Sentinel Node Biopsy for High-Risk Nonmelanoma Cutaneous Malignancy FREE

The presence of regional lymph node metastasis is the most powerfulprognostic factor for predicting recurrence and survival in patients withmost solid tumors. Identification of patients with occult metastasis is vitallyimportant for staging and treatment planning. Lymphatic mapping and sentinelnode biopsy have become the standard of care for melanoma at most major melanomacenters and university melanoma programs.^{1 - 3}

There is currently no adequate noninvasive means for detection of occultmetastases in regional nodes in most malignancies.⁴ Thetrend toward less invasive surgical methods to stage regional lymph nodeshas led to the gradual replacement of radical lymph node dissection proceduresby limited, selective procedures. The rationale for less invasive stagingprocedures is that they can provide prognostic information pertinent to treatmentplanning and therefore lower morbidity and cost. Sentinel node biopsy, themost recent chapter of this minimally invasive revolution, has increasinglybecome the staging procedure of choice for patients with clinically node-negativebreast cancer.^{5 - 6} Several reportson sentinel node staging for squamous cell malignancy of the male and femalegenitalia^{7 - 8} and Merkel cell carcinoma^{9 - 11} suggest that lymphaticmapping and sentinel lymph node biopsy may be used for a variety of nonmelanomacutaneous malignancies.

We postulated that the stepwise progression of early lymphatic metastasesmight occur in certain nonmelanoma cutaneous malignancies with a high riskof nodal metastases. The false-negative rate of histologic sentinel node evaluationin this population is unknown. The objective of this study was to investigatethe feasibility of sentinel node biopsy in selected high-risk skin cancersby determining the false-negative rate and sensitivity of this procedure fordetection of occult regional lymph node metastases.

We reviewed our cumulative experience with sentinel node staging innonmelanoma cutaneous malignancies in a consecutive series of clinically node-negativepatients who underwent this procedure between 1997 and 2002. The study protocolwas reviewed and approved by the institutional review board of Indiana University–PurdueUniversity and all patients signed informed consent to participate in thestudy. The 6 inclusion criteria were primary squamous cell carcinoma (SCC)at least 4 cm in diameter or invasive of deep connective, skeletal, or muscularstructures; locally recurrent SCC; SCC at least 1 cm in diameter on the genitalia;SCC at least 2 cm in diameter in immunosuppressed patients; Merkel cell carcinoma;and primary cutaneous adenocarcinoma. The 3 exclusion criteria were lymphnode metastases or adenopathy; prior wide excision greater than 2 cm or lymphaticsurgery; and if the patient was pregnant or breastfeeding. There were 2 groupsof patients (Figure 1). Group 1consisted of 9 patients who participated in a prospective nonrandomized clinicalpilot study designed to test the preliminary feasibility of sentinel nodestaging in selected high-risk cutaneous malignancies. Group 1 patients underwentsentinel node biopsy, followed by complete lymphadenectomy for all basinsidentified by lymphoscintigraphy.

Group 2 consisted of a consecutive series of 15 patients to whom thesentinel node procedure was offered as a clinical adjunct or an alternativeto staging lymph node dissection because they had high-risk cutaneous malignancies.All patients in this group also had clinically negative, ie, nonpalpable regionallymph nodes. Patient and tumor characteristics were similar in group 1 andgroup 2. Sentinel node–positive patients in group 2 were offered completelymphadenectomy or definitive radiation therapy. Sentinel node–negativepatients were followed up clinically without further therapy or underwentcomplete planned lymphadenectomy after harvesting of the sentinel node ornodes.

All patients underwent preoperative dermatolymphoscintigraphy 1 to 2hours prior to surgery with unfiltered technetium Tc 99m-labeled sulfur colloidinjected intradermally at several points around the tumor or biopsy scar.Imaging was performed for up to 2 hours using a gamma camera with a wide fieldof view to identify focal accumulations of tracer in regional lymph node basins.Intraoperative radiolocalization was performed with a handheld Neoprobe (NeoprobeCorporation, Dublin, Ohio) or C-track (Care Wise Medical Products, MorganHill, Calif) gamma probe detector. The radiolocalization was combined withintradermal injection of 1 to 2 mL of isosulfan blue to aid in visual localizationof the sentinel nodes.

All basins identified by lymphoscintigraphy were explored through limitedincisions directed by the gamma probe and skin markings. The primary criterionfor sentinel node status was blue coloration. All blue nodes were identifiedand removed as sentinel nodes. Ex vivo sentinel node to residual lymph nodebasin radioactivity ratios were determined after removal of the blue nodes.If this ratio was at least 10:1, the procedure was terminated. If the ratiowas less than 10:1, the gamma probe was used to identify and remove additionalradioactive lymph nodes until the ratio was at least 10:1.

