Should We Screen Newborns for Neuroblastoma?


Synopsis: Screening for neuroblastoma increased the incidence of favorable tumors in infancy but did not de crease unfavorable, advanced-stage disease or mortality.

Source: Woods WG, et al. A population-based study of the usefulness of screening for neuroblastoma. Lancet 1996;348:1682-1687.

Between may 1, 1989, and april 30, 1994, 476,603 children born in the province of Quebec were screened for neuroblastoma by measuring catecholamines in urine specimens at three weeks and six months. Compliance with screening was 91% at three weeks and 76% at six months. Children with a positive screen were referred to one of four pediatric cancer centers for uniform evaluations. Standardized Rate of Incidence (SRI) for neuroblastoma was calculated for Quebec and two similar population based controls (the state of Minnesota and the province of Ontario) during the same time.

A total of 118 cases of neuroblastoma were diagnosed. Twenty were diagnosed clinically prior to the three-week screen, possibly due to a "halo effect" of increased awareness related to the screening program. Forty-three were detected pre-clinically by the three-week or six-month screen. Of the 55 cases diagnosed later, 52 had normal early screens.

The incidence of neuroblastoma in Quebec was approximately twice that in the two control areas (SRI: 2.17, Quebec; 1.13, Minnesota; 0.93, Ontario). However, the increase was almost exclusively due to increases in early-stage cases. The incidence of advanced-stage disease and mortality was similar in all three areas.


Neuroblastoma is the most common solid tumor of the first five years of life. The majority of these tumors are hormonally active, secreting catecholmines that can be detected and quantitated in the urine. The prognosis of neuroblastoma is much better when the tumor is clinically diagnosed before one year of age and before distant metastasis, particularly to the skeleton, has occurred. These facts have been the basis for consideration of mass screening of infants. In Japan, there has been a national neuroblastoma screening program for several years but the value of this program is controversial.

To a great extent, the controversy relates to the basic biological nature of neuroblastoma. Clinicians and scientists have long recognized two subgroups of neuroblastoma. The first is characterized by genetic abnormalities (a hyperdiploid karyotype and trk gene expression). This variety accounts for the large majority of all neuroblastomas diagnosed before one year of age, as well as low-stage tumors diagnosed at older ages. These "good" neuroblastomas are often curable with surgery alone, and some even undergo spontaneous regression. The second variety is characterized by a pseudodiploid karyotype, chromosome 1p deletions, and N-myc gene amplification. These usually present after one year of age with disseminated disease and account for almost all neuroblastoma deaths. While the Japanese screening program demonstrated increased detection of "good" neuroblastomas, the way the program had been implemented made it impossible to know if they had resulted in reduced mortality.

The questions remained: 1) Were screening programs just detecting neuroblastomas that would spontaneously regress or have good outcome with surgery alone? Or 2) Did screening prevent deaths from neuroblastoma by diagnosing the disease at a time when it was still curable.

The study by Woods et al provides extremely strong support for the first alternative. The incidence and clinical characteristics of neuroblastomas in the study population of Quebec, where screening was instituted in 1989, were compared to those of the unscreened populations of Ontario and Minnesota, as well as to Quebec before 1989. Screening increased the number of localized, good-prognosis cases, but had no effect on the incidence of advanced-state disease or mortality. It appears that "bad" neuroblastomas arise de novo and do not evolve from pre-existing detectable tumors. Almost all of the aggressive tumors arose in children who had normal screens in infancy. Mass screening of infants at or before six months of age did not reduce mortality from neuroblastoma.

An accompanying commentary outlines some of the problems associated with mass screening in Japan. Because screening is not effective in detecting aggressive tumors at an asymptomatic, curable stage, and because of the dilemmas posed by detecting early cases, some of which may require no therapy, the author advocates calling a halt to mass screening programs for neuroblastoma.1 (Dr. van Hoff is Associate Professor of Pediatrics at Yale University School of Medicine and is a Pediatric Cancer Epidemiologist.)


1. Bessho F. Where should neuroblastoma mass screening go? Lancet 1996;348:1672.