Sentinel Lymph Node Radiolocalization in Head and Neck Squamous Cell Carcinoma

Authors


  • Presented at the Meeting of the Eastern Section of the American Laryngological, Rhinological and Otological Society, Inc., Providence, Rhode Island, January 31, 1999.

    Supported by a grant from the American Cancer Society (no. IRG58-012-H).

Abstract

Objectives: To determine the feasibility of sentinel node radiolocalization in stage N0 in head and neck squamous cell carcinoma and to gain insight as to whether the sentinel node could be prognostic of regional micrometastatic disease.

Study Design: A prospective report on the application sentinel node radiolocalization in eight patients with N0 squamous cell carcinoma of the head and neck region.

Methods: For each patient a peritumoral submucosal injection of filtered technetium (99mTc) prepared with sulfur colloid was performed immediately following intubation. After at least 30 minutes, focal areas of accumulation corresponding to a sentinel node were marked on the skin surface. Complete neck dissections were performed, and the sentinel nodes were identified for later histological evaluation and comparison to the remaining lymphadenectomy specimen.

Results: Sentinel node radiolocalization accurately identified two or more sentinel lymph nodes in all eight cases. In one patient, two of the three lymph nodes containing micrometastatic disease were sentinel lymph nodes. There was no instance in which sentinel node was negative for micrometastatic disease while being positive in a nonsentinel lymph node.

Conclusions: Accurate localization of the sentinel lymph node using radiolabeled sulfur-colloid is feasible in patients with squamous cell carcinoma of the head and neck region. Although sentinel node radiolocalization in head and neck squamous cell cancer may potentially reduce the time, cost, and morbidity of regional lymph node management, more experience with technique is required before its role can be determined.

INTRODUCTION

In oral cavity, oral pharyngeal, and supraglottic squamous cell carcinoma of the head and neck, the incidence of occult metastases can be estimated by several prognostic factors (site, size and depth of tumor, perineural lymphatic vascular invasion, and host tumor interface) and ranges from 15% to 60%. 1 Management of the N0 neck has been advocated when the risk of micrometastases is approximately 30%. 1,2 Treatment modalities include radiation therapy and elective neck dissection.

Radiation therapy has several shortcomings. 2 First, patients without metastases are treated unnecessarily. Second, once irradiated, patients with new primaries or recurrences cannot use this treatment modality. Third, patients are unable to have histological staging of their lymphatics and thus are denied the histological information necessary for precise treatment planning. In contrast, elective neck dissection allows complete pathological evaluation and conclusive assessment of the regional lymph node status. When micrometastases are identified, the patient's planned therapy can be modified to reflect the more serious nature of the disease. The principle disadvantage of elective neck dissection is that a significant number of “at-risk” patients do not have micrometastases and thus are unnecessarily subjected to the time, expense, and morbidity of lymph node resection.

Continually aware of the disadvantages, surgeons over the last century, as they have learned more about the metastatic spread of cancer, have slowly modified their philosophy toward elective lymph node dissection. Progressing from the radical neck dissection as introduced by Crile 3 in 1906, to the modified radical neck dissection as discussed by Suarez 4 and Bocca and Pignataro, 5 and then to the selective neck dissection as introduced by the surgeons at M.D. Anderson Cancer Center, 6 elective neck dissection has evolved from a generalized en block resection to a focused surgery that attempts to remove only those nodes at highest risk.

Currently, surgical judiciousness guides the extent of regional lymph node resection for all at-risk patients. However, if a precise, minimally invasive method of diagnosing metastatic disease were available, treatment would be based on the actual, not the theoretical presence of micrometastases. The surgical emphasis would shift from decreasing the extent of surgical resection to restricting the number of patients who undergo surgery. Only patients with proven metastasis would have treatment, and the 55% to 85% of at-risk but nonaffected patients would avoid the morbidity of presumptive elective treatment.

