Aims: A multivariate analysis was performed to evaluate the impact of various prospectively evaluated risk factors.
Patients and methods: From January 1987 to December 1993, 120 patients were registered in the study. 108/120 patients were eligible. There were 49 girls and 59 boys with a median age at diagnosis of 14 months (range, 0 to 224 months). Patients were classified according to the Evans classification system. LDH, NSE, Ferritin, N-myc amplification, 1p-deletion and ploidy were evaluated at diagnosis. Treatment intensity was based on the results of primary surgery: surgery only for 22 (20%) stage I and IIA patients (macroscopic residue without lymph node involvement), the 9 IIB patients (8,5%) (macroscopic residue with lymph node involvement) had mild chemotherapy in addition (6 x VCR/CYC) and elective radiation (Rx). Stage III patients were divided into 2 groups: IIIA patients (n = 17/16%) had to have ferritin levels under 300 micrograms/ml, NSE lower than 100 ng/ml and age below 2 years at diagnosis and received 6 alternating cycles of DAMO/ MVDOC. If one of these three criteria was not fulfilled, patients were assigned to the more intensive treatment arm stage IIIB (n = 12/11%), i.e. 9 alternating cycles of DAMO/MVDOC/IPE. Stage IV pts (n = 35/32.5%) received 8 MVDOC/IPE cycles and 20 patients received megatherapy followed by stem cell reinfusion in addition. 13 stage IVs patients (12%) were registered and had elective VCR/CYC and/or liver radiation in case of poor clinical condition. The median observation time is 4.2 years (range, 1 to 7.5).
Results: The survival rate at 3 years was excellent for localized disease and stage IVs with survival rates of 100% for stage I/IIA and 92% for stage IVs. Stage IIIA patients had an EFS rate of 81% whereas stage IIB patients achieved only 69%. Stage IV patients reached 51%, however outcome was especially poor for stage IIIB patients (20%) due to treatment related toxicities. The toxic death rate in the study was 13% (2 surgical deaths, 8 infections, 4 multiple organ failures). Univariate analysis demonstrated the following significant unfavorable risk factors: age over 1 year at diagnosis (58/108 pts, p = 0.006), NSE > 100 ng/ml (26/95 pts, p < 0.0001), Ferritin > 300 ng/ml (19/98 pts, p = 0.007), LDH > 300 (400) U/L, 51/87 pts, p = 0.004), presence of N-myc amplification (17/59 pts, p = 0.001), deletion of the short arm of chromosome 1 (19/74 pts, p < 0.0001) and di/tetraploidy (32/72 pts, p = 0.008). The power of these factors was even stronger in patients with localized disease whereas no significant prediction was observed for stage IV patients. Furthermore a significant correlation of the serological (NSE, ferritin, LDH) and biological factors (N-myc, deletion 1p, di/tetraploidy) was detected in this study. Only NSE was identified as an independent prognostic factor (p = 0.018) whereas no independent factor could be identified within the 3 biological parameters due to their high correlations (Kendall's tau for N-myc and deletion 1p:0.7). However, N-myc (p = 0.005) as well as deletion 1p (p = 0.01) were found significantly more important than di/tetraploidy.
Conclusions: Biological classification of neuroblastomas should be mandatory and be the prerequisite for any risk adapted treatment. One serological and 2 biological factors could be a good standard evaluation to identify neuroblastoma patients at risk.