Staging of regional nodes in AJCC stage I and II melanoma: 18FDG PET imaging versus sentinel node detection.

PRIMARY PURPOSE The staging of regional nodes by means of sentinel node detection has been shown to accurately detect subclinical nodal metastases from cutaneous melanoma. On the other hand, the oncological applications of 18F-fluoro-2-deoxy-D-glucose positron emission tomography (18FDG PET) are, nowadays, firmly established. However, the sensitivity of such metabolic imaging for staging the regional nodes in primary melanoma remains debatable. We prospectively assessed the actual value of PET for detecting sentinel node metastases in 21 consecutive patients presenting with early-stage melanoma. MATERIALS AND METHODS Twenty-one melanoma patients scheduled for lymphatic mapping and sentinel lymphadenectomy underwent fully corrected whole-body PET using 18FDG. In all cases, the disease was initially classified as either stage I or II, from the latest version of the American Joint Committee on Cancer staging system. The sentinel node detection was systematically performed within the week following the PET scan. Serial sections of the sentinel nodes were analyzed by both conventional pathology and immunohistochemical staining. Metastatic sentinel nodes were also assessed for the size of tumor deposits and the degree of nodal involvement (focal, partial, or massive). The median follow-up time was 12 months. RESULTS Six of the 21 patients (28.5%) had an involved sentinel node. PET was positive in only one case with a sentinel node >1 cm. In the five other cases, the sentinel nodes missed by PET were <1 cm with focal and/or partial involvements. One patient, free of regional nodal metastases in both sentinel node detection and PET imaging, had, however, a same-basin recurrence 3 months later. In another case, PET had one false positive result. Overall, the sentinel detection of subclinical nodal metastases had a sensitivity of 86%. PET detected only 14% of sentinel node metastases. CONCLUSIONS Sentinel node detection remains the procedure of choice for detecting subclinical lymph node involvement from primary cutaneous melanoma. Owing to its limited spatial resolution, PET appears insufficiently sensitive to identify microscopic nodal metastases. As a practical consequence, metabolic imaging is not recommended as a first-line imaging strategy for staging regional lymph nodes in patients with stage I or II melanoma.