Southern blot analysis of neuroblastoma (NB) cell DNA from 52 patients (58 samples) allowed the detection of an N-myc amplification on three of the 13 BM samples and three of 13 tumor samples at diagnosis, on two of 17 tumor samples taken after induction therapy, on three of seven BM samples and two of the five local tumor samples taken after relapse. N-myc was amplified in two of the 15 patients with stage I to III NB and in 10 of the 36 patients with stage IV neuroblastoma over 1 year of age. Conclusions from the analysis are as follows: first tumor samples obtained on previously treated patients are histologically modified and detection of N-myc amplification is not accurate; consequently N-myc amplification must be defined at diagnosis. Surgical biopsies of the primary tumors can, however, be delayed and malignant cells obtained by ultrasound-guided punctures or analyzed on bone marrow samples if they represent more than 50% of the total population. Second, any attempt to define the prognostic value must be performed on a group of patients treated with a new aggressive protocol of chemotherapy; it will require a multiparametric analysis including, in particular, results of an extensive clinical staging at diagnosis and histological criteria.