Francisco Civantos and Samuel J. Trosman Abstract
The technology and indications for sentinel node dissection have continued to expand over the last several years. It is used routinely for intermediate thickness melanoma and Merkel cell carcinoma, and has been used for cutaneous squamous cell carcinoma of the head and neck. Sentinel node biopsy has been shown to be a reliable staging technique for intermediate thickness oral cavity cancer, with sensitivity and negative predictive value rates mirroring those for melanoma. Its use in pharyngeal and laryngeal cancer continues to be investigated. The use of 99mTc tilmanocept, a CD206 receptor target, allows for more specificity for the first echelon sentinel lymph nodes and may result in improved accuracy and a lower false-negative rate. The widespread availability and accuracy of combined SPECT/CT imaging provides a scrollable three-dimensional analysis of the drainage basins, resulting in better targeted dissections and patient counseling. Despite these technological advances, the use of the gamma probe for sentinel node dissection is not always intuitive and requires specific training and expertise, especially in cutaneous lesions that show frequent metastases to the parotid gland. Serial step sectioning and immunohistochemistry analysis are critical to the detection of micrometastases. Intraoperative frozen section analysis has also shown great promise for identifying sentinel nodes in oral cavity squamous cell carcinoma, allowing for completion lymphadenectomy in the same setting. Sentinel node biopsy and dissection remains an intriguing prognostic and potentially therapeutic option for malignancies with an intermediate risk of regional metastases, with increased diagnostic information compared to watchful waiting with serial imaging but sparing the morbidity of a full elective neck dissection in patients who end up with microscopically negative lymphatic basins.
Keywords: melanoma, oral cavity squamous cell carcinoma, lymphoscintigraphy, gamma probe, tilmanocept, SPECT, immu- nohistochemistry
21.1 Introduction
Management of the clinically node-negative neck in patients with early-stage head and neck cancer remains controversial. A significant body of literature, culminating in a recent randomized trial, has proven there is significant risk of decreased survival and increased morbidity with a watchful waiting/ observational approach for lesions with a relatively high (>15%) rate of regional metastasis.1 However, this must be balanced with the morbidity of systematic cervical lymphadenectomy, which may include temporary or permanent lower lip weakness, shoulder dysfunction, lymphedema, chyle leak, and a long list of rarer neurological and vascular complications. This dilemma is magnified for lesions that approximate the midline, where bilateral neck dissections are under consideration.
Sentinel node dissection and biopsy (SNB) involves the integration of radiologic and surgical techniques. While lymphoscintigraphy can be used alone for anatomic mapping of the drainage pathways of a lesion, evaluation of the true histopa- thological status of the lymph node is the only strategy for detection of micrometastases. Cross-sectional imaging studies such as computed tomography (CT), magnetic resonance imaging (MRI), and fused positron emission tomography (PET)/CT continue to improve in their sensitivity for regional lymphatic spread; however, they generally will not reliably identify metastatic disease less than 1 cm in size, and will almost never detect micrometastases less than 5 mm in size.
Morton et al reintroduced the concept of sentinel node biopsy to surgical practice in publications describing the technical details and their early prospective clinical experience in patients with clinically N0 cutaneous malignant melanoma.23 After injection of vital blue dye at the primary site, 259 sentinel nodes were identified in 194 of 237 lymphatic nodal basins, and the incidence of false-negative sentinel nodes (i.e., the identified sentinel node is found to be disease free when metastatic disease is present in the regional lymphatic vessels) was less than 1%. The model of initial sentinel node biopsy with detailed pathologic analysis followed by completion dissection if positive has been subsequently validated for multiple head and neck subsites.
Based on the work of Krag and Alex et al,4 5 modern sentinel node biopsy has been performed using peritumoral intradermal injections of radiotracers such as unfiltered technetium 99 m (99mTc) sulfur colloid and intraoperative gamma detection probes. This combined modality allows placement of the biopsy incision directly over the radiolabeled sentinel node(s), with dissection aimed directly at the most radioactive node without disturbance of surrounding tissues. Sentinel node biopsy in melanoma using the gamma probe resulted in retrieval of sentinel nodes in over 80% of cases, with a very low incidence of false negatives.6.7 Sentinel node retrieval has continued to improve to approximate 99% in more recent studies. In addition, nearly all authors have reported occasional unusual drainage patterns that would have been missed in standard cervical lymphade- nectomy, particularly unexpected bilateral drainage for lateral- ized lesions, “skip” drainage to a more distant node in a group than might be anticipated, and drainage to multiple lymph node groups in the neck.
As has occurred in clinically node-negative melanoma, sentinel node biopsy ultimately offers the possibility of identifying those patients with clinically N0 carcinomas of the other head and neck subsites, including aerodigestive tract sites, who harbor occult metastases in the cervical lymphatics. At this point, it has primarily been applied in the more accessible oral cavity site. It is important that an established negative predictive value (NPV) and false-negative rate (FNR) be available for each tumor type before incorporating it into routine clinical practice. In particular, the NPV is an important value that might be used in the informed consent process for an appropriate patient who might truly not harbor disease in his neck. The NPV for a particular tumor type should estimate the risk that a patient would recur in the neck after a negative sentinel node biopsy. Since selective neck dissection is an excellent technique for staging the cervical lymphatics, with moderate but generally acceptable morbidity, it is important that quality data be generated before accepting the less invasive sentinel node biopsy approach. On the other hand, particularly for oral cancer, patients presenting with relatively superficial lesions, perhaps estimated to be 1 to 3 mm in thickness, for which the cervical lymphatics would often be observed, sentinel node biopsy may theoretically represent a more aggressive approach for those who desire a more intense evaluation of the lymph nodes relative to what is currently standard.
