| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
NEW APPROACHES TO THE TREATMENT OF HEPATIC MALIGNANCIES |
From the Departments of Surgery (CN, CL-S, JSH, DH, KBC) and Pathology (PF, XS-G), Institut Curie, Paris, France.
Correspondence: Address correspondence and reprint requests to: Claude Nos, MD, Department of Surgery, Institut Curie, 26 Rue d’Ulm, 75005 Paris, France; Fax: 33-14-432-4006; E-mail: claude.nos{at}curie.net
ABSTRACT
Background: One of the problems of sentinel lymph node (SLN) biopsy is the risk of false negatives. At the Institut Curie, to reduce the false-negative rate, we have developed a histological quality control of the SLN performed by blue dye alone, which consists of verification of the SLN blue stain by the pathologist.
Methods: A total of 324 patients underwent an SLN biopsy procedure with patent blue dye only followed by an immediate axillary dissection. Initially, SLNs were checked to ensure that they were blue by macroscopic examination. Finally, a search for immunohistochemistry micrometastasis was performed.
Results: In 277 (85.5%) of 324 patients, an SLN was identified by the surgeon. After standard examination, the false-negative rate was 11.1% (10 of 90). After macroscopic checking of the 197 negative SLNs, 167 of the 197 were confirmed blue, and there were 5 false negatives, which brought the false-negative rate down to 5.6% (5 of 90). Sixty SLNs out of the 167 confirmed blue SLNs were then proved to be immunohistochemically micrometastatic, and there were 3 false negatives, giving a final false-negative rate of 2.2% (2 of 90; P = .002).
Conclusions: In this series, the procedure of pathologic analysis of the SLN has resulted in a significant reduction of the false-negative rate.
Key Words: Sentinel lymph node biopsy • Breast carcinoma • Axillary dissection • Pathologic color quality assessment
According to most experts and medical institutions, the best technique for sentinel lymph node (SLN) detection in breast cancer is the combination of both radiocolloid and blue dye injection.1–7 However, because of the lack of definitive consensus in this matter, a number of surgeons continue to use the blue dye method alone because it is inexpensive and straightforward to use.5,8–12 At the Institut Curie from 1996 to 2000, which was our learning-curve period, we used blue dye injection alone for SLN biopsy. The two problems encountered initially, as in the first series of Giuliano et al.,9 who used blue dye alone, were a poor SLN detection rate of approximately 85% and a relatively high false-negative rate of approximately 10%.
First, a poor detection rate for SLN detection with blue dye alone can be improved by implementing a learning-curve period,3,10,13 by using a subareolar site of injection,12 by adding isotopes to the blue dye,1–7 and, in our experience, by selecting patients according to their morphology (nonobese patients and small- and medium-sized breasts).14 Second, the relatively high false-negative rate of 10% is similar to that found in other studies,3,9 including multicenter trials.5,15–17 By introducing a system of quality control associated with research for micrometastasis, we have noticed that it is possible to reduce the false-negative rate from 10% to <3%. We describe our study hereafter.
PATIENTS AND METHODS
Between December 1997 and August 2000, a study was performed on SLN biopsy with the blue dye–only technique at the Institut Curie. During this period, 324 patients underwent an SLN biopsy immediately followed by a level I and II axillary dissection. The operations were performed by seven different surgeons. The patients’ mean age was 58 years (range, 28–87 years). Thirty-two patients (9.8%) had clinical stage T0 tumors, 167 (51.6%) had T1 tumors, and 125 (38.6%) had T2 tumors. Most patients underwent a lumpectomy (298 patients; 92%), and a small proportion required a mastectomy (26 patients; 8%). The mean size of invasive tumor, measured histologically, was 16 mm (range, 3–50 mm). The invasive tumors were subclassified as ductal (n = 253; 78%), lobular (n = 53; 16%), or other (n = 18; 6%). Tumor grade was I (n = 144; 45%), II (n = 132; 41%), III (n = 37; 12%), or undetermined (n = 6; 2%). Most tumors were estrogen receptor positive (86%) and progesterone receptor positive (69%). Lymphovascular invasion was noted in 25.6% of cases. The mean number of SLNs was 2.1 (range, 1–7). The mean number of nodes identified in the completion dissections was 11.2 (range, 4–29).
