Figure 1: Concept of sentinel lymph node biopsy. Dye or radiocolloid is injected to identify first draining lymph node from tumor.

Table 1: Techniques of SLNB

Study	N	Technique	SLN ID (%)	Accuracy (%)	FN rate (%)
Giuliano et al (1996)	174	Dye	65	96	12
Giuliano et al (1997)	107	Dye	66	100	0
Guenther et al (1997)	145	Dye	71	88	12
Veronesi et al (1997)	163	Radioactive colloid	98	98	5
Feldman et al (1999)	75	Radioactive colloid	93	94	19
Veronesi et al (1999)	376	Radioactive colloid	99	96	7
Moffat et al (1999)	70	Radioactive colloid	93	97	10
Borgstein et al (1998)	130	Radioactive colloid	94	99	2
Krag et al (1998)	443	Radioactive colloid	93	97	11
Viale et al (1999)	155	Radioactive colloid	100	97	7
Schlag and Bembenek (2000)	146	Radioactive colloid	81	93	8
Krag et al (2001)	145	Radioactive colloid	98	98	4
Quan et al (2002)	152	Radioactive colloid	93	100	0
Albertini et al (1997)	62	Both	92	100	0
O’Hea et al (1998)	60	Both	93	95	15
Molland et al (2000)	103	Both	85	98	5
Nwariaku et al (1998)	119	Both	81	99	4
Bass et al (1999)	186	Both	93	99	2
Haigh et al (2000)	283	Both	81	99	3
Nano et al (2002)	328	Both	87	94	8
Shivers et al (2002)	426	Both	86	99	4
Tafra et al (2001)	535	Both	87	96	13
Bergkvist et al (2001)	498	Both	90	n/a	11
Doting et al (2000)	136	Both	93	98	5
McMasters et al (2001)	2206	Both	93	97	8

Figure 2: Blue lymphatic channel

Figure 3: Blue lymph node

2. Radioactive colloid

Krag et al (1993) first described the use of radioactive colloid in lymphatic mapping in breast cancer. The location of greatest radioactivity was found transcutaneously using a hand-held gamma probe. This guided the surgeon to locate the incision in this area. The nodes with the greatest radioactivity were removed as SLNs. Although this technique is thought to be easier to perform than the blue dye approach(Noguchi, 2001) it still has a substantial learning curve. In their initial series of 22 patients, Krag et al (1993) found the SLN successfully in 82% of cases Their subsequent experience with 443 patients improved their success rate to 93% (Krag et al, 1998).

There has been some debate over which radiopharmaceutical is optimal for sentinel lymph node mapping. Although ⁹⁹m-Tc sulfur colloid is the only agent approved by the United States Food and Drug Administration for lymphoscintigraphy (Noguchi, 2001) a number of other materials have been used in Europe and elsewhere. The various agents vary in particle size and opinions regarding the “ideal” size differ. The majority of studies conclude that particles approximately 200 nm work well for lymphatic mapping (Galimberti et al, 2000; Mariani et al, 2001) as particles smaller than 50 nm have very rapid transit and are not retained in the initial draining lymph nodes and those larger than 500 nm take a prolonged period of time to accumulate in the SLN (Galimberti et al, 2000). Some investigators have studied the use of filters which reduce the large particle size of ⁹⁹m-Tc sulfur colloid to particles which are 50-100 nm in diameter, resulting in more rapid transit (Hung et al, 1995). However, a non-randomized comparison between unfiltered and filtered ⁹⁹m-Tc sulfur colloid found a better SLN identification rate in patients injected with the unfiltered compound (Linehan et al, 1999).Similarly, Paganelli et al (1998) found that the SLN could be more easily identified using radiolabeled colloidal albumin (200-1000 nm) rather than small-particle radiotracer (<80 nm), and felt that the radiolabeled albumin technique limited the removal of second echelon, or non-sentinel nodes. The University of Louisville Breast Cancer Sentinel Lymph Node study, however, found no difference in the SLN identification rate, false negative rate, or number of SLN removed whether filtered or unfiltered ⁹⁹m-Tc sulfur colloid was used (unpublished data).

