Comparability of Different Pathologic Protocols in Sentinel Lymph Node Evaluation: An Analysis of Two Step-Sectioning Methods for the Same Patients With Breast Cancer

Sentinel lymph nodes (SLNs) are recognized as the first lymph nodes to receive lymphatic drainage from a primary tumor before spreading to other lymph nodes. The SLN dissection is widely performed for patients with a range of cancers, especially those with breast cancer. The status of SLNs accurately reflects the presence or absence of metastases in the axillary lymph nodes and can also help in determining the tumor stage and appropriate surgical strategies for further axillary lymph node dissection.1 Because of its very high negative predictive value regarding the status of the remaining axillary lymph nodes, a negative SLN dissection can spare patients with breast cancer who have lymphedema pain, numbness, and restricted shoulder motion from undergoing complete axillary lymph node dissection.2 Although SLN evaluation has been accepted as a standard surgical procedure in early breast cancer, there still is no consensus on the pathologic protocol that should be used on an SLN specimen.3 A recent survey of the histopathologic protocols used in Europe for the examination of SLNs of patients with breast cancer showed that 123 different histologic protocols exist across 240 laboratories.4 The existing pathologic procedures for SLNs range from simple approaches to more extensive examinations of step or multilevel sections with or without cytokeratin immunostaining and/or sophisticated molecular techniques.5 The more thorough that pathologists are with SLN evaluations, the greater the number of metastatic foci detected. An examination of a limited number of pathologic sections may result in an underestimation of the metastatic potential of breast carcinoma.6 

The step- or multilevel-sectioning method is currently the agreed-upon “standard” for the pathologic examination of SLNs.7–8 However, at present, no widely accepted pathologic protocol for SLN evaluation is considered the “standard.” Previous studies in the literature have suggested that it may be more efficient to obtain fewer sections at larger intervals than to obtain many sections at small intervals for each slide, as a greater volume of SLN can be examined by the former than the latter method.9 Relatively few studies have compared the different step-sectioning methods in separate groups of patients.7,10 This study aims to statistically compare the yields of 2 different step-sectioning methods. For each patient with breast cancer, a 20-μm section, representative of a smaller interval, and a 150-μm section, representative of a larger interval, were examined.

Our study was approved by the Institutional Research Ethic Committee of the faculty of medicine at Prince of Songkla University (PSU) in Hat Yai, Songkhla, Thailand. Since November 2004, the Songklanagarind Hospital in Thailand has provided SLN dissection for patients with breast cancer. The first phase of the study consisted of labeling and identifying the SLNs during surgery and evaluating the SLNs by comparing their histologic status against the histologic status of all removed axillary lymph nodes. This study yielded satisfactory results (unpublished data).

Between November 2004 and September 2007, a total of 110 patients with T1-T2 breast carcinoma underwent successful SLN dissection at Songklanagarind Hospital. The SLNs were evaluated by using the PSU protocol consisting of a preoperative subareolar injection with 2 mL of in-house blue dye and 1 mL (1 mCi) of Tc 99m nanocolloid radioactive tracer. Intraoperative detection with a gamma probe (Neoprobe Corporation, Johnson & Johnson Medical, Hamburg, Germany) was used.

A set of 186 SLNs from 80 patients who initially had negative results from the standard microscopic examination (single hematoxylin-eosin [H&E] slide) was further evaluated. The pathologic protocol of SLN assessment is described here briefly. All lymph nodes were dissected clearly and fixed in formalin. Each SLN greater than 3 mm was bivalved or serially cut to between 2 to 3 mm along the transverse plane to have the widest surface. All lymph nodes were submitted in their entirety. Each paraffin block was cut at 3 μm and stained with H&E. If the initial slide showed a negative result, each block was then sectioned at an interval of 20 μm for 5 levels and further sectioned at an interval of 150 μm for an additional 5 levels. The sections were sequentially numbered in the order in which they were cut. For each set of slides, the first 2 were stained with H&E and the remaining 3 were prepared by cytokeratin immunostaining with antibody AE1/AE3 (1:500 dilution; Dako, Glostrup, Denmark). All slides were examined independently by the same pathologist (K.S.) without knowledge of the actual SLN result obtained from each method. For an individual SLN, the findings from each method were compared, and all SLNs that contained any cell consistent with a malignant morphology were considered pathologically positive. Metastatic foci were measured by using an Olympus eyepiece micrometer (Thailand) and classified according to the TNM staging scheme as (1) isolated tumor cells (ITCs) of 0.2 mm or smaller; (2) micrometastasis larger than 0.2 mm but not greater than 2 mm; and (3) macrometastasis larger than 2 mm. All axillary dissections and primary tumors were examined as in a routine histologic examination. Tumor grade was assigned by using the modified Bloom-Richardson system.

