Chagpar A, Martin III RC, Chao C, Wong SL, Edwards MJ, Tuttle T, McMasters KM, for the University of Louisville Breast Cancer Sentinel Lymph Node Study Group. Validation of Subareolar and Periareolar Injection Techniques for Breast Sentinel Lymph Node Biopsy. Arch Surg. 2004;139(6):614-620. doi:10.1001/archsurg.139.6.614
Subareolar or periareolar injection of radioactive technetium sulfur colloid is equivalent to other injection techniques for breast cancer sentinel lymph node (SLN) biopsy.
Design and Setting
Prospective, multicenter clinical trial.
A total of 3961 individuals with clinical stage I and II breast cancer.
All patients underwent attempted SLN biopsy followed by completion axillary dissection. Injection technique was determined by the preference of each participating surgeon. Most surgeons had little or no experience with SLN biopsy before participation in this study.
Main Outcome Measures
The SLN identification and false-negative rates.
An SLN biopsy was performed in 3961 patients using blue dye alone or radioactive colloid plus blue dye. Subareolar and periareolar radioactive colloid injection techniques were associated with SLN identification rates of 99.3% and 95.6%, respectively, with false-negative rates of 8.3% and 8.9%, respectively. The identification rates were significantly higher for these 2 techniques than for peritumoral injection of radioactive colloid (91.1%) or the use of blue dye alone (88.5%) (P<.001). The false-negative rates were similar for all techniques.
Although many medical centers have adopted subareolar and periareolar radioactive colloid injections because of their simplicity and convenience, a paucity of data from a few single-institutional studies has existed to substantiate the false-negative rates associated with these techniques. The results of this multicenter study establish the validity of subareolar and periareolar radioactive colloid injections and support the hypothesis that the lymphatic drainage of the entire breast is to the same few SLNs.
During the past decade, sentinel lymph node (SLN) biopsy has emerged as a minimally invasive alternative to axillary lymph node dissection. Despite widespread use of SLN biopsy, many questions remain regarding the technical aspects of this procedure.
One of these controversies surrounds the optimal location for injection of radioactive colloid. Although initial studies1- 3 advocated peritumoral injection of radioactive tracer, and this has remained the most widely used method, it has become evident that there are several disadvantages to this approach. Peritumoral injection requires localization of the tumor before injection of radiotracer. This then mandates injection under radiographic guidance. In early studies, the false-negative rate associated with multifocal and multicentric tumors was high, causing some researchers to argue that such patients are not candidates for this technique.4 In addition, for tumors in the upper outer quadrant, diffusion of the radiotracer causes substantial radioactive background, making identification of the SLN more difficult.5 It would be simpler if the injection could be performed in the same manner for all patients regardless of tumor location or palpability. Other studies6,7 have shown that dermal or subdermal injection of radioactive colloid results in more reliable SLN identification than peritumoral injection.
It has been hypothesized that the lymphatic drainage of the entire breast is to the same few SLNs. This concept stems from the knowledge that the lymphatic drainage of the breast embryologically originates from the subareolar plexus.8 As such, it has been suggested that a central injection technique may be simpler to perform than and just as accurate as the peritumoral technique. Several single-institutional studies have found that a subareolar or periareolar radioactive colloid injection correlates with peritumoral injection of blue dye; in other words, both tracer agents, injected in different locations, drain to the same SLNs most of the time. Fewer studies of the subareolar and periareolar injection techniques have sought to establish the false-negative rate, which requires concomitant completion axillary dissection, and all have come from single institutions. Table 1 summarizes these studies.9- 21 The subareolar and periareolar injection techniques have become popular despite the relative paucity of data to substantiate a low false-negative rate. We sought to determine, in a large, multi-institutional study, whether subareolar and periareolar injection techniques were equivalent to peritumoral injection techniques in terms of the SLN identification and false-negative rates.
