- Email: [email protected]
Neoplasia in childhood -25 years of progress
Annals of Oncology 6 (Suppl. 1): S3-S9, 1995. © 1995 Kluwer Academic Publishers. Printed in the Netherlands.
Symposium article Neoplasia in childhood — 25 years of progress M. G. Mott Department of Paediatric Oncology, Institute of Child Health, Bristol, U.K. Summary
Introduction
The development of neoplasia involves a disturbance in the delicate balance between the twin processes of cellular proliferation and differentiation, the precise regulation of which is vital to normal growth and development. Factors which affect that balance can be either intrinsic or extrinsic (Fig. 1). In the elderly, it is possible to discern environmental influences affecting the development of most neoplasms; in the young, however, such influences are relatively rare, and a growing body of evidence suggests that most cases of paediatric neoplasia are associated with genetic mutations. The identification of such genetic changes has provided an insight into the nature of the development of neoplasia in general, which is likely radically to alter treatment strategies in the future. Such strategies will move away from attempts to kill neoplastic cells, towards more gentle 'physiological' methods of coaxing cells to either differentiate or initiate programmed cell death (apoptosis). Neoplasia in children differs in many respects from
that in adults, and the recognition of the differences in aetiology and natural history has been an essential prerequisite to the dramatically improved results of treat-
Intrinsic
Youth
Extrinsic
Age
Fig. 1. The intrinsic and extrinsic factors which affect the balance between cellular proliferation and differentiation.
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
Background: Several factors have contributed to the improved prog- of specific genetic mutations underlying many childhood rumours nosis for the survival and quality of life of children with cancer. may now provide an 'Achilles heel' to enable the development of Childhood tumours tend to have their origin in intrinsic genetic ab- highly specific therapies that are relatively non-toxic to the normal normalities, and are usually disseminated by the time of diagnosis. tissues undergoing rapid growth and development during childhood. As a result, conventional treatments, such as ablative surgery and/or The rarity of childhood cancer and the need for multidisciplinary radiotherapy (the effects of which are detrimental to the growth and management make it impossible for the ordinary district hospital to development of normal tissues), are rarely successful. The advent of deliver optimal treatment, though 'maintenance' treatment and foleffective combination chemotherapy, given as an adjuvant to eradi- low-up can be delivered locally as part of 'shared care' with a regioncate micrometastases, has led also to the dramatic regression of in- al centre. These centres are members of the U.K. Children's Cancer operable primary tumours. Subsequent surgery often enables com- Study Group which was formed in 1977 and now treats 75% of all plete resection without the need for radiotherapy. Primary surgery is cases of childhood malignancy, with sufficient numbers to run rannow used to obtain sufficient tissue to make a precise diagnosis, and domized clinical trials for most tumour types, often in collaboration recently developed molecular techniques make this possible with with other national paediatric oncology groups in Europe and the great precision - from needle biopsies in many instances. U.S.A. The development of an effective treatment regimen is the single Conclusion: Children with cancer are likely to be major benefimost important prognostic factor. This also permits analysis of the ciaries from the recent advances in the understanding of neoplasia, differences between successfully treated groups and those not cured many of which stem from work on paediatric malignancies. It is by standard treatment. These prognostic factors usually have a bio- therefore important and mutually advantageous to foster and mainlogical basis that is identifiable at diagnosis, allowing stratification of tain close links with mainstream 'adult' oncology and with the cantreatments and further development. Some of the worst prognosis cer research institutions. cases on past standard therapy, such as B-cell acute lymphoblastic leukaemia, have an excellent prognosis when specific treatment regimens are designed to fit their particular characteristics. The finding Key words: cancer, childhood, progress
Table 2. Survival rate at 5 years in children treated for various paediatric malignancies at the Bristol Paediatric Tumour Registry, 1970-79 and 1980-89. Diagnosis
Number of cases
5-Year survival (%)
All cases
392" 771" 155" 228" 28" 34" 12" 26" 20" 55" 34" 56" 20" 39" 20" 41" 10" 17"
52 68 60 76 7 42 92 92 50 74 85 93 20 42 50 69 30 64 33 64
Acute lymphoblastic leukaemia Acute myeloid leukaemia Hodgkin's disease Non-Hodgkin's lymphoma Nephroblastoma Neuroblastoma Rhabdomyosarcoma Ewing's sarcoma Osteosarcoma
21 b " 1970-79. " 1980-89.
Table 1. Five-year survival of children aged under 15 years diagnosed with various forms of cancer between 1971 and 1985 [1]. Type of cancer
Survival (%)
Acute lymphoblastic leukaemia Acute non-lymphoblastic leukaemia Hodgkin's disease Non-Hodgkin's lymphoma CNS tumours Medulloblastoma Neuroblastoma Nephroblastoma Ewing's tumour Osteosarcoma Rhabdomyosarcoma Retinoblastoma All neoplasms
54.2 14.2 85.5 42.8 48.2 35.1 28.7 71.2 37.0 30.1 44.3 88.3 50.4
treated by the group continues to rise at a rate of about 1% per year. The survival rates and quality of life for the minority of children still not treated at specialist paediatric oncology centres are improving more slowly, so the gap continues to widen, emphasizing the importance of these organizational aspects of cancer care. A sophisticated arrangement has evolved at regional, national and international levels which allows the development of randomized clinical trials in even the rarest of malignancies. Effective treatment programmes then allow analysis of prognostic factors which distinguish between responders and non-responders. Treatment responsiveness has often been found to relate to biological differences that were not initially apparent. Knowledge of these biological factors has, in turn, stimulated the development of improved treatment strategies for the subgroups so identified. A few examples may serve to illustrate some of these points.
5-year survival = 68% 10-year survival = 63%
1980-89 (n = 771)
Acute lymphoblastic leukaemia
48
72
120
Time after diagnosis (monttis)
fig. 2. Survival rates at 5 and 10 years for 771 children aged under 15 years at diagnosis and treated for all types of cancer at the Bristol Paediatric Tumour Registry, 1980-89.
Acute lymphoblastic leukaemia (ALL) is the most common type of malignancy in childhood. Major progress was made in the treatment of this disease in the 1960s and 1970s, when it changed from being a disease that was invariably fatal to one in which the majority of children could be expected to be cured in a relatively short period of time (Fig. 3) [3]. As results improved, it became apparent that age was an important prognostic factor, with infants under 1-year-old and adolescents faring worse than children diagnosed
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
ment during the last 25 years [1]. The national data show that around half of the children diagnosed in 1971-85 were survivors at 5 years (Table 1), and thus were potentially cured, compared with only 26% in the period 1962-70. An important step has been the recognition that, as children only make up about 1% of the total population of cancer sufferers, there is a need to centralize their care; multidisciplinary teamwork by experts in paediatric cancer care is essential to deliver optimal management [2]. There is a substantial difference in rates of survival between children treated in specialized centres and those who are not. Thus, in the Bristol Paediatric Tumour Registry the 5- and 10-year survival rates for a cohort of 771 children treated between 1980 and 1989 were 68% and 63%, respectively (Fig. 2, Table 2). The foundation of the U.K. Children's Cancer Study Group (UKCCSG) in 1977 resulted in 50% of all children with malignant disease in the U.K. receiving treatment in association with specialist centres according to uniform policies and protocols. Now, 75% of children with malignancies in the U.K. are treated by the UKCCSG, and the overall 5-year survival rate for all patients
translocation typical of leukaemia in infancy [7] and the (1; 19) translocation of pre-B ALL [8]. The retinoic acid receptor has recently been identified as a partner in the t(15;17) translocation characteristic of acute promyelocytic leukaemia, and it has been demonstrated that affected children can achieve complete remission with ALL-transretinoic acid therapy without the usual coagulopathy, albeit with other unanticipated toxicities, such as pseudotumour cerebri, hyperleucocytosis and the retinoic acid syndrome [9].
