Over 70 per cent of children with neuroblastoma have detectable metastases at the time of diagnosis. The greatest proportion of children with metastatic neuroblastoma at the time of diagnosis are over two years of age.1 Good prognosis has generally been associated with children diagnosed before age two years (preferably under one year) who have a localized tumor.2 Curability decreases as the extent of disease at diagnosis increases. Thus, detection of neuroblastoma in infants and children prior to tumor dissemination is an important obstacle to overcome in order to improve the cure rate. To achieve early detection, we urgently need to define the high-risk populations of children to whom sensitive and specific methods for detection can be applied.
CHILDREN AT HIGH RISK FOR NEUROBLASTOMA
A special effort at diagnosis should be made in susceptible individuals or populations, particularly within the first two years of life. At this age there is increased probability of localized or regional disease and cure. Definitions of individuals or populations at risk are necessarily retrospective; however, with continued accumulation and analysis of data, they may eventually become of prospective value. As a working hypothesis, the following parameters might be used to identify children most susceptible to developing neuroblastoma.
First trimester bleeding. No published data exist on the relationship between the incidence of first trimester bleeding and an increased occurrence of neuroblastoma. However, a review of the pregnancies of 36 mothers of patients referred to Memorial Sloan-Kettering Cancer Center revealed eight (22 per cent) who had such signs, an increased percentage, compared with the usual incidence of 10 per cent of mothers.
Third trimester pseudo-eclampsia. Six mothers described by Voute3 in a retrospective study had hypertension without albuminuria. Voute alluded to the possibility that functional catecholamines were released by the fetus, crossed the placenta, and affected the maternal blood pressure. In a review of the prenatal history of 50 children at Memorial Hospital, two similar case histories were found.
There is no significant correlation between neuroblastoma and hypertension in children with neuroblastoma even in the presence of elevated urinary catecholamines and metabolites. It is not known whether noradrenalininactivating enzymes (monoamine oxidase or catechol-orthomethyl transferases) are functional in the fetus. The inference that excess catecholamines in the fetus are not inactivated intracellularly or near the cell of release remains to be established. Newborn infants with neuroblastoma have not been reported to have high blood pressure, which would mitigate against this possibility. Even in the absence of an explanation, pseudo-eclampsia in the third trimester certainly occurs and can serve to heighten suspicion of neuroblastoma in the unborn fetus.
Maternal exposure to known carcinogens (radiation, nonsteroidal estrogens, and cancer chemotherapy). A causal relationship between maternal exposure to ionizing radiation, drugs, and other chemotherapeutic agents and fetal development of neuroblastoma is conjectural at present. In a review of 36 mothers of children with neuroblastoma, 12 (33 per cent) had a history of diagnostic radiographic examination during their pregnancy. Drugs taken included aspirin, progesterone, estrogen, appetite depressants, antibacterial antibiotics, or vitamins. Because of the evidence of transplacental teratogenesis and carcinogenesis with chemical agents,4 it seems sufficient to consider human exposure to known teratogenic and carcinogenic agents as a hazard. Infants born of such pregnancies should be considered high risks and should be carefully screened during their first two years.
Viral illness during pregnancy. Viral infections (measles, mumps, influenza) have not been associated with neuroblastoma. Upper respiratory and inapparent viral infections are ubiquitous and do not allow any cause and effect relationship to be established. Children born of mothers with documented viral infections during pregnancy would be worthwhile candidates to screen for neuroblastoma as well as other neoplastic conditions. It is probable that the yield of positive cases would be small considering the small total numbers of cases detected annually. In the absence of nationwide cancer control programs involving vast numbers of potential high-risk patients at present, it seems best to concentrate on children born of mothers with documented gestational infections. Although there appears to be a tendency for neuroblastoma to occur with greater frequency in the first-born (Figure 1), when this observation is analyzed using the Green wood- Veile method for analyzing birth orders within siblings, a lower than expected frequency resulted in patients in birth orders one and two and a higher than expected frequency occurred in birth orders five through seven. This indicates an environmental component in the disease etiology and would suggest examining possible physical, chemical, and viral agents operating during the intrauterine and postnatal periods.
