Schuman S, Walker G, Avisar E. Processing Sentinel Nodes in Breast CancerWhen and How Many?. Arch Surg. 2011;146(4):389-393. doi:10.1001/archsurg.2011.29
To analyze a series of sentinel nodes (SNs) from patients with node-positive breast cancer to determine their diagnostic value, to delineate a working algorithm, and to assess the clinical value of our common practice
A prospectively collected database.
Tertiary referral center.
One hundred five patients with node-positive breast cancer who underwent SN biopsy.
Main Outcome Measures
The diagnostic value of SNs by analyzing the sensitivity of processing the hottest, 2 hottest, hot and blue, or hot, blue, and suspicious SNs.
Three hundred fifty-three axillary SNs were recorded in the database. An analysis of the 282 radioactive axillary nodes for which the 10-second count was recorded reveals that the most radioactive node was positive in 73 of 94 analyzable patients (77.7%). Consideration of the 2 most intense axillary nodes was sufficient to diagnose nodal disease in an additional 12 patients, representing a significant increase in sensitivity to 90.4% (P < .001). Examination of all other radioactive nodes did not diagnose any additional cases. On the basis of all 105 patients, consideration of nonradioactive blue axillary nodes did not add significant diagnostic value relative to testing only radioactive nodes: sensitivity of 86.7% vs 88.6% (P = .50), whereas consideration of all hot, blue, and suspicious nodes improved sensitivity to 96.2% (P = .002).
Processing of the 2 hottest nodes, along with suspicious but nonhot and nonblue nodes, is sufficient for initial axillary staging. Additional radioactive SNs should be processed only in the presence of nodal disease.
Sentinel lymph node (SN) biopsy has gained acceptance in the stratified therapeutic approach to breast cancer. This allows the axilla to be staged more accurately without the morbidity of a full axillary dissection.1- 5 In this procedure, a radioactive tracer, a blue dye, or both are injected around the tumor. The lymph nodes primarily draining the injection site are identified in the axilla by the blue color and the radioactive signal recorded using a small detecting probe.
After SNs are removed, the pathologist performs multiple sections on each node looking for metastatic disease. Sometimes, cytokeratin stains are added to increase the detection rate.This tedious examination allows the identification of micrometastases in the SNs and leads to upstaging of approximately 9% to 10% of them.6- 9
It has been shown previously10,11 that more than 1 SN is commonly found in the axilla. The blue dye defines an SN in a rather simple way. Any lymph node with blue color in it or in a lymphatic entering it is considered an SN.12 This definition is, however, not so simple for radioactive “hot” nodes, where different amounts of radioactivity might be present in different lymph nodes. The commonly accepted way to measure the radioactivity has been to perform a 10-second radioactivity count ex vivo after the lymph node is removed from the surgical field.
Previous studies13,14 have demonstrated that the first encountered node is not necessarily the hottest node and that cancer cells might be found in the second or third to the hottest node. It is, therefore, common practice to continue dissecting lymph nodes until the counts are less than 10% of the hottest node.
A better understanding of the diagnostic value of multiple SNs dissection is, however, of primary importance to minimize operating time and possible morbidity and to decrease cost. Once a positive node is identified, standard axillary dissection is routinely performed. In that case, there is no additional diagnostic value in dissecting more SNs before the axillary dissection. The objective of this study was to analyze a series of SNs from patients with node-positive breast cancer to determine their diagnostic value, to delineate a working algorithm, and to assess the clinical value of our common practice.
