| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
1 Department of Pathology, Hospital Clinic, C/Villarroel 170, 08036, Barcelona, Spain
2 Department of Nuclear Medicine, Hospital Clinic, 08036, Barcelona, Spain
3 Department of Gynecology, Hospital Clinic, 08036, Barcelona, Spain
4 Department of Radiodiagnostic, Hospital Clinic, 08036, Barcelona, Spain
Correspondence: Address correspondence and reprint requests to: Pedro L. Fernandez, MD, PhD; E-mail: plfernan{at}clinic.ub.es
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: One hundred fifty-two SLNs were collected from 86 patients. Lymphoscintigraphy, blue dye, and gamma camera intraoperative controls were used for localization. Each SLN was sliced 2 mm thick and was intraoperatively evaluated by using the combination of frozen section and imprint cytology. The final examination included standard hematoxylin and eosin staining, and, in case of persistent negativity, further sectioning, including hematoxylin and eosin combined with immunohistochemistry (CAM5.2 cytokeratin), was performed.
Results: The combination of frozen section and imprint cytology for intraoperative SLN evaluation yielded an intraoperative sensitivity of 78% and a specificity of 100%. All macrometastases (>2 mm) were detected during surgery, as were 2 micrometastases. Final examination detected seven more micrometastases, six of which consisted of isolated tumor cells.
Conclusions: We propose a fast, cost-effective, and accurate procedure for SLN evaluation that is useful for making intraoperative decisions, feasible for most institutions, and reliable because of its high sensitivity (100% for macrometastases) and specificity.
Key Words: Breast • Frozen section • Immunohistochemistry • Intraoperative • Micrometastases • Sentinel lymph node
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
SLN Localization
All patients involved in the study underwent lymphoscintigraphy the day before surgery. The tracer used was a colloid labeled with 99m Tc (Nanocoll; Nycomed Amersham PLC, Little Chalfont, Buckinghamshire, UK; 111 MBq in .5 mL of serum) injected intratumorally because, in our experience, it provides a more successful lymphatic mapping than other modalities (subdermal or peritumoral) and allows a high degree of internal mammary SLN detection. After the injection, serial gammagraphic controls were performed with a gamma camera fitted with a low-energy, high-resolution collimator. Static images of 300 seconds were performed at 30, 120, and 240 minutes, or later if necessary. SLNs were located by identifying the zones that accumulated the tracer and depended on a lymph channel or the first apparent lymph node in the lymphoscintigraphy. Once located, the overlying skin was marked with indelible ink. After general anesthesia, a mixture of 1 mL of methylene blue and 1 mL of saline serum was injected intratumorally, and a gentle massage was performed afterward to visualize SLNs during the surgical procedure. After the tumor was removed, a handheld gamma probe device (Navigator; USCC, Norwalk, CT) was used to identify the highest-activity zone, which was frequently coincidental with the one previously discovered by lymphoscintigraphy. SLN dissection was performed after tumorectomy to remove isotopic activity, which can interfere with SLN location, from the tumoral bed. Our group had previously conducted a validation study for SLN identification in our institution with 105 consecutive cases, as well as another 70 in the application phase. The rate of correct SLN identification in the application phase21,22 was 98.5% with the combination of not only preoperative lymphoscintigraphy and the dye-guided method,23,24 but also the gamma probe. This triple technique is widely used and in a recent review article was described as the most accurate.1
Intraoperative Histopathologic Evaluation of SLNs
The fresh tissue containing the SLN (one or more fragments) was submitted for pathologic examination. All nodal structures isolated were serially cross-sectioned at 2-mm intervals perpendicular to the longitudinal axis (Fig. 1a
). Thus, a 1-cm lymph node became five 2-mm-thick sections. In cases in which the SLN was <5 mm, bisection along the longest axis was considered acceptable. Great care was taken to preserve the correct orientation of all the tissue surfaces and to perform flat sections in order to thereby waste a minimal amount of material.
|
The entire material was then submitted to FS after optimal compound temperature (OCTTM) embedding (Fig. 1b
). Two consecutive 4-µm sections, each containing all tissue surfaces, were cut and intraoperatively evaluated by the pathologist. The first was toluidine blue–stained, and the second was fast hematoxylin and eosin (H&E)–stained (Fig. 1c). By following the example of a 1-cm lymph node, 10 sections (5 tissue surfaces per slide) were evaluated during surgery.