3. Both blue dye and radioactive colloid

A number of studies demonstrated an advantage to using both blue dye and radioactive colloid to identify the SLN. Albertini et al (1996) injected filtered ⁹⁹m-Tc sulfur colloid into the breast parenchyma 2-4 hours prior to surgery. After induction of general anesthesia, 1% isosulfan blue was injected. A hand-held gamma probe was used to detect the location of the SLN prior to making the incision. The radioactivity was useful in directing them to the SLN when a blue stained lymphatic channel was difficult to identify. At the same time, the dye was useful in finding the SLN when there was considerable “shine through” from radioactive material injected around the primary tumor. Using this combined technique, they were able to achieve a 92% SLN identification (Albertini et al, 1996). Similarly, in a large multicenter trial, McMasters et al (2000) concluded that the false negative rate was significantly higher when lymphatic mapping used either blue dye or radioactive colloid than when both were used concurrently (11.8% vs. 5.8%, p<0.05).

B. Radioactive colloid: where to inject?

1. Peritumoral injection

The initial studies of breast SLNB used peritumoral injection of either blue dye, radioactive colloid, or both. This was based on the reasonable assumption that injection of the breast tissue immediately surrounding the tumor would result in reliable and accurate identification of the sentinel nodes to which the breast cancer would drain. Peritumoral blue dye injection remains the “gold standard” for SLN identification—an afferent blue-stained lymphatic channel entering a blue lymph node indicates an unequivocal direct lymphatic drainage pathway from the site of the primary tumor. However, it has become apparent that peritumoral injection of radioactive colloid is sometimes not always optimal for SLN identification. When peritumoral injection is used, typically 4 to 8 mL of radioactive colloid is injected. This causes significant diffusion of the radioactive tracer. This background radioactivity, or “shine through,” can obscure the often slightly radioactive sentinel nodes in the axilla. This is especially problematic for upper outer quadrant tumors, which represent half of all breast cancers. Failure to identify SLN in all cases using this technique, even when combined with blue dye injection, led some investigators to evaluate other techniques.

2. Dermal and subdermal injection

Dermal injection of radioactive colloid for SLNB in melanoma is highly reliable and results in nearly a 100% SLN identification rate. Borgstein and colleagues (1997) first demonstrated that the skin overlying the breast parenchyma shares the same lymphatic drainage pathway as the breast tissue beneath it. At the same time, Veronesi and colleagues (1997) found that subdermal injection of radioactive colloid was associated with a 98% identification rate and a low false negative rate. Linehan et al (1999) showed concordance of peritumoral blue dye and dermal radioactive colloid in 95% of cases. In the University of Louisville study, dermal radioactive colloid injection was associated with significant improvement in the SLN identification rate compared to peritumoral injection. Furthermore, dermal injection was associated with improved ability of surgeons learning to perform SLNB to accurately identify SLN (McMasters et al, 2001). While data from several centers substantiates the value of dermal injection for identification of axillary SLN (Linehan et al, 1999; Cody et al, 2001; Kersey et al, 2001, McMasters et al, 2001) it is not clear whether dermal injection will accurately identify internal mammary (IM) or extra-axillary nodal drainage, for those surgeons who are interested in searching for them. For example, Uren et al (1995, 1998) found significant variation between lymphatic drainage from intradermal and intraparenchymal injection sites.

3. Subareolar and periareolar injection

Peritumoral, dermal, or subdermal injection of radioactive colloid require localization of non-palpable tumors by ultrasound or mammogram in order to guide the injection. It would be simpler if all patients with breast cancer could undergo the same injection procedure, regardless of tumor location. In fact, significant evidence now suggests that the lymphatic drainage of the entire breast is to the same few sentinel nodes. This has prompted a number of investigators to evaluate subareolar or periareolar injection techniques (Klimberg et al, 1999; Beitsch et al, 2001; Donahue, 2001; Bauer et al, 2002; Tuttle et al, 2002).