The University of Louisville Breast Cancer Sentinel Lymph Node Study is a prospective multicenter study evaluating SLN biopsy in women with clinical stage I and II invasive breast cancer. More than 300 surgeons, most of whom had little or no experience with SLN biopsy before the study, participated. All patients underwent SLN biopsy followed by completion axillary lymph node dissection. The technique of SLN biopsy, including the location of injection, was left to the discretion of the individual surgeon. Information regarding patient demographics, injection technique, number of SLNs identified, number of remaining lymph nodes harvested, and pathologic results were recorded. Patient confidentiality was maintained, and the study was approved by each medical center's institutional review board. Statistical analysis of the SLN identification and false-negative rates of subareolar and periareolar injection vs peritumoral injection techniques was performed using χ2 analysis and a statistical software program (SPSS version 10.1; SPSS Inc, Chicago, Ill).
Between November 1, 1996, and April 30, 2003, 3961 patients were enrolled in the University of Louisville Breast Cancer Sentinel Lymph Node Study. The median ± SD patient age was 60 ± 1.7 years (range, 27-100 years). The median ± SD tumor size was 1.5 ± 1.2 cm (range, <0.1-11.0 cm). A median of 2 SLNs were removed per patient.
Of the 3961 patients, 348 had injection with blue dye alone. Of the 1762 patients who had injection of technetium sulfur colloid, 1431 had peritumoral injection, 148 had subareolar injection, and 183 had periareolar injection. There were 1078 and 773 patients who were classified as having dermal and subdermal injections, respectively, in the skin overlying the tumor. Table 2 gives the clinicopathologic features of each of the 6 groups of patients (blue dye only, peritumoral, subareolar, periareolar, dermal, and subdermal) based on injection technique. All 6 groups were similar in patient age, tumor size, and tumor palpability. Compared with peritumoral injection, the periareolar technique was more often used in lesions in the upper outer quadrant (P = .001). Periareolar injection was also used more commonly than peritumoral injection in patients undergoing mastectomy (P = .02).
The SLN identification rate was defined as the number of patients seen in whom an SLN was found divided by the number of patients in whom identification of an SLN was attempted. An SLN was defined as a lymph node that was either a blue node, the most radioactive node, or any node that contained 10% or greater of the radioactive counts of the hottest node.22 The false-negative rate was defined as the number of patients in whom all of the SLNs harvested were found to be negative and a non-SLN was found to harbor metastatic disease divided by the total number of patients with positive axillary lymph nodes.23 The SLN identification and false-negative rates for each of the injection techniques are given in Table 3. The SLN identification rates were statistically significantly different among the 6 groups, with the highest and lowest identification rates being found with the subareolar technique (99.3%) and with the use of blue dye alone (88.5%), respectively. In pairwise comparisons, a significantly different identification rate was noted between the peritumoral technique and the subareolar and periareolar techniques (P<.001 and P = .04, respectively), but no difference was detected between the peritumoral technetium injection technique and the use of blue dye alone (P = .15). In addition, the subareolar and periareolar techniques were associated with significantly better identification rates than the use of blue dye alone (P<.001 and P = .007, respectively). Subareolar injection was superior to periareolar injection in terms of identification rate (P = .004), and dermal injection was better than subdermal injection (P = .03). There were no statistically significant differences in the false-negative rates among the 6 groups.
In a multivariate analysis to determine the factors that affect the SLN identification rate using patient age, tumor size, histologic findings, location, palpability, and injection type, only type of injection, patient age, and tumor palpability remained statistically significant in the model (Table 4).
As SLN biopsy has become increasingly popular in the management of breast cancer, questions have been raised as to the optimal injection technique. Various permutations of the SLN biopsy technique have been reported using blue dye, radioactive colloid, or both. Initial studies1- 3 of SLN biopsy in breast cancer used a peritumoral injection technique based on the concept that the lymphatics draining the tumor would then be the conduits through which the blue dye or radiotracer would travel to accurately identify the first draining lymph node.