Thirteen pathologists of our department, 11 of whom are not specifically trained in breast pathology or cytology, were involved in the intraoperative evaluation of both the IC and FS included in the study. IC and FS were both available for intraoperative evaluation in all cases. The pathologists were blind as to which of these types of samples would be considered more reliable. A checklist to assess what types of procedures (IC, FS, or both) were considered positive and most valuable for final diagnosis in each particular case was distributed to the participating pathologists.
Definitive Histopathologic Evaluation
For definitive evaluation, SLNs were fixed in neutral buffered formalin, embedded in paraffin blocks, and routinely processed for H&E examination. Sectioning of the block required some degree of deepening: in our department, two or three 4-µm-thick sections are usually discarded for obtaining an adequate slide. One or more slides, depending on the number of isolated SLNs, including all tissue surfaces, were H&E-stained. In the example of a 1-cm SLN, five new tissue surfaces (8–12 µm from the intraoperative surfaces) were evaluated on a glass slide.
When the pathologist did not identify tumor cells, IHC with the monoclonal antibody CAM5.2 directed against low-molecular-weight cytokeratins present in most glandular epithelia, including normal and malignant breast (monoclonal, 1/100; Becton-Dickinson, San Jose, CA) tissue, was performed by using the envision method (DAKO, Dako, Denmark) for development. This procedure implied further profundization in the paraffin block. This time, only two or three 2-µm sections were discarded, and the next two were stained by IHC and H&E. The gold standard used in this study included the final H&E and IHC results.
In summary, the complete SLN study consisted of three intraoperative slides (one IC and two FS) and three more slides for definitive evaluation: one standard H&E and, in negative cases, a new H&E plus IHC. Therefore, for a 1-cm SLN, only 6 slides are needed during the entire protocol, including 1 IC and 25 tissue levels (5 IHC stained) in only 5 more slides. Figure 2 shows a schematic representation of the intraoperative and final evaluation of an SLN.
|
Statistical Analysis
The sensitivity (Se) in the detection of metastases in our study was defined as the ratio between the true-positive (TP) biopsy results detected during surgery and the total number of positive biopsy results, including the final study (true positive plus false intraoperative negative [FN]). The statistical formula was Se = TP/(TP + FN). The specificity (Sp) was defined as the probability of an intraoperatively true negative (TN) result (TN) versus all negative results detected after the final study (FP indicates false positive): Sp = TN/(TN + FP).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), and no case of ITCs was reported during surgery. Only one intramammary SLN was detected and was reported as negative both during surgery and on final diagnosis. Because the clinical relevance of neoplastic involvement in SLNs found in this location remains unclear26 and considering the difficulty involved in deciding whether the node is truly intramammary or axillary, we grouped our only case with the axillary SLN. In 21 of the 25 intraoperatively positive SLN biopsy samples, malignant cells were detected in both IC and FS. Four biopsy samples showed tumor cells only in the FS, and none was positive only in the IC. When questioned as to their preference regarding the type of procedure used (IC vs. FS) for intraoperative SLN evaluation, six pathologists considered that both procedures were equally important and necessary, six believed FS to be more useful, and only one, a cytopathologist, expressed a preference for IC.
Three biopsy samples (3.4%) considered as SLN negative in the intraoperative study showed metastases in the final H&E stain. Two corresponded to ITC, and the third was a .5-mm micrometastasis.
The IHC stain revealed four (4.6%) more SLN-positive biopsies. All four metastases were <.2 mm and were therefore categorized as ITCs.
Using the final H&E with IHC stain as the gold standard, seven false intraoperative negative biopsy results were detected. It is remarkable that no false-positive result was reported during surgery and that the initial percentage of 28.7% intraoperatively SLN-positive patients increased to 36.7% at the end of the study.
The sensitivity of the intraoperative detection of metastases in our study was .78 (78%; false-negative rate, 22%), including all sizes of positive SLNs, and was 100% for macrometastases. The specificity in our study was 1 (100%), because no false-positive results were reported.