71 ± 4% Era 4 (1984-68)
53 ±2% Era 3 (1979-83) 36 ±2% Era 2 (1967-78)
9 ±3%
2
3
4
5
Era 1(1962-6)
6
7
8
9
10
20
Time after diagnosis (years)
Nephroblastoma (Whins' tumour) [10]
Fig. 3. Survival rates in the St Jude studies of acute lymphoblastic leukaemia during four eras between 1962 and 1988.
Table 3. Key chromosomal translocations in acute leukaemia. Disease
Cytogenetics
Frequency
Molecular defect
Acute lymphoblastic leukaemia
t(9;22) t(l;19) t(4;ll)
5% 5% 2%
BCR/ABL E2A/PBX MLL/AF-4 (HRX/FEL)
t(l;14) t(7;9) t(7;19) t(ll;14) t(10;14)
-20% < 10% <5% <10% 7%
T-cell receptor (7 and 14) fused to various transcriptional regulators
t(8;21) inv 16 t(15;17) t(9;ll) t(ll;V)
12% 12% 7% 7% 5%
AMLI/ETO CBFB/SMMHC PML/RARa MLL/AF-9 MLL/V
T-cell acute lymphoblastic leukaemia
Acute myeoloid leukaemia
Male (50 cases. 0 deaths)
100
All Female (58 cases, 10 deaths)
I « M
40
0
12
24
36
48
60
72
84
96
108
120
Time after diagnosis (months)
Fig. 4. Survival rates in males and females diagnosed with Wilms' tumour between 1973 and 1994 at the Bristol Paediatric Tumour Registry. Ten-year survival rates were 100% for men, 80% for women and 90% for all patients studied (males vs. female, p = 0.002).
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
between the ages of 2 and 9 years. Those with low white blood cell counts at presentation did better than those with high counts, and girls did better than boys [4]. The prognostic importance of biological factors was first appreciated morphologically with the development of the French-American-British classification. The prognostic relevance of T and B lymphocyte subsets as lineage-specific cell surface markers was discovered and monoclonal antibodies for these markers were developed. Technical developments in cytogenetics established the prognostic importance of favourable (hyperdiploid) [5] and unfavourable (specific translocation) subsets. Currently, there is a relatively sophisticated understanding of the genetic basis for ALL in childhood (Table 3). A number of fusion chimaeric proteins arising from specific translocations act as aberrant transcription factors and are known to herald such a poor prognosis with standard therapy that patients with these findings at diagnosis may currently be offered bone marrow transplantation as a first treatment option. Typical examples are the Philadelphia chromosome-positive t(9;22) ALL [6], the t(4; 11)
The cooperation of three large children's cancer groups in the U.S.A. in 1969 led to the formation of the National Wilms' Tumor Study [11] and to a series of trials that have revolutionized the prognosis for this disease. In Europe, this effort was matched by the International Society of Paediatric Oncology (SIOP) [12]. As a result it is possible to cure 90% of children with Wilms' tumour, while only exposing a small minority of patients to the hazards of radiation therapy (Fig. 4). The difference in prognosis between boys and girls has not been noted in the large multicentre studies, and has only become apparent due to the improvement in cure rate; it may relate to a number of interesting biological factors in this disease. Considerable progress has been made in elucidating the biology of nephroblastoma, largely as a consequence of two rare syndromes where there is a predisposition to its development: • sporadic aniridia and the Wilms'-aniridia-genitourinary malformation (WAGR) syndrome; • Beckwith-Wiedemann syndrome. The genetic locus for the WAGR syndrome is within a deletion at 11 pi3, and the Wilms' tumour suppressor gene at this locus (WTI) was identified in 1990 [13]. Its product is a zinc finger protein and it acts as a tran-
Neuroblastoma
The traditional treatment for neuroblastoma with surgery, radiation and chemotherapy showed that prognosis was clearly related to the clinical stage of disease [24]. A special situation was recognized at an early point for infants who presented with widespread metastatic disease confined to the skin, liver and bone marrow, as these infants have the capacity to undergo spontaneous regression (stage 4S) [25]. Intensive investigations have attempted to elucidate the cause of this process. The more typical patients with disseminated neuroblastoma (stage 4) would often have evidence of gene amplification in the form of homogeneously staining regions (HSR) and double minutes [26] in the tumour cells. The amplified gene was shown to be the Nmyc oncogene [27], and a clear-cut relationship between prognosis and amplification of Nmyc was established [28]. Loss of heterozygosity at chromosome Ip36 is found in about one-third of neuroblastomas, the loss being of preferential maternal origin, and all tumours with Nmyc amplification appear to have this abnormality which is the strongest unfavourable prognostic factor identified to date [29]. The receptor for nerve growth factor has, as its primary component, a glycoprotein which is the product of the TRK-A gene, a transmembrane tyrosine kinase. Expression of TRK-A is inversely related to Nmyc amplification, but low mRNA expression may also be
Table 4. Maternal allele loss specificity in paediatric neoplasia. Tumour
Chromosome loss
Nephroblastoma Rhabdomyosarcoma Neuroblastoma Retinoblastoma Osteosarcoma
llpllplp13q13q-
associated with disease progression in tumours without Nmyc amplification, which suggests that TRK-A expression is a very powerful prognostic marker [30]. Failure to express the NGF receptor TRK-A may be an early step in the tumorigenesis of immature neuroblasts, rendering them unable to respond to the normal developmental signals to differentiate or die by apoptosis, and allowing continued growth in an embryonic mode while sustaining subsequent genetic mutations or rearrangements that ultimately lead to complete malignant transformation. The clinical implications of these biological factors are profound. It is now possible to define a subgroup of patients with neuroblastoma who will have a high chance of spontaneous regression if offered minimal therapy, or who can be cured even with incomplete surgery without additional radiation and/or chemotherapy. Conversely, there are groups with a poor prognosis despite megatherapy (to the extent of requiring bone marrow transplant support), for whom the development of biological therapies tailored to the precise genetic lesions present must be a high priority. Molecular techniques are already being refined to provide a rapid distinction of prognostic groups at the time of diagnosis, based on genetic findings [31], and this will allow early treatment stratification. The delineation of specific genetic lesions is not confined to nephroblastoma and neuroblastoma but is a general phenomenon [32] which will have a major impact on clinical developments in paediatric oncology in the next decade. Medulloblastoma aetiology
Tumours of the brain and spinal cord account for a quarter of all paediatric neoplasia (Fig. 5) [33]. In our own practice in south-west England we have been struck by the apparent lack of cases of medulloblastoma recently, and we have now confirmed that this trend is also apparent elsewhere in the U.K. (Fig. 6) [34]. For more than a decade it has become increasingly common for pregnant women to take multivitamin supplements including folate in order to reduce the incidence of neural tube defects [35], and a recent report from the U.S.A. has shown that folate, early multivitamin and iron supplements during pregnancy are protective against primitive neuroectodermal brain tumours, including medulloblastoma [36]. The possibility that this tumour might be prevented by periconceptional and early pregnancy vitamin supplementa-
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
scription factor to control downstream genes that are intimately involved in the process of normal nephrogenesis [14]. It functions as a typical tumour suppressor gene so that loss of one allele and mutation of the remaining allele results in neoplasia [15, 16]. In sporadic tumours, it is preferentially the maternal allele that is lost [17], as is the case in the development of many other types of paediatric malignancy associated with malfunction of tumour-suppressor genes (Table 4). The propensity to develop Wilms' and other embryonal tumours in Beckwith-Wiedemann syndrome is associated with a locus at 11 pi5. There is strong evidence that genes at this locus are normally imprinted [18,19] and that relaxation of this imprinting may be important in pathogenesis [20]. Duplication of the paternal allele by unbalanced translocation or uniparental disomy, for example, may upset the balance between the maternal and paternal alleles [21]. A likely candidate gene is IGF2, which is heavily methylated, or its close partner H19 [22]. Methylation of CpG dinucleotides with spontaneous deamination of methylcytosine to thiamine occasionally leads to incorrect DNA repair. The mutation rate of methylated cytosine appears to be 30-40 times higher than at other bases, so these are likely hot spots for mutation. Although 5-methyl-cytosine comprises less than 1% of bases in human DNA, it is estimated that 35% of point mutations occur at CpG dinucleotides and more than 90% of these are GC-AT transitions [23].
