Kinney H, Faix R, Brazy J (1980) The fetal alcohol syndrome and neuroblastoma. Pediatrics 66:130-132 Kramer S,Ward E, Meadows AT,Malone KE (1987) Medical and drug risk factors associated with neuroblastoma: a case-control study. J Natl Cancer Inst 78:797-804 Maris JM, Kyemba SM, Rebbeck TR, White PS, Sulman EP, Kensen SJ,Allen C, Biegel JA, Brodeur GM (1997) Molecular genetic analysis of familial neuroblastoma. Eur J Cancer


Michalek AM, Buck GM, Nasca PC, Freedman AN, Baptiste MS, Mahoney MC (1996) Gravid health status, medication use, and risk of neuroblastoma. Am J Epidemiol 143:996-1001 Olshan AF, Bunin GR (2000) Epidemiology of neuroblastoma. In: Brodeur GM, Sawada T, Tsuchida Y, Voute PA (eds) Neuroblastoma. Elsevier, Amsterdam, pp 33-39 Olshan AF, Smith J, Cook MN, Grufferman S, Pollock BH, Stram DO, Seeger RC, Look AT, Cohn SL, Castleberry RP, Bondy ML (1999a) Hormone and fertility drug use and the risk of neuroblastoma: a report from the Children's Cancer Group and the Pediatric Oncology Group. Am J Epidemiol 150: 930-938

Olshan AF, DeRoos AJ, Teschke K, Neglia JP, Stram DO, Pollock BH, Castleberry RP (1999b) Neuroblastoma and parental occupation. Cancer Causes Control 10:539-549 Olshan F, Smith JC, Bondy ML, Neglia JP, Pollock BH (2002) Maternal vitamin use and reduced risk of neuroblastoma. Epidemiology 13:575-580 Schuz J, Kaletsch U, Meinert R, Kaatsch P, Spix C, Michaelis J (2001) Risk factors for neuroblastoma at different stages of disease. Results from a population-based case-control study in Germany. J Clin Epidemiol 54:702-709 SEER (2003) Surveillance, Epidemiology, and End Results (SEER) Program ( SEER*Stat Database: Incidence - SEER 9 Regs, Nov 2002 Sub (1973-2000), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2003, based on the November 2002 submission Stiller CA, Parkin DM (1992) International variations in the incidence of neuroblastoma. Int J Cancer 52:538-543 Yang Q, Olshan AF, Bondy ML, Shah NR, Pollock BH, Seeger RC, Look AT, Cohn SL (2000) Parental smoking and alcohol consumption and risk of neuroblastoma. Cancer Epidemiol Biomark Prevent 9:967-972

Screening for Neuroblastoma

William G.Woods


2.1 Introduction 7

2.2 The Rationale for Neuroblastoma Screening . . 7

2.3 Early Pioneering Studies Investigating Neuroblastoma Screening in Japan 8

2.4 Initial North American and European Neuroblastoma Screening Trials . 9

2.5 Follow-up Studies from Japan and Europe . ... 11

2.6 Definitive Controlled Trials from Quebec and Germany 11

2.6.1 Studies, Designs, and Logistics 11

2.6.2 Studies'Results 15

2.7 Biologic, Psychologic, Economic, and Clinical Aspects of Neuroblastoma Screening 17

2.7.1 Biologic Aspects 17

2.7.2 Psychologic Aspects 17

2.7.3 Economic Aspects 17

2.7.4 Clinical Implications 18

2.8 Conclusions 18

References 18

2.1 Introduction

Pediatricians are by training the most prevention oriented of all the primary care physicians. Immunizations for various potentially life-threatening infectious diseases and early screening for inborn errors of metabolism are two shining examples of childhood disease prevention. Prior to 1970,no one had attempted to reduce the morbidity, or more importantly, the mortality, of any childhood cancer through preclinical detection, specifically by mass screening for this disease. Over the past 30 years there has been much effort put into better understanding the role of preclinical detection of neuroblastoma, and potentially lowering mortality from this most challenging of childhood solid tumors. This chapter addresses various aspects of screening for neuroblastoma in children.

2.2 The Rationale for Neuroblastoma Screening

Neuroblastoma has an incidence of about 10 per million children 0-14 per year throughout the developed world (Young et al. 1986; Bernstein et al. 1992). In North America neuroblastoma will develop in approximately one in 7000 children before the age of 5 years, and over 700 cases are expected to be diagnosed annually. The incidence of neuroblastoma is about twice that of phenylketonuria, almost tenfold higher than that of galactosemia, and slightly less common than neonatal hypothyroidism (Woods and Tuchman 1987), all diseases which are mandated by most neonatal screening programs throughout the U.S.