21.2 Sentinel Node Biopsy for Cutaneous Malignancies
21.2.1 Melanoma
Reports from multiple large-volume centers have shown the benefit of sentinel node biopsy for patients with invasive, intermediate-risk clinically node-negative melanoma. This led to the creation of a phase III prospective trial termed the “Multicenter Selective Lymphadenectomy Trial (MSLT-1),” which randomized 2,001 patients with cutaneous melanoma of all subsites to either (1) wide excision and observation, with lymphadenectomy for nodal recurrence, or (2) wide excision and sentinel node biopsy, with immediate lymphadenectomy for positive nodal disease. Among patients whose lymphatic basins turned out to harbor micrometastases, in those with intermediate thickness (Breslow depth: 1.2-3.5mm) melanomas, sentinel node biopsy significantly improved mean 10-year disease-free survival (71.3 vs. 64.7%), as well as rates of 10-year distant disease-free survival (hazard ratio: 0.62) and melanoma-specific survival (hazard ratio: 0.56). In addition, again only for those who actually harbored lymphatic metastases, sentinel node biopsy improved mean 10-year disease-free survival in patients with thick (> 3.5 mm) cutaneous melanoma (50.7 vs. 40.5%).8 National Comprehensive Cancer Network (NCCN) guidelines currently recommend sentinel node biopsy for N0 patients with cutaneous melanoma thickness of 0.76 to 1 mm and either ulceration or mitotic rate of > 1 per mm2, and for all N0 patients with melanoma of thickness of greater than 1 mm.9
More recently, the second Multicenter Selective Lymphade- nectomy Trial (MSLT-2) was performed in an effort to evaluate the utility of completion lymphadenectomy in patients with intermediate thickness cutaneous melanoma and positive SNB. A total of 1,934 patients with positive SNB were randomized to either (1) immediate completion lymphadenectomy or (2) observation with ultrasound. Patients with immediate lymphade- nectomy had an increased rate of regional control (mean: 92 vs. 77%) and a slightly higher disease-free survival (mean: 68 vs. 63%), but had a nearly equivalent rate of melanoma-specific survival (86% in both groups). The authors suggest that any survival benefit from early surgery may occur in patients with disease limited to the sentinel nodes. The prognostic value of completion lymphadenectomy has to be weighed against the risk of complications, as MSLT-II showed a significant increase in the incidence of lymphedema (24.1 vs. 6.3%) in the lymph node dissection group.10 No anatomic subsite data were provided in this study. However, given the importance of avoiding the local effects of regional recurrence in the head and neck, and to a lesser degree the lymphedema that occurs relative to extremity melanoma, most authors still advocate completion lymphadenectomy for positive sentinel nodes in the head and neck region.
21.2.2 Merkel Cell Carcinoma
Merkel cell carcinoma is a rare, aggressive neuroendocrine cutaneous malignancy that frequently arises in the head and neck. While increasing tumor size appears to be the main risk factor for lymphatic spread, even small (< 2 cm) lesions have high rates of regional metastases. Although there is conflicting evidence regarding a survival benefit to sentinel node biopsy in Merkel cell carcinoma, NCCN guidelines currently recommend SNB for all clinical N0 tumors due to improved prognostic information on regional control.11 Adjuvant radiation therapy appears to be associated with improved survival, and can be considered as an alternative to completion lymphadenectomy in sentinel nodepositive patients.12 For larger, deeper lesions, elective lympha- denectomy should be considered, as the rate of sentinel node positivity is exceedingly high.13
21.2.3 Cutaneous Squamous Cell Carcinoma
The role of sentinel node biopsy in the treatment of cutaneous squamous cell carcinoma remains unclear. The NCCN recommends the consideration of SNB in certain high-risk lesions, “although the benefit of and indication for this technique has yet to be proven.”14 Systematic reviews have shown positive and negative predictive values of over 90%, with similar FNRs to that of melanoma.15.16 However, larger prospective trials are needed to validate the technique in this patient population. Most importantly, appropriate selection criteria delineating which squamous cell carcinomas are appropriate for the sentinel node procedure need to be identified.
21.3 Sentinel Node Biopsy in Oral Cavity and Aerodigestive Tract Malignancies
Aerodigestive tract malignancies show an increased propensity for regional metastases, especially lesions with high-risk features such as size greater than 2 cm, thickness greater than 3 to 4 mm, angiolymphatic invasion, and perineural invasion. Several analyses have shown a survival benefit to the prophylactic treatment of the N0 neck in early-stage head and neck squamous cell carcinoma (HNSCC), including a recent prospective trial that showed improved survival in patients with stage I/II lateralized oral cavity cancers who received unilateral elective neck dissection versus those who were observed and underwent therapeutic neck dissection if/when they developed N + disease.1 The risk of occult metastases must be weighed against the morbidity of neck dissections for patients, many of who will not have pathologic neck disease. Sentinel node biopsy offers a diagnostic technique to identify subclinical cervical metastases and is ideal for situations where the expected risk of metastases falls in the 5 to 15% range, which might be too high to feel comfortable with observation but too low to justify a full selective neck dissection. It also may be particularly useful for patients with stage 1 lesions of borderline size and thickness that approximate the midline, for whom the morbidity of bilateral neck dissection for clinically and radiologically negative cervical lymphatic basins is difficult to justify.
Multiple centers in the United States and Europe initiated single-institution trials of sentinel node biopsy for upper aerodigestive tract malignancies in the 1990s and 2000s, mostly for oral cavity cancer, as this is the most accessible subsite, though some included oropharyngeal lesions as well. Studies consistently reported NPVs of 95 to 100%, similar to the rates seen in cutaneous melanoma. They also described unexpected patterns of lymphatic drainage, including unexpected contralateral drainage, and upstaging made possible by the identification of micrometastases via fine sectioning and immunohistochemistry (IHC) in 10 to 20% of cases.17
The importance of appropriate step sectioning and IHC of sentinel nodes to the accuracy of the technique has been emphasized in nearly every study. Christensen et al in 2010 performed an exhaustive step sectioning and IHC at 150 μm of the numerous nonsentinel nodes in their completion neck dissections from their validation trial after initial SNB and subsequent lymphadenectomy showed extremely low rates of additional occult metastases. In no case, with the additional fine sectioning of nonsentinel nodes, did they find a micrometastasis in a nonsentinel node in a patient with negative sentinel nodes.18 Only in one case with a positive sentinel node and positive nonsentinel node by routine technique was an additional micrometastasis found. This study validates the current algorithm of fine sectioning of sentinel nodes with IHC and routine analysis of nonsentinel nodes. In addition, in another review, Jefferson et al performed reexamination of 35 negative sentinel nodes from 10 patients and sectioned the nodes at 150 pm rather than the original 2-mm sections. In this study, they did not find any additional microcarcinomas, suggesting that the additional yield of finer than 2-mm sectioning is small.19 Clearly an additional micrometastasis may occasionally be found, and this infrequent occurrence could make a major difference for a patient. Currently, 150 μm sectioning is standard practice in Europe, but in North America the labor intensity of this practice is felt unjustifiable by most pathology departments, and 2-mm thick sections are more common.
