| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Medical Imaging (A.M.F., S.Ku., C.W.) and Endocrinology and Diabetes (C.C.P., S.Ka., F.J.C.), Royal Childrens Hospital, Parkville, Victoria 3052, Australia; Clinical Epidemiology and Biostatistics Unit, Murdoch Childrens Research Institute (S.V., J.B.C.), Parkville, Victoria 3052, Australia; and Department of Radiology (A.M.F.) and Pediatrics (C.C.P., S.K., F.J.C., S.V., J.B.C.), University of Melbourne, Parkville, Victoria 3052, Australia
Address all correspondence and requests for reprints to: Dr. A. Michelle Fink, Department of Medical Imaging, Royal Childrens Hospital, Flemington Road, Parkville, Victoria 3052, Australia. E-mail: michelle.fink{at}wch.org.au.
| Abstract |
|---|
|
|
|---|
Methods: Children under the age of 10 yr undergoing brain MRI for seizures or idiopathic developmental delay and who had no endocrine abnormality were recruited prospectively over 2 yr. All MRI studies included a three-dimensional sequence. Only subjects with normal studies were included. One hundred thirty-nine children were eligible (mean age, 5.2 yr). Direct pituitary volumes were measured from contiguous 1-mm thick reconstructed coronal and sagittal images. Estimated pituitary volumes were calculated using pituitary height, width, and length.
Results: Volumes obtained from reconstructions in either plane were essentially identical. There was a linear increase in log-transformed pituitary volume with age, but relatively weak correlations with height or body mass index. There was no gender difference and only weak correlations between pituitary height and pituitary volume and between estimated pituitary volume calculation and measured pituitary volume. We provide age-related reference ranges for pituitary volumes in graphical and tabular forms.
| Introduction |
|---|
|
|
|---|
MRI of the pituitary is advocated as part of the baseline investigations to be performed on children with poor growth or suspected endocrinopathy (1, 2, 3, 4, 5, 6, 7, 8, 9). Notwithstanding its frequent clinical use, recognition of abnormalities in pituitary size can be difficult, and radiological reports of pituitary gland size are frequently based upon subjective judgment. The size and shape of the normal pituitary gland vary considerably and are also affected by age, sex, and hormonal environment (10, 11, 12, 13, 14, 15, 16, 17, 18). This is also true throughout childhood, with the most dramatic changes seen in the neonatal period and in puberty (18, 19, 20, 21, 22, 23). The variation in shape of the pituitary between individuals means that any assessment of its size will be subject to a high degree of imprecision unless a true volume is measured.
Most reports to date on the size or volume of the normal pituitary gland on MRI come from data acquired in two dimensions, and most are in the adult population. They are mainly based on measurements in a single plane (usually gland height) (11, 12, 13, 14, 15, 16, 17, 18) or on rough approximations of volumes calculated from two-dimensional (2D) data (10). In only one study, which included 17 normal adolescent controls, were volumes directly measured from 3D datasets (24).
There are only two publications of which we are aware that describe the size of the normal whole pituitary gland in children relative to age. One is a study of the height of the pituitary gland as a function of age (20), and this has frequently been quoted as a standard reference (2, 4, 7, 17, 25, 26, 27). The other is the only study in children to use 3D magnetic resonance volumetry (23). This study includes prepubertal and postpubertal children and is limited to a Japanese population.
The aim of our study was to establish a normal reference range for pituitary volumes according to age in prepubertal children based on true volumetric measurements in an ethnically diverse population.
| Subjects and Methods |
|---|
|
|
|---|
After receiving ethics approval from the ethics in human research committee of Royal Childrens Hospital, we recruited 254 children up to the age of 10 yr. The study period extended from July 2001 to September 2003. The children included in the study were to have MRI of the brain as part of the investigation of possible seizures or idiopathic developmental delay. The study population reflected the wide ethnic mix of the catchment area of our institution, which is predominantly European and southeast Asian. Written informed consent was obtained from all parents/guardians before the childs participation.
Exclusion criteria
Any patient with a history or clinical evidence of endocrine abnormality, genetic syndrome, preterm delivery at less than 35 wk gestation, birth asphyxia, hospitalization for head injury, craniospinal irradiation, or abnormal MRI of the brain was excluded from the study. Any patient who was peri- or postpubertal on examination was also excluded from the study. After application of these exclusion criteria, 143 children were considered for inclusion.