complication. New chemotherapy agents, and the capacity to give more intensive chemotherapy as supportive care improved, have resulted in a substantial in—L crease in the proportion of long-term survivors; at the ^ V ^ / Soft tissue same time, however, unwanted side-effects of chemo/ >/ sarcoma (7%) ' \ / Bone therapy have become more of a problem, such as y \ (4%) nephrotoxicity and ototoxicity (from cisplatin) and cardiotoxicity (from anthracyclines). Nevertheless, it has ,C ^ \— Kidney been possible to circumvent many of these problems by ^ (7%) paying meticulous attention to detail. Thus, a recent / - Neuroblastoma pilot treatment protocol for osteosarcoma has enabled / (4%) us to achieve major shrinkage of primary lesions allowing endoprosthetic replacement rather than amputation in most instances, while at the same time eradicating micrometastases. The use of carboplatin rather than cisplatin, and epirubicin rather than doxorubicin, has Brain and spinal minimized the late sequelae [38]. Prospective studies (24%) Lymphoma (10%) have suggested that it is possible to predict at an early Fig. 5. Tumour types among children aged under 15 years in the stage those patients who are likely to develop significant anthracycline cardiotoxicity [39], thus offering the south-west of the U.K., 1976-85. prospect of cardioprotection for such high-risk patients with agents like ICRF187 [40]. Leukaemia (36%)
Other (8%)
76
77 78
79
80
Fig. 6. Incidence rates in the south-west and northern regions of the U.K. for medulloblastoma in children aged under 15 years, diagnosed between 1976 and 1993 (with 95% C.I.). Incidence for 1976-84 was 5.5 (range 4.2-7.0) and 1985-93 was 2.9 (range 1.94.1) (p-0.004).
Acknowledgements
tion is exciting and appealing, not only clinically but also because of the biological implications.
The clinical results from the Bristol Paediatric Oncology Unit are the joint work of the entire multidisciplinary team in Bristol and the many regional consultants with whom we share the care of such patients. Our clinical and research work is supported by a number of charities, particularly the Cancer and Leukaemia in Childhood (CLIC) Trust.
Late effects
References
In the 1970s the small proportion of long-term survivors of childhood malignancy were considered lucky to be alive. As success rates improved, however, attention inevitably began to focus on some of the late sequelae of treatment, and in particular the long-term effects of radiation to growing and developing tissues. Growth failure following radiation to the hypothalamic pituitary axis (as in prophylactic cranial radiation for ALL) was a particular concern [37], and led to alternative forms of therapy, with a reduction in that particular
1. Robertson CM, Hawkins MM, Kingston JE. Late deaths and survival after childhood cancer: Implications for cure. BMJ 1994; 309:162-6. 2. Stiller CA. Centralised treatment, entry to trials and survival. Br J Cancer 1994; 70: 352-62. 3. Rivera GK, Pinkel D, Simone JV et al. Treatment of acute lymphoblastic leukemia. N Engl J Med 1993; 329:1289-95. 4. Baumer JH, Mott MG. Sex and prognosis in childhood ALL. Lancet 1978; ii: 128. 5. Seeker Walker LM. The prognostic implications of chromosomal findings in ALL. Cancer Genet Cytogenet 1984; 11: 233.
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
The most worrying late effect is the development of second malignancies. Some cases are due to an inherited predisposition to neoplasia, as in familial retinoblastoma or the Li-Fraumeni syndrome (with a mutant p53 tumour suppressor gene), but others are clearly related to therapy, either with radiation and/or chemotherapy, the primary culprits being alkylating agents and epipodophyllotoxins. Combinations of radiation and chemotherapy in particular for children with lymphomas seem to lead to second malignancies [41], and we have therefore developed a mutation assay which hopefully will help to identify those patients at mostrisk[42].
31.
32. 33. 34. 35. 36. 37.
38. 39.
40. 41.
42.
prognostic marker with biological and clinical significance. J Nat Cancer Inst 1993; 85: 344-5. Taylor CPF, McGuckin AG, Brown NP et al. Rapid detection of prognostic genetic factors in neuroblastoma using fluorescence in situ hybridisation on tumour imprints and bone marrow smears. Br J Cancer 1994; 69: 445-51. Diffin F, Porter H, Mott MG et al. Rapid and specific diagnosis of t(ll;22) translocation in paediatric Ewing's sarcoma and PNET by RNA-PCR. J Clin Pathol 1994; 47: 562-4. Foreman NK, Thome R, Berry PJ et al. Childhood malignancies in the south-west region of England 1976-1985. Med Pediatr Oncol 1994; 23:14-19. Thorne RN, Pearson ADJ, Nicoll JAR et al. Decline in the incidence rate of medulloblastoma. Cancer 1994; 74: 324-4. Smithells RW, Sheppard S, Schorah CJ et al. Apparent prevention of neural tube defects by periconceptional vitamin supplementation. Arch Dis Child 1981; 56: 911-18. Bunin GR, Kuijten RR, Buckley JD et al. Relation between maternal diet and subsequent primitive neuroectodermal brain tumours in young children. N Engl J Med 1993; 329: 536-41. Mott MG. Neurotoxic complications of CNS prophylaxis in childhood leukaemia. In Mastrangelo R, Poplack DG, Riccardi R (eds): Central Nervous System Leukaemia. Prevention and Treatment. Boston, U.S.A.: Martinus Nijhoff 1983; 83-93. Mott MG, Oakhill A, Foreman N et al. Treatment of osteosarcoma with a regimen designed to minimise late sequelae. Program/Proceedings of ASCO 1994 (Abstr 1447). Bu'Lock FA, Martin RP, Oakhill A, Mott MG. Early identification of patients with significant anthracycline cardiotoxicity: Potential for prevention of late sequelae. Med Pediatr Oncol 1994; 23: 238. Bu'Lock FA, Gabriel HM, Oakhill A et al. Cardioprotection by ICRF187 from high dose anthracycline toxicity in children with malignant disease. Br Heart J 1993; 70: 185-8. Ingram L, Mott MG, Mann JR et al. Second malignancy in children treated for non-Hodgkin's lymphoma and T-cell leukaemia with the UKCCSG regimen. Br J Cancer 1987; 55: 463-6. Mott MG, Boyse J, Hewitt M, Radford M. Do mutations at the GPA locus in patients treated for childhood Hodgkin's disease predict for risk of secondary leukaemia? Lancet 1994; 343: 828-9.