Such an approach, however, has been criticized for a variety of reasons. First, the peritumoral injection technique mandates precise localization of the primary tumor. With the increasing use of screening mammography and heightened public awareness, the incidence of nonpalpable tumors has steadily increased. In light of this fact, peritumoral injection requires radiographic guidance for injection, which is sometimes cumbersome to coordinate. In addition, peritumoral injection may be problematic in the setting of multiple tumors in the breast.4 Finally, because most tumors are located in the upper outer quadrant of the breast, injection of radioactive colloid in this region often causes diffusion of radioactivity, or "shine through," which may make identification of SLNs more difficult.5
As a result of these concerns, some researchers have argued that a central injection (with a subareolar or periareolar technique) may be beneficial. Such an approach would simplify injection for nonpalpable or multicentric tumors and would minimize the shine-through effect, with injection further away from the axilla. The rationale for such an approach is based on the embryologic origin of the breast and on early studies8,24,25 that mapped the lymphatic pathways of the breast. As the lymphatics of the breast develop embryologically, their connection to the nipple-areola complex is maintained. The flow of lymph is centripetal from the central subareolar plexus out to the regional lymphatics. This implies that the breast is drained uniformly to a few SLNs and that a subareolar or periareolar injection technique should be accurate in identifying the sentinel node.
Indeed, several studies have found this to be the case and, in fact, have demonstrated that identification rates may actually be improved using this simpler technique (Table 1). The SLN identification rates reported with the subareolar and periareolar injection techniques have ranged from 94% to 100%.9- 21 In a study10 of 332 patients who underwent peritumoral injection of radioactive colloid, with 83 having peritumoral injection of blue dye and 249 having subareolar injection of blue dye, the identification rates between the 2 techniques were similar—95.2% vs 96.8%, respectively. Given the comparable identification rates, the subareolar and periareolar injection techniques have gained popularity as they are reportedly simple to learn and reliable in terms of SLN identification. Although few studies have incorporated completion axillary lymph node dissection to evaluate the false-negative rate associated with this technique, those that did found a false-negative rate of 0% to 9.1%.12,15,16,18,19 Several studies evaluated the use of blue dye and radioactive colloid in combination, with each being injected into a different location. They then evaluated the number of nodes that were found to be sentinel by both techniques, that is, both blue and hot. Concordance rates between the peritumoral and central techniques ranged from 75.0%, in a small study18 of 8 patients with multicentric tumors, to 95.4%.9- 14,17,20,21
These studies, although adding to the evidence supporting the use of central injection techniques in SLN biopsy, are all relatively small, single-institutional studies. Only 150 patients have previously been studied with concomitant axillary dissection to establish the false-negative rates for subareolar and periareolar injection. The results of the present study validate this initial work in a multicenter prospective trial. This study demonstrates not only that central injection techniques are feasible but that they are associated with a higher SLN identification rate than peritumoral injection or use of blue dye alone. The fact that excellent identification rates were achieved in this study using subareolar and periareolar injection techniques despite the relative inexperience of the participating surgeons in performing SLN biopsy adds credence to the simplicity of this technique. In addition, one of the criticisms of previous data has been that completion axillary dissection was not performed in most studies and, therefore, that the reliability of this technique could not be assumed.26 In our study, completion axillary dissection was performed in all patients, and, therefore, a true estimation of the false-negative rate could be obtained. All techniques of SLN biopsy seem to be associated with a similar false-negative rate, implying that the lymphatic drainage of the breast identifies the same SLNs regardless of injection technique.