The first large European multicenter trial evaluating the accuracy of SNB for HNSCC was published in 2010. A total of 227 SNB procedures were carried out on 134 patients; 79 patients underwent SNB alone, while 55 underwent sentinel node-assisted neck dissection. The authors reported a 93% successful sentinel node identification rate, with a 95% NPV for sentinel node biopsy. Patients receiving SNB did not have significantly different long-term survival compared to patients receiving elective neck dissection, prompting the authors to conclude that SNB is a viable alternative to lymphadenectomy in N0 patients with early-stage HNSCC. There was concern that the technique was problematic for floor of the mouth tumors, as four of the five false negatives came from these patients.20
In order to formally validate sentinel node biopsy for oral cavity cancer, an NCI-funded trial was completed in North America under the auspices of the American College of Surgeons Oncology Group (ACOSOG). One-hundred forty patients with T1 or T2 clinically N0 oral cancer underwent lymphoscintigraphy with 99mTc sulfur colloid, nuclear imaging, and narrow-exposure SNB followed by immediate completion neck dissection. SNB with step sectioning and IHC had an NPV of 96% for all tumors, and an NPV of 100% for T1 tumors. The FNR was 9.8%. Investigators were categorized as experts who came in with significant numbers of cases versus intermediate users who took the animal course and were trained on the protocol. None of the false negatives occurred in the expert group of investigators on this study.21
Based on the strong validation data, it has been felt that it was appropriate to transition to sentinel node biopsy as primary neck management, in a research setting, for appropriately selected cases. A meta-analysis of 21 pooled validation studies and 847 patients showed an overall sensitivity of 93% for SNB.22 Techniques have been reported for accurate SNB for squamous cell carcinoma in the floor of mouth, alleviating a prior concern. These techniques do involve more extensive level I dissection to separate the lymphatic basins from the primary tumor. In 2015, the results of the EORTC-approved Sentinel European Node Trial (SENT) with SNB as primary neck management in oral cancer, with neck dissection for positive sentinel nodes only, was published. This was the first prospective trial to use SNB as the sole staging procedure in oral cancer, without concurrent END. Four-hundred fifteen patients with T1-2N0 HNSCC underwent SNB followed by neck dissection within 3 weeks if positive. A sentinel node was detected in 99.5% of cases; 109 of the 415 patients (26%) had occult metastases. SNB showed a sensitivity of 86% and an NPV of 95%. The disease-free survival for the cohort was 92%, prompting the authors to confirm SNB as a safe oncologic procedure. Unexpected contralateral lymphatic drainage occurred in 12% of cases, with seven positive contralateral sentinel nodes that would have been missed with an en-bloc unilateral neck dissection. The FNR was higher than expected at 14%, which is likely related to technical issues in a multi-institutional trial consisting of a large group of surgeons with a variety of techniques and experience levels. The authors in their discussion imply that this number would come down with continued refinements in technique.23
Multiple single institution studies looking at SNB with completion neck dissection only for positive sentinel nodes have also been completed concurrent with this trial and are detailed in Table 21.1.24,25,26,27 All studies report acceptable NPV, and several larger trials report significantly lower FNRs, in the range of 6 to 9%, indicating that technical factors may explain some of the variations in data. It should be mentioned that in smaller series with few positive sentinel nodes, one false negative could skew the FNR tremendously.
Table 21.1 Studies with sentinel nodes as primary neck management (n >50 included)
|
Study |
n |
NPV (%) |
FNR (%) |
|
Pezier et al24 (20 1 2) |
59 |
97.5 |
6 |
|
SENT trial23 (20 1 5) |
415 |
95 |
14 |
|
Pedersen et al25 et al (2016) |
258 |
95 |
5 |
|
Flach et al26 (2014) |
62 |
88 |
20 |
|
Broglie et al27 (2011) |
51 |
96 |
6 |
Abbreviations: FNR, false-negative rate; NPV, negative predictive value.
21.4 Surgical Technique: Practical Guidelines
21.4.1 Patient Selection
An appropriate patient for sentinel node dissection is one with a small, injectable lesion that carries a significant risk of lymphatic micrometastases but a relatively low risk ofdistant meta- stases. In the head and neck region, SNB is standard for intermediate-thickness melanoma. It can be used for stage 1 melanoma, but the yield of micro metastases is low and it usually is not justified in this group without other risk factors. On the other hand, for deeply invasive melanoma with greater than 4 mm depth of invasion, the traditional concept is that high risk of distant disease makes prophylactic treatment of lymphatic micrometastases irrelevant. The MSLT-1 trial calls the latter concept into question, as significant benefits in terms of regional control, reduced morbidity, and increased disease-free survival may occur even in this group. For cutaneous lesions of other histologies, the criteria are less well defined in the literature, and surgeons tend to extrapolate from the experience in melanoma.
In cancer of the oral cavity, criteria for sentinel node biopsy remain controversial. The standard approach in 2017 is selective neck dissection for all significantly invasive lateralized lesions (stage 1 and above) and watchful waiting for extremely thin stage 1 lesions. Thickness estimates are based on physical examination and generous biopsy of the thickest appearing portion of the lesion. Management of midline stage 1 lesions of intermediate thickness may represent an area of continued controversy. The “most aggressive” advocates of sentinel node biopsy argue for SNB for all N0 patients, stages I and II. We believe a reasonable way to apply this technology is to use it in those situations where watchful waiting has been applied frequently, but where a reasonable argument could also be made for selective neck dissection. This would certainly include stage I and early stage II tumors with an intermediate depth of invasion, corresponding to estimates of 2 to 4 mm depth of invasion, and stage 1 midline lesions with even thicker depths. It could be offered as an alternative to neck dissection for lateralized lesions with estimated depths up to 8 mm, for patients who are concerned about the morbidity of neck dissection, with caveats related to small false omission rates (the NPV subtracted from 100), and the need to perform two stages of surgery when sentinel nodes are detected to be positive on final pathology. The bulk of patients in the ACO- SOG validation trial were in this last group. With greater experience and larger published clinical trials on SNB, the proper group for selection will be better delineated.