Correspondence to: Professor Martin G. Mott CLIC Professor of Paediatric Oncology Institute of Child Health Southwell Street St Michael's Hill Bristol BS2 8EG U.K.
Discussion Dr. Ian Smith (RoyalMarsden Hospital): Adult oncologists would love to show such improvements in prognosis. The difference is due in part to chemosensitivity but, as a paediatric oncologist, do you have a message to adult oncologists about how we should be going about things? Professor Mott: The biology of childhood tumours is very different, but the general strategies we have adopted would be familiar to the adult oncologist. In paediatric oncology the organization of randomized,
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
6. Nowell PC, Hungerford DA. A minute human chromosome in human granulocytic leukemia. Science 1960; 132:1497. 7. Pui CH, Behm FG, Downing JR et al. Ilq23/MLL rearrangement confers a poor prognosis in infants with acute lymphoblastic leukemia. J Clin Oncol 1994; 12:909-15. 8. Mellentin JD, Murre C, Donlon TA et al. The gene for enhancer binding proteins E12/E47 lies at the t(l: 19) breakpoint in acute leukemias. Science 1989; 246: 379-82. 9. Mahmoud HH, Hurwitz CA, Roberts WM et al. Tretinoin toxicity in children with acute promyelocytic leukaemia. Lancet 1993; 342:1394-5. 10. Mott MG. Nephroblastoma (Wilms' tumour). Br J Hosp Med 1975; 13:161-80. 11. D'Angio GJ, Evans AE, Breslow N et al. The treatment of Wilms' tumor: Results of the National Wilms' Tumor Study. Cancer 1976; 38: 633-46. 12. Jereb B, Burgers JMV, Tournade MF et al. Radiotherapy in the SIOP (International Society of Paediatric Oncology) Nephroblastoma studies: A review. Med Ped Oncol 1994; 22: 221-7. 13. Call KM, Glaser T, Ito CY. Isolation and characterization of a zinc finger polypeptide gene at the human chromosome 11 Wilms' tumor locus. Cell 1990; 60: 509-20. 14. Pritchard-Jones K, Fleming S, Davidson D et al. The candidate Wilms' tumour gene is involved in genitourinary development. Nature 1990; 346:194-7. 15. Brown KW, Watson JE, Poirier V et al. Inactivation of the remaining allele of the WT1 gene in a Wilms' tumour from a WAGR patient. Oncogene 1992; 7: 763-8. 16. Brown KW, Wilmore HP, Watson JE et al. Low frequency of mutations in the WT1 coding region in Wilms' tumor. Genes Chromosom Cancer 1993; 8: 74-9. 17. Williams JC, Brown KW, Mott MG, Maitland NJ. Maternal allele loss in Wilms' tumour. Lancet 1989; i: 283-4. 18. Brown KW, Williams JC, Maitland NJ, Mott MG. Genomic imprinting and the Beckwith-Wiedemann syndrome. Am J Hum Genet 1990; 46: 1000-1. 19. Williams JC, Maitland NJ, Mott MG, Brown KW. Methylation of chromosome l i p genes in Wilms' tumour and kidney tissue. IntJ Oncol 1992; 1:743-6. 20. Ogawa O, Eccles MR, Szeto J et al. Relaxation of insulin-like growth factor II gene imprinting implicated in Wilms' tumour. Nature 1993; 362: 749-51. 21. Brown KW, Gardner A, Williams JC et al. Paternal origin of I l p l 5 duplications in the Beckwith-Wiedemann syndrome: A new case and review of the literature. Cancer Genet Cytogenet 1992; 58: 66-70. 22. Reik W, Brown KW, Slatter RE et al. Allelic methylation of H19 and IGF2 in the Beckwith-Wiedemann Syndrome. Hum Mol Genet 1994; 3:1297-301. 23. Buckley JD. The aetiology of cancer in the very young. Br J Cancer 1992; 66 (Suppl XVIII): S8-12. 24. Evans AE, D'Angio GJ, Randolph J. A proposed staging for children with neuroblastoma. Children's Cancer Study Group A. Cancer 1971; 27: 374-8. 25. Brodeur GM, Pritchard J, Berthold F et al. Revisions of the International Criteria for neuroblastoma diagnosis, staging, and response to treatment. J Clin Oncol 1993; 11:1466-77. 26. Cox D, Yuncken C, Spriggs A. Minute chromatin bodies in malignant tumours of childhood. Lancet 1965; ii: 55-8. 27. Schwab M. Amplification of Nmyc as a prognostic marker for patients with neuroblastoma. Cancer Biol 1993; 4: 13-18. 28. Seeger RC, Brodeur GM, Sathe H et al. Association of multiple copies of the Nmyc oncogene with rapid progression of neuroblastomas. N Engl J Med 1985; 313: 1111-16. 29. Caron HN, van Sluis P, van Hoeve M et al. Allelic loss of chromosome Ip36 in neuroblastoma is of preferential maternal origin and correlates with Nmyc amplification. Nature Genet 1993; 4:187-90. 30. Brodeur GM. TRK-A expression in neuroblastomas: A new
collaborative clinical trials has ensured that all children with cancer in the U.K. have the best chance of effective therapy. It is rather shameful how few adult patients are offered the opportunity of taking part in such trials.
subtle interplay between the primary, basic defects that drive the cells to become neoplastic and the secondary and tertiary events.
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
Dr. Nicholas Reed (Western Infirmary, Glasgow): As anProfessor James Malpas (St. Bartholomew's Hospital): adult oncologist I occasionally see adult patients with In paediatric oncology we are probably reaching a what are normally childhood cancers. Is there anything plateau in cure rate. In the next 10 years, where would different at the molecular level in these patients, and you put the money that we might get for research so as does this offer any insight into future strategies? to start improving survival rates again? Professor Mott: Such exceptions can teach us a lot and Professor Mott: We are reaching a plateau, now that we we ought to maximize the amount of information that have done the easy things. My money would go on we attempt to get from these patients. The short answer trying to understand the biology of the disease better. A to your question is that things on the whole tend to look good example of the way forward is the one I gave of the same, but there are also clear differences. Second promyelocytic leukaemia and trans-retinoic acid; as we events, which occur more often in adult than in paedi- understand what is driving the cell to neoplasia, we atric malignancy, tend to become more common as shall be able to direct therapy in a much simpler, a patients get older. We see this even in children pre- much less toxic way, to the precise lesion that is causing senting aged 15-20 rather than 2-3 years. There is a the problem.