Clinical assessment
All subjects were clinically assessed by one of two endocrinologists (C.C.P. or S.K.) before the MRI. Information regarding signs and symptoms of endocrine disorders, use of endocrine therapies, genital abnormalities, and genetic syndromes was recorded. Weight was assessed by digital scale, and height was determined by Harpenden stadiometer. Body mass index (BMI) was calculated as the ratio of weight to height2 (kg/m2). Patients with extreme values of height were excluded, because height was considered the most appropriate auxological indicator of subclinical endocrinopathy. In addition to this, all patients studied were clinically euthyroid and had a normal (prepubertal) genital appearance.
MRI evaluation
MRI examinations were performed using a GE 1.5 T Echospeed LX system. Thin section volumetric studies were obtained using T1-weighted sagittal 3D fast spoiled gradient echo (SPGR) data acquisition according to our protocol for the investigation of seizures and congenital brain abnormalities. The following parameters were used: minimum echo time and repetition time, inversion time of 350 msec for children over 2 yr increasing up to 850 msec for infants under 6 months of age, 320 x 192 matrix (zero filled to 512 x 512), 25° flip angle, 8.33-Hz band width, 1- to 1.4-mm slice thickness (interpolated to 0.50.7 mm), and 21- to 23-cm of field of view. For infants under 6 months of age, the parameters were as follows: 3D T1-weighted SPGR with minimum full echo time, 35-msec repetition time, 45° flip angle, 8.93-Hz bandwidth, 20- to 26-cm field of view, 0.8- to 1.2-mm slice thickness, and 256 x 192 matrix. This conventional spoiled gradient echo (SPGR) sequence was used because it was determined in clinical studies that the inherent tissue contrast of the fast SPGR was not appropriate for gray/white matter differentiation in this younger age group.
To obtain the pituitary volumes, additionally targeted sagittal and coronal reformats with 1-mm contiguous slices were obtained from the 3D dataset. The cross-sectional areas of the pituitary gland were measured manually using a surface area mark-up tool on the contiguous 1-mm sections in both sagittal and coronal planes. The volume (in cubic millimeters) was obtained by adding the areas (in square millimeters) for each plane. In addition, the maximum height and length of the gland were measured from the midsagittal image, and the maximum width of the gland was measured from the coronal images. The hyperintense posterior pituitary was included in all measurements of the entire gland.
Indirect volume calculations were performed using the formula most commonly used by previous researchers: V = (length x height x width)/2 (2, 8, 10, 28, 29). This formula was originally derived from the formula of an ellipsoid [V = (4
/3) x (length x height x width)/2] to estimate the volume of the sella turcica on plain radiographs in 1960 (30).
All volume measurements were performed by one radiologist (S.K.), who was blinded to demographic data of the subjects. Volume measurements for 21 children were repeated after an interval of a minimum of 3 months to assess intraobserver reliability.
Statistical analysis
Height and BMI were converted to z-scores, adjusting for age and sex using the 1990 British growth reference (31). The concordance correlation coefficient (CCC) was calculated to assess the agreement between pituitary volume measured on coronal and sagittal planes (32). Spearmans correlation was used to measure the association between pituitary volume and pituitary height, width, and thickness. Reference values with 95% ranges were obtained by regression analysis, using fractional polynomials where necessary to handle nonlinearity (33, 34). Because the distribution of pituitary volumes was skewed, analysis was performed on the logarithms, with resulting summaries given as geometric means for the volume. Stata software were used for all statistical analyses (release 8.0, Stata Corp., College Station, TX).
| Results |
|---|
|
|
|---|
Of the 143 children thought to be eligible for inclusion, four children with height z-score absolute values greater than 2.2 were excluded; in the remaining sample, 5.0% (seven of 139) had heights outside the normal 95% range. The remaining total of 139 children up to the age of 10 yr, with 82 males and 57 females, were included in the study. Table 1
summarizes the clinical characteristics of the children.
|
We observed extremely wide variation in the morphology of the pituitary gland on high resolution MRI regardless of subject age. No two pituitaries were identical in shape or distribution of the isointense anterior gland and the hyperintense posterior gland. They varied in their relative antero-posterior length and relative height, whereas the shape varied from crescent-like to hemispherical and near-spherical. In some, the anterior height was greater; in others, the posterior height was greater; and some were dumbbell-shaped. The posterior pituitary bright spot could be elongated or flattened and extended variably in the anterior direction, often beneath the anterior portion of the gland (Fig. 1
).
|
Intraobserver reliability was high in the 21 repeated volume measurements, with a CCC of 0.986 [95% confidence interval (CI), 0.9670.994].
There was very strong agreement between the volumes measured from the coronal and sagittal reconstructions (CCC, 0.986; 95% CI, 0.9810.990; Fig. 2
). Because there was no substantial difference between the volumes obtained from the coronal and sagittal reconstructions, the coronal value was used for subsequent analysis.
|
|
|
|
|
| Discussion |
|---|
|
|
|---|
Prepubertal pituitary volumes were found to increase with age and appeared to be independent of gender. Independent of age, there was no correlation found between either height or BMI and pituitary volume.
