of the 13 patients originally classified as node positive but with negative SNs, two were positive by these criteria. When patients who were characterized as node negative by the treating institution were reclassified according to this definition, 38% were upstaged. Unfortunately, this method also altered the sensitivity of SN examination, yielding a false-negative rate of 32% and a sensitivity for the detection of node-positive disease of only 40% (Fig 3 and Table 7).
Comparison of MMD in SNs and Non-SNs To further test the hypothesis that SNs are more likely to harbor occult metastatic disease than non-SNs, we examined the overall frequency of MMD detected by IHC in LNs from patients classified as node negative after multilevel sectioning and conventional histologic examination. This included a total of 55 SNs and 46 non-SNs examined by IHC at each of three separate levels (Table 3). Any IHC positivity (defined by at least single positive cells) was detected in 32 (58%) of 55 SNs and 23 (50%) of 46 non-SNs. The more stringent IHC definition of five or more positive cells or clusters of two or more positive cells was met by 14 (26%) of 55 SNs and 10 (22%) of 46 non-SNs. These findings indicate that MMD is no more likely to be detected in SN than in non-SN.
In reported series, the use of IHC to detect CEA or cytokeratin antigens in patients with colorectal cancer upstaged 25% to 68% of patients.6,7,15-19 These techniques are difficult to standardize because of differences in antibodies, staining techniques, and scoring systems. Perhaps for these reasons, most of these retrospective studies failed to find prognostic significance for metastases detected by IHC only.6,7,15-19 In this study, we used two scoring systems to define MMD, and these methods were compared with conventional histopathologic staging with the addition of multilevel sectioning. In agreement with results from other SN studies in breast and colon cancer, we found that multilevel sectioning had little impact on staging of patients with colon cancer.7,20 In contrast to results from melanoma and breast cancer trials, however, where the sensitivity and specificity of SN examination is approximately 95%, we found that evaluation of colon SN according to conventional criteria yielded an unacceptably high false-negative rate of 54%.5 This rate is consistent with the rates from several other SN studies in colorectal cancer.4,21,22 SLNS is only useful for studies of MMD if it accurately predicts the status of the nodal basin based on a uniformly applied method to characterize micrometastases. Using a definition of MMD that assigns a positive status to nodes for which IHC reveals a single stained cell, we showed that SN analysis has a low false-negative rate of 12%. By this definition, 70% of patients who were N0 by conventional criteria were upstaged. Because we know that the tumor recurrence rate in conventionally staged N0 patients is approximately 25%, SN analysis by this method would lead to an unacceptably high number of patients classified as having high-risk disease. An additional problem with this criterion is that it yielded a false-positive rate of 20% for benign conditions. We conclude that the presence of single CEA- or cytokeratin-positive cells is not a valid indicator of clinically relevant MMD. This conclusion is supported by other studies.7,23,24 We are left with the challenge of determining what level of tumor expression within a regional LN corresponds to disease that has acquired the ability to survive and establish independent metastatic foci. One approach is to assign a threshold to the amount or character of MMD. Theoretically, tumor cell clusters are more likely to represent clinically significant micrometastases because these constitute a possible colony of locoregional disease. Review of our data showed a high correlation between the presence of five or more single IHC-positive cells and clusters of positive cells. Using this definition to assign a threshold for node positivity, 38% of previously node-negative patients were upstaged, which is a figure that is more consistent with the recurrence rate for this disease. Unfortunately, if patients were staged using only SN examination, this method increased the false-negative rate to 32%, which is a value too high to make SN examination useful in assigning clinical risk. It would be possible to use other intermediate levels of nodal disease; however, this would result in upstaging more patients than would be expected to develop recurrences (false positives). In addition, the false-negative rate of SN analysis could never be reduced to less than 12%, which is the value obtained by defining a positive node positive by the presence of a single positively stained cell. Proponents of SNLS for colon cancer report that this technique is highly accurate, with values ranging from 89% to 97% for the detection of node-positive disease.3,25-27 Most, if not all, of the studies that reach this conclusion apply different diagnostic criteria to the SNs than those used for non-SNs. For example, it is common for SNs but not non-SNs to be scored positive based on IHC results. If it is the case that MMD detected by this method carries the same prognostic significance as metastases that are detected by conventional histology, then this practice is valid. However, if this definition of MMD is not clinically relevant, our data indicate that this practice introduces error and artificially lowers the false-negative rate of SN examination. In addition to benefit achieved by improvements in clinical staging, understanding the differences in character between MMD and nonmetastasizing precursors would provide important targets for improved diagnostic and therapeutic agents. In the future, it is possible that imaging techniques, such as preoperative magnetic resonance28 or intravenously administered near-infrared fluorescence probes, will accurately detect LN metastases from colon cancer. It remains to be seen whether any of these methods are sensitive enough to be applied to studies of MMD. Finally, other methodologies, such as the use of high-throughput polymerase chain reaction–based techniques to identify tumor-specific mutations, may be applied to all nodes removed, thereby eliminating the need for node selection.