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
Symposium article Neoplasia in childhood — 25 years of progress M. G. Mott Department of Paediatric Oncology, Institute of Child Health, Bristol, U.K. Summary
Introduction
The development of neoplasia involves a disturbance in the delicate balance between the twin processes of cellular proliferation and differentiation, the precise regulation of which is vital to normal growth and development. Factors which affect that balance can be either intrinsic or extrinsic (Fig. 1). In the elderly, it is possible to discern environmental influences affecting the development of most neoplasms; in the young, however, such influences are relatively rare, and a growing body of evidence suggests that most cases of paediatric neoplasia are associated with genetic mutations. The identification of such genetic changes has provided an insight into the nature of the development of neoplasia in general, which is likely radically to alter treatment strategies in the future. Such strategies will move away from attempts to kill neoplastic cells, towards more gentle 'physiological' methods of coaxing cells to either differentiate or initiate programmed cell death (apoptosis). Neoplasia in children differs in many respects from
that in adults, and the recognition of the differences in aetiology and natural history has been an essential prerequisite to the dramatically improved results of treat-
Intrinsic
Youth
Extrinsic
Age
Fig. 1. The intrinsic and extrinsic factors which affect the balance between cellular proliferation and differentiation.
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
Background: Several factors have contributed to the improved prog- of specific genetic mutations underlying many childhood rumours nosis for the survival and quality of life of children with cancer. may now provide an 'Achilles heel' to enable the development of Childhood tumours tend to have their origin in intrinsic genetic ab- highly specific therapies that are relatively non-toxic to the normal normalities, and are usually disseminated by the time of diagnosis. tissues undergoing rapid growth and development during childhood. As a result, conventional treatments, such as ablative surgery and/or The rarity of childhood cancer and the need for multidisciplinary radiotherapy (the effects of which are detrimental to the growth and management make it impossible for the ordinary district hospital to development of normal tissues), are rarely successful. The advent of deliver optimal treatment, though 'maintenance' treatment and foleffective combination chemotherapy, given as an adjuvant to eradi- low-up can be delivered locally as part of 'shared care' with a regioncate micrometastases, has led also to the dramatic regression of in- al centre. These centres are members of the U.K. Children's Cancer operable primary tumours. Subsequent surgery often enables com- Study Group which was formed in 1977 and now treats 75% of all plete resection without the need for radiotherapy. Primary surgery is cases of childhood malignancy, with sufficient numbers to run rannow used to obtain sufficient tissue to make a precise diagnosis, and domized clinical trials for most tumour types, often in collaboration recently developed molecular techniques make this possible with with other national paediatric oncology groups in Europe and the great precision - from needle biopsies in many instances. U.S.A. The development of an effective treatment regimen is the single Conclusion: Children with cancer are likely to be major benefimost important prognostic factor. This also permits analysis of the ciaries from the recent advances in the understanding of neoplasia, differences between successfully treated groups and those not cured many of which stem from work on paediatric malignancies. It is by standard treatment. These prognostic factors usually have a bio- therefore important and mutually advantageous to foster and mainlogical basis that is identifiable at diagnosis, allowing stratification of tain close links with mainstream 'adult' oncology and with the cantreatments and further development. Some of the worst prognosis cer research institutions. cases on past standard therapy, such as B-cell acute lymphoblastic leukaemia, have an excellent prognosis when specific treatment regimens are designed to fit their particular characteristics. The finding Key words: cancer, childhood, progress
Table 2. Survival rate at 5 years in children treated for various paediatric malignancies at the Bristol Paediatric Tumour Registry, 1970-79 and 1980-89. Diagnosis
Number of cases
5-Year survival (%)
All cases
392" 771" 155" 228" 28" 34" 12" 26" 20" 55" 34" 56" 20" 39" 20" 41" 10" 17"
52 68 60 76 7 42 92 92 50 74 85 93 20 42 50 69 30 64 33 64
Acute lymphoblastic leukaemia Acute myeloid leukaemia Hodgkin's disease Non-Hodgkin's lymphoma Nephroblastoma Neuroblastoma Rhabdomyosarcoma Ewing's sarcoma Osteosarcoma
21 b " 1970-79. " 1980-89.
Table 1. Five-year survival of children aged under 15 years diagnosed with various forms of cancer between 1971 and 1985 [1]. Type of cancer
Survival (%)
Acute lymphoblastic leukaemia Acute non-lymphoblastic leukaemia Hodgkin's disease Non-Hodgkin's lymphoma CNS tumours Medulloblastoma Neuroblastoma Nephroblastoma Ewing's tumour Osteosarcoma Rhabdomyosarcoma Retinoblastoma All neoplasms
54.2 14.2 85.5 42.8 48.2 35.1 28.7 71.2 37.0 30.1 44.3 88.3 50.4
treated by the group continues to rise at a rate of about 1% per year. The survival rates and quality of life for the minority of children still not treated at specialist paediatric oncology centres are improving more slowly, so the gap continues to widen, emphasizing the importance of these organizational aspects of cancer care. A sophisticated arrangement has evolved at regional, national and international levels which allows the development of randomized clinical trials in even the rarest of malignancies. Effective treatment programmes then allow analysis of prognostic factors which distinguish between responders and non-responders. Treatment responsiveness has often been found to relate to biological differences that were not initially apparent. Knowledge of these biological factors has, in turn, stimulated the development of improved treatment strategies for the subgroups so identified. A few examples may serve to illustrate some of these points.
5-year survival = 68% 10-year survival = 63%
1980-89 (n = 771)
Acute lymphoblastic leukaemia
48
72
120
Time after diagnosis (monttis)
fig. 2. Survival rates at 5 and 10 years for 771 children aged under 15 years at diagnosis and treated for all types of cancer at the Bristol Paediatric Tumour Registry, 1980-89.
Acute lymphoblastic leukaemia (ALL) is the most common type of malignancy in childhood. Major progress was made in the treatment of this disease in the 1960s and 1970s, when it changed from being a disease that was invariably fatal to one in which the majority of children could be expected to be cured in a relatively short period of time (Fig. 3) [3]. As results improved, it became apparent that age was an important prognostic factor, with infants under 1-year-old and adolescents faring worse than children diagnosed
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
ment during the last 25 years [1]. The national data show that around half of the children diagnosed in 1971-85 were survivors at 5 years (Table 1), and thus were potentially cured, compared with only 26% in the period 1962-70. An important step has been the recognition that, as children only make up about 1% of the total population of cancer sufferers, there is a need to centralize their care; multidisciplinary teamwork by experts in paediatric cancer care is essential to deliver optimal management [2]. There is a substantial difference in rates of survival between children treated in specialized centres and those who are not. Thus, in the Bristol Paediatric Tumour Registry the 5- and 10-year survival rates for a cohort of 771 children treated between 1980 and 1989 were 68% and 63%, respectively (Fig. 2, Table 2). The foundation of the U.K. Children's Cancer Study Group (UKCCSG) in 1977 resulted in 50% of all children with malignant disease in the U.K. receiving treatment in association with specialist centres according to uniform policies and protocols. Now, 75% of children with malignancies in the U.K. are treated by the UKCCSG, and the overall 5-year survival rate for all patients
translocation typical of leukaemia in infancy [7] and the (1; 19) translocation of pre-B ALL [8]. The retinoic acid receptor has recently been identified as a partner in the t(15;17) translocation characteristic of acute promyelocytic leukaemia, and it has been demonstrated that affected children can achieve complete remission with ALL-transretinoic acid therapy without the usual coagulopathy, albeit with other unanticipated toxicities, such as pseudotumour cerebri, hyperleucocytosis and the retinoic acid syndrome [9].