Method for Identification of the SLN
The method for identifying the SLN consisted of an injection of 2 mL of patent blue dye (Laboratoire Guerbet, Villepinte, France). This was injected into the immediate area surrounding the tumor (72%) or, if the tumor had previously been removed, into the tumor bed (28%). The details of the technique have been previously published.14 The blue dye was injected into the breast parenchyma next to the tumor on the axillary side of the tumor. The tumor, the lymphatic, and the SLN were colored blue, thus demonstrating the lymphatic continuum. The intradermal and subareolar routes for dye injection were not used.
Histological Method
Standard Examination
The SLN and completion axillary lymph node dissection were sent to the pathologists. No frozen section was performed on the SLN. The SLN and all axillary lymph nodes were bisected and fixed in AFA (5% acetic acid, 75% absolute ethyl alcohol, 18% distilled water, and 2% formalin), and the two half-nodes were embedded in paraffin separately. One or two levels of hematoxylin and eosin (H&E)-stained slide from each node were examined for the presence of metastatic disease.
Quality Control
The negative SLN underwent a pathologic color quality assessment (PCQA) for the presence of dye. This was performed by the pathologist, who checked the paraffin blocks macroscopically and assessed whether at least one SLN identified by the surgeon was blue. If one SLN was blue (in the opinion of the pathologist), the PCQA was "confirmed blue." If none of the SLNs was blue, PCQA was "not blue" (Fig. 1). This PCQA method was performed only on SLNs that were negative, because correctly identified positive (metastatic) nodes do not affect the false-negative rate, by definition. The PCQA results were then expressed as a percentage. 
|
| (1) |
|
The false-negative rate was calculated by the following formula:4.
|
| (2) |
RESULTS
One or more SLNs were retrieved in 277 of 324 patients. This gave a surgical identification rate of 85.5%. Histological examination with standard H&E staining revealed 80 SLNs to be positive (metastatic), and 197 SLNs showed no metastatic disease. There were 10 false-negative SLNs, there being at least 1 metastatic node identified in the completion axillary dissection specimen. This gave a false-negative rate of 11.1% (10/80 + 10). Details of the false-negative nodes are listed in Table 1.
|
The next step was to search for micrometastasis by using IHC in the 167 negative on H&E and confirmed blue SLNs. Sixty SLNs of the 167 showed IHC micrometastasis. Three of the five false-negative results were reclassified by examination for IHC micrometastasis. After quality control and the search for IHC micrometastasis, the final false-negative rate was significantly (P = .002) reduced to 2.2% (2 of 90; standard H&E vs. standard H&E plus PCQA plus search for IHC micrometastasis).
DISCUSSION
We have set up a procedure for quality control of SLN biopsy performed with patent blue dye only. This method halved the false-negative rate in this series. The procedure consists of macroscopic examination of the negative SLNs (ensuring that they are blue) by someone independent from the operating team. We decided to instigate this procedure for quality control after the observation that on macroscopic review, some SLNs were not blue. This process is incorporated into the histopathologist’s examination of the SLN. Confirmation of the blue color might be performed during surgery with a frozen-section biopsy or later by examination of the histopathological paraffin blocks, as was the case in this study. The AFA fixative used to prepare the lymph nodes at the Institut Curie is colorless and does not interfere with the interpretation as to whether an imbedded node is blue or not.18 Furthermore, there is no danger that a node might be erroneously colored blue.18 AFA fixative is a simple preparation that was formulated as described previously and is commonly used in most cancer centers. We did not use other fixatives or test isosulfan blue, which is not commercialized in France. Currently, this control is used during surgery for blue dye–only SLN detection. The pathologist checks the blue color of the SLN sent by the surgeon during its macroscopic examination, just before the frozen section.