Neuroblastoma is a fascinating neoplasm because of several clinical and biologic characteristics. The tumor is unique biochemically because it possesses metabolic pathways for catecholamine synthesis and metabolism. Homovanillic acid (HVA), the main metabolite of dopamine, and vanillylmandelic acid (VMA), the main metabolite of adrenalin and nora-drenalin, are sensitive and convenient markers of neuroblastoma since they are excreted in excessive amounts in a patient's urine (Hinterberger and Bartholomew 1969). Homovanillic acid and VMA have routinely been measured in patients with neu-roblastoma for the past 30 years, and have been found to be invaluable aids in both neuroblastoma diagnosis and follow-up.

The treatment and outcome of neuroblastoma are highly age- and stage dependent. Children who are diagnosed with early-stage localized disease or under 1 year of age, irrespective of stage, can often be treated with limited therapy and have excellent survival (Bernstein et al. 1992; Matthay et al. 1989). In contrast, children over the age of 1 year who present with advanced-stage disease have a very poor survival despite aggressive chemotherapeutic treatment regimens, including bone marrow transplantation (Bernstein et al. 1992; Matthay et al. 1989). While some authors have hypothesized that neuroblastoma presents as at least two discreet clinical pathologic entities (Woods et al. 1992; Brodeur and Nakagawara 1992), others believe that malignant progression is a natural transition from benign-acting neuroblastoma in an infant to advanced-stage disease in a child. We presently know that favorable prognosis is strongly associated with specific tumor cellular characteristics (see Chaps. 4,5, and 8) But in the 1980s, in the absence of this molecular genetic information, it was hypothesized that this natural transition may be interrupted by early detection to eradicate preclinical neuroblastomas.

2.3 Early Pioneering Studies Investigating Neuroblastoma Screening in Japan

The identification of elevated urinary catechola-mines in infants with neuroblastoma was first made in 1957 (Mason et al. 1957). Over the next 15 years, methods for measuring the main urinary metabolic byproducts of dopamine and epinephrine, HVA, and

VMA were refined. Twenty-four-hour collections were the rule, and elevated urinary catecholamines became extremely important in aiding in the diagnosis of children with "small round cell tumors" and subsequent follow-up of catecholamine-secreting neuroblastomas (Tuchman et al. 1987). A urinary VMA "spot test" based on the reaction of phenolic acids with diazotized p-nitroaniline became commonplace in pediatric oncology practice (LaBrosse 1968). In the early 1970s, Sawada and colleagues from the Kyoto Prefectural University of Medicine began pilot studies which led to implementing a mass screening program for 6-month-old children in eight cities and prefectures in Japan using the VMA spot test on random urine samples (Sawada et al. 1984). The annual incidence of neuroblastoma in Japan, 8 per million children, was similar to that reported in the U.S. at the time. Originally, 282,000 infants were screened by Sawada et al., representing 50-75 % of all births in the areas studied (Sawada et al. 1984). Because of a positive test or logistic problems with the initial sample, almost 11,000 infants (3.8%) were retested. Among 264 infants (1 in 1000) who required clinical evaluation for neuroblastoma at a medical center because they had three consecutive positive urinary tests, 16 cases of neuroblastoma were subsequently identified, giving an incidence by screening of 1 in 17,600 infants. As opposed to the high expected incidence of metastatic disease at diagnosis, 5 patients were found with Evans stage-I tumor, 4 with stage II, 2 with stage III, 5 with stage IV-S, and none with stage IV. The 16 patients were treated with surgery and limited chemotherapy, 15 of whom were alive more than 5 years after diagnosis. The only death occurred 1 month after surgery in a patient with stage-II disease. Of the original screened cohort, an additional 6 children were found to have neurob-lastoma 14-29 months after their urinary spot tests gave negative results. Hence,the false-negative rate of the Kyoto screening program was 6 of 22, or 27%, similar to what one would have expected using a VMA spot test (Sawada et al. 1984).