71 ± 4% Era 4 (1984-68)
53 ±2% Era 3 (1979-83) 36 ±2% Era 2 (1967-78)
9 ±3%
2
3
4
5
Era 1(1962-6)
6
7
8
9
10
20
Time after diagnosis (years)
Nephroblastoma (Whins' tumour) [10]
Fig. 3. Survival rates in the St Jude studies of acute lymphoblastic leukaemia during four eras between 1962 and 1988.
Table 3. Key chromosomal translocations in acute leukaemia. Disease
Cytogenetics
Frequency
Molecular defect
Acute lymphoblastic leukaemia
t(9;22) t(l;19) t(4;ll)
5% 5% 2%
BCR/ABL E2A/PBX MLL/AF-4 (HRX/FEL)
t(l;14) t(7;9) t(7;19) t(ll;14) t(10;14)
-20% < 10% <5% <10% 7%
T-cell receptor (7 and 14) fused to various transcriptional regulators
t(8;21) inv 16 t(15;17) t(9;ll) t(ll;V)
12% 12% 7% 7% 5%
AMLI/ETO CBFB/SMMHC PML/RARa MLL/AF-9 MLL/V
T-cell acute lymphoblastic leukaemia
Acute myeoloid leukaemia
Male (50 cases. 0 deaths)
100
All Female (58 cases, 10 deaths)
I « M
40
0
12
24
36
48
60
72
84
96
108
120
Time after diagnosis (months)
Fig. 4. Survival rates in males and females diagnosed with Wilms' tumour between 1973 and 1994 at the Bristol Paediatric Tumour Registry. Ten-year survival rates were 100% for men, 80% for women and 90% for all patients studied (males vs. female, p = 0.002).
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
between the ages of 2 and 9 years. Those with low white blood cell counts at presentation did better than those with high counts, and girls did better than boys [4]. The prognostic importance of biological factors was first appreciated morphologically with the development of the French-American-British classification. The prognostic relevance of T and B lymphocyte subsets as lineage-specific cell surface markers was discovered and monoclonal antibodies for these markers were developed. Technical developments in cytogenetics established the prognostic importance of favourable (hyperdiploid) [5] and unfavourable (specific translocation) subsets. Currently, there is a relatively sophisticated understanding of the genetic basis for ALL in childhood (Table 3). A number of fusion chimaeric proteins arising from specific translocations act as aberrant transcription factors and are known to herald such a poor prognosis with standard therapy that patients with these findings at diagnosis may currently be offered bone marrow transplantation as a first treatment option. Typical examples are the Philadelphia chromosome-positive t(9;22) ALL [6], the t(4; 11)
The cooperation of three large children's cancer groups in the U.S.A. in 1969 led to the formation of the National Wilms' Tumor Study [11] and to a series of trials that have revolutionized the prognosis for this disease. In Europe, this effort was matched by the International Society of Paediatric Oncology (SIOP) [12]. As a result it is possible to cure 90% of children with Wilms' tumour, while only exposing a small minority of patients to the hazards of radiation therapy (Fig. 4). The difference in prognosis between boys and girls has not been noted in the large multicentre studies, and has only become apparent due to the improvement in cure rate; it may relate to a number of interesting biological factors in this disease. Considerable progress has been made in elucidating the biology of nephroblastoma, largely as a consequence of two rare syndromes where there is a predisposition to its development: • sporadic aniridia and the Wilms'-aniridia-genitourinary malformation (WAGR) syndrome; • Beckwith-Wiedemann syndrome. The genetic locus for the WAGR syndrome is within a deletion at 11 pi3, and the Wilms' tumour suppressor gene at this locus (WTI) was identified in 1990 [13]. Its product is a zinc finger protein and it acts as a tran-
Neuroblastoma
The traditional treatment for neuroblastoma with surgery, radiation and chemotherapy showed that prognosis was clearly related to the clinical stage of disease [24]. A special situation was recognized at an early point for infants who presented with widespread metastatic disease confined to the skin, liver and bone marrow, as these infants have the capacity to undergo spontaneous regression (stage 4S) [25]. Intensive investigations have attempted to elucidate the cause of this process. The more typical patients with disseminated neuroblastoma (stage 4) would often have evidence of gene amplification in the form of homogeneously staining regions (HSR) and double minutes [26] in the tumour cells. The amplified gene was shown to be the Nmyc oncogene [27], and a clear-cut relationship between prognosis and amplification of Nmyc was established [28]. Loss of heterozygosity at chromosome Ip36 is found in about one-third of neuroblastomas, the loss being of preferential maternal origin, and all tumours with Nmyc amplification appear to have this abnormality which is the strongest unfavourable prognostic factor identified to date [29]. The receptor for nerve growth factor has, as its primary component, a glycoprotein which is the product of the TRK-A gene, a transmembrane tyrosine kinase. Expression of TRK-A is inversely related to Nmyc amplification, but low mRNA expression may also be
Table 4. Maternal allele loss specificity in paediatric neoplasia. Tumour
Chromosome loss
Nephroblastoma Rhabdomyosarcoma Neuroblastoma Retinoblastoma Osteosarcoma
llpllplp13q13q-
associated with disease progression in tumours without Nmyc amplification, which suggests that TRK-A expression is a very powerful prognostic marker [30]. Failure to express the NGF receptor TRK-A may be an early step in the tumorigenesis of immature neuroblasts, rendering them unable to respond to the normal developmental signals to differentiate or die by apoptosis, and allowing continued growth in an embryonic mode while sustaining subsequent genetic mutations or rearrangements that ultimately lead to complete malignant transformation. The clinical implications of these biological factors are profound. It is now possible to define a subgroup of patients with neuroblastoma who will have a high chance of spontaneous regression if offered minimal therapy, or who can be cured even with incomplete surgery without additional radiation and/or chemotherapy. Conversely, there are groups with a poor prognosis despite megatherapy (to the extent of requiring bone marrow transplant support), for whom the development of biological therapies tailored to the precise genetic lesions present must be a high priority. Molecular techniques are already being refined to provide a rapid distinction of prognostic groups at the time of diagnosis, based on genetic findings [31], and this will allow early treatment stratification. The delineation of specific genetic lesions is not confined to nephroblastoma and neuroblastoma but is a general phenomenon [32] which will have a major impact on clinical developments in paediatric oncology in the next decade. Medulloblastoma aetiology
Tumours of the brain and spinal cord account for a quarter of all paediatric neoplasia (Fig. 5) [33]. In our own practice in south-west England we have been struck by the apparent lack of cases of medulloblastoma recently, and we have now confirmed that this trend is also apparent elsewhere in the U.K. (Fig. 6) [34]. For more than a decade it has become increasingly common for pregnant women to take multivitamin supplements including folate in order to reduce the incidence of neural tube defects [35], and a recent report from the U.S.A. has shown that folate, early multivitamin and iron supplements during pregnancy are protective against primitive neuroectodermal brain tumours, including medulloblastoma [36]. The possibility that this tumour might be prevented by periconceptional and early pregnancy vitamin supplementa-
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
scription factor to control downstream genes that are intimately involved in the process of normal nephrogenesis [14]. It functions as a typical tumour suppressor gene so that loss of one allele and mutation of the remaining allele results in neoplasia [15, 16]. In sporadic tumours, it is preferentially the maternal allele that is lost [17], as is the case in the development of many other types of paediatric malignancy associated with malfunction of tumour-suppressor genes (Table 4). The propensity to develop Wilms' and other embryonal tumours in Beckwith-Wiedemann syndrome is associated with a locus at 11 pi5. There is strong evidence that genes at this locus are normally imprinted [18,19] and that relaxation of this imprinting may be important in pathogenesis [20]. Duplication of the paternal allele by unbalanced translocation or uniparental disomy, for example, may upset the balance between the maternal and paternal alleles [21]. A likely candidate gene is IGF2, which is heavily methylated, or its close partner H19 [22]. Methylation of CpG dinucleotides with spontaneous deamination of methylcytosine to thiamine occasionally leads to incorrect DNA repair. The mutation rate of methylated cytosine appears to be 30-40 times higher than at other bases, so these are likely hot spots for mutation. Although 5-methyl-cytosine comprises less than 1% of bases in human DNA, it is estimated that 35% of point mutations occur at CpG dinucleotides and more than 90% of these are GC-AT transitions [23].