Since early cancer of the oral cavity has a smaller risk of distant metastases than melanoma, the manner in which these patients need to be approached is somewhat different from the way we approach melanoma patients. We do not want to lose the opportunity for cure with timely lymphadenectomy, but need to balance this against the morbidity of formal neck dissection in numerous patients who do not harbor lymphatic metastases. Multiple series have confirmed that the technique is not appropriate for T3 and T4 primary tumors due to the significant volume of tissue that would need to be injected, the excessively large number of radioactive nodes generated, the greater risk that grossly positive nodes exist, the potential for false negatives due to incomplete injection, and the technical futility of removing a large number of nodes in piecemeal fashion. Furthermore, since the risk of having a positive node increases with increasing stage, there comes a point where the percentage of patients who will convert to neck dissection is too large to justify the SNB procedure. To the contrary, the technique is best applied toT1 lesions and smaller T2 lesions. If a lesion is less than 3 cm in maximal diameter but has significant fixation of the tongue or other manifestation of deep invasion, then this lesion is truly a T4 lesion and results with sentinel node biopsy are unlikely to prove accurate and useful.
If the primary tumor meets criteria, the next issue is to determine whether the cervical lymph nodes are clinically involved. While SNB is an excellent technique for detecting micrometa- stases, it is less useful for detecting nonpalpable but grossly involved lymph nodes. This appears to be particularly true with squamous cell carcinoma. It is postulated that when a large percentage of the lymph node is replaced by cancer, physiologic obstruction can occur and alternative patterns of lymphatic drainage develop. It is important to detect the presence of such gross disease on preoperative imaging and physical exam and, as a last opportunity, at the time of intraoperative palpation. We should avoid applying SNB to this group of patients in order to avoid false positives. Generally, contrast-enhanced CT or MRI are the cross-sectional imaging modalities most commonly used. These should be strictly interpreted, and patients should be excluded if there are nodes greater than 1.2 to 1.5 cm in size with central necrosis, irregular enhancement, or a poorly defined or irregular capsular border, or with groups of three or more asymmetrically located lymph nodes with a minimal axial diameter of 8 mm or more in the suspected tumor drainage area. Fused PET/CT is also useful in ruling out regional metasta- ses greater than 1 cm, but remains plagued by false positives and should be evaluated with some skepticism, particularly if the PET scan is performed shortly after an oral biopsy or if superinfection of the tumor is suspected. Inflammatory nodes can have elevated SUV values and patients may be inappropriately denied the opportunity for a SNB approach.
21.4.2 Injection of the Primary Tumor
The injection is performed prior to the surgical procedure, generally on the morning of surgery. Injection can also sometimes be performed late on the day before, although the effect of this on the success rate of sentinel node identification is still unclear. While awake injection and imaging in radiology is the most commonly used technique, as we extend this procedure to endoscopi- cally accessible oropharyngeal, supraglottic, and hypopharyngeal lesions, it is likely that cooperative efforts with the nuclear radiologist and the use of portable cameras will allow for intraoperative endoscopic injection and gamma probe-guided sentinel node biopsy without the need for uncomfortable injections in an awake patient. Theoretical advantages of injecting under general anesthesia include better exposure of the primary and avoidance of motion of the patient related to discomfort. This may eventually increase the reliability of this method. Taking into account that the radio localization of the detected hot spots does not represent the drainage of the primary but rather the drainage of the tracer deposits, which act as a surrogate for the lymphatic drainage of the primary, the impact of a thorough and representative tracer injection is evident. Due to the density and direction of the head and neck lymphatics, the primary may drain into several alternative lymphatic pathways, all representing first echelon “sentinel” lymph nodes (Fig. 21.1).
Nevertheless, due to regulatory issues related to the injection of radioactive substances and the scarcity of widely available portable nuclear imaging, awake injection remains the most commonly used technique. It is important to ensure that the patient is comfortable so that an adequate preoperative injection is obtained. We use topical anesthetic, mild oral sedation, and/or lingual, inferior alveolar, or sphenopalatine nerve blocks to ensure patient comfort during manipulation and injection of the primary tumor. Direct injection of the tumor with local anesthetic should not be performed as it may affect uptake of the radionuclide and reportedly may even cause it to precipitate in the tissues. The injection technique involves narrow injection with a fine 25-gauge needle on a tuberculin syringe, circumferentially encompassing the leading edge of the lesion and, for thicker lesions, an additional injection in the center of the lesion. Typically, we have used five tuberculin syringes with 1 mL aliquots of technetium 99 sulfur colloid, similar to doses used for melanoma, with a total radioactivity of 400 millicuries representing a standard dose for the morning of surgery. A higher dose would be used the night before.
More recently, we have begun using 99mTc tilmanocept, a CD206 receptor targeting pharmaceutical, instead of technetium sulfur colloid. The receptor targeting nature of tilmanocept eliminates the shine-through effect that may be seen with radiolabeled colloids that are retained for long periods of time at the injection site. Also, tilmanocept appears to have greater retention in first echelon nodes with less movement downstream due to its receptor binding properties, and appears to have greater flexibility in the timing of imaging and surgery (same day vs. next day). One phase III multi-institutional trial has shown lower FNRs and greater accuracy with 99mTc tilmanocept when compared to 99mTc sulfur colloid.28 Peritumoral injection with 50 pg of 99mTc-tilmanocept within 15 hours (same day) or between 15 and 30 hours of surgery (next day) is accomplished in a similar manner as described earlier.
