The clinical significance of serum neuron-specific enolase (NSE) was evaluated at diagnosis (before orchiectomy) in 132 patients, and in the follow-up of 236 men with germ cell tumours of the testis. The purpose of the study was to determine whether serum NSE was a more clinically useful tumour marker than the existing tumour markers beta-hCG, AFP and LDH. The mean observation time was 4.3 years. At orchiectomy, serum NSE levels were raised in 38 (29%) of all the patients, and in 8/53 (15%) of the patients with clinical stage 1 seminoma. NSE had the same sensitivity as beta-hCG in seminoma. There were very few false-positive values for the classic markers beta-hCG and AFP, but 117 out of 219 patients (53.5%) did experience false-positive serum NSE levels at least once during follow-up. The false-positive values were wide ranging [13-70 microg/L]. None of the four evaluable patients with seminoma who experienced relapse had increased levels of serum NSE at relapse. Two out of 14 non-seminoma patients who experienced relapse had elevated serum NSE levels at the time of relapse, but increased NSE was not the first sign of relapse in any of them. It is concluded that serum NSE is of no clinical value in monitoring patientswith testicular cancer.