In treating the regional lymphatics in cutaneous melanoma, such a shift in the surgical philosophy has occurred. Several authors were able to show that if the first draining lymph node (sentinel node) had no evidence of micrometastases on histological examination, there was a less than 5% incidence of micrometastases to the remaining lymphatic basin. 7–9 Based on this proven predictive relationship between the sentinel node and its lymphatic basin, melanoma patients receive lymphadenectomy only when the biopsied sentinel node is positive for micrometastatic disease on histological examination.

In 1993, in response to the need for a minimally invasive method of sentinel node localization, we developed a technique called γ-probe radiolocalization, which could transcutaneously identify the sentinel node and facilitate biopsy through a small incision. 10,11 γ-probe radiolocalization uses a peritumoral injection of a radionuclide/sulfur compound that percolates through the lymphatics to the first draining lymph node, where it accumulates, producing a “hot” node. This hot node can then be identified on the skin surface and a biopsy specimen obtained through a small incision using the handheld γ-probe. Since its introduction 6 years ago, γ-probe radiolocalization and biopsy of the sentinel node have been documented by several authors to be successful approximately 95% of the time. 7–11 In head and neck melanoma, we recently reported a 96% successful localization rate. 12

In 1996, Alex and Krag 13 first described successful radiolocalization of sentinel node in the aerodigestive system in a patient with supraglottic squamous cell cancer. To this case, we add the following case reports and our experience with this modality over the past 6 years.

MATERIALS AND METHODS

The patients in this study were treated in a teaching hospital in a tertiary care setting in accordance with the Human Investigative Committee and Radiation Safety Committee of each author's respective institution. The eight patients presented in this study had T1 to T4, N0 squamous cell carcinoma of the oral cavity, oropharynx, supraglottis, or larynx and at the time of this writing are part of an ongoing prospective, consecutive series study. Pregnant patients are not eligible, and if clinically indicated, a pregnancy test is administered to rule out pregnancy.

Only patients with a diagnosis of N0 squamous cell cancer of the head and neck as determined by biopsy of the primary tumor and clinical and computed tomography evaluation of the neck were eligible for the study. All patients who met the study criteria had radionuclide localization performed on the morning of surgery.

The radioactive tracer used was technetium 99m (99mTc) prepared with sulfur colloid (CIS-US, Bedford, MA). The tracer (0.1 mCi in 1–2.5 mL) was injected intradermally around the circumference of the primary tumor 10,11 (Fig. 1). After approximately 30 minutes, focal areas of accumulation (“hot spots”) corresponding to sentinel nodes were marked on the skin surface (Fig. 2). Measurements of the accumulated radioactivity in the radiolabeled lymph nodes were performed with a handheld γ-detector (C-Trak, Care Wise Medical Products, Morgan Hill, CA). The γ-probe is a scintillation probe consisting of a sodium iodide crystal coupled to a photomultiplier tube, both of which are contained within a tungsten alloy. The detector element is connected to a preamplifier and a signal processor with a digital and analogue read-out of the radioactive emissions. A focal area of radionuclide accumulation was readily discerned by listening to the γ-detector audio signal, which increases in pitch as the emission level increases. The point of maximal emission identified the sentinel node.

Figure Fig. 1..

Peritumoral injection of radionuclide tracer.

Figure Fig. 2..

Survey of the lymphatic basin for sentinel lymph nodes and localization on the skin surface.

The significance of a given emission measurement is derived from the Poisson distribution, which relates the SD to the mean number of disintegrations according to the formula below:MATH where NL = counts of the hot spot or sentinel lymph node and NX = mean counts of the adjacent tissue or the lymphatic bed after sentinel node biopsy. To achieve an emission measurement precision of 98%, the counts for a given hot spot or radiolabeled lymph node must be greater than background counts by three SDs. In practical terms, to qualify as a hot spot, a discrete area was at least 15 counts in 10 seconds and three times the count of the background. 10,11 For each patient, the following γ-count measurements were obtained: 1) the hot spot/node in vivo before incision, 2) the hot spot/node ex vivo, 3) the lymphatic bed after hot spot/node removal, and 4) the background in the operating room. After localization on the skin surface, the skin flaps for neck dissection were raised and the handheld γ-probe, sterilely wrapped, was used to precisely identify any radiolabeled nodes. The patient then received elective neck dissection and removal of the primary tumor according to current accepted treatment standards. After removal, the lymphadenectomy specimen was placed on a side table and the radiolabeled sentinel lymph nodes were dissected from the specimen. The sentinel lymph nodes and the lymphadenectomy specimen were submitted separately for histological evaluation (Fig. 3). The incidence of micrometastases in the sentinel lymph nodes and the remaining lymphadenectomy specimen was compared.

equation image(1)
Figure Fig. 3..