Although consistently related congenital abnormalities have never been associated with neuroblastoma,5'6'7'* all children with such stigmata should be considered high-risk individuals for any cancer. A large number of birth defects have been observed to occur in association with cancer: congenital heart disease, osteogenesis imperfecta, congenital absence of rectum and anus, anencephaly, microencephaly, hydrocephaly, spina bifida, craniosynostosis, scaphoencephaly, maple sugar urine disease, rib malformations, vascular anomalies in the brain, intrauterine growth retardation, cleft lip and palate, buccopharyngeal membrane, tracheo-esophageal fistula, aganglionosis, Polydactyly, syndactyly, club foot, hemi-hypertrophy, neurofibromatosis, and inguinal hernia. This lack of specificity leads to the conclusion that any associated birth defect is probably fortuitous.
There may be an increased prevalence of neuroblastoma in children of mothers having a history of fetal wastage. In one study, 29 per cent of mothers of children with neuroblastoma had a previous history of stillbirths or spontaneous abortions.10 This is slightly higher than the 20 per cent spontaneous abortion rate reported for all pregnancies. Because there is a possibility that more than one child with neuroblastoma or symptoms of neuroblastoma may occur in a family,9'10 fetal deaths may constitute previously undiagnosed cases. Infants born to mothers with such antenatal histories ought to be considered to have increased risk of neuroblastoma. The known high incidence of neuroblastoma in situ11'12 suggests a need for postmortem examination of abortuses to determine the actual incidence of gestational neuroblastoma.
FAMILY HISTORY OF CANCER
The risk of cancer appears to be increased in certain families.13 A family history (previously born siblings, aunts, uncles, grandparents, and parents) of multiple cases of cancer should be considered sufficient cause to include all children born into the family in a special risk category. Sixty-six parents of children with neuroblastoma interviewed at Memorial SloanKettering Cancer Center revealed an 80 per cent incidence of cancer in close relatives.
VALUE AND APPLICATION OF METHODS OF DETECTION
This consists of "screening" (to be defined later) high-risk individuals previously mentioned and any child with a possible diagnosis of cancer. Although cancer has become the second most common cause of death in children between one and fifteen years of age, the total number of cancer cases is low. The probability of one pediatrician in practice diagnosing a child with neuroblastoma is small. The average pediatrician diagnoses two to four cases of cancer per lifetime of pediatric practice. This means that any effort applied by pediatricians or general practitioners to conscientiously screen all their patients gives an exceedingly small and, therefore, expensive yield.
Figure 1. Histogram showing birth order of children with neuroblastoma.
One possible answer to this dilemma is to emphasize neuroblastoma within the physicians' diagnoses during routine practice, especially during the course of well-baby visits and general health examinations. Careful physical examinations within the first two years of life, including palpation of the abdomen, rectal examination for pelvic masses, and so forth, yield only a small proportion of cases. This is because the primary tumor arises in the sympathetic ganglia or adrenal gland, both of which are retroperitoneal or retromediastinal structures and are difficult to palpate. The physician needs to be alerted to some more exacting clues.
IMPORTANT SIGNS AND SYMPTOMS
An evaluation of the clinical signs and symptoms of 87 neuroblastoma patients at diagnosis has revealed a tendency for certain redundant signs and symptoms.1 The four main modes of presentation are:
1. Mass, representing the primary or metastatic distribution, usually abdominal or in lymph nodes
2. Neurologic signs, such as weakness in an extremity, limp, or paralysis
3. Pain, usually in the bone or joints (this is most infrequent within the first two years of age)
4. Orbital signs, ecchymosis and proptosis (Figure 2).
Additional but less specific presenting signs and symptoms occur less frequently and should serve only as a caveat to the primary physician. These include respiratory tract symptoms, diarrhea, epistaxis, hematuria, and hematochezia. Particular attention should be focused on the small percentage (2.5 per cent) of patients with diarrhea at diagnosis. Nonspecific anemia without leukopenia, fever, and lymphadenopathy is present in many patients. However, with the exception of orbital signs associated with an abdominal mass, no particular constellation of signs and symptoms is of sufficient frequency to constitute a pathognomonic syndrome.
All these signs and symptoms can serve to alert the primary physician to suspect neuroblastoma. Once the disease is suspected, a series of diagnostic tests should be done. These can be carried out in the following sequence: determination of urinary excretion of catecholamines, histologic examination of tumor tissue, and identification of the locations and thereby extent of disease.
SCREENING FOR NEUROBLASTOMA
Examinations or procedures used in diagnostic screening for neuroblastoma should be those that are most specific, sensitive, and least injurious to the child; no single test fulfills all these requirements. Urine examination is by far the most innocuous procedure and can be taken advantage of in neuroblastoma because of its characteristic secretory activity. In 7 out of 10 cases the tumor secretes excess quantities of catecholamines or metabolites or both into the blood. These accumulate in the urine and readily can be quantitated in most laboratory facilities.
