SCREENING INFANTS FOR NEUROBLASTOMA DOES NOT REDUCE THE INCIDENCE OF POOR-PROGNOSIS DISEASE

The Quebec Neuroblastoma Screening Project (QNSP) was initiated to det ermine whether mass screening would reduce mortality from this tumor i n a large cohort of infants. A total of 476,603 children were born in the province of Quebec (PQ) from May 1, 1989 through April 30, 1994 an d were eligible for urinary homovaniliic acid (HVA) and vanillylmandel ic acid (VMA) determination at 3 weeks and 6 months of age, infants wi th positive screening were referred to one of four pediatric cancer ce nters in Quebec for uniform evaluation and treatment. Standardized inc idence ratios (SIRs) were calculated for PQ (screened cohort) and for two prospective population-based (nonscreened) controls: the state of Minnesota (MN) and the province of Ontario (ONT). In this study, 91% o f the PQ cohort (n = 425,816) was screened at 3 weeks and 74% at 6 mon ths (n = 349,706). Through July 31, 1997 with a follow-up of the cohor t for 39-99 months, 128 cases of neuroblastoma were diagnosed in the s creened cohort: 45 were detected by screening, 20 were detected clinic ally before 3-week screening, and 62 were detected clinically having n ormal screens (58) or never screened (4). Three of 128 cohort patients moved outside PQ and were omitted from the calculations; 63 cases of neuroblastoma would have been expected in PQ during the study period, while 125 were observed (SIR = 1.98, 95% confidence interval [CI] = 1. 64-2.34, p < 0.001). In the controls from MN and ONT, 54 and 92 cases of neuroblastoma have been detected, respectively, compared with 44 an d 99 expected (SIR = 1.24 and 0.93). SIR for PQ by age at diagnosis sh owed a markedly increased incidence at less than 1 year of age, with n o reduction in incidence in older age. Limiting analysis to only patie nts diagnosed at older than 1 year of age with advanced-stage disease, 25 cases were clinically detected in PQ versus 27 expected. Data from the two control groups showed no significant change in stage of disea se in children younger than or older than 1 year of age. Preliminary m ortality data from a subset of the PQ, ONT, and MN cohorts showed stan dardized mortality ratios (SMRs) of 1.12, 0.72, and 1.01, respectively . Screening for neuroblastoma markedly increases the incidence in infa nts younger than 1 year of age, without decreasing the incidence of un favorable advanced-stage disease in older children. Preliminary mortal ity data is reported. However, it is unlikely that screening for neuro blastoma will reduce mortality from this disease. Med. Pediatr. Oncol. 31:450-454, 1998. (C) 1998 Wiley-Liss, Inc.