Fig. 21.1 A 1.5-inch needle and tuberculin syringe are used to inject gingerly, completely encompassing the lesion. Excess force should be avoided so that nonphysiologic drainage patterns will not be opened. Sedation, topical anesthetics, and nerve blocks can be used, but lidocaine should not be infiltrated into the bed of injection. Avoid injecting more widely or deeply than necessary, as the injected colloid will extravasate more widely than is apparent.
We have used the aforementioned technique for visible oral lesions. For cutaneous lesions, it is well documented that a scar from a previous excisional biopsy can be injected to allow for accurate sentinel node biopsy. Whether a previously excised oral lesion could undergo sentinel node excision by injection of an intraoral scar is not proven, but logically similar principles should apply. It is important to inject narrowly, and not to inject the deep tissues. The radionuclide will diffuse more widely in the oral cavity around the site of injection than occurs in the skin and will usually go to the neck more quickly. There is no benefit to trying to inject a margin around the tumor, as this will lead to an unmanageable excess of radioactive nodes.
The use of blue dye concurrently with radionuclide particles has become popular in sentinel node biopsy for cutaneous lesions. This is a reasonable technique for skin, and certainly can help during the learning phase of the procedure, as the subtle blue dyed lymphatic vessels can be traced toward the sentinel node. Furthermore, reinjection of the blue dye can be performed in the operating room under anesthesia, and can provide a measure of security against inadequacy of preoperative injection due to patient discomfort. Particularly for oral cancer, our preference is to use the radionuclide alone. Numerous publications indicate that it is extremely unlikely to have blue sentinel nodes that are not radioactive. Anaphylactic allergic reactions to isosulfan blue dye, though rare, can occur. For oral lesions, we prefer to remove the primary first to eliminate radioactive background at the primary site. When the technique is performed in this sequence, the blue dye has usually run through to the distal lymphatics by the time the oral resection is completed and margins are sent, making the dye less useful. Finally, oral cavity resections have functional implications that force the surgeon to obtain adequate but closer margins compared to those obtained for cutaneous melanoma. Blue staining of the oral tissues can lead to loss of the visual cues the surgeon uses to guide decisions regarding whether there is tumor involvement at a margin. For all of these reasons, we prefer not to use adjunctive blue dye, particularly for oral lesions. However, other surgeons prefer to use adjunctive blue dye and obtain excellent results. The removal of sentinel nodes based solely on the blue dye technique is less accurate and should not be performed without also injecting radionuclide.
21.4.3 Radiologic Lymphatic Mapping
After injection, nuclear imaging of the lymphatics is obtained. We now routinely use fused single photon emission computed tomography-computed tomography (SPECT/CT) due to its superior three-dimensional localization ability and a potential to detect more sentinel nodes in unexpected drainage basins. Scrollable images allow for improved targeting and patient counseling prior to the sentinel node dissection (Fig. 21.2).
A dynamic phase should be acquired with serial images for at least one-half hour following injection. These images should be acquired for 1 minute each. Transmission images should be acquired for 1 to 2 minutes in each new movement of the camera (transmission images are obtained using a flat source placed on the detector opposite the working detector. This is done to achieve better localization of any lymph nodes that take up the radiopharmaceutical by superimposing them on a body image). While it is possible to perform sentinel node biopsy with the intraoperative gamma probe alone, the radiologic image can be useful in providing a guide to the location of the sentinel node, allowing for a more complete informed consent process by predicting unexpected drainage to the contralateral neck or other areas that were not expected to be involved. It can also allow for greater surgical efficiency, and likely shorter surgery times, by more exactly delineating the anatomic location of sentinel nodes. In the modern practice of head and neck sentinel node biopsy, most surgeons advocate for the use of fused preoperative SPECT/CT.
21.4.4 Removal of the Primary Tumor
As mentioned earlier, we prefer to resect the primary tumor first. If the injection field is sufficiently narrow, this usually eliminates or greatly reduces background radioactivity at the primary site that can confound the sentinel node identification. The usual, appropriate surgical margins with frozen section control should be obtained. The sentinel node technique for aerodigestive tract malignancies can also be performed in conjunction with a mandibulotomy or other technique of exposure, as long as the primary tumor stage is T2 or less, and as we extend the procedure to the oropharynx this may become more common. In some situations, it may be necessary to perform the nodal biopsy prior to primary resection. However, for most appropriate oral and cutaneous lesions, the lesion will be accessible for resection prior to addressing the lymphatics. Removal of the primary tumor first is less important when the lymphatic basin is distant from the primary tumor, and is often less of an issue for skin lesions (i.e., auricular scalp or nasal lesions). However, this can be important for lesions of the neck or preauricular skin where the lymphatics are immediately deep to the primary tumor. It is especially important in cancer of the floor of the mouth due to the immediate proximity of the submandibular lymphatics.
21.4.5 Gamma Probe-Guided Sentinel Node Biopsy
The hand-held gamma probe is now used to confirm the location of the sentinel lymph nodes, which previously were determined by lymphoscintigraphy and subsequent SPECT-CT. The skin is marked with the location of the nodes. Background readings should be taken of the precordium (Fig. 21.3) as a lower limit background measurement, as well as at the resected primary site as an upper limit of background. The latter is important in avoiding errors due to shine through from the primary injection site.
If the patient is to undergo sentinel node biopsy alone, with neck dissection planned only for positive findings intraoperatively or on permanent histopathology, the incision can be drawn narrowly over the node. However, the incision must be consistent with the possibility of subsequent neck dissection, and the planned incision for the formal lymphadenectomy should be considered. Alternatively, the incision can be drawn in the line of that to be used for neck dissection, although shorter in length, and flaps can be elevated. This latter approach is always used when immediate gamma probe-guided completion neck dissection is the plan, either due to patient preference for combined lymphatic mapping and selective neck dissection or as part of a validation trial. After the incision is made, subplatysmal flaps are elevated sufficiently to provide access to the hot area. The neck should first be carefully palpated in order to identify palpable gross lymphatic disease that may not be physiologically functional, and hence may not take up radioactivity. The finding of gross cancer involvement would, of course, contraindicate sentinel node biopsy and mandate formal lymphadenectomy.
Fig. 21.3 Background reading is taken at the precordium.