Patients with positive results for tumor cells in axillary SLNs detected during surgery underwent ALDN in the same surgical set or delayed ALDN if the diagnoses of metastases were not intraoperative. The protocol in our institution includes performing ALDN even if the metastases were detected only by IHC or as ITCs. Of the 32 SLN-positive biopsy results (1 bilateral case), only 27 ALDNs were performed: in 5 instances the positive biopsy result was an internal mammary SLN, and these patients received adjuvant therapy but no ALDN. A total of 409 axillary non-SLNs were examined with standard H&E stain, with a mean of 15.1 per patient (range, 2–30). One or more positive lymph nodes were present in 7 of 27 ALDNs. In six of the seven tumors with metastatic axillary lymph nodes, the SLN metastases ranged from 4 to 10.5 mm. The remaining case corresponded to a .5-mm SLN micrometastasis. No case classified as having ITC SLN micrometastasis had positive axillary lymph nodes. These results are summarized in Table 1.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Intraoperative SLN Evaluation
If we accept SLN biopsy as a valid staging method for breast cancer, pathologic intraoperative SLN examination can be envisioned as essential to make a rapid, accurate decision regarding the performance or not of surgical axillary dissection during the same surgical procedure. Indeed, when tumor cells are identified, axillary adipose tissue can be removed in one surgical procedure; conversely, a negative result will spare the patient from undergoing ALDN and the associated secondary morbidity (pain, numbness, and paresthesia of the upper arm, as well as later motor dysfunction and lymphedema).10 Some of the methods of assessing SLN during surgery are listed in Table 2.
|
IC alone has been proposed as an intraoperative method to assess the SLN11,14–16 because its preparation is rapid and easy and it involves no tissue loss or frozen tissue artifact,2,14,30 but it is more difficult to interpret than FS for non–cytology-trained pathologists. Experienced cytopathologists may be necessary to evaluate the glass slides,14,15 and in some hospitals this expertise is not always available. The number of touch prints would also depend on the method of SLN slicing and on whether they were obtained from all cut surfaces or not. IC is also associated with variable sensitivity, ranging from 53% to 97.1%, and has a greater chance of indeterminate results.2,14–16
It has been proposed that frozen sectioning allows better diagnostic confidence than IC alone, with fewer indeterminate results, but it is technically more difficult and consumes tissue.30 In this sense, we believe that the benefit of avoiding a second surgical procedure is greater than the minimal risk of a false-negative result in the final study because of tissue consumption derived from the loss of material in FS. FS alone has thus been proposed as an intraoperative method.12,13,17 The sensitivities reported for this technique also vary widely, ranging from 67.9% to 93.7%.12,17 The highest sensitivity corresponds to Veronesi et al.,17 but their intraoperative method involves high costs and a lengthy procedure, and the aim of reducing false-negative SLN results necessitates step sectioning of the complete sample. This may be accomplished with great economical and personal effort and leaves little tissue after the intraoperative study. We think that the risk of prolonged anesthesia, together with the impracticality of its application in most institutions, is greater than the benefit of discovering a few more micrometastases, the significance of which is still debatable.
Turner et al.30 compared FS and IC separately by using H&E-stained paraffin sections as a standard. FS showed a slightly higher sensitivity than IC (80% vs. 60%–70%). The authors also concluded that both IC and FS often failed to detect micrometastases and that neither was clearly superior.31
Unfortunately, many hospitals have insufficient personnel for specific training in breast pathology or cytology. There is therefore a need for a combined method that can be widely adopted and that can provide sufficient information to achieve a reliable intraoperative report on SLN status.
Evidence to date suggests that intraoperative cytological examination should be used in conjunction with FS to enhance the efficiency of SLN intraoperative evaluation.1,10,29 Motomura et al.10 made intraoperative IC from all cut surfaces, but only one portion of the entire SLN was submitted for FS, and when there was more than one SLN, only a portion of the largest was submitted. The reported separate values of sensitivity for IC were better than for FS, but this may be explained by the limited samples for the latter.
In our study, 2-mm-thick serial sectioning of the entire SLN and focused analysis combining both intraoperative IC and FS spared 78% of the patients a second costly surgical procedure. It is remarkable that all macrometastases were discovered during surgery, because every tumor deposit 
2 mm appeared in the tissue surfaces, and even two non-ITC micrometastases were detected during surgery. An increase in the number of tissue surfaces on each glass slide also helped to evaluate more levels efficiently. Only a previously unnoticed .5-mm micrometastasis and six ITCs appeared in the final study, thus indicating a significant sensitivity of 66% (two of three) for non-ITC micrometastases on intraoperative evaluation. No false-positive result was reported in our study, contrary to other authors who used FS12 or IC2,15 alone.
It is interesting to note that malignant cells were detected only in the FS, with negative IC, in four intraoperatively positive SLNs. This suggests a higher sensitivity for the former procedure when there is no specific cytopathology training. Only one participant, a cytopathologist, considered IC as the most valuable method for this type of evaluation, whereas the remaining pathologists preferred either FS or a combination of both for effective detection of metastases.
