OBJECTIVE To evaluate the Sheffield Trophoblastic Tumour Centre protocol for central nervous system (CNS) involvement in high-risk patients with gestational trophoblastic neoplasia (GTN) and determine the impact of brain imaging and lumbar puncture (LP) results on subsequent clinical care. METHODS The trophoblastic tumor database was searched for patients fitting any of the following criteria registered between January 1, 1988, and December 31, 2008: hCG levels > 50,000 IU/L, high risk, > or = 2 for metastases. Placental site trophoblastic tumors (PSTTs) were excluded, and all patients with signs or symptoms suggestive of CNS involvement were investigated. Patients were to have computed tomography (CT) scan of the head and, if not contraindicated, LP to determine the ratio of cerebrospinal fluid to blood hCG level. RESULTS A total of 154 patients met > or = 1 of the defined criteria for CNS investigation. In 7 patients there was evidence of CNS involvement on CT. Only 2 cases had no clinical evidence of CNS disease-both had very-high-risk choriocarcinoma. No diagnosis of CNS disease was made on LP alone. CONCLUSIONS We propose that in the absence of neurologic symptoms or signs, only patients with choriocarcinoma need be screened. Magnetic resonance imaging head scan is preferred as the most sensitive and safe technology available.