After receiving institutional review board approval from the University of Miami, Miami, Florida, we created a prospective database of 105 patients with node-positive breast cancer undergoing SN biopsy during a 6-year period. All the patients signed informed consent forms. One attending surgeon (E.A.) at 2 institutions performed the SN biopsies. The patients were previously diagnosed by either core or excisional biopsy. In the first phase, all the patients underwent SN biopsy immediately followed by axillary dissection in a validation study. In the second phase, SN biopsy was offered as a stand-alone procedure. Axillary dissection was performed only if the SN was positive or as part of the National Surgical Adjuvant Breast and Bowel Project B-35 study. The SN biopsies were performed with technetium sulfur colloid, blue dye, or both. A peritumoral injection of technetium sulfur colloid, 400 μCi, divided into 4 equal doses was performed within 24 hours of surgery. Two milliliters of patent blue (Laboratoire Guerbet, Paris, France) or isosulfan blue (Ben Venue Laboratories Inc, Bedford, Ohio) was injected peritumorally in 4 equal doses within 30 minutes of the skin incision. The SNs were defined as any blue, suspicious, or radioactive nodes using the radiodetector probe and up to 10% of the hottest node.
Node location, SN definition criteria (hot/blue/suspicious), 10-second radioactive counts, and pathologic status were recorded. Demographic information, tumor size and side, and histologic findings were recorded prospectively, along with the SN results and the final axillary nodal status when available. Specific data on each SN were also entered prospectively, including serial number, hot/blue/suspicious status, and 10-second counts. For diagnostic value analysis, the SNs were reorganized for each case by radioactivity intensity so that the hottest node was labeled as “first,” the second hottest node as “second,” etc.
All the SNs were preserved in a formalin solution and were processed as follows. The SN was sectioned perpendicular to the long axis at 2- to 3-mm intervals and was submitted entirely in cassettes. Each block was serially sectioned at 50-μm intervals to produce 15 levels of 3 μm each. Pancytokeratin immunoperoxidase staining was performed on levels 4, 8, and 12. The other 12 levels were stained with hematoxylin-eosin. All the slides were examined sequentially from level 1 to 15. In each case, the value of dissecting less radioactively intense nodes was assessed in comparison with the information available from lower numbered nodes. In addition, the added diagnostic value of the blue dye injection was compared with the radioactivity injection. The sensitivity of the diagnostic tests based on various SN criteria was estimated by the 95% exact confidence interval, and comparisons were made using the McNemar test for paired data, exact version.15,16
A total of 105 patients with node-positive breast cancer were enrolled in this study. The median patient age was 57 years (age range, 25-86 years); 103 were women and 2 were men. Fifty-three tumors were located in the right breast, 51 in the left breast, and 1 bilaterally. One hundred patients had their SNs located in level I, and 5 patients had their SNs located in levels I and II. All the patients had invasive tumors, of which 11 were lobular, 1 mucinous, and 1 metaplastic. All other tumors (n = 92) were invasive ductal carcinomas (Table 1).
Three hundred fifty-three axillary SNs were recorded in the database. The number of nodes per patient ranged from 1 to 15 (median, 3). Radioactivity was found in 301 SNs (85.3%) and blue dye in 103 (29.2%), with 90 nodes (25.5%) being hot and blue. Thirty-nine SNs (11.0%) were removed for suspicious appearance only. One hundred forty-seven of the 353 SNs (42.6%) tested positive for disease. Of the 147 positive nodes, 127 (86.4%) were radioactive and 51 (34.7%) were blue. Forty-five nodes (30.6%) were hot and blue (Table 2).
An analysis of the 282 radioactive axillary nodes for which the 10-second count was recorded reveals that the most radioactive node was positive in 73 of 94 analyzable patients (77.7%). Consideration of the 2 most intense axillary nodes was sufficient to diagnose nodal disease in an additional 12 patients, representing a significant increase in sensitivity to 90.4% (P < .001). Examination of all other radioactive nodes did not diagnose any additional cases (Table 3). On the basis of all 105 patients, consideration of nonradioactive blue axillary nodes did not add significant diagnostic value relative to testing only radioactive nodes: sensitivity of 86.7% vs 88.6% (P = .50). However, inclusion of suspicious nodes, which increased sensitivity to 96.2%, represented a significant gain relative to radioactive nodes (P = .002) (Table 4).