The concept of subareolar injection builds upon the embryology of the breast, and early studies which mapped its lymphatic drainage. The breast develops from an ectodermal streak, which becomes the nipple-areola complex. As the lymphatics of the breast elongate, they maintain their connection to the subareolar complex (Gray, 1939). This concept was confirmed by Sappey in 1834. His experiments of mercury injection into breast lymphatic channels demonstrated a centripetal flow of lymph into the subareolar plexus and then onto regional lymphatics. Turner-Warwick (1959) illustrated, through studies using autoradiography of surgical specimens, that the lymphatic flow of the breast is from superficial to deep, arriving at the regional lymphatics through channels in the interlobular spaces and along lactiferous ducts (Turner-Warwick, 1959). As these channels would be in close proximity to each other in the subareolar space, proponents of subareolar injection argue that this technique should identify the sentinel lymph node regardless of tumor location in the breast.

A number of authors have demonstrated that subareolar injection is not only feasible, but may be as accurate as peritumoral injection (Klimberg et al, 1999; Beitsch et al, 2001; Donahue, 2001; Bauer et al, 2002; Tuttle et al, 2002). The subareolar technique has advantages in avoiding the need for image guided injection for non-palpable tumors, avoiding shine through in upper outer quadrant lesions, and may be useful in patients with multicentric disease. However, at present, there is insufficient evidence of a low false negative rate for subareolar or periareolar injection techniques to recommend their routine use.

C. Preoperative lymphoscintigraphy

Another area of debate is in the use of preoperative lymphoscintigraphy. Although this nuclear medicine technique of obtaining a road map of lymphatic drainage preoperatively is employed routinely for localization of SLN in melanoma (Haddad et al, 1999), its utility in the setting of breast cancer is less clear given the nearly universal drainage of breast cancers to the axilla. It has been demonstrated that preoperative lymphoscintigraphy, while increasing the cost and time required to treat patients, does not improve the ability to accurately identify axillary SLN (McMasters et al, 2000). On the other hand, extra-axillary SLNs can be found in 10% of patients with preoperative imaging (Dupont et al, 2001). The majority of these cases have dual drainage to both the axilla and the IM chains, however approximately 2% of patients have drainage the IM nodes alone (Cserni and Szekeres, 2001). These extra-axillary drainage sites may be missed using the hand-held gamma probe if preoperative imaging directing the surgeon to the alternate site was not done (Noguchi et al, 2000). Unless biopsy of the IM nodes is performed, however, the utility of preoperative knowledge of this drainage pattern is debatable (Rubio and Klimberge, 2001).

Several authors have demonstrated that SLNB of the IM nodes is feasible (Johnson et al, 2000; Sacchini et al, 2001). Proponents of this technique argue that metastatic disease found in IM SLNs may alter decisions regarding adjuvant radiation and systemic therapy. The current American Joint Committee on Cancer (AJCC) staging system for breast cancer has noted the prognostic implication of IM nodal involvement, factoring this in to the staging system (Singletary et al, 2002). However, the technique of IM SLNB has not gained wide acceptance. As evaluation of the axillary nodes has been the cornerstone of breast cancer staging for the past century, it appears to be a reasonable viewpoint that SLNB of the axillary nodes alone would be a less invasive alternative to ALND.

III. Patient Issues

SLNB has become commonplace in the surgical management of invasive breast cancers. There are several scenarios, however, in which the use of SLNB has been controversial.

A. Ductal carcinoma-in-situ

In theory, axillary nodal staging for ductal carcinoma-in-situ (DCIS) is not necessary. By definition, DCIS is not an invasive malignancy and does not metastasize to lymph nodes or elsewhere. It is known, however, that axillary metastases may be present in 1 to 2% of these patients, and is generally attributed to microinvasion which is undetected in the breast specimen (Cox et al, 2001). While ALND is not recommended in these patients given the morbidity associated with this procedure, the minimal risk of SLNB has brought the question of axillary staging in patients with pure DCIS back to the fore.