Ex vivo identification and separation of the sentinel lymph nodes from the lymphadenectomy specimen for histological evaluation.

CASE REPORTS

Case 1

The patient was a 72-year-old man with a T2N0M0 squamous cell carcinoma of the midline anterior floor of mouth. Two milliliters of 99mTc sulfur colloid (1 mCi) in 0.25–mL aliquots was circumferentially injected around the tumor. γ-Probe radiolocalization and identification of the sentinel lymph nodes were performed on the skin surface and in vivo. Bilateral type III modified radical neck dissections and excision of the primary were performed. The sentinel lymph nodes were dissected from the lymphadenectomy specimen and submitted separately. Both right-side and left-side lymphadenectomy specimens were negative for micrometastases.

Case 2

The patient was a 54-year-old woman with T2N0M0 supraglottic carcinoma of the left aryepiglottic fold and the left false cord with extension to the midline. Using direct laryngoscopy and a modified Bruening pistol-grip syringe, 2.5 mL 99mTc sulfur colloid (1 mCi) in 0.5-mL aliquots was circumferentially injected around the tumor. γ-Probe radiolocalization of the sentinel lymph node was performed on the skin surface and in vivo. Bilateral type III modified radical neck dissections and a supraglottic laryngectomy were performed. The sentinel lymph nodes were dissected from the lymphadenectomy specimen and submitted separately. Histological evaluation of the right-side neck lymphadenectomy specimen, which had no areas of focal accumulation, revealed that 0 of 17 nodes were positive for metastatic tumor. In the left-side neck lymphadenectomy specimen 3 of 21 nodes were positive for squamous cell carcinoma. Of the three nodes with micrometastases, two were radiolabeled sentinel lymph nodes.

Case 3

The patient was a 58-year-old man with a T2N0M0 squamous cell carcinoma of the left side of the floor of the mouth. One milliliter of 99mTc sulfur colloid (1 mCi) in 0.1-mL aliquots was circumferentially injected around the tumor. γ-Probe radiolocalization of the sentinel lymph node was performed on the skin surface and in vivo. Bilateral type III modified radical neck dissections and excision of the primary were performed. The sentinel lymph nodes were dissected from the lymphadenectomy specimen and submitted separately. Both right and left lymphadenectomy specimens were negative for micrometastases.

Case 4

The patient was a 67-year-old man with a T2N0M0 squamous cell carcinoma of the right-side anterolateral tongue. One milliliter of 99mTc sulfur colloid (1 mCi) in 0.1-mL aliquots was circumferentially injected around the tumor. γ-Probe radiolocalization of the sentinel lymph node was performed on the skin surface and in vivo. Bilateral type III modified radical neck dissections and excision of the primary were performed. The sentinel lymph nodes were dissected from the lymphadenectomy specimen and submitted separately. Both right and left lymphadenectomy specimens were negative for micrometastases.

Case 5

The patient was a 62-year-old man with a T2N0M0 squamous cell carcinoma of the laryngeal surface of the epiglottis and the superior portion of the left false cord. One milliliter of 99mTc sulfur colloid (1 mCi) in 0.1-mL aliquots was circumferentially injected around the tumor using a 21-gauge “butterfly” needle attached to an upbiting laryngeal biopsy forceps. γ-Probe radiolocalization of the sentinel lymph node was performed on the skin surface and in vivo. Bilateral type III modified radical neck dissections and supraglottic laryngectomy were performed. The sentinel lymph nodes were dissected from the lymphadenectomy specimen and submitted separately. Both right and left lymphadenectomy specimens were negative for micrometastases.