The issue of dermal vs subdermal injection has also been widely debated in the literature. It has been demonstrated that the skin overlying the breast parenchyma shares the same lymphatic drainage pathway as the breast tissue beneath it.27 This prompted interest in a dermal injection technique that may be simpler to perform than the intraparenchymal technique. Linehan et al28 showed that in 95% of cases, the same SLN was identified when peritumoral blue dye and dermal radioactive colloid were used. The results of the present study further validate the findings of previous works29,30 substantiating the value of dermal injection for the identification of axillary SLNs. Although dermal injection is easy for surgeons to learn and results in reproducible SLN identification, its one drawback is that it requires preoperative localization of nonpalpable tumors for which an excisional biopsy was not performed. Although this is not usually an issue for patients who are undergoing breast conservation, since localization of nonpalpable tumors is required to perform the lumpectomy, central injection techniques certainly may simplify the procedure for patients undergoing mastectomy or for patients with multicentric tumors.31,32
Sentinel lymph node biopsy has revolutionized the manner in which breast cancer is treated. No longer is it required to inflict undue morbidity on patients with node-negative disease. Although many controversies continue to exist regarding the use of this technique, central injection techniques have been demonstrated to accurately stage the axilla, with a high identification rate and a low false-negative rate. This central injection technique is simple, convenient, and widely applicable and may well emerge as the technique of choice for SLN biopsy in patients with breast cancer.
Accepted for publication January 28, 2004.
This paper was presented at the 111th Scientific Session of the Western Surgical Association; November 11, 2003; Tuscon, Ariz; and is published after peer review and revision. The discussions that follow this article are based on the originally submitted manuscript and not the revised manuscript.
Corresponding author: Anees Chagpar, MD, MSc, Division of Surgical Oncology, University of Louisville, 315 E Broadway, Suite 312, Louisville, KY 40202 (e-mail: firstname.lastname@example.org).
Nora M. Hansen, MD, Santa Monica, Calif: This multicenter prospective study evaluated the role of sentinel lymph node biopsy in women with clinical stage I and II breast cancer. Over 300 surgeons participated in the study, most of whom had little or no experience with sentinel node biopsy prior to the study. The various technical aspects of sentinel lymph node biopsy have been reported in many previous publications, but some of the techniques, especially the periareolar and subareolar injection techniques, had not been validated adequately with concomitant axillary node dissections. This study has clearly demonstrated the feasibility of several injection techniques, including periareolar, subareolar, dermal, and subdermal, in accurately identifying sentinel node. They have done this by performing ALND [axillary lymph node dissection] in all patients, so that the true identification rate and false-negative rates are identified.
I do have several questions for the authors. (1) Although it may be obvious, can you explain the difference between periareolar and subareolar injection techniques, and why do you think there is a difference between these identification rates between the 2 techniques since they are basically utilizing the same lymphatic pathway? (2) The initial studies describe the peritumoral approach; however, that injection technique has been criticized because some feel, as stated by your authors, that it mandates the precise location of the primary tumor. However, as we have become more familiar with the technique, it has become clear that we are mapping the lymphatics of the breast, and the lymphatic drainage of the entire breast is likely to drain to the same few sentinel nodes. So if we believe this, do you think it really makes a difference where we inject the blue dye or radioactive tracer? Or do you think it is probably more related to the comfort level and expertise of the individual surgeon rather than the injection site? I do agree with the authors that the shine-through phenomenon of the radioactive tracer for upper outer quadrant lesions can be diminished by a periareolar or subareolar approach. This problem is not seen with blue dye alone.
(3) My last question relates to the timeline for each participating surgeon. Do you have any data regarding the "learning curve" of each participating surgeon? Did they perform only one type of injection technique for all of the patients that entered into the study, or did they use multiple techniques? If they did use multiple techniques, did they start with the peritumoral injection approach and then as they became more familiar with identifying the sentinel node, did they switch to the periareolar, subareolar, dermal, or subdermal approaches? If this was the case, the improvement in the sentinel node identification rates might be falsely elevated and related to experience with the technique rather than the type of injection technique.
Many institutions have adapted the techniques already, but few have adequately substantiated the identification of false-negative rates. The authors of this study have done that.
Dr McMasters: Dr Hansen first asked if I could explain the difference between subareolar and periareolar injection. Subareolar injection, of course, is injection directly underneath the nipple areolar complex. Some surgeons use a periareolar injection technique in which they inject either in the dermis or subdermis around the edge of the areola—that is the difference.