The appendix is included in the full-text version of this article, available online at www.jco.org. It is not included in the PDF (via Adobe® Acrobat Reader®) version.
The authors indicated no potential conflicts of interest.
Supported in part by Grant No. CA31946 from the National Cancer Institute to the Cancer and Leukemia Group B (CALGB) and Grant No. CA59594 to the CALGB Surgery Committee. Presented at the 40th Annual Meeting of the American Society of Clinical Oncology, New Orleans, LA, June 5-8, 2004. The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Morton DL, Thompson JF, Essner R, et al: Validation of the accuracy of intraoperative lymphatic mapping and sentinel lymphadenectomy for early-stage melanoma: A multicenter trial—Multicenter Selective Lymphadenectomy Trial Group. Ann Surg 230:453-463, 1999[CrossRef][Medline] 2. Giuliano AE, Dale PS, Turner RR, et al: Improved axillary staging of breast cancer with sentinel lymphadenectomy. Ann Surg 222:394-399, 1995[Medline] 3. Saha S, Wiese D, Badin J, et al: Technical details of sentinel lymph node mapping in colorectal cancer and its impact on staging. Ann Surg Oncol 7:120-124, 2000[CrossRef][Medline] 4. Merrie AEH, van Rij AM, Phillips LV, et al: Diagnostic use of the sentinel node in colon cancer. Dis Colon Rectum 44:410-417, 2001[CrossRef][Medline] 5. Bertagnolli MM, Miedema B, Redston M, et al: Sentinel node staging of resectable colon cancer: Results of a multicenter study. Ann Surg 240:624-628, 2004[Medline] 6. Calaluce R, Miedema BW, Yesus YW: Micrometastasis in colorectal carcinoma: A review. J Surg Oncol 67:194-202, 1998[CrossRef][Medline] 7. Tschmelitsch J, Klimstra DS, Cohen AM: Lymph node micrometastases do not predict relapse in stage II colon cancer. Ann Surg Oncol 7:601-608, 2000[Medline] 8. Jemal A, Murray T, Ward E, et al: Cancer statistics, 2005. CA Cancer J Clin 55:10-30, 2005 9. Singletary SE, Greene FL, Sobin LH: Classification of isolated tumor cells: Clarification of the 6th edition of the American Joint Committee on Cancer Staging Manual. Cancer 98:2740-2741, 2003[CrossRef][Medline] 10. Greene FL, Page DL, Fleming ID: AJCC Cancer Staging Manual (ed 6). New York, NY, Springer, 2002 11. Goldstein NS, Sanford W, Copffey M, et al: Lymph node recovery from colorectal resection specimens removed for adenocarcinoma. Am J Clin Pathol 106:209-216, 1996[Medline] 12. Compton CC, Greene FL: The staging of colorectal cancer: 2004 and beyond. CA Cancer J Clin 54:295-308, 2004 13. Saha S, Dan AG, Beutler T, et al: Sentinel lymph node mapping technique in colon cancer. Semin Oncol 31:374-381, 2004[CrossRef][Medline] 14. Cserni G, Bianchi S, Boecker W, et al: Improving the reproducibility of diagnosing micrometastases and isolated tumor cells. Cancer 103:358-367, 2005[CrossRef][Medline] 15. Broll R, Schauer V, Schimmelpenning H, et al: