Use of Preoperative PET/CT Staging in Sentinel Lymph Node–Positive Melanoma

Sentinel lymph node (SLN) mapping is an integral part of the staging of cutaneous malignant melanoma. If regional metastases are identified on SLN biopsy, patients undergo completion lymph node dissection (LND) with curative intent. Microscopic regional lymph node involvement, however, is a strong predictor of disease-specific survival and delineates a population at high risk of developing metastatic melanoma.¹ The National Comprehensive Cancer Network (NCCN) makes a category 2B recommendation based on low-level evidence for staging by positron emission tomography coupled with computed tomography (PET/CT) prior to LND.² Discovery of occult synchronous metastases informs prognosis, may preclude LND, and presents an opportunity for metastasectomy, which can improve survival in patients with low-volume stage IV disease.³

Existing data suggests that PET/CT has use in the detection of metastases from multiple primary tumor types.⁴ However, PET/CT lacks data supporting its use in staging asymptomatic patients with early-stage melanoma, may inconsistently impact treatment decisions, and carries a false-positive finding risk that may detract from its use. To evaluate an evolving practice, this study aims to assess the use of PET/CT in detecting occult metastases in SLN-positive melanoma.

A retrospective evaluation of patients with melanoma and clinically silent regional lymph nodes treated at the University of Michigan between April 2013 and September 2014 was performed. Patients with palpable lymph nodes or symptoms of metastatic disease were excluded. We identified 78 patients with positive SLN mapping, 46 of whom underwent PET/CT prior to LND. Remaining patients either underwent no staging or body computed tomography with brain magnetic resonance imaging as ordered by the referring physician. Relevant clinical data were obtained from the medical record (Table 1). Accuracy of PET/CT findings was assessed according to follow-up biopsies. Outcomes measured include changes in clinical management and incidence of false-positive findings defined as biopsy-evaluated PET/CT findings nondiagnostic for melanoma. Fisher exact tests were used to compare clinical factors with PET/CT use.

The University of Michigan institutional review board approved this study.

Of the 46 patients who underwent a preoperative PET/CT, 15 (33%) had intense uptake distant from the primary tumor and local lymph node basin. Nine of those 15 patients (60%) had abnormalities biopsied prior to LND. Three of the 9 biopsies yielded metastatic melanoma, a false-positive rate of 67% for PET/CT in identifying distant metastases in asymptomatic patients. Of the 46 patients staged with PET/CT, 3 (7%) had PET/CT findings that ultimately identified metastatic melanoma and precluded LND (Table 2). There was not a statistically significant association between T stage (P = .12) and N stage (P = .85) and whether or not patients underwent PET/CT imaging. There was a statistically significant association between undergoing PET/CT and tumor ulceration (P = .004). No patients undergoing computed tomography with brain magnetic resonance imaging had findings requiring further evaluation prior to treatment.

Staging patients for the detection of clinically occult metastatic melanoma is critical because its presence may preclude morbid interventions aimed at locoregional control (eg, a completion LND). It also informs prognosis. In our retrospective review, however, we found that PET/CT has a high false-positive rate and a minimal effect on patient management in this setting, and is best reserved in its established role evaluating treatment response and recurrence. These data also suggest that a significant majority of asymptomatic patients with newly discovered microscopic SLN-positive melanoma do not have synchronous metastases detectable by PET/CT. Our review also identified the potential for treatment delays, increased exposure to risk from additional procedures, and increased health care resource usage that patients incur following detection of asymptomatic abnormalities by preoperative PET/CT. We recommend that PET/CT staging be reevaluated as a category 2B recommendation by the NCCN and its use further evaluated prospectively in a clinical trial.

Corresponding Author: Benjamin Y. Scheier, MD, 1500 E Medical Center Dr, C343 Med Inn, 5848, Ann Arbor, MI 48109-5848 (bscheier@med.umich.edu).

Published Online: September 24, 2015. doi:10.1001/jamaoncol.2015.3664.

Conflict of Interest Disclosures: None reported.