complication. New chemotherapy agents, and the capacity to give more intensive chemotherapy as supportive care improved, have resulted in a substantial in—L crease in the proportion of long-term survivors; at the ^ V ^ / Soft tissue same time, however, unwanted side-effects of chemo/ >/ sarcoma (7%) ' \ / Bone therapy have become more of a problem, such as y \ (4%) nephrotoxicity and ototoxicity (from cisplatin) and cardiotoxicity (from anthracyclines). Nevertheless, it has ,C ^ \— Kidney been possible to circumvent many of these problems by ^ (7%) paying meticulous attention to detail. Thus, a recent / - Neuroblastoma pilot treatment protocol for osteosarcoma has enabled / (4%) us to achieve major shrinkage of primary lesions allowing endoprosthetic replacement rather than amputation in most instances, while at the same time eradicating micrometastases. The use of carboplatin rather than cisplatin, and epirubicin rather than doxorubicin, has Brain and spinal minimized the late sequelae [38]. Prospective studies (24%) Lymphoma (10%) have suggested that it is possible to predict at an early Fig. 5. Tumour types among children aged under 15 years in the stage those patients who are likely to develop significant anthracycline cardiotoxicity [39], thus offering the south-west of the U.K., 1976-85. prospect of cardioprotection for such high-risk patients with agents like ICRF187 [40]. Leukaemia (36%)
Other (8%)
76
77 78
79
80
Fig. 6. Incidence rates in the south-west and northern regions of the U.K. for medulloblastoma in children aged under 15 years, diagnosed between 1976 and 1993 (with 95% C.I.). Incidence for 1976-84 was 5.5 (range 4.2-7.0) and 1985-93 was 2.9 (range 1.94.1) (p-0.004).
Acknowledgements
tion is exciting and appealing, not only clinically but also because of the biological implications.
The clinical results from the Bristol Paediatric Oncology Unit are the joint work of the entire multidisciplinary team in Bristol and the many regional consultants with whom we share the care of such patients. Our clinical and research work is supported by a number of charities, particularly the Cancer and Leukaemia in Childhood (CLIC) Trust.
Late effects
References
In the 1970s the small proportion of long-term survivors of childhood malignancy were considered lucky to be alive. As success rates improved, however, attention inevitably began to focus on some of the late sequelae of treatment, and in particular the long-term effects of radiation to growing and developing tissues. Growth failure following radiation to the hypothalamic pituitary axis (as in prophylactic cranial radiation for ALL) was a particular concern [37], and led to alternative forms of therapy, with a reduction in that particular
1. Robertson CM, Hawkins MM, Kingston JE. Late deaths and survival after childhood cancer: Implications for cure. BMJ 1994; 309:162-6. 2. Stiller CA. Centralised treatment, entry to trials and survival. Br J Cancer 1994; 70: 352-62. 3. Rivera GK, Pinkel D, Simone JV et al. Treatment of acute lymphoblastic leukemia. N Engl J Med 1993; 329:1289-95. 4. Baumer JH, Mott MG. Sex and prognosis in childhood ALL. Lancet 1978; ii: 128. 5. Seeker Walker LM. The prognostic implications of chromosomal findings in ALL. Cancer Genet Cytogenet 1984; 11: 233.
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
The most worrying late effect is the development of second malignancies. Some cases are due to an inherited predisposition to neoplasia, as in familial retinoblastoma or the Li-Fraumeni syndrome (with a mutant p53 tumour suppressor gene), but others are clearly related to therapy, either with radiation and/or chemotherapy, the primary culprits being alkylating agents and epipodophyllotoxins. Combinations of radiation and chemotherapy in particular for children with lymphomas seem to lead to second malignancies [41], and we have therefore developed a mutation assay which hopefully will help to identify those patients at mostrisk[42].
31.
32. 33. 34. 35. 36. 37.
38. 39.
40. 41.
42.
prognostic marker with biological and clinical significance. J Nat Cancer Inst 1993; 85: 344-5. Taylor CPF, McGuckin AG, Brown NP et al. Rapid detection of prognostic genetic factors in neuroblastoma using fluorescence in situ hybridisation on tumour imprints and bone marrow smears. Br J Cancer 1994; 69: 445-51. Diffin F, Porter H, Mott MG et al. Rapid and specific diagnosis of t(ll;22) translocation in paediatric Ewing's sarcoma and PNET by RNA-PCR. J Clin Pathol 1994; 47: 562-4. Foreman NK, Thome R, Berry PJ et al. Childhood malignancies in the south-west region of England 1976-1985. Med Pediatr Oncol 1994; 23:14-19. Thorne RN, Pearson ADJ, Nicoll JAR et al. Decline in the incidence rate of medulloblastoma. Cancer 1994; 74: 324-4. Smithells RW, Sheppard S, Schorah CJ et al. Apparent prevention of neural tube defects by periconceptional vitamin supplementation. Arch Dis Child 1981; 56: 911-18. Bunin GR, Kuijten RR, Buckley JD et al. Relation between maternal diet and subsequent primitive neuroectodermal brain tumours in young children. N Engl J Med 1993; 329: 536-41. Mott MG. Neurotoxic complications of CNS prophylaxis in childhood leukaemia. In Mastrangelo R, Poplack DG, Riccardi R (eds): Central Nervous System Leukaemia. Prevention and Treatment. Boston, U.S.A.: Martinus Nijhoff 1983; 83-93. Mott MG, Oakhill A, Foreman N et al. Treatment of osteosarcoma with a regimen designed to minimise late sequelae. Program/Proceedings of ASCO 1994 (Abstr 1447). Bu'Lock FA, Martin RP, Oakhill A, Mott MG. Early identification of patients with significant anthracycline cardiotoxicity: Potential for prevention of late sequelae. Med Pediatr Oncol 1994; 23: 238. Bu'Lock FA, Gabriel HM, Oakhill A et al. Cardioprotection by ICRF187 from high dose anthracycline toxicity in children with malignant disease. Br Heart J 1993; 70: 185-8. Ingram L, Mott MG, Mann JR et al. Second malignancy in children treated for non-Hodgkin's lymphoma and T-cell leukaemia with the UKCCSG regimen. Br J Cancer 1987; 55: 463-6. Mott MG, Boyse J, Hewitt M, Radford M. Do mutations at the GPA locus in patients treated for childhood Hodgkin's disease predict for risk of secondary leukaemia? Lancet 1994; 343: 828-9.