Data were collected and stored with EpiData 3.1 software. All analyses were performed using R statistical software version 2.7.0. The McNemar χ² test was used to compare the lymph node results obtained by the 20-μm and 150-μm step-sectioning methods. Differences were considered statistically significant when the P value was .05 or less.

The clinicopathologic parameters of the study group are summarized in Table 1. The size of the invasive tumors ranged from 0.7 cm to 5 cm (median, 2 cm).

From 80 patients, 186 originally negative SLNs were evaluated. The median number of SLNs per patient was 2 (range, 1–7) and ranged in size from 0.3 to 3 cm (mean, 1.1 cm). The results of SLNs obtained from the 20-μm and 150-μm intervals are shown in Tables 2 and 3. The overall detection rate for metastatic foci was 27.5% (22/80) at the 20-μm interval and 20% (16/80) at the 150-μm interval. Across the 2 section types, for example, at 20-μm and 150-μm intervals, the findings for 12 of the 80 cases (15%) were positive for lymph node metastasis and 54 (67.5%) were negative (Tables 4 and 5). Neither sectioning method missed any macrometastasis. However, 2 cases of micrometastasis were underdetected at the 20-μm interval, whereas 10 cases with ITCs (requiring cytokeratin staining) were underdetected at the 150-μm interval. When considering only macrometastases, only 1.2% (1/80) of cases were understaged by both methods; if micrometastases were additionally taken into account, 2.5% (2/80) of cases were understaged by the 20-μm sectioning method. However, the SLN results at 20-μm and 150-μm intervals are not statistically different (McNemar test; P = .18 for case-based results and P = .052 for nodal-based results).

The sentinel lymph node biopsy is already accepted as the standard staging procedure for patients with early breast cancer (stages I and II). It has been claimed that this technique improves staging by allowing the pathologist to focus on only 1 or a few selected lymph nodes.8 As long as the surgeon performs a reliable SLN biopsy procedure, the determination of whether lymph node metastases are present depends on the pathologist.11 The lack of a universal protocol for the pathologic examination of SLNs has resulted in wide heterogeneity in the procedures adopted by different institutions.4 Thus, the detection rates as well as the predictive values derived from the protocols used in these studies are varied and incomparable.

Traditionally, histopathologic examination of SLNs was the same as the routine assessment of the axillary lymph node. However, the false-negative rate was too high, resulting in many patients receiving delayed adjuvant therapy or none at all. The serial- or multilevel-sectioning technique has thus become the accepted standard protocol for SLN processing. However, the “intra-protocol” methods including handling, sectioning, and staining used in these studies vary widely. Many studies have tried to address the question of optimal pathologic examination of SLNs by comparing different protocols. Turner et al examined 42 SLNs and found 15% of SLN metastases in the first 2 levels at a 40-μm interval, but only 3% revealed tumor cells in the remaining deeper sections.7 Zhang and associates reported that serial sectioning at levels of 25%, 50%, and 75% within the tissue blocks detected almost all metastases.10 A later study by Torrenga and colleagues suggested 4 additional levels, separated by a 250-μm interval.11 It is not altogether surprising that the authors of different studies of individual sets of patients have presumed that their pathologic protocols produce the highest sensitivity. From this perspective, a direct comparison of the results obtained with these protocols may be inappropriate because the biologic characteristics of patients are nonhomogeneous. In contrast, the results from our study are comparable because we conducted the study “internally,” for example, by using the same SLNs from the same patients, so the issue of biologic variation can be ignored. The McNemar statistical test used in this study facilitated a reasonable comparison between the paired outcomes of detection for sections at the 20-μm and 150-μm intervals.