Case 6

The patient was a 63-year-old woman with a T2N0M0 squamous cell carcinoma of the laryngeal surface of the epiglottis. One milliliter of 99mTc sulfur colloid (1 mCi) in 0.1-mL aliquots was circumferentially injected around the tumor using a 21-gauge butterfly needle attached to an upbiting laryngeal biopsy forceps. γ-Probe radiolocalization of the sentinel lymph node was performed on the skin surface and in vivo. Bilateral type III modified radical neck dissections and supraglottic laryngectomy were performed. The sentinel lymph nodes were dissected from the lymphadenectomy specimen and submitted separately. Both right and left lymphadenectomy specimens were negative for micrometastases.

Case 7

The patient was a 77-year-old white man with a T4N0M0 squamous cell carcinoma of the anterior larynx with extension into the subglottis. One milliliter of 99mTc sulfur colloid (1 mCi) in 0.1-mL aliquots was circumferentially injected around the tumor using a 21-gauge butterfly needle attached to an upbiting laryngeal biopsy forceps. γ-Probe radiolocalization of the sentinel lymph node was performed on the skin surface and in vivo. Bilateral type III modified radical neck dissections and a total laryngectomy were performed. The sentinel lymph nodes were dissected from the lymphadenectomy specimen and submitted separately. Both right and left lymphadenectomy specimens were negative for micrometastases.

Case 8

The patient was a 50-year-old white man with a T3N0M0 squamous cell carcinoma of the right retromolar trigone. One milliliter of 99mTc sulfur colloid (1 mCi) in 0.1-mL aliquots was circumferentially injected around the tumor. γ-Probe radiolocalization of the sentinel lymph node was performed on the skin surface and in vivo. Right-side type III modified radical neck dissections and excision of the primary were performed. The sentinel lymph nodes were dissected from the lymphadenectomy specimen and submitted separately. Both right and left lymphadenectomy specimens were negative for micrometastases.

RESULTS

Sentinel node radiolocalization accurately identified two or more sentinel lymph nodes in all eight cases. In one patient, two of the three lymph nodes containing micrometastatic disease were sentinel lymph nodes. There was no instance in which sentinel node was negative for micrometastatic disease while being positive in a nonsentinel lymph node (Table I).

Table Table 1.. Sentinel Lymph Node Results.
original image

DISCUSSION

The theory that certain lymph nodes within a regional lymphatic basin are prognostic of regional tumor spread has several precedents. One of the more colorful descriptions was made by a fourth-year Harvard medical student, Raymond H. Randal, in 1948. Under the tutelage of J.H. Means, Randal noted the potential relationship between a prelaryngeal node and the clinical behavior of laryngeal tumors. He creatively dubbed this node the “delphian node” after the prophetic oracle at Delphi. 14 In similar fashion, the concept of the sentinel lymph node is critical to understanding the relationship of a primary tumor to its regional lymphatics. As first described by Cabanas 15 in 1977, only a limited number of lymph nodes (sentinel lymph nodes) are the initial recipients of micrometastases. Cancer cells, as they leave the primary tumor, move sequentially down the lymphatic ducts to the first draining lymph nodes (sentinel lymph node), then to second and third draining lymph nodes. The sentinel node, being the initial recipient of metastatic tumor cells, predicts the histological features of its remaining lymphatic basin. In cutaneous melanoma, sentinel lymph nodes accurately reflect whether regional metastases have occurred with a demonstrated false-negative rate of 4% and a predictive value of a negative sentinel lymph node of 98.5%. 10,11 Critics of the sentinel node hypothesis and of γ-probe radiolocalization with biopsy as a staging tool for squamous cell carcinoma of the head and neck cite the same concern as that for selective neck dissection. Namely, as evidenced by skip metastases, nodal metastases do not always spread in a predictable and sequential manner. Such concerns would be valid if a generalized anatomical chart could describe each person's unique lymphatic drainage system. However, recent studies using lymphoscintigraphy have challenged the accuracy of traditional lymphatic charts and shed new light on the theory of “skip metastases.” According to these studies, 16,17 a given individual's unique lymphatic drainage pattern, as determined by lymphoscintigraphy, can be discordant from accepted anatomical charts as much as 60% of the time. Thus skip metastases may in fact represent nonstandard but sequentially draining lymphatic pathways and/or multiple draining lymphatic groups in parallel array. According to the sentinel node theory, regardless of how nonstandard or unique a patient's lymphatic drainage system may be, the sentinel node will always be the initial site of micrometastases. Therefore, as with cutaneous melanoma, evaluation of the histological status of the sentinel node may predict the incidence of regional micrometastases in head and neck squamous cell carcinoma.