Although the goal of SLNB is to identify the rare patients with DCIS who have nodal metastases to potentially improve their outcome, SLNB may actually give us misleading information that adversely affects patient care. Reported rates of SLN metastasis for patients with DCIS (without microinvasion) have been as high as 12-23% when immunohistochemistry (IHC) for cytokeratins is used for histopathologic analysis of the SLN (Klauber-DeMore et al, 2000; Cox et al, 2001; Tamhane et al, 2002). This has generated enormous controversy regarding the prognostic implications of IHC-detected micrometastases for patients with breast cancer. Some of these patients with DCIS and IHC-detected micrometastases have undergone completion ALND and have received adjuvant chemotherapy. Given that historically over 98% of patients with DCIS have been cured with appropriate surgical therapy, it is difficult to justify the risk of such therapy in nearly 20% of patients who were found to have “micrometastases” which may have, in reality, only been isolated tumor cells which would not have impacted their outcome.

SLNB for pure DCIS is justifiable, however, when performing mastectomy based on a core needle biopsy diagnosis of DCIS. In this situation, the core needle biopsy can underestimate the presence of invasive cancer in up to 35% of cases (Lieberman, 2000; Jackman et al, 2001). Because SLNB cannot be performed later, after the breast has been removed, SLNB is warranted to avoid the need for ALND at a later date. When SLNB is performed for DCIS, however, caution should be exercised in the interpretation of the results, and patients with isolated tumor cells should not be treated as if they had macrometasis (McMasters et al, 2002). IHC should not be used for routine analysis of SLN for DCIS, or invasive cancer.

B. Neoadjuvant chemotherapy

Increasingly, patients are being treated with neoadjuvant chemotherapy as a method of monitoring in vivo tumor response to cytotoxic agents (Fisher et al, 1997). The response of axillary lymph nodes to neoadjuvant therapy is not necessarily uniform, and hence, the accuracy of SLNB in these patients has been questioned. Several small single-institution studies addressing this issue have now been performed with varying results. Two studies have found very high false negative rates (up to 33%) and conclude that SLNB is not accurate after neoadjuvant chemotherapy (Nason et al, 2000; Fernandez et al, 2001). Four other studies, however, found SLNB to have an acceptable false negative rate except in patients with inflammatory breast cancer (Breslin et al, 2000; Tafra et al, 2001; Stearns et al, 2002; Piato et al, 2003). All of these studies are small, and therefore the use of SLNB after neoadjuvant chemotherapy is an issue which continues to be debated in the literature.

Mamounas et al (2003) recently reported their multi-institutional experience using SLNB after neoadjuvant chemotherapy in patients participating in the NSABP B-27 study. Of 343 patients who underwent SLNB followed by ALND, they found a false negative rate of 7% and an overall accuracy of 96%. While this study may suggest that there may be a role for SLNB after neoadjuvant chemotherapy, this remains an area of controversy. Many do not accept SLNB in this setting to be completely reliable and, therefore, some have advocated the strategy of doing a SLNB prior to the initiation of neoadjuvant chemotherapy to stage the axilla, and proceeding to axillary dissection in those with a positive SLN at the time of definitive surgery (Sabel et al, 2003).

C. Immediate reconstruction

The oncologic safety of skin sparing mastectomy and the psychological benefits of immediate reconstruction have been well documented in the literature (Kroll et al, 1991). With the introduction of SLNB, however, the question of the accuracy of intra-operative assessment of the SLN has become pivotal. False negative intraoperative SLN pathology results may have important ramifications in mandating an ALND around a microvascular anastamosis in patients who have had immediate free flap reconstruction with anastomosis to the thoracodorsal vessels (Kronowitz et al, 2002). Some authors have suggested timing the SLNB prior to the mastectomy in order to allow for more accurate pathologic assessment based on permanent section histology (Brady et al, 2003). Whether a full axillary node dissection is required in the setting of a positive SLN is still under active investigation (Grube and Guiliano, 2001).