Despite its widespread use, the technical aspects of SN biopsy are still in evolution in an effort to maximize its effectiveness and to decrease the false-negative rate. Initially, SN biopsy was performed with either radioactive isotope or blue dye alone. Further studies have shown that radioactive tracer and blue dye in combination reduces the number of false-negative biopsy results.10,11
Much debate has been focused on the best definition of a radioactive or “hot” SN. Should this be a definition based on absolute radioactive counts or, rather, a relative definition where the SN is compared with the background or with non-SNs. It was quickly realized that the absolute radioactivity counts differ between cases based on the specific lymphatic drainage, the injection technique, and the exact time elapsed between injection and surgery. It also became clear that there often is more than 1 hot node and that those hot nodes have different radioactive counts. Furthermore, it was observed that tumor can be found in a less radioactive node, whereas the more active node is negative.13,14 Two potential explanations exist for this phenomenon. The first relies on the hypothesis that some tumors might have dual lymphatic drainage to 2 separate lymph nodes and, therefore, one might be contaminated with tumor cells while the other is not. The second explanation relies on the hypothesis that in a metastatic lymph node, further lymphatic flow into the node might be partially blocked by tumor cells, leading to a bypass toward a secondary node. Therefore, the hottest node is not really the SN and might not have cancer cells in it while the true SN, which is contaminated with tumor cells, might have absorbed less tracer and is less hot.17
The question remains, however, how many hot nodes need to be sampled to get the best results? In a previous study using only radioactive colloid, Borgstein et al13 demonstrated a low false-negative rate of 1.7% by removal of all SNs in 50% of the hottest node. In a large retrospective analysis, McMasters et al18 showed that tumor cells can be found also in the third or fourth SN. They concluded that all nodes in 10% of the hottest node should be removed. The true question, however, is, “What is the diagnostic benefit of removing and analyzing the second, third, and fourth SNs after a stronger node has already proved the absence of tumor cells in the axilla?” It would be unnecessary to request from the pathologists a full SN protocol if there are no clinical implications. In addition, although SNs are associated with less morbidity,19,20 it seems logical that removal of more nodes would lead to more adverse effects.
In this study, we ranked the radioactive SNs by their ex vivo 10-second counts. We observed that although removal of the hottest node correctly diagnosed a positive axilla in 77.7% of cases, removal of the second hottest node diagnosed an additional 12 patients, representing a significant increase in sensitivity to 90.4%. Examination of all other radioactive nodes did not diagnose any additional cases. Therefore, although metastatic disease can, indeed, be found in the third, fourth, or even fifth hottest SN, there is no additional diagnostic benefit to the dissection and analysis of those nodes as SNs. Regarding SN criteria based on radioactivity relative to the hottest node (percentage), we initially noted that all positive axilla were diagnosed after removing nodes in 30% of the hottest node. As we added more patients to the database, we found a case diagnosed by the second hottest node, which was 25% only from the hottest node. Therefore, we believe that a larger data set will probably lower this value even more. Based on the 2 potential explanations for the nonpositivity of the hottest node, the exact percentage value of the second node would be highly variable depending on the amount of tumor blockage of the afferent lymphatics or depending on the injection spot around the tumor, which has dual drainage. Conceptually, therefore, we believe that the 2 hottest nodes concept carries a much more valid rationale.
A major difference exists between the first and second SNs encountered in the axilla and the hottest, second hottest, etc. From a practical standpoint, it is not always possible to dissect the hottest node first, the second hottest node second, and so on. Sometimes the hottest node lies deep in the axilla and is identified only after less radioactive nodes have been removed. In those cases, the benefit of this study would be for the pathologic analysis of the nodes and could save time and money. When the remaining lymph nodes in the axilla have an in vivo 10-second count that is less than the ex vivo 10-second counts of 2 previously dissected nodes, the present results would not support the need to dissect those nodes.
In this series, pathologic evaluation of the 2 hottest nodes reduced the number of radioactive SNs analyzed in 94 patients from 282 to 169 (40% reduction). A cost analysis reveals that Medicare technical and professional reimbursement for pathologic analysis of 1 SN is $191.82. On the basis of a median of 3.5 nodes per patient, the cost is $671.37. Data from the University of Miami reveal that approximately 600 patients with breast cancer undergo SN biopsy annually, resulting in $402 822 in cost. Analysis of the 2 hottest nodes only would result in a yearly cost reduction of $172 638 (3.5 × 191.82 × 600 minus 2 × 191.82 × 600) or slightly more when considering that some patients will have only 1 hot SN.