Identification of a blue SLN is sometimes challenging for the surgeon. In principle, when performing an SLN biopsy by the blue dye technique, the surgeon has to find a blue node in the axilla. In practice, the situation is more complicated. From an ideal perspective, for the functioning of the technique, the lymphatic drainage for all tumors would be to the axilla. In reality, it is not always possible to obtain a blue node. The mean rate for SLN identification is 76%, with a maximum rate of 98% reported by Giuliano et al.,9,10,19 who have the most experience with this technique. Certain circumstances can interfere with an axillary sentinel node being blue. When a node is massively invaded or replaced by tumor, it creates an obstacle to the progression of the blue, and the node may only be only weakly colored or not blue at all. This is usually readily identified by the surgeon as a blue lymphatic duct that comes to an end abutting an enlarged hard node. This node should have been excised with its blue lymphatic duct and subjected to a frozen-section histopathologic analysis. Such nodes are easily assessed in this manner, enabling confirmation that they are malignant, and a block dissection can then be performed under the same anesthetic. This problem has already been described for SLN biopsy performed with radioactive isotopes.13,20 In other cases in which no blue SLN is found, the explanation is more difficult. Sometimes there is a blue lymphatic tract in the axilla, but the lymph node to which it drains remains not blue. It is often possible to make this node blue by waiting longer and massaging the breast. A lymph node can undergo involution with fatty replacement and may be quite large. A lymphatic may pass through such a node, which is functionless as a filter, and in SLN biopsy, this node would not become colored. In this situation, a higher node is the true SLN, and if this is not sought, it may result in errors. It is important that the surgeon carefully clip the afferent and efferent lymphatics to the SLN to prevent the leakage of blue from the node, which might then be judged not blue at quality control. Another cause for technical error when there is no scrub nurse present is the blue coloring of surgical instruments that have been used for lumpectomy and then used immediately for axillary surgery, which may color the nodes. Finally, above all, surgeons early in their learning curve or those who are too confident may think they have identified a node that is blue when it is not in fact an SLN. The desire to avoid an axillary dissection for the patient, especially when the tumor is small and the axillary nodes are not obviously involved, results in the procedure degenerating into a random node sampling. The principal aims of the quality control are to enable such errors to be corrected and to ensure that the surgical technique is performed with the degree of rigor required to make this technique of sentinel node biopsy reliable and reproducible.
In our series, surgeon E, for example, had a very good identification rate. The only two SLNs that failed quality control (i.e., no blue on assessment by the pathologist) were false negatives. These were cases 23 and 34 in the surgeon’s series (Table 1). At this level, the notion of learning curve per the recommendations of the American Society of Breast Surgeons is only partially applicable to our series.13 Furthermore, 70% of false negatives occurred after the 30th case for any given surgeon.
When we used our method of quality control of the SLN, we observed that it was responsible for a decrease in the identification rate of the SLN by approximately 10% (from 85.5% to 76.2%) when the blue dye–only method was used to identify the SLN. This rate would be low for a technique, were it applied to all tumors.
There are three strategies to improve the identification rate: first, by restricting the indications relating to patient selection; second, use an alternative injection technique, e.g., subareolar or intradermal, for the dye; and, finally, combine the blue dye technique with a second, i.e., isotopic, method of SLN identification. First, regarding patient selection, in our experience, the patent blue dye–only technique works better on nonobese patients (body mass index <30 kg/m2) with small to medium-sized breasts (A–C bra cup size) and when the injection of blue dye is performed with the tumor still in place. Under these conditions, the surgical identification rate was 93.5% and the PCQA rate was 87% in our series.14 Blue dye is also associated with a better SLN detection rate when the tumor is located in the outer part of the breast.8,11,14 At the Institut Curie, we now restrict the blue dye–only technique to patients who satisfy these criteria (i.e., body mass index <30 kg/m2, breast size of C cup or smaller, and lateral tumors that are still in situ).