This encouraging early trial was reconfirmed by Sawada in longer-term follow-up of the screened population (Sawada 1986). Subsequently, many other screening trials were initiated in Japan, increasingly using quantitative assays for measuring both VMA and HVA. In an important trial from Sapporo City, Nishi, Takeda and colleagues demonstrated markedly improved survival in children in that city offered screening compared with neighboring rural areas in Hokkaido Prefecture, in which no screening was available (Nishi et al. 1987). Several childhood cancer experts throughout Europe and North America called for the institution of neuroblastoma screening programs on their continents. Many other Japanese investigators began trials in their own prefectures, and by 1986, screening for neuroblastoma was mandated by law in Japan.

A more careful analysis of the Japanese neuroblastoma screening studies revealed many methodologic limitations (Tuchman et al. 1990). Firstly, there was no utilization of a population-based cohort of infants: studies were generally performed in prefectures which did not have the ability to guarantee ascertainment of all neuroblastoma cases occurring in that region, either detected by screening, or missed and subsequently clinically found. Secondly, the data were all based on survival rather than mortality. To the untrained observer, one would surmise that mortality is the reverse of survival (or "one minus survival"). In fact, mortality represents the number of deaths in a given population, and is not affected by the incidence of a disease in that given population. This difference from "survival"becomes most important in evaluating neuroblastoma screening trials. For example, as the Japanese increasingly used more sensitive and specific quantitative assays for measuring VMA and HVA in their trials, there were increasing data suggesting a rise in neuroblastoma incidence (Yamamoto et al. 2002). As can be seen graphically in Fig. 2.1, if one has an incidence in a disease of 1X, with a survival of 50 %, of 100 children, 50 % will die. If one artificially raises the incidence to 2X, or 200 individuals in this case, and one maintains the same mortality (50 deaths), there is an artificial increase in the survival to 75% (150 of 200). Hence, looking at survival only, when the actual relevant end point is death rate, can greatly mislead an investigator. In addition, the early Japanese studies utilized no control groups other than historical controls; therefore, potential declines in neuroblastoma mortality could

Figure 2.1

Effect of incidence on survival

Figure 2.1

Effect of incidence on survival have been attributed to improvements in therapy, rather than preclinical detection. Finally, without the utilization of a population-based cohort trial, several other classic methodologic issues, such as lead time or length bias, could produce falsely optimistic results.

2.4 Initial North American and European Neuroblastoma Screening Trials

In the context of the potentially exciting results coming out of Japan mixed with the realities of those studies' limitations, several groups throughout Europe and North America began early pilot studies looking at the potential effectiveness of neuroblastoma screening. Small exploratory studies were initiated in Quebec (Scriver et al. 1987), Minnesota (Tuchman et al. 1989), northern England (Craft et al. 1989), Germany (Schilling et al. 1991), France (Mathieu et al. 1996), Austria (Kerbl et al. 1997), and elsewhere (Bergeron et al. 1998). Newcastle hosted the first International Symposium on Neuroblastoma Screening in 1988, where investigators had the opportunity to share logistical challenges and early results. Several important methodologic aspects of

Table 2.1. Principles to be considered for a cancer screening program (Adapted from Prorok and Connor 1986)

1. The disease should be a "common"serious health problem,with substantial morbidity and mortality

2. The target population should be clearly defined and have a reasonable disease prevalence

3. The target population should be accessible,with reasonable compliance to screening expected

4. The screening test should be acceptable in its performance (sensitivity, specificity) and acceptable to those screened

5. Effective treatment should exist for the disease to be detected by screening

6. There should be a reasonable expectation that patients with positive screening will comply with recommended work-up, diagnosis, therapy, and follow-up

7. Sufficient resources should be available to perform the screening

8. Develop policies for early recall of patients testing positive and follow-up of those testing negative

9. Quality control procedures to maintain sensitivity and specificity of the screening test should be in placeO

neuroblastoma were revealed. These aspects deserve some comment, given that they represent challenges of any population-based screening approaches for any diseases:

— Sample collection. In a series of important studies, Tuchman and colleagues from Minnesota demonstrated that measuring spot urines to determine HVA and VMA were as valid as 24-hour sample collections, thus obviating long collections for children in whom neuroblastoma was suspected clinically (Tuchman et al. 1985). In Japan, urine was squeezed out of diapers into plastic soy sauce bottles which held 5-10 ml of urine. North American and European investigators began collecting urine on diapers blotted against a 10x10 cm piece of filter paper which, when dried, could be mailed by regular mail to the screening laboratory for accurate determination of VMA and HVA, with urinary creatinine as the internal standard.

— Assays. Multiple laboratory assays for measuring catecholamines were debated, from the totally qualitative VMA spot test and the semi-quantitative thin layer chromatographic approach, through high performance liquid chromatography (HPLC), ELISA immuno-assays, and ultimately gas chro-matography/mass spectroscopy (GC-MS) as the gold standard.

— Compliance. No adequate population-based screening trial can be done without a high compli ance rate among the participants. Early studies in Minnesota (Tuchman et al. 1989), Texas (Ater et al. 1998), and Austria (Kerbl et al. 1997) found compliance rates of returning filter papers by parents of 6 month olds to be as low as 9 percent, pointing out the need for a massive public health infrastructure to support adequate compliance, even for something as simple as collecting urine from a diaper. Because of such issues, investigators in Minnesota joined forces with those in Quebec, combining clinical trials expertise (Bernstein et al. 1992), an infrastructure already in place for collecting urine in a large majority of 3-week olds, as part of a urinary metabolic screening program for various inborn errors of metabolism (Scriver et al. 1987); and a rapid GC-MS assay for VMA and HVA determination (Tuchman et al. 1983).

— Sample sizes. It became rapidly clear that to adequately study neuroblastoma screening, one might need a trial studying up to a million children or more to get meaningful results (Esteve et al. 1995). This sobering reality led to major modifications in many trials, some of which were abandoned due to the cost, and others that waited years for adequate funding before they proceeded.

— Case and control ascertainment. In-place state and country-wide tumor registries collecting incidence and mortality data with greater than 90-95 % ascertainment are an important requirement for an adequate prevention study.

Table 2.1, from Prorok and Connor (1986), lists principles to be considered for a cancer screening program. One could argue that any childhood disease is not a "common" serious health problem with substantial morbidity and mortality, to warrant the expense of a screening program; however,based on past precedent and the fact that children represent the future of the world, should screening of any childhood disease lower mortality, implementation would be seriously considered.

2.5 Follow-up Studies from Japan and Europe

Since 1986, when screening for neuroblastoma in Japan was mandated by law, compliance with the Japanese screening program has been much greater than 80% nationwide (Sawada and Takeda 2000). Although highly successful in recruiting parents to participate in this program, such widespread mass screening also led to less ability to measure the efficacy of this approach, for example, by comparing mortality from neuroblastoma in a population offered screening versus that not offered screening; however, subsequent attempts to document screening efficacy were performed in Japan on relatively small populations. Investigators in general found no diminution in the incidence of late-stage disease, an early marker of potential screening success, in the incidence of the disease with unfavorable biologic features, or in mortality (Yamamoto et al. 2002; Bessho et al. 1991;Yamamoto et la. 1995;Kaneko et al. 1990; Suita et al. 1998).

During the 1990s, as noted above, several smaller studies were also performed in Europe, usually without controls, with preliminary results suggesting that screening increased the incidence of the disease (Mathieu et al. 1996; Bergeron et al. 1998). Ultimately, only two prospective population-based controlled trials examining the role of neuroblastoma screening in reducing mortality from this disease were implemented that had adequate funding to guarantee a high screening compliance rate; uniform neuroblas-toma evaluation, staging, treatment, and follow-up; and optimum ascertainment procedures for determining incidence and mortality. These were the Que bec Neuroblastoma Screening Project (Woods et al. 1996, 2002) and the German Project on Neuroblastoma Screening (Schilling et al. 1998, 2002). Both of these studies deserve special mention, noting similarities and differences.

2.6 Definitive Controlled Trials from Quebec and Germany

2.6.1 Studies, Designs, and Logistics

The greatest strength of both the Quebec and German trials was that they were prospective, population-based controlled studies in which neuroblas-toma mortality was the definitive end point, rather than survival (vide supra). Both studies had considered a randomized trial approach, but "randomized controlled trials in population-based intervention studies are not always feasible" (Woods et al. 1999) as pointed out by the Quebec researchers. To clarify, the North American group had to decide what they were actually studying by introducing a new screening procedure in an infant population. Were they going to evaluate the screening test itself (urine sampling of 6-month-old babies by parents at home), or were they going to study the entire public health intervention which included introducing a new screening test? They decided that the latter question was much more relevant to improving scientific knowledge, and that to achieve a reasonable compliance rate multiple population-based education methods would be necessary, as noted below. If these measures led to a "halo effect," with an increased incidence in the non-screened population, the study results would have been viewed with skepticism. There were also practical matters including the fact that there were no other infant urinary screening programs in place in North America with a high compliance rate. Hence, control populations were picked throughout North America where no public health interventions were performed, and where had such been attempted they would have required millions of dollars in resources to be successful. These areas included the states of Minnesota and Florida, the Greater Delaware Valley, and the Province of Ontario (Woods et al. 1996). German investigators faced a similar problem. They im plemented screening in six German states selected on the basis of the "feasibility of implementing the screening program" (Schilling et al. 2002).

The Quebec Neuroblastoma Screening Project was a joint collaboration of 31 investigators throughout North America. The Quebec trial was designed specifically to answer the question of whether screening infants at or before 6 months of age (and the public health interventions associated with it) would lower mortality from this disease.After appropriate sample-size estimates were performed, geared at lowering overall mortality by 40 %, it was decided to offer screening to a 5-year birth cohort in the Province beginning 1 May 1989, once NIH funding was secured. The only screening data available around the world at that time was for infants screened at 6 months of age in Japan. Investigators hence decided that they would screen at the same age, to be able to confirm or refute results from Japan. It was furthermore decided that infants would be screened at two ages: once at 3 weeks to take advantage of the urinary screening infrastructure which had been in place for well over 10 years (Scriver et al. 1987), and again at 6 months of age with a new public health intervention. Parents were given a "screening kit" at the birth of their child. The kit included filter paper collection instructions and a bilingual consent form with a "passive" informed consent process specifically explained, approved by an NIH-certified review board. Parents knew that if they did not want to screen their infants for either inborn errors or neuroblastoma, they did not need to return the filter paper. On the other hand, if they wanted their infants screened for the already-in-place program for metabolic abnormalities but not for neuroblastoma, they simply needed to check a box indicating refusal to participate in the "cancer test," mailing the consent form with the filter paper. Greater than 90 % compliance was expected with the 3-week test, as compliance had consistently been above that level for several years for the metabolic screen (Scriver et al. 1987). Because the 6-month screen represented a new public health measure, multiple mechanisms were put in place to achieve compliance of about 75%. Some of these mechanisms included radio/television appearances and public service announcements, newspaper and magazine arti cles, posters in physicians' office and health clinics, information given to parents at birth, notices included with the Provincial "subsistence checks" which generally were mailed to all parents of infants, and even reminder inserts in diaper boxes.

Initial analyses of filters from both time periods were done in Sherbrooke utilizing thin-layer chro-matography. The assays were geared towards the highest sensitivity and accepted a lower specificity: all positive filters, representing between 5 and 10% of infants screened, were then sent to Minneapolis where definitive, highly specific GC-MS assays were performed on the same sample. If the results were positive, parents were contacted and a second sample was requested, which was again studied by GC-MS. All children with a second positive sample were referred to one of the four Quebec pediatric cancer centers for uniform neuroblastoma evaluation (Table 2.2).

In the German Project on Neuroblastoma Screening, initial pilot studies examined the feasibility of performing a screening study in infants at 6 months of age (Schilling et al. 1991). Subsequently, pilot studies in Japan were instituted looking at screening at a later age; and preliminary data were emerging from Quebec suggesting a greatly increased incidence of the disease by screening at or earlier than 6 months, with no evidence of lowering the incidence of advanced-stage disease (Woods et al. 1996). Investigators worldwide believed that any reduction of mortality from a screening approach would be potentially heralded by a lower incidence of children "destined" to do poorly. Stuttgart and Hamburg researchers hypothesized that if neuroblastoma screening at 6 months of age was not going to lower mortality, perhaps screening at 1 year would be more successful, as well as potentially lower the incidence of disease by not detecting cases which would have spontaneously regressed before that age. After securing funding from the German government, screening was offered to all children at 1 year of age born in six German states, between 1 July 1994 and 31 October 1999.

German investigators hoped to achieve a compliance of over 70 % to insure accurate sample-size estimates geared at lowering mortality. Unfortunately,

Table 2.2. Comparison of trials. NA not applicable


Quebec trial

German trial

Screening birth cohort period

1 May 1989 to 30 April 1994

1 July 1994 to 31 October 1999


All of Quebec

Six German states

Number in cohort offered screening



Age at screening

3 weeks and 6 months

1 year

Screening compliance

89% at 3weeks J 92% overall

61 %

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook

Post a comment