Correspondence to: Professor Martin G. Mott CLIC Professor of Paediatric Oncology Institute of Child Health Southwell Street St Michael's Hill Bristol BS2 8EG U.K.
Discussion Dr. Ian Smith (RoyalMarsden Hospital): Adult oncologists would love to show such improvements in prognosis. The difference is due in part to chemosensitivity but, as a paediatric oncologist, do you have a message to adult oncologists about how we should be going about things? Professor Mott: The biology of childhood tumours is very different, but the general strategies we have adopted would be familiar to the adult oncologist. In paediatric oncology the organization of randomized,
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
6. Nowell PC, Hungerford DA. A minute human chromosome in human granulocytic leukemia. Science 1960; 132:1497. 7. Pui CH, Behm FG, Downing JR et al. Ilq23/MLL rearrangement confers a poor prognosis in infants with acute lymphoblastic leukemia. J Clin Oncol 1994; 12:909-15. 8. Mellentin JD, Murre C, Donlon TA et al. The gene for enhancer binding proteins E12/E47 lies at the t(l: 19) breakpoint in acute leukemias. Science 1989; 246: 379-82. 9. Mahmoud HH, Hurwitz CA, Roberts WM et al. Tretinoin toxicity in children with acute promyelocytic leukaemia. Lancet 1993; 342:1394-5. 10. Mott MG. Nephroblastoma (Wilms' tumour). Br J Hosp Med 1975; 13:161-80. 11. D'Angio GJ, Evans AE, Breslow N et al. The treatment of Wilms' tumor: Results of the National Wilms' Tumor Study. Cancer 1976; 38: 633-46. 12. Jereb B, Burgers JMV, Tournade MF et al. Radiotherapy in the SIOP (International Society of Paediatric Oncology) Nephroblastoma studies: A review. Med Ped Oncol 1994; 22: 221-7. 13. Call KM, Glaser T, Ito CY. Isolation and characterization of a zinc finger polypeptide gene at the human chromosome 11 Wilms' tumor locus. Cell 1990; 60: 509-20. 14. Pritchard-Jones K, Fleming S, Davidson D et al. The candidate Wilms' tumour gene is involved in genitourinary development. Nature 1990; 346:194-7. 15. Brown KW, Watson JE, Poirier V et al. Inactivation of the remaining allele of the WT1 gene in a Wilms' tumour from a WAGR patient. Oncogene 1992; 7: 763-8. 16. Brown KW, Wilmore HP, Watson JE et al. Low frequency of mutations in the WT1 coding region in Wilms' tumor. Genes Chromosom Cancer 1993; 8: 74-9. 17. Williams JC, Brown KW, Mott MG, Maitland NJ. Maternal allele loss in Wilms' tumour. Lancet 1989; i: 283-4. 18. Brown KW, Williams JC, Maitland NJ, Mott MG. Genomic imprinting and the Beckwith-Wiedemann syndrome. Am J Hum Genet 1990; 46: 1000-1. 19. Williams JC, Maitland NJ, Mott MG, Brown KW. Methylation of chromosome l i p genes in Wilms' tumour and kidney tissue. IntJ Oncol 1992; 1:743-6. 20. Ogawa O, Eccles MR, Szeto J et al. Relaxation of insulin-like growth factor II gene imprinting implicated in Wilms' tumour. Nature 1993; 362: 749-51. 21. Brown KW, Gardner A, Williams JC et al. Paternal origin of I l p l 5 duplications in the Beckwith-Wiedemann syndrome: A new case and review of the literature. Cancer Genet Cytogenet 1992; 58: 66-70. 22. Reik W, Brown KW, Slatter RE et al. Allelic methylation of H19 and IGF2 in the Beckwith-Wiedemann Syndrome. Hum Mol Genet 1994; 3:1297-301. 23. Buckley JD. The aetiology of cancer in the very young. Br J Cancer 1992; 66 (Suppl XVIII): S8-12. 24. Evans AE, D'Angio GJ, Randolph J. A proposed staging for children with neuroblastoma. Children's Cancer Study Group A. Cancer 1971; 27: 374-8. 25. Brodeur GM, Pritchard J, Berthold F et al. Revisions of the International Criteria for neuroblastoma diagnosis, staging, and response to treatment. J Clin Oncol 1993; 11:1466-77. 26. Cox D, Yuncken C, Spriggs A. Minute chromatin bodies in malignant tumours of childhood. Lancet 1965; ii: 55-8. 27. Schwab M. Amplification of Nmyc as a prognostic marker for patients with neuroblastoma. Cancer Biol 1993; 4: 13-18. 28. Seeger RC, Brodeur GM, Sathe H et al. Association of multiple copies of the Nmyc oncogene with rapid progression of neuroblastomas. N Engl J Med 1985; 313: 1111-16. 29. Caron HN, van Sluis P, van Hoeve M et al. Allelic loss of chromosome Ip36 in neuroblastoma is of preferential maternal origin and correlates with Nmyc amplification. Nature Genet 1993; 4:187-90. 30. Brodeur GM. TRK-A expression in neuroblastomas: A new
collaborative clinical trials has ensured that all children with cancer in the U.K. have the best chance of effective therapy. It is rather shameful how few adult patients are offered the opportunity of taking part in such trials.
subtle interplay between the primary, basic defects that drive the cells to become neoplastic and the secondary and tertiary events.
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015
Dr. Nicholas Reed (Western Infirmary, Glasgow): As anProfessor James Malpas (St. Bartholomew's Hospital): adult oncologist I occasionally see adult patients with In paediatric oncology we are probably reaching a what are normally childhood cancers. Is there anything plateau in cure rate. In the next 10 years, where would different at the molecular level in these patients, and you put the money that we might get for research so as does this offer any insight into future strategies? to start improving survival rates again? Professor Mott: Such exceptions can teach us a lot and Professor Mott: We are reaching a plateau, now that we we ought to maximize the amount of information that have done the easy things. My money would go on we attempt to get from these patients. The short answer trying to understand the biology of the disease better. A to your question is that things on the whole tend to look good example of the way forward is the one I gave of the same, but there are also clear differences. Second promyelocytic leukaemia and trans-retinoic acid; as we events, which occur more often in adult than in paedi- understand what is driving the cell to neoplasia, we atric malignancy, tend to become more common as shall be able to direct therapy in a much simpler, a patients get older. We see this even in children pre- much less toxic way, to the precise lesion that is causing senting aged 15-20 rather than 2-3 years. There is a the problem.
Downloaded from http://annonc.oxfordjournals.org/ at Cornell University Library on July 26, 2015