A second possibility is an alternative injection method. Some authors have reported high success rates with the cutaneous lymphatic drainage pathways by dermal injection.21,22 Others have shown high rates of identification with subareolar injection, although there are no large published series for this technique.12,23
Finally, the solution adopted by most of the teams is the use of a radioactive isotope in combination with the blue dye for all patients to improve the detection rate.1–7,13,16,17,21 Clearly, this combined technique also has the potential to reduce the false-negative rate.1–3,5,6 At the Institut Curie, we use the combined technique for patients who do not satisfy our criteria for the blue-only procedure. However, the combined technique does not show a perfect correlation between SLNs identified by either radioactive or blue dye. In the study by Cox et al.2 on 466 patients, the identification rate of SLNs was 94.4%. Only 27.6% of the SLNs were both blue and radioactive, 40.2% were only radioactive, and 32.2% were only blue. In the study by the Memorial Sloan-Kettering Cancer Center on 500 cases, the identification rate was 93%; 77% of the SLNs were blue and radioactive, 14% were radioactive only, and 9% were blue only.3 Further refinement is required before the two techniques achieve perfect concordance with each other. Ideally, the SLNs should be both radioactive and blue. Because of the discordance in the results and from a quality control perspective, verification of the blue dye–only SLN should be checked by the pathologist. The verification of the color of the SLN could then be included in the histopathologic report so that the technique is thus not operator dependent.
The application of an assessment for IHC micrometastasis, a well-established technique,20,24–26 also enabled us to improve the sensitivity of the technique. In this series, three of five false-negative SLNs on H&E and PCQA blue were found to be positive for IHC micrometastasis. None of the five false-negative SLNs on H&E and PCQA "not blue" was IHC positive (Table 1). This IHC technique is expensive and time consuming. It is practical only when relatively few nodes are sent as SLNs (random node sampling should be avoided). At the Institut Curie, we have searched for occult metastasis by means of IHC on many serial sections after research of metastases by one or two levels of H&E. This led to discovering many more occult events we have called thereafter IHC micrometastasis. We consider micrometastasis (H&E or IHC) in SLNs an indication for treating the rest of the axilla, but their prognostic implications in terms of either local recurrence or distant metastasis continue to be the subject of debate.26–28 In our protocol, discovery of micrometastasis (H&E or IHC) is never in itself an indication for systemic treatment. Therefore, in this series we have not tried to differentiate between H&E micrometastasis and IHC micrometastasis. Where the mean tumor size was 16 mm, there were 29% (80 of 277) positive SLNs after standard histological examination with H&E staining. If we take into account the results of the research for micrometastases by IHC, 60 (21%) more patients out of 277 would have been considered as IHC positive and considered as candidates for axillary clearance.
CONCLUSION
Sentinel node biopsy is an alternative to axillary dissection for operable breast cancers without clinically involved axillary nodes. Surgeons who use the blue dye–only technique must ensure that the SLNs that they identify are indeed stained blue. However, the technique is operator dependent, and it is essential that quality control mechanisms are put in place to make the technique reliable and reproducible. Checking the paraffin blocks to ensure that the embedded nodes are blue is a simple method of ensuring quality. This procedure renders the technique of SLN biopsy by the blue dye–only method more reliable. In our series, this method, in combination with routine IHC, for initially tumor-negative SLNs enabled us to reduce the false-negative rate to 2.2%.
Acknowledgments
The authors thank M. C. Falcou for her technical assistance. The following surgeons took part in the study: Dr. Didier Bourgeois, Dr. Jean-Yves Charvolin, Dr. Krishna Clough, Dr. Jean-Noel Guglielmina, Dr. Jean-Pierre Hamelin, Dr. Vincent Jacquin de Margerie, and Dr. Claude Nos.
The acknowledgments are available online at www.annalssurgicaloncology.org.
Footnotes
Confirmation that the tumor-negative sentinel lymph nodes were definitely blue by macroscopic assessment could halve the false-negative rate in our series of 324 procedures.
Received for publication October 2, 2002. Accepted for publication February 21, 2003.
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  K. A. Kern Achieving the Lowest False-Negative Rate in Peritumoral Breast Lymphatic Mapping: The Oncologic Search for the Holy Grail Ann. Surg. Oncol., June 1, 2003; 10(5): 486 - 487. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