In the present study, the routine pathologic examination of a single H&E section, when compared to the 5-level H&E and cytokeratin sections, failed to detect macrometastatic or micrometastatic tumor deposits in 8 (10%) of the 20-μm sections and 10 (12.6%) of the 150-μm sections. Our study has also demonstrated that routine H&E staining, as well as H&E staining in serial sections, underdetects some micrometastases and all ITCs. It is clear that in addition to the pathologic evaluation of the sentinel lymph node, immunohistochemical staining increases the metastatic tumor detection rate. The detection rate of micrometastasis in our study was small when compared to a previous study.12 This can be explained by both the higher fraction of small tumors (pT1) in our study and the biologic variability of tumors. A higher percentage of grade III invasive ductal carcinoma suggests a variation among ethnicities. Two micrometastases, which were missed by the 20-μm interval technique, were in larger lymph nodes (15 and 17 mm) and their sizes, 0.4 and 0.7 mm, were in the lower range of micrometastasis. Nevertheless, from the statistical analysis, the results of the 20-μm and 150-μm interval methods were comparable, possibly because we tried to cut across the hilum of all the SLNs at the contralateral entrance of lymph vessels. The metastatic foci forming in lymph nodes usually enter via afferent lymphatic vessels located opposite to the lymph node hilum. Thus, a considerable proportion of the metastases are generally identified in the first section or in the first few sections obtained subsequently away from the center of SLNs.

The general consensus on metastatic detection in SLNs is that, regardless of the method used, the aim of SLN evaluation should be the identification of all macrometastases of at least 2.0 mm.3 However, the clinical significance of micrometastasis is also becoming increasingly obvious.13 It has been suggested that micrometastases larger than 1 mm have a higher prevalence of associated non-SLN metastases than do smaller micrometastases (30.2 versus 17; P = .01).4,14 The International (Ludwig) Breast Cancer Study15 showed that both disease-free survival and overall survival were significantly affected by the presence of micrometastatic disease at 5 and 6 years of follow-up, and Rosen et al have found that patients with T1 tumors with nodal metastasis smaller than 2 mm had poor prognosis for disease-free interval and overall survival at 10 years.16 A recent recommendation from the American Society of Clinical Oncology Expert Panel suggests an axillary lymph node dissection for patients with SLN micrometastasis, regardless of the method of detection.13 Although the conclusions from the Consensus Conference Committee in Philadelphia in 2001 did not recommend using immunohistochemistry or molecular biology assays in SLN examination,17 our study has shown that immunohistochemistry identifies some micrometastases that may have been missed by H&E staining. It should be re-emphasized that immunohistochemistry will identify more ITCs than H&E staining alone. Although there is still no evidence that ITCs have a prognostic relevance, a study by Viale showed that, in a series of 116 patients, 14.7% of patients with ITCs had further axillary involvement.18 

Because of the statistical comparability of the detection rates for metastases with the 20-μm and 150-μm interval step-sectioning methods, a pathologist who deals with SLN examination could choose either of them with confidence. The advantage of our current protocol, which produces sections at a 20-μm interval, is that less material is consumed in tissue blocks (about 120 μm per block) and that it is more practical for technicians to produce a set of slides from the smaller-step interval. We recommend obtaining 2 additional H&E slides and a further 3 immunohistochemical slides for step sectioning, to be separated by a 20-μm interval when the original H&E section is negative. Despite the fact that our protocol might still miss some small metastases, we believe that it guarantees a complete detection of all meaningful metastases known to date and also creates a more acceptable workload for the pathologist. Institutional consensus among pathologists and surgeons, as well as oncologists, is required.

The present study was supported by a grant from the Faculty of Medicine Research Fund. The authors wish to thank associate professor Hutcha Sriplung for comments on research methodology and statistical analysis and Mr. David Brown, an English facilitator, for his assistance with manuscript preparation.