IV. Pathologic issues

The concept of SLNB which maintains that the SLN is the first node to which tumor cells spread puts a heavy responsibility on the pathologist to carefully evaluate this node for metastatic disease. The optimal technique for evaluation of the SLN, both intraoperatively and on permanent sections, has not been well established. Furthermore, the meticulous search for microscopic foci of disease has led to further debate as to the implications for such miniscule metastases.

A. Intraoperative evaluation

Intraoperative SLN evaluation has the potential to guide surgical decision-making while sparing patients a potentially unnecessary second procedure. Two techniques have been evaluated in the literature, both having advantages and disadvantages. Frozen section has been found to have a negative predictive value of 90-95% (Zurrida et al, 2001; Chao et al, 2002).Most pathologists are comfortable with this technique, which allows visualization of histologic architecture. However, using frozen section is often purported to waste valuable tissue in the cryostat. Touch imprint cytology, on the other hand, preserves all of the tissue, but relies on cytologic interpretation of cells touched to a glass slide. This technique has been found to be an acceptable alternative to frozen section analysis with negative predictive values of 87-99% (Kane et al, 2001; Henry-Tillman et al, 2002; Shiver et al, 2002). The preferred technique for intraoperative analysis of SLN depends on the expertise at individual centers.

Because touch prep cytology and frozen section are both associated with the potential for both false-negative and false-positive results, many centers prefer to wait for permanent section pathology results before deciding to perform ALND. In a formal decision analysis of this topic, we found no clear overall patient benefit for intraoperative SLN analysis (Chao et al, 2003).

B. Histological analysis and immunohistochemistry

The question of optimal tissue processing is no less in the setting of the final pathologic evaluation of SLNs. Given the importance of the SLN in staging, pathologists have become increasingly rigorous in their search for metastatic disease, using serial sectioning, IHC, and in some cases, reverse transcriptase-polymerase chain reaction (RT-PCR) to improve their yield. These techniques have all been found to upstage patients who were previously thought to be node negative on routing hematoxylin and eosin staining (Table 2). The implications of such findings, however, are unclear. Some studies have found that micrometastases found using IHC impact survival (Bettelheim et al, 1990; Nasser et al, 1993; McGuckin et al, 1995)while others do not demonstrate any difference (Wilkinson et al, 1982; Friedman et al, 1988; Cote et al, 1999). The current AJCC staging system has established the definition of micrometastases to be metastatic deposits between 0.2 – 2.0 mm (Singletary et al, 2002). The significance of such minimal nodal disease, and indeed of isolated tumor deposits less than 0.2 mm, has yet to be established. The current recommendation from the College of American Pathologists, the American College of Surgeons Oncology Group, and the clear message from the studies of SLNB for DCIS is that IHC of SLN should not be used for making patient care decisions until further evidence is available.

V. Conclusions

SLNB has become an accepted minimally invasive alternative to routine level I and II ALND for patients with breast cancer. While it is widely practiced, and considered state-of-the-art, there remain a number of controversies surrounding the technical, clinical and pathologic aspects of this procedure. Further investigation is warranted to address these issues, and participation in trials which seek to clarify these questions should be encouraged.

Table 2: Detection of micrometastases

Study	N	Technique	No. upstaged (%)
Wong et al (2001)	869	IHC	58 (7)
Pendas et al (1999)	385	IHC	41 (11)
Schreiber et al (1999)	210	IHC	17 (9)
McIntosh et al (1999)	52	IHC	8 (14)
Weaver et al (2000)	489	IHC/SS	20 (4)
Pargaonkar et al (2003)	64	IHC/SS	5 (8)
Stitzenberg et al (2002)	55	IHC/SS

Review Article

Received: 1 September 2003; Accepted: 18 September 2003; electronically published: October 2003

I. Introduction

III. Patient Issues

A. Ductal carcinoma-in-situ

IV. Pathologic issues

B. Histological analysis and immunohistochemistry