Evaluation of S-100 beta assays for the prediction of recurrence and prognosis in patients with AJCC stage I-III melanoma

We performed prospective serial studies on 266 melanoma patients undergoing surgery for high risk primary or lymph node metastases to assess the predi ctive value for recurrence of melanoma of S-100 beta detection in the serum using immunoradiometric assay (IRMA) or immunoluminometric assay (LIA-mat) . The tests for S-100 beta were most frequently positive in the first 3 mon ths after surgery. Results of the assays did not show a strong correlation with clinical stage of the disease. Studies on a cohort of 147 patients who had been followed for a minimum of 2 years post-surgery were carried out t o assess the sensitivity of the assays for identifying patients who develop recurrence of their melanoma. The LIA-mat and IRMA assays detected S-100 b eta in the serum of 47.5% and 23%, respectively, of the subset of 61 patien ts who developed recurrence of disease. By Kaplan Meier analysis, patients positive for S-100 beta by LIA-mat assay in the first 3 months post-surgery were shown to have an approximately 2.7 times shorter disease-free surviva l (DFS) period than patients negative for S-100 beta. This was significant by log rank analysis. Multivariate analysis indicated that the presence of S-100 beta was an independent prognostic determinant of DFS, and was indepe ndent of tumour thickness and lymph node status. This prospective analysis in a large number of patients suggests that assays for S-100 beta in patien ts following surgical resection of AJCC stage I-III melanoma are of limited value for identifying patients who will develop recurrence of their diseas e. The results of the LIA-mat assays, but not the IRMA assays, do however p rovide an additional measure of prognosis that is independent of existing m easures. (C) 1999 Lippincott Williams & Wilkins.