Fig. 21.4 Blunt dissection towards the hotspot is followed by reinsertion of the gamma probe into the path of dissection and angulation in various directions, seeking the radioactively tagged lymph node.
If no gross disease is identified, the surgeon will now localize the sentinel node(s). Use of the probe to locate the nodes in a three-dimensional location is not intuitive and it is best learned through instruction by a surgeon with experience in the technique. Initial readings are taken of the precordium and back table in order to assess the level of hematogenous radioactivity. Readings are also obtained from the resected tumor specimen and the bed of resection. The probe is slowly passed over the neck at a steady rate, assessing the auditory input for radioactivity generated by the gamma probe. Care is taken to aim away from the primary resection bed. Since the probe measures radioactivity over time, rapid or unsteady movement will lead to higher readings and louder auditory input, and should be avoided. Using steady constant motion, the probe is moved radially across each hot spot allowing the surgeon to determine the direction in which to proceed, in three dimensions, in order to locate the sentinel node. The surgeon then bluntly dissects toward the sentinel node. Bipolar cautery can be used to divide the tissues to provide wider exposure. We recommend avoidance of paralysis and caution in using unipolar electrocautery as the neurovascular structures in the neck are not specifically identified, although the spinal accessory and marginal nerves may often be visualized during the course of the procedure.
Fig. 21.5 The sentinel node is excised using a combination of blunt dissection and division of tissue using bipolar cautery. Unipolar cautery should be avoided when the proximity of neurovascular structures is not known.
As a dissection cavity is opened, the gamma probe is introduced into this space along the plane of dissection and angled in various directions in order to guide the surgeon to the sentinel node (Fig. 21.4). The sentinel node is bluntly excised (Fig. 21.5). Probe readings (counts per minute) are recorded for initial readings taken while the node is in the patient, as well as for “ex vivo” readings of the extracted node, away from the patient (Fig. 21.6). Repeat readings are taken of the resection bed to ensure that there are no adjacent hot nodes that also need to be removed. Any node exhibiting 10% or more of the radioactivity of the most radioactive node in the same anatomic area will be considered an additional sentinel node and will be harvested separately. If there are a large number of very radioactive nodes (i.e., more than six), this essentially represents a failure of the technique and piecemeal removal of a large number of nodes is illogical. The surgeon should remove the four most radioactive nodes or proceed to selective neck dissection if indicated. In the case where there is a very “hot” sentinel node in a specific area, there may be a relatively “hot” node in a completely separate anatomic region (i.e., submental region vs. level II jugular region) that does not reach 10% of the radioactivity of the hottest node. If this second node is truly in a separate area and is significantly greater than background (two or more times background readings), it should still be harvested as a sentinel node, as it may represent a separate drainage pattern from a different portion of the tumor. Review of the imaging and knowledge of basic anatomic principles will allow the surgeon to judge whether such additional areas of borderline radioactivity need to be excised. When the SLN dissections are performed prior to resection of an oral cavity cancer, or if significant radioactivity persists in the bed of resection, the use of intraoral lead shields can be helpful. The presence of a collimator on the gamma probe is recommended to reduce background signal from the primary tumor, and most modern probes have fine tips with collimators. With posterior tongue tumors, background activity can be avoided by using a transoral suture on the tongue to pull the primary bed away from the lymphatics.
The issue of dealing with background activity at the primary site is most marked for level I nodes with cancer of the floor of the mouth. In this situation, the surgeon may need to perform some initial dissection, below the level of the marginal mandibular nerve, transecting the tissues down to the level of the mylohyoid muscle. In this manner, the lymph nodes are mobilized away from the oral cavity, allowing for more accurate identification of the SLN(s) by placing the gamma probe into the tunnel thus created and directing the probe inferiorly away from the background radioactivity at the floor of the mouth injection site. Each SLN is labeled, measured, described, and recorded separately as to location and total ex vivo counts per second.
Fig. 21.6 Ex vivo readings pointing away from the patient should confirm whether this is the radioactive node.
Skin lesions of the periauricular region and scalp represent a unique technical challenge as they commonly drain to the parotid lymph nodes. The proximity of the facial nerve branches and the dense nature of parotid tissue often make blunt dissection problematic. In our experience, in many cases, nodes can be identified on the capsule of the parotid gland, which can be safely excised with a capsular dissection. These may be located on the tail of the parotid, anterior edge of the parotid, or region around the temporal vessels (for scalp lesions). Electromyographic intraoperative facial nerve monitoring is useful in such cases. If the gamma probe directs the surgeon to nodes deep within the parotid, then in many cases it is best to identify a distal branch of the facial nerve, or occasionally even the main trunk, and perform a localized excision of the hot portion of the parotid gland, with identification of the facial nerve. Such a procedure might be termed a “gamma probe-guided partial parotidectomy” and still simplifies the procedure relative to a formal lateral lobe parotidectomy with concurrent neck dissection. In this situation, it is particularly important to tag the tissue adjacent to facial nerve branches with permanent suture in order to provide for easy reexploration if the sentinel node is positive. In high-risk situations for metastasis, consider proceeding with formal parotidectomy, given the difficulties of reexploration in the parotid area. Similar marking of the location of the spinal accessory nerve would be advantageous if it is identified in the course of an upper jugular sentinel node biopsy.
21.4.6 Rigorous Histopathologic Assessment of the Sentinel Node
In any situation where sentinel node biopsy alone is performed, exhaustive histopathologic evaluation of the sentinel node with fine step sectioning as well as concurrent IHC should be performed to rule out microscopic foci of cancer and to allow for therapeutic neck dissection or radiation. While intraoperative frozen sections often prove difficult in cases of melanoma or neuroendocrine carcinomas, large-scale reviews have shown step-section frozen section analysis to be cost-effective and accurate in breast axillary lymph node biopsies.29 At least one institutional review of frozen section analysis of sentinel nodes in oral cavity squamous cell carcinoma has shown a high degree of accuracy, with an NPV of 99% when compared to the permanent specimen analysis.30 It is our practice to send intraoperative frozen sections on sentinel nodes suspicious for squamous cell carcinoma metastases, evaluating three- or four-step sections within the node.