Although all eight cases presented above demonstrate that γ-probe radiolocalization of the sentinel node in squamous cell carcinoma of the head and neck is technically feasible, the first case study demonstrates the potential benefit of this technique. Of the three nodes containing micrometastases, two were the radiolabeled lymph nodes that were identified using the handheld γ-probe. In this particular case, the sentinel node was reflective of the remaining basin. Also, in the eight cases presented, there was no instance in which a nonsentinel lymph node was positive for micrometastases when the sentinel node was negative.

These findings raise the possibility that the sentinel node could be prognostic in squamous cell carcinoma of the head and neck. If true, staging could be achieved by a minimally invasive biopsy. The histological results would provide the precise indications for surgery or radiation therapy (Fig. 4). Patients without micrometastases may require only excision of the primary and sentinel node evaluation. Although as yet unproven, the potential for selectively tailoring treatment of the patient with N0 disease and of reducing the current cost, time, and morbidity of treatment is intriguing and warrants continued study.

Figure Fig. 4..

Potential algorithm for the management of N0 squamous cell carcinoma of the head and neck.

The technique of sentinel node localization, although relatively easy to learn, does have several notable points. The choice of lymph node tracer—blue dye or 99mTc sulfur colloid—is critical. Based on our 6-year experience with melanoma, blue dye mapping technique has the following drawbacks 10–13 : 1) It is difficult to determine the precise location of the blue-stained nodes before skin incision, resulting in unnecessary dissection; 2) substantial experience is required to develop the technical skill necessary to achieve a high localization success rate; 3) it is difficult to verify complete removal of all first-draining lymph nodes; 4) the lymph nodes are stained for only 15 minutes, which often necessitates reinjection of the blue dye; and 5) the blue dye characteristically causes staining of local tissues and is cumbersome to use in the head and neck region. With aerodigestive tumors, the lattermost drawback is particularly important. Unlike with the skin, surgical access to the aerodigestive system is limited and the nature of the anatomy predisposes to pooling of the dye and staining of adjacent tissues. In a recent study using blue dye to label lymph nodes in squamous cell carcinoma of the head and neck, the authors, unable to identify a single sentinel node, terminated their study after 16 patients. 18

Another problem occurs when the lymphatic bed being surveyed is adjacent to the radionuclide injected primary (i.e., the evaluation of level I nodes that are next to a floor-of-mouth lesion). Shielding the handpiece from scatter by adding a collimator, changing the angulation of the handpiece, shielding the primary with lead, and adjusting the window and threshold parameters of the scintillation counter have all been effective maneuvers when identifying the sentinel lymph node in this particularly difficult situation.

CONCLUSION

In melanoma, γ-probe radiolocalization allows noninvasive identification of the sentinel node on the skin surface and biopsy through a small incision using the handheld γ-probe. If applicable to squamous cell carcinoma of the head and neck, γ-probe radiolocalization of the sentinel node can potentially elucidate each person's unique drainage pattern, provide a cosmetically appealing and minimally invasive method of sentinel node localization and biopsy, assist in the staging of the regional lymphatics, and aid in determining whether a regional lymphadenectomy or radiation therapy would be indicated.

This study demonstrates that identification of the sentinel lymph node is feasible using radiolocalization and provides the first suggestion that the sentinel node is prognostic in squamous cell carcinoma of the aerodigestive system. Because of the potential for selectively tailoring treatment of the patient with N0 disease and for reducing the current cost, time, and morbidity of treatment, further study of this approach is warranted.

Ancillary