The sentinel nodes were step sectioned at 1-mm intervals and analyzedby light microscopy using hematoxylin-eosin stain. Any lymph node with atleast 1 positive section was counted as positive for metastasis. All nonsentinelnodes were sent for standard histologic analysis.

The sensitivity of the procedure was determined by comparing the incidenceof micrometastasis in the sentinel node with the histologic findings of thecomplete lymphadenectomy in 12 patients. The 12 patients who underwent sentinellymph node biopsy as the sole procedure had a follow-up physical examinationof the residual basins in which biopsies were performed.

Characteristics of the study population are shown in Table 1. Twenty-four patients (15 men and 9 women; mean age, 61.4years [range, 32-93 years]) underwent lymphatic mapping and sentinel nodebiopsy. The histologic diagnoses were squamous cell carcinoma (n = 17), Merkelcell carcinoma (n = 5), and adenocarcinoma (n = 2). Tumor locations were thehead/neck (n = 6), extremities (n = 11), trunk (n = 1), and genitalia (n =6). Twenty-nine lymph node basins were identified by lymphoscintigraphy andstaged by sentinel node biopsy, and 55 sentinel nodes were removed (mean numberof sentinel nodes per basin, 1.9 [range, 1-4]).

Seven of 24 patients had at least 1 tumor-positive sentinel node (29%).Sentinel node biopsy findings by tumor type are shown in Table 2. Twelve patients had completion lymphadenectomy to removeall nodes in the basin where a biopsy was performed. Additional tumor-containingnonsentinel lymph nodes were noted in 3 (43%) of the 7 sentinel node–positivepatients. One sentinel node–negative patient had a positive nonsentinellymph node in a lymphadenectomy specimen. This false-negative sentinel nodebiopsy finding occurred in patient 9 (Table 1) who had locally recurrent squamous cell carcinoma of the scalpand had undergone prior resection and radiation.

Sentinel lymph node biopsy without complete lymphadenectomy was performedin 12 patients. Biopsy findings revealed positive sentinel nodes in the cervicallymph nodes of 2 (17%) of these patients who were then offered radiation therapy.The median clinical follow-up for 10 sentinel node–negative patientswas 11 months (range, 4-41 months). No nodal recurrences were noted in anysentinel node–negative basin. Compared with all clinical information,the sensitivity of the sentinel node procedure was 88% (7/8), its specificitywas 100% (16/16), its positive predictive value was 1.0 (7/7), and its negativepredictive value was 0.94 (16/17).

Nonmelanoma skin cancer is the leading cause of cancer in the UnitedStates, with more than 1 million new cases diagnosed annually.^{12 - 16} Basalcell carcinoma, which very rarely metastasizes, accounts for most nonmelanomacutaneous carcinomas¹⁷ while SCC, which accountsfor about 20% of nonmelanoma cutaneous carcinomas, is increasing in incidence.^{13 - 16 ,18} Approximately80% of initial treatment failures in cutaneous SCC occur in regional lymphnodes.^{19 - 20} The overall incidenceof regional nodal metastases is relatively low (only about 0.5%) in all cutaneousSCCs, but the presence of metastasis carries a poor prognosis.^{19 - 20} Merkelcell carcinoma and primary cutaneous adenocarcinomas are uncommon skin malignancies,with a relatively poor prognosis and a significant chance of regional nodemetastases.^{9 - 11 ,21} The5-year survival rate for stage III disease is approximately 25% to 35%, dependingon histologic findings and tumor burden.^{19 - 22}

The optimal approach to occult nodal metastases in patients with high-riskcutaneous nonmelanoma malignancies has not been standardized by prospectiverandomized clinical trials. Elective surgical lymphadenectomy or nodal irradiationare frequently considered in the treatment planning for these malignancies.Nodal staging may contribute prognostic information useful in preventing unnecessarysurgery or other therapies.

The present study demonstrates the usefulness of sentinel node biopsyfor the staging of clinically node-negative patients with several high-riskskin malignancies. These results parallel our larger experience with treatmentof melanoma²³ and suggest a clinical role forminimally invasive staging in selected nonmelanoma cutaneous malignancies.The study's small size and limited follow-up time preclude definitive conclusionsregarding the sensitivity and false-negative rates to be expected from theprocedure. Nonetheless, it appears reasonable to offer the procedure to selectedpatients as an alternative to observation or to elective complete lymphadenectomy.