It is also important for the pathologist to classify the type of metastatic lymph node involvement. Currently, the nomenclature used in sentinel lymph node pathologic examination is extrapolated from AJCC breast cancer staging: namely, the distinction between micrometastases (0.2-2 mm; Fig. 21.7) and macrometastases (>2 mm). The discovery of isolated tumor cells in lymph node specimens by morphologic (IHC) or nonmorphologic (flow cytometry or PCR) means should be differentiated from micrometastases, as the prognostic implications of these clusters of cells in oral cavity cancer are unknown. Additionally, there may be mummified keratin positive or anu- cleate cells that are not considered a positive outcome for sentinel node analysis.
21.5 Risks and Complications of Sentinel Node Biopsy
Complications are relatively rare with the aforementioned technique, as the minimally invasive nature of the procedure relative to the moderate morbidity of neck dissection is the impetus for the development of the technique. Nonetheless, like any procedure, complications can occur. One concern with SNB in the head and neck region is the theoretical risk of injury to the facial and spinal accessory nerves during blunt dissection through a narrow exposure. It is also possible that surgeons employing this technique might leave behind sentinel nodes deep in the parotid gland, which allows the avoidance of facial nerve dissection but can imply an oncologic risk. It is important, once the procedure is initiated, to do what is necessary to make the most of the radionuclide injection and ensure that all sentinel nodes of appropriate radioactivity, representing real first echelon drainage pathways, are removed (Fig. 21.8).
Fig. 21.7 Micrometastasis seen on hematoxylin and eosin stain (a) and immunohistochemistry (IHC) for cytokeratin (b). With IHC micro- metastases are more easily identified.
Although there remains a paucity of data on the exact complication rate of SNB, reported incidences of even minor complications are less than 1%. Theoretically in the hands of an inexperienced operator, the risk of injury to the facial or spinal accessory nerves may be greater with SNB than with formal parotidectomy and selective neck dissection. The SNB procedure in the head and neck should only be performed by surgeons who perform large volumes of neck dissections and parotidectomies and are comfortable with the anatomy.
Another issue is our frequent inability to achieve immediate diagnosis of positive sentinel nodes, especially in melanoma. Frozen section, while useful in squamous cell carcinoma, is absolutely not accurate for melanoma, and tissue is often better preserved for permanent sections. The only exception to this is if gross clinically positive metastases are encountered, which rarely occurs. A touch preparation for cytology can be performed on the sectioned normal-appearing node.
Fig. 21.8 Deep intraparotid sentinel node biopsy. The frontalis branch has been identified and is indicated by the arrow.
Thus, for the minority of patients with micrometastases identified in the sentinel nodes, we are sometimes dealing with issues of reexploration and dissection in functionally delicate areas. The potential risk of nerve injury related to the reexploration of an inflamed, recently operated wound needs to be considered. Rushing the pathological analysis to allow for early reexploration and tagging adjacent to any visualized nerves with blue Prolene sutures at the time of the sentinel node procedure are steps that can make secondary lymphade- nectomy easier.
21.6 Conclusion
Sentinel node biopsy offers a minimally invasive prognostic and potentially therapeutic surgical option for patients with intermediate risk, clinically node-negative disease. It has become the standard of care for intermediate thickness N0 melanoma of the head and neck, is recommended by the NCCN for all N0 Merkel cell carcinomas, and is an emerging technique for the staging of lymphatics in cutaneous squamous cell carcinoma. It is also an option to be considered in the NCCN guidelines for early oral cavity cancer as alternative to neck dissection for intermediate thickness lesions.
Selective neck dissection remains the standard approach for the majority of oral cancers, particularly for larger and significantly invasive T2, T3, and T4 lesions. Significant data already exist, however, to advocate SLNB as a reasonable alternative to selective neck dissection for smaller, thinner oral cancers, including those of the floor of the mouth albeit with a slightly modified technique, that fall in a category where watchful waiting might reasonably be chosen as an alternative but where the tumors are not so minimally invasive that the risk of metastases is negligible.
The sentinel node concept has been discarded by some based on the misconception that selective neck dissection has no significant morbidity. Coming from a tradition of more radical neck procedures, the selective neck dissection is generally viewed as an intervention with negligible morbidity by many head and neck surgeons. Although the morbidity of selective neck dissection is significantly less than that of modified radical and radical dissections, there is measurable morbidity in a variable percentage of patients, including issues with shoulder function secondary to temporary trapezius weakness followed by adhesive capsulitis of the shoulder, pain syndromes, contour changes, and lower lip mobility. There is associated visible neck asymmetry. This has been demonstrated in numerous quality-of-life studies and objective functional assessments. The moderate morbidity of selective neck dissection has led some to suggest watchful waiting as an alternative for patients of lower risk. SNB has developed as an intermediate option in response to this controversy, with all of the aforementioned complications observed much less frequently.
The sentinel node technique is likely to have an increasing role in the management of head and neck cancer in the future. Surgeons can gain experience in the use of this technique for cutaneous malignancies and early-stage oral cancers with minimal to intermediate invasiveness. The technique can also be practiced in the context of a gamma probe-guided neck dissection for more invasive cancers, preferably in the context of a clinical trial, allowing the patient to benefit from improved mapping of drainage patterns and more accurate staging through better identification of micrometastases. Pilot data on the use of this technique in the pharynx and larynx will continue to emerge. We hope to provide a guide to surgeons as they evaluate the neck both clinically and radiographically, and to describe the developing role of lymphoscintigraphy and sentinel node biopsy in detecting microscopic lymphatic metastases.
