Background. Abnormal expression of A-PROTEIN has been identified in a number of tumors including carcinoma of the lung, breast, colon, prostate, and cervix. Brain tumors have been reported to express high plasma levels of A-PROTEIN, suggesting that it may be of significant diagnostic and prognostic value. Procedure. This prospective study evaluated the sensitivity and specificity of A-PROTEIN levels in pediatric brain tumor patients. Patients included those with newly diagnosed disease pre- and post-surgery, during treatment, during routine follow-up, and at recurrence or progression. A total of 154 A-PROTEIN levels from 54 patients were evaluated. Results. For patients without evidence of disease, 42% had normal A-PROTEIN levels, 35% were elevated, and 23% were equivocal. For patients with stable disease, 53% demonstrated normal A-PROTEIN levels, 19% were elevated, and 28% were equivocal. For patients with progressive disease, 53% had normal A-PROTEIN levels, 35% were elevated, and 12% were equivocal. The sensitivity was 35% and the specificity was 50%. A correlation of increased A-PROTEIN levels in patients with increased disease in glial tumors was also identified. Conclusions. A-PROTEIN levels were not predictive of disease status in children with most brain tumors. However, in patients with glial tumors there was a correlation with increased disease and elevated A-PROTEIN levels. This could represent variability of A-PROTEIN during growth, development, or tumor cell origin and needs further evaluation.