The next question was whether or not we think there is a difference if we inject the blue dye or the radioactive tracer in different locations. I agree with Dr Hansen. I believe that the entire breast likely drains to the same 1 or 2 sentinel nodes regardless of where we perform the injection. If that is the case, then the next question is: what is the easiest technique for surgeons to learn how to perform? We know from our previous studies that dermal or subdermal injection of the radioactive colloid makes it much easier for surgeons to find the sentinel nodes reliably, even early in their experience. It appears from these data that subareolar or periareolar injection does the same thing. These central injection techniques make it easier for surgeons to find the sentinel nodes because, as with dermal or subdermal injection, the radioactivity in the sentinel nodes is much greater. The nodes are hotter, easier to find with the gamma probe before you ever make an incision, and the shine through, or the background radioactivity that prevents adequate use of the gamma probe, is eliminated because you can retract the breast away from the axilla and get separation between the breast injection site and the site of the sentinel nodes.
The next question was whether or not we have data on the learning curve of each individual surgeon. Did some surgeons switch over to doing subareolar or periareolar injections after they had already used other injection techniques, and that is why the data looked so good for the central injection techniques? We previously reported the learning curve data for subdermal and dermal injection. We don't have those particular data for this study, but it's a mixture, I'm certain, of surgeons who started using the same central injection technique and never changed, and some who switched to different techniques during their experience. Overall, however, the excellent identification rates and the low false-negative rates really point to the fact that these central injection techniques are easy for inexperienced surgeons to learn, and perhaps preferable to the peritumoral injection technique, with which it is much more difficult to achieve a near 100% identification rate. Blue dye alone does not suffer from the problems of the background radioactivity as Dr Hansen alluded to, but we found in our study that blue dye alone is difficult for most surgeons to master until they have probably done many hundreds of cases. Many of us who have performed sentinel node biopsy for many years now would never consider performing this procedure without the radioactive tracer and gamma probe.
Howard Silberman, MD, Pacific Palisades, Calif: The most disturbing feature of the sentinel node concept is that 1 in 12 patients with nodal metastases will have a negative sentinel node. In this 8% to 9% of patients in whom a false-negative was found, was any subset analysis done that indicates what features might be present to identify patients in whom the sentinel node may not be an accurate predictor (eg, lobular vs ductal cell type, tumor size, tumor location)?
Dr McMasters: In fact, the false-negative rate is not calculated with the denominator of patients being the total number in which it was attempted—the denominator is the number of patients who had positive lymph nodes. So, therefore, an 8% false-negative rate among a group of patients that has about a 30% lymph node positivity rate is much lower than that; overall, 2.4 patients out of 100 in this study would have had an incorrect nodal staging procedure. We have also shown previously that surgeon experience is the single most important factor that will predict a lower false-negative rate. After 20 cases, the false-negative rate drops off precipitously such that it is much less than 5%. So realize that these numbers, while they are very consistent with an 8% false-negative rate overall, are for mainly private practice surgeons in a variety of practice environments across the United States, very early in their sentinel lymph node learning curve. After they have performed more than 20 cases, the false-negative rates dramatically improve.
Gary B. Talpos, MD, Detroit, Mich: Some authors have pointed out that the method of diagnosis can impact the accuracy of results. For instance, was the diagnosis of malignancy in the breast achieved with a needle biopsy, which would not disrupt lymphatic drainage, or with previous open biopsy, which might result in altered lymphatic drainage? Have you analyzed this in your paper?
Dr McMasters: We have previously performed that analysis and published those results. There is really no difference between patients who have undergone a needle biopsy or an excisional biopsy in terms of the identification or false-negative rates, although, intuitively, we would think the disruption of the breast lymphatic pathway by an excisional biopsy might lead to less optimal results. We and others have not been able to demonstrate that with hard data.
Anton Joel Bilchick, MD, Santa Monica: Data from the Netherlands Cancer Institute suggest that evaluation of the internal mammary nodes is very important in staging as well. How effective is subareolar or periareolar injection for finding internal mammary nodes? Secondly, how are your surgeons able to get permission to use radiopharmaceuticals intraoperatively?
