Gender differences in the incidence of astrocytoma?

Brain tumors are the second most common childhood malignancy, accounting for about 20% of all tumors in children less than 15 years of age. Astrocytoma, medulloblastoma, and ependymoma are the most common histologic subtypes. Several international registries have demonstrated increasing incidence rates for astrocytoma, but the reason(s) for this remains unclear. Moreover, despite advances in therapy, morbidity and mortality remain high. A new report [Hjalmars et al. Cancer 85:2077-2090, 1999] describes an increased incidence of malignant brain tumors and in particular, astrocytoma, in Sweden for the period 1973-1992. In 1992, the overall incidence for malignant brain tumors was 28.6 per million children, representing an average annual increase of about 2.6% (95% CI 1.5-3.8). For astrocytomas, a higher incidence was noted for grade 1-2 compared to grade 3-4 with the highest incidence rates in both groups occurring in females; 15.2 cases per million and 9.8 cases per million in the grade 3-4, respectively. Time trend analysis showed an increase in incidence rates among the astrocytoma group, average annual change of 3% (95% CI 1.6-4.4), and an increase in females by 3.9% (95% CI 2.0-5.8). This increase was most pronounced in the age groups older than 6 years, and was essentially confined to females. In contrast, medulloblastoma and ependymoma were more common among boys with a male/female ratio of 1.4 and 1.1, respectively. There were no statistically significant increasing trends among either of these two subgroups. 

COMMENT: This study demonstrated statistically significant increasing incidence for malignant brain tumors (particularly, astrocytoma). An increase in childhood brain tumor incidence rates is consistent with observations from other studies [Lannering et al, Cancer 1990; 66:604-9; Stiller et al, Cancer 1995; 76:709-713]. As noted, improved diagnostic capabilities may help to explain some of this increase in incidence (See also Smith et al, JNCI 1998; 90:1269-77; C3, Vol 9, No 5). Nonetheless, the striking gender differences reported here for astrocytoma, along with the increasing incidence, suggest that environmental factors could be important. This warrants further investigation. Mary Eapen 

Monosomy 7 - now you see it, now you don't!

Monosomy 7 (along with deletions of the long arm of chromosome 7) are recurring, non-random chromosomal abnormalities frequently found in children and adults with de novo myeloid disorders. Monosomy 7 is also frequently reported in patients who develop treatment-related myelodysplastic syndrome (MDS) following alkylating agents. There have been three cases reported in the literature where children with monosomy 7-related MDS (de novo and treatment-related) achieved spontaneous hematologic and cytogenetic remission. Mantadakis et al [Cancer 1999; 85:2655-61] contribute 5 additional cases (all boys) to this literature. Three of these children had de novo MDS diagnosed between 8 and 15 months of age. The other two children had treatment-related MDS diagnosed at an age of 8 and 10 years. Four of these five children have achieved spontaneous and durable hematologic disease remission and survived 14 months to 11 years from diagnosis. One child, a 15 month old with de novo MDS, initially achieved clinical and cytogenetic disease remission but subsequently developed disease recurrence and died. The authors suggest that, similar to transient myeloproliferative disorder in children with Down syndrome, there exists a subset of children who present with transient MDS which can resolve spontaneously. 

COMMENT: Mantadakis et al are to be commended on pulling the literature together and adding additional data to report on such a fascinating occurrence. It is important to note that all the children with de novo MDS and a spontaneous remission were very young (15 months or less). The biology behind these spontaneous remissions may hold a key to understanding how some myeloid leukemias develop. The authors suggest two possible explanations for these spontaneous remissions involving monosomy 7. One explanation may be that monosomy 7 in a cell subpopulation may be insufficient to obtain a proliferative advantage and the clone eventually dies out. Secondly, spontaneous regression in some patients with monosomy 7 may be due to an initiating mutation in a stem cell pool of limited self-renewing capacity (i.e., although important in disease development the stem cell pool eventually dies out). Additional experiments on defined cell sub-populations could help answer these important questions. In the rare case of an infant with MDS, caution and very careful observation should be used in planning therapy. Julie A. Ross

The in utero experience- is birthweight telling us something?