Prognostic relevance of occult tumor cells in lymph nodes of colorectal carcinomas: An immunohistochemical study. Dis Colon Rectum 40:1465-1471, 1997[CrossRef][Medline] 16. Isaka N, Nozue M, Doy M, et al: Prognostic significance of perirectal lymph node micrometastases in Dukes' B rectal carcinoma: An immunohistochemical study by CAM5.2. Clin Cancer Res 5:2065-2068, 1999 17. Oberg A, Stenling R, Tavelin B, et al: Are lymph node micrometastases of any clinical significance in Dukes Stages A and B colorectal cancer? Dis Colon Rectum 41:1244-1249, 1998[CrossRef][Medline] 18. Nakanishi Y, Ochiai A, Yamauchi Y, et al: Clinical implications of lymph node micrometastases in patients with colorectal cancers: A case control study. Oncology 57:276-280, 1999[CrossRef][Medline] 19. Clarke G, Ryan E, O'Keane JC, et al: The detection of cytokeratins in lymph nodes of Duke's B colorectal cancer subjects predicts a poor outcome. Eur J Gastroenterol Hepatol 12:549-552, 2000[Medline] 20. Lilleng PK, Hartveit F: ‘Missed’ micrometastases: The extent of the problem. Acta Oncol 39:313-317, 2000[CrossRef][Medline] 21. Joosten JJA, Strobbe LJA, Wauters CAP, et al: Intraoperative lymphatic mapping and the sentinel node concept in colorectal carcinoma. Br J Surg 86:482-486, 1999[CrossRef][Medline] 22. Read TE, Fleshman JW, Caushaj PF: Sentinel lymph node mapping for adenocarcinoma of the colon does not improve staging accuracy. Dis Colon Rectum 48:80-85, 2005[CrossRef][Medline] 23. Millis RR, Springall R, Lee AH, et al: Occult axillary lymph node metastases are of no prognostic significance. Br J Cancer 86:396-401, 2002[CrossRef][Medline] 24. Susnik B, Frkovic-Grazio S, Bracko M: Occult micrometastases in axillary lymph nodes predict subsequent distant metastases in stage I breast cancer: A case-control study with 15 year follow-up. Ann Surg Oncol 11:568-572, 2004[CrossRef][Medline] 25. Paramo JC, Summerall J, Poppiti R, et al: Validation of sentinel node mapping in patients with colon cancer. Ann Surg Oncol 9:550-554, 2002[CrossRef][Medline] 26. Saha S, Bilchik A, Wiese D, et al: Ultrastaging of colorectal cancer by sentinel lymph node mapping technique: A multicenter trial. Ann Surg Oncol 8:94S-98S, 2001 (suppl 9)[Medline] 27. Bilchik AJ, Saha S, Tsioulias GJ, et al: Aberrant drainage and missed micrometastases: The value of lymphatic mapping and focused analysis of sentinel lymph nodes in gastrointestinal neoplasms. Ann Surg Oncol 8:8S2-85S, 2001 (suppl 9) 28. Harisinghani MG, Barentsz J, Hahn PF, et al: Noninvasive detection of clinically occult lymph-node metastases in prostate cancer. N Engl J Med 348:2491-2499, 2003 Submitted July 25, 2005; accepted October 21, 2005.
Related Editorial
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2006 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
200 µm scored as positive. 