In this study, the use of blue dye identified 2 additional node-positive patients; however, the gain in sensitivity is small (2%) and is not statistically significant. In addition, few SNs are identified by blue dye only when the radioactive tracer does not migrate to the nodes. In those rare cases, there is a clear benefit to blue dye in avoiding an unnecessary axillary dissection for SN-negative patients. Moreover, blue dye makes the identification of hot and blue nodes easier.
Removal of suspicious, nonblue, and nonhot nodes was beneficial, diagnosing 8 more cases than a test based on hot or blue nodes and increasing sensitivity from 88.6% to 96.2% (P < .008). Thus, we believe that palpation of the axilla should still be routinely performed.
At the University of Miami, we routinely dissect the internal mammary SNs identified preoperatively via lymphoscintigraphy or intraoperatively via the gamma probe. This analysis was repeated including internal mammary SNs, which resulted in a modest improvement in sensitivity that attained marginal significance for the test based on radioactive nodes (91.4% vs 86.7%, P = .06) and when both radioactive and blue nodes are considered (93.3% vs 88.6%, P = .06). Internal mammary SNs did not add diagnostic value, however, in relation to testing all axillary nodes (100% vs 96.2%, P = .13). Nevertheless, 4 patients were positive in the internal mammary SN only and negative in the axilla. Thus, they were upstaged and their treatment plan was changed. However, we did not include the internal mammary SNs in the final analysis for 2 reasons. First, one may argue that we should not compare the intensity of radioactive tracer uptake between 2 different lymph node basins. The drainage patterns may be different between the 2 basins, leading to differential uptake and intensity of the radioactive tracer in the axillary SNs vs the internal mammary SNs. In such circumstances, the concept of analysis of the hottest, second hottest, etc, may no longer be valid. Second, we aimed for analysis to be applicable to the routine practice of SN biopsy. Because most surgeons do not routinely look for or biopsy the internal mammary SNs, we excluded the internal mammary SNs from the final analysis. Table 5 explains the false negatives in this study when the analysis was performed on the basis of axillary SNs only or both axillary and internal mammary SNs. Those results are based on a single institution's relatively small database. Before recommending implementation of the 2 hottest nodes practice, those results should be validated in a larger multi-institutional database.
In conclusion, although the technique of SN sampling in breast cancer should not be altered, intraoperative pathologic processing may be limited to the 2 hottest nodes, together with suspicious but nonhot and nonblue nodes. This is sufficient for the initial axillary staging. Additional radioactive SNs should be processed only in the presence of nodal disease. When radioactive SNs were present, blue dye did not add diagnostic value but made SN identification easier.
Correspondence: Samer Schuman, MD, Division of Gynecologic Oncology, University of Miami–Miller School of Medicine, 1475 NW 12th Ave, Sylvester Comprehensive Cancer Center 3500, Miami, FL 33136 (firstname.lastname@example.org).
Accepted for Publication: November 1, 2011.
Author Contributions: Dr Schuman had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Schuman and Avisar. Acquisition of data: Schuman and Avisar. Analysis and interpretation of data: Schuman, Walker, and Avisar. Drafting of the manuscript: Schuman and Avisar. Critical revision of the manuscript for important intellectual content: Schuman, Walker, and Avisar. Statistical analysis: Walker. Administrative, technical, and material support: Schuman and Avisar. Study supervision: Schuman and Avisar.
Previous Presentations: This study was presented as a poster at the American Society of Clinical Oncology 2009 Breast Cancer Symposium; October 8-10, 2009; San Francisco, California (American Society of Clinical Oncology Merit Award); and at the San Antonio Breast Cancer Symposium; December 9-13, 2009; San Antonio, Texas.