Several studies have reported statistically significant associations between birth weight and risk of specific childhood cancers. Leukemia, Wilms' tumor, and neuroblastoma have been associated with higher birth weights [Yeazel et al, J Pediatr 1997; 131:621-627]. In contrast, hepatoblastoma has been linked with prematurity and low birth weight (see C3, Vol 9, No 3). Intriguingly, recent studies suggest that birthweight may also be predictive of adult-onset cancers. Mogren et al [Cancer Causes & Control, 1999; 10:85-94] recently analyzed population-based data from Sweden and explored characteristics of pregnancy and birth with malignancy in the offspring (both young and old). Data from over 248,000 deliveries during the period 1955-1990 were abstracted and linked with data from the Cancer Registry (cancer diagnosed from infancy up to 39 years of age). Variables explored included maternal age, parity, child's birthweight, gestational age, presence of Down syndrome, malformations (other than Down syndrome), diabetes mellitus (mother), single/multiple birth, intrauterine growth retardation, and pre-eclampsia. As expected, an association was found between Down syndrome and childhood leukemia (SIR 50.0 95%CI 13.5-99.9). Moreover, high birthweight (> 4500 grams) was associated with a 4-fold increased risk of leukemia in children less than 4 years of age (SIR 4.3, 95% CI 1.6-9.3). Although there were few cases, diabetes mellitus in the mother was associated with an increased risk of testicular cancer in offspring [SIR 17.8, 95% CI 3.6-52.1]. Finally, advanced maternal age was associated with an increased risk of both childhood leukemia and breast cancer in offspring. No other statistically significant associations were found. The authors conclude that in utero factors (e.g., hormones) likely contribute to risk of specific malignancies in offspring. 

COMMENT: Considerable interest exists in understanding the contribution of the intrauterine environment to childhood and adult-onset disease. Some countries have the tools available to link population data (neonatal records, cancer registry data, etc) and explore important questions. However, without biological samples, it is impossible to definitively test hypotheses (hormones, etc). For example, the consistent association between high birthweight and childhood leukemia [reviewed in Ross et al, Cancer Causes & Control, 1996; 7:553-559] suggests that the relationship is real, but hypothesis-testing biological studies are needed. Several large pregnancy cohort studies are being established (which incorporate questionnaires and biological samples) to investigate the contribution of the in utero environment to early childhood as well as adult-onset diseases. Hopefully, these types of studies will help us address some of these important associations. Julie A. Ross 

RET rearrangements in thyroid cancer revisited

As reported previously (see C3 Vol 7, no 1), there has been a marked increase in the occurrence of papillary thyroid carcinoma (PTC) in children from Belarus who were exposed to high levels of radiation from Chernobyl. A large proportion (approximately 60%) of PTC tumors from these children display rearrangements involving the RET proto-oncogene (resulting in activation to an oncogene), suggesting that radiation exposure may be important in RET activation. At least three different ret rearrangements have been identified in thyroid carcinomas (including RET/PTC1, RET/PTC2, and RET/PTC3), and studies have investigated whether one type of rearrangement may be more frequent in radiation-association PTC when compared with sporadic PTC. In a recent study by Smida et al, [Int J Cancer, 80:32-38, 1999], 51 PTC from children in Belarus were selected for analysis of specific RET rearrangements. All children were less the 9 years of age at the time of the Chernobyl accident (1986) and all were less than 16 years of age at the time of surgery for PTC. In addition, 16 sporadic (i.e., no radiation history) adult PTCs from Germany were examined, as well as 16 PTCs from adults who were living in the contaminated area of Belarus. Analysis of specific RET rearrangements was performed using RT-PCR and direct sequencing. Of the adult PTCs, only RET/PTC1 rearrangements were detected (69% of the Belarussian cases and 19% of the German cases). In contrast, for the children from Belarus, 13 cases (26%) had RET rearrangements involving PTC3 and 12 (24%) had rearrangements involving PTC1. No RET-PTC2 rearrangements were detected in either the adult or childhood PTCs. Since earlier reports have suggested a predominance of RET-PTC3 rearrangements in children from Chernobyl, the authors suggest that the latency period (between radiation exposure and diagnosis of the tumors) may be important in determining the type of rearrangement that predominates. In this study, the mean latency period was 9 years, but when the authors restricted the analysis to earlier-diagnosed tumors, the ratio of RET-PCT3/RET-PTC1 was approximately 3:1 (in agreement with earlier published studies). 

COMMENT: The RET-PTC1 and PTC3 rearrangements involve fusion of the RET gene with either the H4 gene or the ELE1 gene, respectively (all of these genes are on chromosome 10). Intrachromosomal rearrangements (inversions) lead to the fusion of the two different genes. It is interesting to speculate that tumors of shorter latency associated with radiation exposure in early childhood would preferentially involve RET rearrangements with the ELE1 gene, whereas tumors that take longer to develop involve the H4 gene. It is surprising, however, that the authors did not observe any RET/PTC2 rearrangements in their series, since these, too, have been observed in PTCs from children exposed to Chernobyl [Fugazzola et al, Cancer Res: 55, 5617-20, 1995]. Studies of RET rearrangements in PTC should also be applied to children exposed to therapeutic radiation who develop secondary PTCs (e.g., survivors of childhood cancer). Julie A. Ross 

C3 Quarterly Newsletter
Children's Cancer Research Fund
Epidemiology Research Unit
Division of Pediatric Epidemiology
Clinical Research
University of Minnesota
420 Delaware St. SE, Box 422
Minneapolis, MN 55455
pedsepi@umn.edu

Editors: 
Stella M. Davies, MD, PhD, and Julie A. Ross, PhD