There are a variety of methods of urine analysis and differences of opinion as to what should be considered an adequate specimen. Concerning methods, when a laboratory elects to use a particular procedure, urine specimens from normal children should be determined with that procedure and should be used as a standard, or comparisons should be made with duplicate aliquotes assayed in established laboratories. The importance of determining normal values for infants from birth to adolescence is indicated by the higher urinary vanilly mandelic acid (VMA) values found during the first year of life than in older children.
There is controversy concerning the adequacy of spot samples versus 24-hour specimens. The proponents of spot sampling argue that collection by their method is easier and facilitates screening. The assumptions made are that excretion of metabolites is constant during the day and that the level of sensitivity of the test is designed to detect small to large increments above normal. Studies in our laboratory of sequential four-hour urinary aliquotes collected from individual patients during a 24-hour period revealed variations in excretion levels of VMA or catecholamines or both ranging from normal to abnormal. Forty per cent of 63 patients with elevated 24-hour urinary VMA excretion at the time of diagnosis had values ranging from 20 to 30 per cent above the upper limits of normal. Spot samplings in such patients, which miss peak VMA excretion, could cause erroneous interpretation of diagnosis. The remaining 60 per cent of patients would have been detected since their excretion values remain constantly elevated above normal. The normally high urinary VMA values detected during the first year of life may cause a large number of false positives if a particularly sensitive spot test is used. This may be constructive in terms of overall detection yields if it would lead to repeated testing and eventually 24-hour quantitation of VMA and catecholamines.
The recently promoted test strip14 for simplified detection is a modification of the diazotized p-nitroaniline spot test. Its screening value remains to be proved. There are some patients with elevated levels of VMA in the urine whose urine was found to be negative with the test strip.15 Coupled with the fact that only 70 per cent of patients excrete abnormal levels of VMA at diagnosis, caution should be applied to this particular screening method when considering its efficacy for large-scale studies. In the absence of a more rapid and simple screening test, and with modifications such as increased sensitivity enabling detection of urinary VMA at concentration of 5 meg. /ml., it could be valuable as a first maneuver in suspected cases.16 But any child with suspected tumor should have a 24-hour urine specimen, collected under appropriate conditions of diet (Table 1) and assayed for VMA and catecholamines.
Figure 2. Proptosis and ecchymosis due to metastatic neuroblastoma.
INSTRUCTION FOR COLLECTION OF URINE FOR CATECHOLAMINES
Figure 3. Bone marrow aspiration with classical, but rarely found, rosette formations of neuroblastoma cells.
Urinary cystathionine excretion as a diagnostic tool17 is a sensitive test but lacks specificity. It is also elevated in hepatic cancer,1 congenital cystathioniuria, and a variety of other malignant tumors in childhood.19 If elevated values are obtained with concomitant elevated cathecholamine or VMA excretion, the diagnosis of neuroblastoma is virtually confirmed. Excess cystathionine excretion suggests the presence of metastases; normal urinary levels would tend to exclude metastatic neuroblastoma.
Catecholamines, VMA, and cystathionine can all be determined from the same 24-hour urine specimen, and the assays can be obtained commercially. The major criticism of the 24-hour collection has been the difficulty of obtaining complete 24-hour urine specimens, in particular from female infants. This has been overcome by using stoma bags #1020 and by educating the parents to apply the sticking surfaces carefully over the perineum. Table 1 contains instructions for urine collection given to the parents of patients with known or suspected neuroblastoma.
Anemia is not a corroborative finding; however, it is present in about half of patients at the time of diagnosis. The white cell count and platelet count usually are normal. Total lymphocyte count has been reported to be low at diagnosis in neuroblastoma patients with poor prognosis.20 This finding has not been consistent in our patients. Although there is a significant correlation between lymphocyte infiltration of the tumor and survival, there is no correlation with the number of lymphocytes circulating in the blood and survival of the patient.21
Figure 4. Intravenous pyelogram demonstrating displacement of right kidney downwards and laterally in an eight-month-old child whose presenting sign was proptosis.
Figures. Phase contrast microphotograph of partially differentiated neuroblastoma cells in culture (X 300).
In the absence of previous confirmation of neuroblastoma, bone marrow aspiration and examination of smears can be used as a diagnostic procedure. Caution must be used in interpretation since the occurrence of classical "rosettes" is not very common (Figure 3). More usual are clumps of extrinsicappearing cells in the marrow. The bone marrow is positive in about 50 per cent of cases at diagnosis; it is an indication of widespread disease and portends a poor prognosis. But because of the spotty distribution of neuroblastoma metastases, it is not unusual to have negative smears in the presence of lytic bone lesions.
If the diagnosis has been assured by previous study, it would be advantageous to permit the bone marrow aspiration to be done in the referral centers where full advantage of research techniques can be taken. Thus, the bone marrow aspirate can serve as a source for cell culture, which itself can be a useful diagnostic test for cell kinetic studies with labeled thymidine to explore the growth rate of the tumor in the bone marrow. The bone marrow can also be used to determine, by formaldehyde-induced fluorescence, the presence of catecholamines in tumor cells.22 Radiographic examinations for neuroblastoma diagnosis include the intravenous pyelogram and the skeletal survey. Following palpation of an abdominal mass the IVP is abnormal in 78 per cent of cases (Figure 4). If all patients with neuroblastoma are considered, 68 per cent have abnormal IVP's at the time of diagnosis. The percentage of abnormal radiographs is less elevated for metastatic bone lesions (48 per cent).
When a palpable or observable mass is first detected, additional diagnostic and histologic proof are necessary before therapy is initiated. The problem oí surgical resection or biopsy can be resolved on a geographical basis. If the patient has an apparently easily resectable subcutaneous mass or lives at a great distance from a cancer center, resection or biopsy should be undertaken. In contrast, if there is a cancer center within a reasonable distance, the patient should be referred and primary surgery should be done there. The cancer center has the advantage of having the combined resources of chemotherapist, surgeon, immunologist, and radiotherapist trained in pediatric oncology.
Results obtained with chemotherapy and radiotherapy of neuroblastoma to date have been poor.23 Improved methods for early detection and improved cure rates will come from basic and applied clinical research on children referred to cancer centers where special technical and research capabilities are available. Ideally, sophisticated studies should be done (Table 2) to determine the extent of disease, to elicit markers of disease activity, and to evaluate the patient's potential for experimental chemotherapy and immunotherapy as these modalities are developed.
The diagnosis of neuroblastoma rests within the minds and hands of practicing physicians. In the absence of inexpensive, sensitive, and specific screening tools, only an awareness of the soft signs and symptoms of tumor mass or neurologic changes in the child under two years of age lead to earlier diagnosis. Once a tumor is suspected, a sequence of studies should be undertaken, including a hemogram, urinary assay for catecholamines, intravenous pyelogram, skeletal survey, and bone marrow aspiration.
SPECIAL STUDIES DONE AT THE MEMORIAL SLOAN-KETTERING CANCER CENTER, IN ADDITION TO ROUTINE STUDIES, FOR WORK UP OF PATIENTS WITH KNOWN OR SUSPECTED NEUROBLASTOMA
Detectable masses should be biopsied or preferably totally resected. The latter is best done in a regional pediatric cancer center, when feasible, where advantage can be taken of the availability of specialized oncologists and research personnel. Through the combined efforts of the alerted physician in the community and clinical and laboratory research at the cancer center, progress towards increasing curability and developing methods of early diagnosis of neuroblastoma may be attained.
1. Helson, L. and Namerow, D.M. Clinical observations on neuroblastoma diagnosis. Clin. Bull. Memorial Sloan-Kettering Cancer Center 2 (1972), 43.
2. Pinkel. D., Pratt, C. Holton, C, et al. Survival of children with neuroblastoma treated with combination chemotherapy. J. Pediat. 73 (1968), 928.
3. Voute, P.A.. Wadman, S.K., and Putten, W.J. Congenital neuroblastoma. Clin. Pediat. 9 (1970), 206.
4. Herbst, A. L., UHelder, H., and Poskanzer, D. C. Adenocarcinoma of the vagina. New Engl. J. Med. 284(1971), 878.
5. Tubergen, D. and Heyn, R. In situ neuroblastoma associated with an adrenal cyst. J. Pediat. 76(1970), 451.
6. Warkany. J. Congenital Malformations. Chicago: Year Book Medical Publishers, Inc.. 1971.
7. Reisman, M., Goldenberg, E. D.. and Gordon, J. Congenital heart disease and neuroblastoma. Amer. J. Dis. Child. 111 (1966), 308.
8. Kouyoumdjian, A. O. and McDonald. J.J. Association of congenital adrenal neuroblastoma with multiple anomalies including an unusual oropharyngeal cavity (imperforate buccopharyngeal membrane). Cancer 4 (1951). 784.
