Discussion
In a detailed and comprehensive account of the Pediatric Official Positions of the International Society of Clinical Densitometry24 on quantitative computer tomography, Adams et al addressed the preferred choice of the non-dominant limb and the selection of sampling sites, including the 4% and 20% radius in addition to the 4%, 38% and 66% tibia as used in this study.
While the findings reported here suggest that, overall, the subjects, long term survivors of ALL in childhood and adolescence, exhibit mainly normal bone health, comparison with data from healthy populations has to be made with due consideration for important differences in their characteristics. For example, as reported by Leonard and colleagues at the Children’s Hospital of Philadelphia (CHOP)16 in a large group of healthy children, adolescents and young adults in the US, the racial composition of the study sample is critical. African Americans have higher values than Whites for numerous cortical measures at the 38% tibia which are reflected in a 13% greater average for the polar section modulus metric of bone strength. Similar differences exist between males and females.16 Comparable findings have been published from South Africa with bone strength measured by polar SSI.30 These authors have reviewed the international experience of the influence of ethnicity on bone.31Consequently, the comparison of our study sample with a reference group of healthy subjects of the same ethnicity provides a more accurate assessment of their bone health. This reveals deficits in the mineralization of trabecular bone at the distal radial metaphysis in more than one third of subjects with a less consistent pattern in the corresponding tibia. Deficits in the mid-shaft of the tibia, except for cortical vBMD, are associated with reduced bone strength in almost one third of subjects.
Our findings of great total and trabecular vBMD in males than in females at the metaphyseal sites have been reported in healthy older adolescents.21,22
A seemingly counter-intuitive finding in the current study is the greater cortical thickness at the 38% than at the 66% tibial sites. However, in a remarkably detailed investigation conducted in 60 healthy young adults, pQCT was performed at 14 levels ranging in 5% increments from the 10% to the 80% sites on the tibia (and the fibula).32 The cortical thickness was substantially greater at the 40% than at the 65% tibial sites. Again, in concert with the current findings, the cross-sectional moments of inertia, measures of the stiffness of bone in relation to bending, were progressively greater proximally along the tibial shaft beyond the 30% site.
Although there are few reports of the use of pQCT to assess bone health in survivors of ALL in childhood, two are especially informative. Investigators in Manchester, UK studied 53 school-aged children and adolescents who had completed treatment, without cranial irradiation, 1.5 to 8.3 years earlier.33 Measurements were made at the 4% and 50% sites of the non-dominant radius and comparisons were made with close to 200 healthy subjects in the same age group; a cohort which grew to 500 in later years,34 affording the opportunity to develop Z scores (though not applied earlier to the ALL group). The subjects who had survived ALL had lower trabecular but not total vBMD than the control group and the deficit was less the longer the interval from the completion of therapy in this cross-sectional study. Findings at the diaphyseal site suggested endosteal bone loss during treatment.
In rather the reverse order, the investigators at CHOP built a reference cohort of more than 650 individuals, affording the opportunity to provide Z scores for clinical populations.18Subsequently they were able to study 50 survivors of ALL, ages 5-22 years, within two years of completing treatment which did not include cranial irradiation.35 Measurements were made at the 3%, 38% and 66% sites of the non-dominant tibia. Trabecular vBMD (at the 3% site) and cortical vBMD (at the 38% site) were significantly lower than in the reference group; no results were reported for total vBMD at either site. When studied one year later both measures showed improvement but were still significantly lower than in the reference group. The finding at the metaphysis is comparable to that reported by Brennan et al,13 but the diaphyseal measurements are not comparable because the sites are much different.
Our findings with post-processing of images point to greater connectivity in and more compact trabecular bone in SR than HR survivors.This may reflect the higher intensity of osteotoxic chemotherapy in the HR group.
The limitations of this cross-sectional observational study include a small sample size, although larger than the other published studies of pQCT in children who had ALL. Again, the subjects had a wide age span (15 years) and a wide range of time from diagnosis, averaging 15 years, although all were more than 10 years from that date. A longitudinal study would be useful, building on the experience of Leonard and colleagues.17 Addressing the lack of ethnic diversity in our subjects would require a much bigger sample size.
A particular strength of the study is the use of post-processing with custom software which expands information on apparent bone architecture compared to conventional pQCT. A prospective serial study of children with ALL, along their cancer journey from diagnosis to long term survivorship, using this technology would provide a detailed picture of osteotoxicity and recovery while offering opportunities for early therapeutic intervention to maximize ultimate bone health in this young population.
Conflict of Interest Statement. The authors have no conflict of interest to declare