In a series of 130 consecutive patients suffering from small cell lung cancer (SCLC), we compared response evaluations according to standard criteria of the WHO with response evaluations according to changes in the neuron-specific enolase (NSE) levels during systemic therapy. For assessment by changes in the marker levels, the difference between two consecutive levels must exceed 30%. This value is based on the formula: Dc = 2(square root 2) x CV (CV: inter-assay coefficient of variation of the NSE test, set at 10 %). Of the 130 patients who entered this study, 18 patients received best supportive care and were excluded from the therapy monitoring. In the remaining 112 patients, 502 evaluations for response to therapy by both methods were performed, ie. 4.5 observations per patient. We found a concordance between the response evaluations according to the marker criteria and the clinical assessment in 69.7 % of the observations when including cases with positive lead-time and those with a temporary drop in the NSE levels due to a short-term response to therapy. The latter cases met the criteria consistent with the clinical evaluations at the next observation. The concordance with the clinical response evaluation increased to 84.2% when considering only those changes in the NSE levels where at least one of the consecutive marker levels was in the pathological range (> 14.5 ng/ml). Most discordant results were due to insufficient changes in the NSE levels at clinical remission or progression. A further limitation to the general use of NSE for therapy monitoring was founded on the marker negativity throughout the follow-up period, despite tumor progression or relapse. Changes in the levels between pretreatment NSE and after the first cycle of chemotherapy were shown to provide prognostic information. Patients with a drop in the NSE levels proved to have a significantly better survival probability than those with unchanging or rising marker values (p = 0.004). It is concluded that in the majority of evaluations, changes in the NSE levels are consistent with clinical findings based on imaging techniques but remain of doubtful utility in an individualpatient. NSE measurement can only be recommended as an adjunct to the clinical assessment in the follow-up of SCLC patients.