The false-negative rate in this series was 12%, which is considerablyhigher than rates reported for large series of patients with melanoma.²³ It is likely due to the small size of this series,and would likely be lower with larger numbers and more precise patient selectionfor the procedure. The patient with the false-negative result of sentinelnode biopsy had a relatively large recurrent scalp SCC after excision andradiation therapy. Lymphatic mapping is known to be somewhat less reliablewhen the normal lymphatic drainage has been altered, eg, after previous wideexcisions, flap reconstruction, and lymphadenectomy. The effect of radiationon the reliability of lymphatic mapping is not known, but is likely similar.And although sentinel node biopsy is generally accepted in the managementof melanomas in the head and neck region, it is more difficult owing to complexlymphatic drainage patterns. These factors may have contributed to the singlefalse-negative observation.

Because of the rarity of high-risk nonmelanoma cutaneous tumors, experiencewith sentinel node biopsy is limited and large clinical trials may not befeasible. Lymphatic mapping has been investigated in SCC of the vulva. Earlyexperiences using only blue dye showed a relatively low sensitivity,²⁴ but more recent reports using the combined techniquesof blue dye and radiolocalization have been highly accurate in establishingthe pathological status of the inguinofemoral node.⁸ Goodresults have also been achieved in the staging of early penile cancer.⁷ There are several reports of successful sentinel nodebiopsy for Merkel cell carcinoma and the procedure seems to be applicableto this rare malignancy.^{9 - 11} Earlyreports suggest that sentinel node biopsy may have an emerging role for stagingof head and neck mucosal cancer.^{25 - 26}

To date, only very small or anecdotal reports of sentinel node stagingfor cutaneous SCC exist.^{27 - 28} Thefrequency of regional nodal metastases varies with anatomic site, histologicfeatures such as increasing tumor thickness and histologic dedifferentiation,tumor size, host immune competency, perineural invasion, and prior treatment.^{19 ,28 - 35} Squamouscell tumors arising in nonglabrous mucocutaneous sites such as the lip, vulva,penis, and perianal area are also more likely to metastasize than those involvingother areas of the skin.²⁹ Squamous cell carcinomaarising in areas of chronic inflammation, nonhealing wounds, chronic osteomyelitis,and in irradiated fields are known to be particularly aggressive, with ratesof nodal metastases between 18% and 30%.^{31 - 34} Immunosuppressedpatients also have a higher incidence of nonmelanoma skin cancer metastases.The presence of perineural infiltration is another recognized determinantof metastatic potential.^{22 ,34}

Based on this series and the prognostic factors in the literature, itis possible to formulate rational recommendations for nodal staging in patientswith nonmelanoma cutaneous malignancies. Sentinel node biopsy should be consideredin all patients with Merkel cell carcinoma clinically localized to the skin.We also advocate considering sentinel node staging for cutaneous SCC in patientswith a primary tumor greater than 2 cm in diameter, any tumor invasive ofunderlying skeletal structures, and any SCC in a chronic wound, with perineuralinvasion, or a thickness greater than 4 mm. Squamous cell carcinomas thatare less than 2 cm in diameter that are poorly differentiated, develop onthe genitalia, recur after therapy, and develop in immunocompromised patientsmay also be considered for nodal staging with sentinel node biopsy. The rarityand clinical diversity of primary cutaneous adenocarcinomas and the smallsize of this series prevent any meaningful conclusions. However, sebaceouscarcinomas, malignancies of eccrine origin, and extramammary Paget diseasehave the potential to metastasize to regional lymph nodes and may be consideredfor sentinel node staging. These preliminary recommendations are empiric andawait confirmation in larger numbers of patients.

Sentinel node biopsy has the potential to improve the clinical managementof selected patients with nonmelanoma cutaneous malignancies. Sentinel nodestaging theoretically could spare node-negative patients the morbidity ofcomplete lymphadenectomy—or empiric radiotherapy. Because observationof clinically negative regional lymph nodes is an accepted management option,patients without disease in sentinel nodes could be selected for close monitoringwithout further therapy. Conversely, patients with micrometastatic diseasein lymph nodes could be appropriately treated with lymphadenectomy or radiationtherapy. Finally, more accurate staging information permits more precise designand analysis of clinical trial results.

In conclusion, minimally invasive staging of clinically node-negativeregional lymphatic basins with sentinel node biopsy is applicable to a varietyof nonmelanoma cutaneous malignancies. Further studies are indicated to verifythese findings and refine guidelines for sentinel node staging in nonmelanomaskin malignancies.

Corresponding author: Jeffrey D. Wagner, MD, RT 471, 535 BarnhillDr, Indianapolis, IN 46202 (e-mail: jdwagner@iupui.edu).

Accepted for publication September 24, 2003.