Evidence of screening effect derives from descriptive studies of local and national programs in Japan, uncontrolled pilot experiences at a number of sites in Europe and the United States, and population-based studies in Canada and Germany.[1,2,3,4,5,6] [7]

An increase in survival rates among screen-detected cases would be expected if screening was detecting neuroblastoma at an earlier and more curable stage. While improved survival rates after initiation of screening have been reported,[8,9] these observations should be viewed cautiously because improvements could be caused by lead-time bias, length bias, and identification of cases through screening that would have spontaneously regressed.

Screening results in an increased incidence of early-stage disease. The cases detected by screening almost exclusively have biologically favorable properties (unamplified N-myc oncogene, near triploidy, and favorable histology), and this type of favorable neuroblastoma has a high survival rate, whether detected by screening or detected clinically.[1,6,7,10,11,12,13,14,15,16,17] There is evidence that some tumors regress spontaneously in the absence of treatment.[18,19,20,21]

Some authors have argued that the Japanese experience shows that the number of children older than 1 year who are diagnosed with neuroblastoma may have decreased since the inception of screening [22] and that overall mortality has declined during this period.[12,23] A true reduction in neuroblastoma mortality may reflect improvements in treatment efficacy as much as a benefit of treating earlier-stage disease. Mortality has decreased in other countries where screening does not occur.[24] In another study of regional comparisons, disease rates were compared between Osaka, Japan, where screenings were initiated in 1985, and Great Britain, where screening was not done.[25] There was little change over time in the cumulative mortality rates in either region; 52 versus 57.5 per million between 1970 and 1979 versus 1991 and 1994 in Osaka, compared with 78.6 versus 70.1 in the corresponding periods in Great Britain. In any case, the majority of cases detected by screening at 6 months appear to have biologically favorable prognoses independent of stage. [1,26,27,28,29] Furthermore, despite the shift in stage distribution of cases detected by screening compared with those that are routinely detected, the evidence of reduction in the incidence of advanced-stage cancers in the Japanese experience has been disputed;[3,11,30] in the Quebec Project, as noted below, no such reduction is observed.[1]

The Quebec Neuroblastoma Screening Project compared neuroblastoma incidence and mortality in a 5-year birth cohort (n = 476,603) from Quebec (where urinary screening was offered at 3 weeks and 6 months [overall compliance, 92%]) to various North American birth cohorts in which no screening took place. In this study, the incidence of early-stage disease in children younger than 1 year in the screened population more than doubled that expected, while in the control population, it approximated that expected (standardized incidence ratio 3.03 (95% confidence interval [CI], 2.30–3.86) in Quebec versus 0.82 in Minnesota (95% CI, 0.41–1.38) and Ontario (95% CI, 0.53–1.17).[1] The incidence of advanced stage disease (stages III and IV) in older children in Quebec showed a statistically nonsignificant increase over that which would have been expected (standard incidence ratio 1.52 [95% CI, 0.95–2.23]).[1] After approximately 8 years of follow-up (range 6–11 years) the neuroblastoma death rate in the screened population was not significantly different from rates in unscreened populations (standardized mortality ratio 1.11 [95% CI, 0.64–1.92] for the Quebec cohort compared with Ontario children).[7] Similar findings were observed in the German neuroblastoma study.[31] Although final mortality rates are expected in 2008, an interim analysis shows that the death rate from neuroblastoma is similar in screened and control populations (1.6 vs. 1.9 deaths per 100,000 children). A study in Austria yielded a similar conclusion, though screening was performed at age 7 to 12 months. In the screening cohort, neuroblastoma incidence was statistically significantly higher than in children who were not screened (18.2 vs. 11.2 per 100,000 births), while mortality was not statistically significantly different (0.96 vs. 1.57 per 100,000 births).[32]