We nevertheless believe that it is not reasonable to expect that intraoperative evaluation will detect every metastatic deposit eventually discovered in permanent serial sections or IHC stains.12 Therefore, a balance must be achieved between accuracy and feasibility.
To further increase such accuracy, which in our methods has a false-negative rate of 22%, rapid IHC methods have been described for FS and for IC intraoperative SLN evaluation, but they add cost and time.12,19,32 Some authors8 have designed intraoperative IHC methods that take approximately 10 minutes and are probably not feasible for most institutions, whereas others12 believe that they cannot be recommended on a routine basis because even IHC on permanent sections is still debatable. For instance, Veronesi et al.17 reported that rapid IHC staining for cytokeratins did not increase the detection rate of SLN metastases with their exhaustive method. Llatjos et al.14 also had available rapid cytokeratin staining in their intraoperative IC method, but 30 minutes was required, and it was considered a delayed procedure. Rapid cytokeratin immunostaining, therefore, does not seem to significantly contribute to an improvement in intraoperative SLN evaluation.
We have estimated that when a histopathologic intraoperative SLN evaluation is performed with our method, arrival from the operating room to diagnosis takes approximately 15 minutes when only one node is detected and takes 20 to 25 minutes for more than one. In the first case, this implies approximately 7 to 8 minutes of technician time and approximately the same for pathologist time, with minimal costs derived from three glass slides, DiL-Quick and H&E stainings, OCT freezing compound, one plastic cryomold, mounting media, and cover-slips. An estimation of the cost of the intraoperative evaluation of one SLN is US$135 in our institution. We therefore believe that this cost does not add a significant economical load to patient treatment if we consider that we are saving many very expensive second surgical procedures.
Definitive Study of the SLN
The proper handling of the SLN in definitive study has not been standardized.19 The number of levels from the paraffin block and the addition or not of IHC varies widely from author to author. It is widely accepted that the addition of IHC to the H&E sections increases the percentage of positive results from 13% to 28%.5,18,33,34
Turner13 intraoperatively examined both FS of 60 SLNs from 42 consecutive patients and paraffin sections at two levels separated by 40 µm that were H&E- and cytokeratin-stained at every level. He then performed eight more levels in the cases that were still negative for tumor cells and concluded that additional step sections with cytokeratin IHC did not significantly increase the number of patients with positive SLNs.
This has also been concluded by others,17 and another interpretation of their results could be that the smaller the sectioning intervals, the higher the rate of metastatic cells or micrometastases that can be detected on H&E and the lower the added value of IHC.34
In other studies, four or five more levels were performed when the H&E was negative, and the number of IHC ranged from two to six.5,12,34,35 The median size of micrometastases detected on cytokeratin IHC stains was approximately .1 mm, in contrast to 1.0 mm for H&E stains.36
Although some authors state that general surgical pathologists prefer routine cytokeratin IHC, which can be used in most institutions, for accurate diagnosis, some breast pathologists recommend against it, because the meaning of micrometastases or ITCs remains unclear.12 Conversely, other authors27,37,38 conclude that patients with SLN micrometastases are more likely to have carcinoma cells in non-SLNs than SLN-negative patients; therefore, a careful search for micrometastases is warranted, and IHC is recommended.
The final examination in our study included standard H&E and, in the case of persistent negativity, IHC of all tissue sections in one or more slides plus a new H&E. This careful, although not extensive, procedure allowed detection of one micrometastasis and six ITCs that had been overlooked during surgery. To further decrease the discrepancy between the intraoperative and final diagnoses by increasing the rate of micrometastasis and ITC detection in FS (two of nine in our study), the addition of IHC or other costly and lengthy procedures, such as polymerase chain reaction–based methods, would be necessary. Until the definitive meaning of occult micrometastases is established, arguments for and against IHC or other sensitive techniques will be frequent.
In conclusion, we believe that, although validation and improvement are necessary in larger series from different institutions, the time has come for obtaining a standard intraoperative evaluation of the SLN and that the method described in this article constitutes a fast, cost-effective, and accurate procedure that is useful for making intraoperative decisions, is feasible for most institutions, and is reliable because of its high sensitivity and specificity.
|
ACKNOWLEDGMENTS |
|---|
Received for publication July 1, 2004. Accepted for publication November 29, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  G.-C. Hsu, C.-H. Ku, J.-C. Yu, C.-B. Hsieh, C.-P. Yu, and T.-Y. Chao Application of intraoperative ultrasound to nonsentinel node assessment in primary breast cancer. Clin. Cancer Res., June 15, 2006; 12(12): 3746 - 3753. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
