FC7 Role of lymph-node ultrasonography (US) in the follow-up of melanoma patients to detect nodal recurrence after sentinel lymph node biopsy (SNLB): a prospective cohort study

Abstract

Nodal recurrence occurs in ♯ 10% of pts after negative sentinel lymph node biopsy (SNLB), and best follow-up strategy after SNLB has not been established. We studied prospectively the diagnostic value of US, a technique that has been shown in pts without prior SNLB (but not after SNLB) to be more sensitive and specific than palpation to detect nodal recurrence after SNLB. Methods From 9/03 to 12/08, 1256 new melanoma pts from Paris university hospitals were entered prospectively into MelanCohort (33). SNLB procedures have been standardized among participating hospitals. Every clinical and paraclinical nodal examinations were recorded as normal, doubtful, or highly suggestive of recurrence. We analyzed follow-up records of all patients who had SNLB procedure until nodal recurrence or to the 1/11/09. Gold standard for nodal status was histology or uneventful follow-up. Results Follow-up visits were recorded for 80% of the 452 pts who had SNLB procedure (22% with nodal invasion). Average melanoma thickness was 2.7±2.7 mm; 34% were ulcerated. 61% of the pts had ≥1 nodal US during follow-up (mean 3.4±2.7/pt). Nodal recurrence occurred in 13% of the pts, synchronous of distant metastasis in 17% only. Using the criteria ‘highly suggestive of recurrence’, US detected nodal recurrence with 88% sensitivity, 96% specificity; and positive and negative likelihood ratio of 21, and 0.12, respectively. 23% of the pts with US had ≥1 doubtful US result, justifying a control (usually negative) within 1–2 m or cytology. The initial mode of discovery of nodal recurrence was the patient himself (17%), clinical (38%) or paraclinical (30%) exam, or not specified (15%). Pts followed with US had similar melanoma thickness, ulceration or SNLB invasion rate than those without US surveillance. Using a Cox model in pts with negative SNLB, overall survival was better for lower AJCC stages (P<0.0001) and marginally better when US was used during follow-up (P=0.1). Conclusion US detected 30% of nodal recurrences after SNLB. As surgery is indicated in most of these pts, this may translate into better pt survival. This study is not randomized. The impact of US on pt's anxiety was not analyzed.