References
[1] D'Cruz AK, Vaish R, Kapre N, et al. Head and Neck Disease Management Group. Elective versus therapeutic neck dissection in node-negative oral cancer. N Engl J Med. 2015; 373(6):521-529
[2] Morton DL, Wen DR, Wong JH, et al. Technical details of intraoperative lymphatic mapping for early stage melanoma. Arch Surg. 1992; 127(4):392-399
[3] Morton DL, Wen DR, Foshag LJ, Essner R, Cochran A. Intraoperative lymphatic mapping and selective cervical lymphadenectomy for early-stage melanomas of the head and neck. J Clin Oncol. 1993; 11(9):1751-1756
[4] Alex JC, Krag DN. Gamma-probe guided localization of lymph nodes. Surg Oncol. 1993; 2(3):137-143
[5] Alex JC, Weaver DL, Fairbank JT, Rankin BS, Krag DN. Gamma-probe-guided lymph node localization in malignant melanoma. Surg Oncol. 1993; 2 (5):303-308
[6] Krag DN, Meijer SJ, Weaver DL, et al. Minimal-access surgery for staging of malignant melanoma. Arch Surg. 1995; 130(6):654-658, discussion 659-660
[7] Glass LF, Messina JL, Cruse W, et al. The use of intraoperative radiolymphoscintigraphy for sentinel node biopsy in patients with malignant melanoma. Dermatol Surg. 1996; 22(8):715-720
[8] Morton DL, Thompson JF, Cochran AJ, et al. MSLT Group. Final trial report of sentinel-node biopsy versus nodal observation in melanoma. N Engl J Med. 2014; 370(7):599-609
[9] National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology. Melanoma. Available at: https://www.nccn.org/profes- sionals/physicians_gls/PDF/melanoma.pdf. Published 2016. Accessed September 30, 2017
[10] Faries MB, Thompson JF, Cochran AJ, et al. Completion dissection or observation for sentinel-node metastasis in melanoma. N Engl J Med. 2017; 376 (23):2211-2222
[11] National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology. Merkel Cell Carcinoma. Available at: https://www.nccn.org/pro- fessionals/physician_gls/PDF/mcc.pdf. Published 2017. Accessed September 30,2017
[12] Mojica P, Smith D, Ellenhorn JD. Adjuvant radiation therapy is associated with improved survival in Merkel cell carcinoma of the skin. J Clin Oncol. 2007; 25 (9):1043-1047
[13] Shnayder Y, Weed DT, Arnold DJ, et al. Management of the neck in Merkel cell carcinoma of the head and neck: University of Miami experience. Head Neck. 2008; 30(12):1559-1565
[14] National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology. Squamous Cell Skin Cancer. Available at: https://www.nccn.org/ professionals/physician_gls/pdf/squamous.pdf. Published 2017. Accessed September 30, 2017
[15] Ross AS, Schmults CD. Sentinel lymph node biopsy in cutaneous squamous cell carcinoma: a systematic review of the English literature. Dermatol Surg. 2006; 32(11):1309-1321
[16] Ahmed MM, Moore BA, Schmalbach CE. Utility of head and neck cutaneous squamous cell carcinoma sentinel node biopsy: a systematic review. Otolaryngol Head Neck Surg. 2014; 150(2):180-187
[17] Stoeckli SJ. Sentinel node biopsy for oral and oropharyngeal squamous cell carcinoma of the head and neck. Laryngoscope. 2007; 117(9):1539-1551
[18] Christensen A, Bilde A, Therkildsen MH, et al. The prevalence of occult meta- stases in nonsentinel lymph nodes after step-serial sectioning and immuno- histochemistry in cN0 oral squamous cell carcinoma. Laryngoscope. 2011; 121(2):294-298
[19] Jefferson GD, Sollaccio D, Gomez-Fernandez CR, Civantos F, Jr. Evaluation of immunohistochemical fine sectioning for sentinel lymph node biopsy in oral squamous cell carcinoma. Otolaryngol Head Neck Surg. 2011; 144 (2):216-219
[20] Alkureishi LW, Ross GL, Shoaib T, et al. Sentinel node biopsy in head and neck squamous cell cancer: 5-year follow-up of a European multicenter trial. Ann Surg Oncol. 2010; 17(9):2459-2464
[21] Civantos FJ, Zitsch RP, Schuller DE, et al. Sentinel lymph node biopsy accurately stages the regional lymph nodes for T1-T2 oral squamous cell carcinomas: results of a prospective multi-institutional trial. J Clin Oncol. 2010; 28 (8):1395-1400
[22] Govers TM, Hannink G, Merkx MA, Takes RP, Rovers MM. Sentinel node biopsy for squamous cell carcinoma of the oral cavity and oropharynx: a diagnostic meta-analysis. Oral Oncol. 2013; 49(8):726-732
[23] Schilling C, Stoeckli SJ, Haerle SK, et al. Sentinel European Node Trial (SENT): 3-year results of sentinel node biopsy in oral cancer. Eur J Cancer. 2015; 51 (18):2777-2784
[24] Pezier T, Nixon IJ, Gurney B, et al. Sentinel lymph node biopsy for T1/T2 oral cavity squamous cell carcinoma-a prospective case series. Ann Surg Oncol. 2012; 19(11):3528-3533
[25] Pedersen NJ, Jensen DH, Hedbàck N, et al. Staging of early lymph node meta- stases with the sentinel lymph node technique and predictive factors in T1/ T2 oral cavity cancer: A retrospective single-center study. Head Neck. 2016; 38 Suppl 1:E1033-E1040
[26] Flach GB, Bloemena E, Klop WM, et al. Sentinel lymph node biopsy in clinically N0 T1-T2 staged oral cancer: the Dutch multicenter trial. Oral Oncol. 2014; 50(10):1020-1024
[27] Broglie MA, Haile SR, Stoeckli SJ. Long-term experience in sentinel node biopsy for early oral and oropharyngeal squamous cell carcinoma. Ann Surg Oncol. 2011; 18(10):2732-2738
[28] Agrawal A, Civantos FJ, Brumund KT, et al. [(99m)Tc]Tilmanocept accurately detects sentinel lymph nodes and predicts node pathology status in patients with oral squamous cell carcinoma of the head and neck: results of a phase III multi-institutional trial. Ann Surg Oncol. 2015; 22(11):3708-3715
[29] Lim J, Govindarajulu S, Sahu A, Ibrahim N, Magdub S, Cawthorn S. Multiple Step-section Frozen Section sentinel lymph node biopsy-a review of 717 patients. Breast. 2013; 22(5):639-642
[30] Terada A, Hasegawa Y, Yatabe Y, et al. Intraoperative diagnosis of cancer metastasis in sentinel lymph node of oral cancer patients. Oral Oncol. 2008; 44(9):838-843