Dr McMasters: We don't have good data about the identification of internal mammary nodes with these techniques because very few surgeons in our study actually attempted to identify those. In other words, most surgeons, I believe, in the United States and in our study, use the sentinel node biopsy technique as an alternative to axillary lymph node dissection, as a means of staging the axillary nodes. While the data from the Netherlands is intriguing, other data from the United States suggest that at best 2 to 3 patients in 1000 could potentially benefit from the internal mammary staging information. In other words, the incidence of having a negative axillary node and a concomitant positive internal mammary node is very low. I did not mean to imply that the surgeons necessarily injected these agents in the operating room; most of these injections were performed in radiology.
Mary C. McCarthy, MD, Dayton, Ohio: I have 2 questions. First, how was the histology of your positive nodes defined? Did you use simple H & E [hematoxylin-eosin] or immunostaining to identify positivity? And secondly, once you have done your 20 cases and you know where these 1 or 2 nodes are, has anyone given consideration to just doing the sentinel node biopsy without any injections?
Dr McMasters: The study underwent some evolution, as did the whole field of sentinel node biopsy, during this time period. At first, we were all strong advocates of serial sections of the sentinel nodes with immunohistochemical stains for cytokeratin. However, during this time period, it became apparent that immunostains may be misleading and may be finding small deposits of cells in the sentinel node that are not clinically significant. The significance of this is being tested in 2 large prospective randomized trials. So, somewhere in the middle of this study, we discouraged people actually from using immunostains, and many fewer centers used those routinely. But, all of the sentinel nodes underwent serial sections at a minimum of 2-mm intervals.
I don't think, based upon my experience and that of others with whom I have discussed this, that you can reliably predict anatomically the position of the sentinel nodes, although, very frequently, the first sentinel node is low in level 1 of the axilla at the bottom of the hairline, as Dr Hansen, Dr Giuliano, and others have taught us. But often there are other sentinel nodes in different locations that are not anatomically constant, so I don't think that we could reliably do this without the tracer agents.
Max Schmidt, MD, Indianapolis, Ind: You have shown a statistically significant difference between the subareolar or periareolar injection and the peritumoral injection, and you mentioned in a previous answer to a question that this might be due to how easy it is for the surgeons who are mostly inexperienced with the technique to employ. I would ask you to consider: is subareolar or periareolar injection more anatomically efficient in terms of how it drains to the sentinel lymph node, that is, the lymphatic drainage around the tumor may be altered due to therapy, previous biopsy, etc? Furthermore, when you inject around the tumor, you may be going into the tumor, around the tumor, or around a previous biopsy site.
Dr McMasters: I think it is the same phenomenon. Surgeons can identify the sentinel nodes more easily early in their experience of performing this procedure because the sentinel nodes are easier to find because they are more radioactive. The problem with peritumoral injection as we have shown previously is that the sentinel nodes are 5 times less radioactive than if you used dermal or subdermal injection. The same thing is true with periareolar or subareolar injection. The subareolar plexus of lymphatics appears to be very rich and capable of taking up more of the tracer agent that is delivered to the sentinel nodes. So when you use the gamma probe, not only is the shine-through effect diminished because you now have the ability to retract the breast away from the axilla (and because half of the tumors are in the upper outer quadrant, it is a common problem), but now the sentinel nodes are 5 times more radioactive and easier to identify with a gamma probe even before you ever make an incision. This is what makes it easy. Once you can take the gamma probe and identify the sentinel node before you make an incision, you know you are going to find that first sentinel node.
James E. Goodnight, Jr, MD, Sacramento, Calif: Given the possibility that this is improved lymphatic mapping for the breast, any implications of a positive axillary node? Does this technique alter the likelihood that the positive axillary node will predict or not predict other axillary nodes being positive?
Dr McMasters: Because the false-negative rate appears to be similar to that of other techniques, we really don't have any evidence that this will help us to predict the status of nonsentinel nodes any better than any other technique.