9. Chatten, J. and Voorhess. M. Familial neuroblastoma. New Engl. J. Med. 277 (1967), 1230.
10. Helson. L., Blasco, P., and Murphy, M. L. Familial neuroblastoma. Clin. Res. 17 (1969). 614.
11. Beckwith. JB. and Perrin. E.V. In situ neuroblastomas ; a contribution to the natural history of neural crest tumors. Amer. J. Path. 43 (1963), 1089.
12. Guin. G. H., Gilbert, E. F., and Jones, B. Incidental neuroblastoma in infants. Amer. J. Clin. Path. 51 (1969), 126.
13. Hardy. P.C. and Nesbit. M. E. Familial neuroblastoma; report of a kindred with a high incidence of infantile tumors. J. Pediat. 80 (1972), 74.
14. Leonard. A. S., Roback, S.. Nesbit, M. E., and Freier. E. The VMA test strip; a new tool for mass screening, diagnosis, and management of catecholamine-secreting tumors. J. Pediat. Surg. 7 (1972), 528-531.
15. Helson, L, Bethune, V.. and Schwartz, M. K. Clinical evaluation of the VMA test strip. Pediatrics 51 ( 1 973) . 1 53.
16. Leonard. AS. Personal communication.
17. Helson. L . Fleisher. M.. Bethune, V., et al. Urinary cystathionine, catecholamine, and metabolites in patients with neuroblastoma. Clin. Chem. 78(1972), 613.
18. Helson. L., Peterson, R., and Schwartz. MK. Cystathionine excess in children with hepatic cancer. Cancer Res. 33. In press.
19. Helson, L. Unpublished observations.
20. Bill. A. H. and Morgan, A, Evidence for immune reaction in neuroblastoma and future possibilities for investigation. J. Pediat. Surg. 5 (1970), 111.
21. Lauder, I. and Aherne, A. The significance of lymphocyte infiltrations in neuroblastoma. Brit. J Cancer 26 (1 972), 321.
22. Helson, L. and Biedler, J. L. Catecholamines in neuroblastoma cells from human bone marrow, tissue culture, and murine C-1300 tumor. Cancer 31 (1973). 1087.
23. Helson. L., Vanichayangkul, P., Tan. CC. et al. Combination intermittent chemotherapy for patients with disseminated neuroblastoma. Cancer Chemother. Rep. 56 (1972). 499.
24. Vaartaja. T., Helson. L.. Baren, A, et al. Cytogenetic observations in children with neuroblastoma. Pediatrics 47 (1971), 839.
25. Spengler. B. A., Biedler, J. L.. Helson, L.. and Freedman, L. S. Morphology and Growth, Tumorigenicity, and Cytogenetics of Human Neuroblastoma Cells Established In Vitro. Abstract). Tissue Culture Association, Inc. 24th annual meeting. Boston, Mass.. June 1973.
26. Goldstein. M.. Freedman, L. S.. Bohoun. A.C., and Guerinot, F. Serum dopamine-Bhydroxylase activity in neuroblastoma patients. New Engl. J. Med. 286(1972), 1123.
27. Imashuku. S. Tyrosine hydroxylase in neuroblastoma. Biochem. Med. 5(1971), 22.
28. Bill, AH.. Seibert, E. S., Beckwith. J. B., and Hartmann, JB. Nerve growth stimulating activity in sera from normal and neuroblastoma patients. J. Nat. Cancer Inst. 43(1969), 1221.
29. Helson, L., Ramos, C, Oettgen, H., and Murphy, M. L. DNCB Reactivity in Children with Neuroblastoma. Abstract). Proc. Amer. Assoc, for Cane. Res. April 1971 .
30. D'Angio. GJ. . Loken. M., and Nesbit, M. Uptake ot selenium-75 methionine in tumor sites of neuroblastoma patients. Ann. Radiol. 14 (1971). 351.
31. Helson, L.. Watson. R.. Benua. R.. and Murphy, M. L. 18F radioisotope scanning of metastatic bone lesions in children with neuroblastoma. Amer. J. Roentgenol. J 75(1972). 191.
INSTRUCTION FOR COLLECTION OF URINE FOR CATECHOLAMINES
SPECIAL STUDIES DONE AT THE MEMORIAL SLOAN-KETTERING CANCER CENTER, IN ADDITION TO ROUTINE STUDIES, FOR WORK UP OF PATIENTS WITH KNOWN OR SUSPECTED NEUROBLASTOMA