Direct 3D assessment of pituitary volume represents a significant advance over methods used in the past with 2D datasets. In these studies, pituitary size was estimated using a single linear measurement (usually pituitary height) (11, 12, 13, 14, 16, 17, 18, 20) or by volumes crudely calculated using formulas adapted from the formula for the volume of an ellipsoid and originally used to estimate the volume of the sella turcica on plain films (10, 30). The vast variation in the shape of the normal pituitary gland, which we confirmed, means that neither linear measurements nor calculated volumes can give a true estimate of pituitary size. This is substantiated by the weak correlation we found between pituitary height and measured pituitary volume and between calculated volume and measured pituitary volume.
To date, the only other published normal data on whole pituitary volumes measured directly from volumetric data are the study by Takano et al. (23). This study included 199 Japanese children, 121 of whom were under the age of 10 yr. The mean prepubertal values and SDs given were reported in relatively broad age groupings of 01, 14, and 59 yr. There appears to be a slight discrepancy compared with our results, in that the measured pituitary volume values reported by Takano et al. are slightly higher, particularly in the 59 yr age group. We are unable to explain the variability between these two studies other than to note that there were several differences between the two studies, namely, the different ethnic compositions of the study populations and the grouped age intervals. Data for anterior pituitary volumes have also recently been published (35) in a European cohort. This cross-sectional study by Marziali et al. (35) included 95 children under 10 yr of age, again plotted according to age. The volumes reported by Marziali et al. (35) are only slightly greater than the pituitary volumes described in this study, notwithstanding the fact that only anterior pituitary volumes were measured. The significance of this is unknown given the limited data presented by Marziali et al. (35) about the number of patients in each yearly age grouping and the variability noted in their volume by age (i.e. pituitary volume was reported to decrease between ages 23 and 78 yr). The difference may simply reflect wide CIs due to low patient numbers when substratified at yearly age intervals.
When using normal reference values, it is important to ensure that study data acquisition has been performed in a similar fashion as reference data. A recent publication (2) calculated pituitary volumes from 2D images using the Di Chiro-derived formula and compared these crude estimates with the 3D true volumetric normative data reported by Takano et al. (23). We advocate using only volumetric MRI techniques and direct volume measurements when evaluating pituitary gland size, because these give the most reliable results.
In summary, we have presented normative, directly measured, 3D volumetric data for pituitary size in prepubertal children. Pituitary gland shape in this age group is highly variable; hence, we found that direct measures of pituitary volume do not correlate with either one-dimensional estimates (height) or 2D, indirectly calculated volumes. Our approach to direct measurement of pituitary volume appears to be robust, with assessment using either sagittal or coronal data reconstructions giving practically identical results. We recommend that with the advent of improved MRI technology and 3D data acquisition, indirect estimates of pituitary size and volume should be replaced by direct volumetric analysis.
| Acknowledgments |
|---|
| Footnotes |
|---|
First Published Online March 22, 2005
Abbreviations: BMI, Body mass index; CCC, concordance correlation coefficient; CI, confidence interval; 2D, two-dimensional; 3D, three-dimensional; MRI, magnetic resonance imaging; SPGR, spoiled gradient echo.
Received August 13, 2004.
Accepted March 15, 2005.
| References |
|---|
|
|
|---|
lu I, Kutlu H, Isgüven P, Tokus F, Isik K 2001 Diagnostic value of pituitary MRI in differentiation of children with normal growth hormone secretion, isolated growth hormone deficiency and multiple pituitary hormone deficiency. J Pediatr Endocrinol Metab 14:517523[Medline]
This article has been cited by other articles:
![]() |
R. W. Pfaeffle, J. J. Savage, C. S. Hunter, C. Palme, M. Ahlmann, P. Kumar, J. Bellone, E. Schoenau, E. Korsch, J. H. Bramswig, et al. Four Novel Mutations of the LHX3 Gene Cause Combined Pituitary Hormone Deficiencies with or without Limited Neck Rotation J. Clin. Endocrinol. Metab., May 1, 2007; 92(5): 1909 - 1919. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. P. S. Bhangoo, C. S. Hunter, J. J. Savage, H. Anhalt, S. Pavlakis, E. C. Walvoord, S. Ten, and S. J. Rhodes A Novel LHX3 Mutation Presenting as Combined Pituitary Hormonal Deficiency J. Clin. Endocrinol. Metab., March 1, 2006; 91(3): 747 - 753. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |