n
NEURORADIOLOGY
Hormone Defi ciency in the
Morphologically Normal Pituitary
Gland Is Associated with
Perfusion Delay
1
Chao-Ying Wang , BS Hsiao-Wen Chung , PhD Nai-Yu Cho , MS Hua-Shan Liu , PhD Ming-Chung Chou , PhD Hung-Wen Kao , MD Chun-Jung Juan , MD , PhD Meei-Shyuan Lee , PhD Guo-Shu Huang , MD Cheng-Yu Chen , MD
Purpose: To investigate quantitatively the topographic perfusion
characteristics of the adenohypophysis by using dynamic contrast material–enhanced magnetic resonance (MR) imaging in a subgroup of patients with idiopathic growth hormone defi ciency (IGHD) and with normal-appearing pituitary morphology on MR images.
Materials and Methods:
This HIPAA-compliant, prospective study was approved by an institutional review board, and informed consent was obtained for all patients. Twenty-fi ve patients (mean age, 10.6 years 6 3.3 [standard deviation]) with clinical growth retardation, proved IGHD, and normal pituitary morphology on MR images were included for analysis. Sixteen children (mean age, 10.8 years 6 5.5) were in-cluded as control subjects. Time to peak (TTP) perfusion properties of the adenohypophysis in 10 regions of inter-est from multisection coronal dynamic contrast-enhanced T1-weighted MR images were quantitatively derived by using the Brix pharmacokinetic model. Signifi cant differ-ence was determined with a two-tailed Student t test. The Pearson correlation coeffi cient was used to correlate the perfusion parameters, including maximal enhancement peak and slope, with serum growth hormone levels in the IGHD group.
Results: TTP for the IGHD group was signifi cantly prolonged
com-pared with that for the control group ( P , .005). The prolonged TTP in the IGHD group was found to be diffuse. The levels of growth hormone defi ciency were negatively correlated with the peak enhancement and the slope of the wash-in phase, which suggests increased blood volume in IGHD within the pituitary gland.
Conclusion: IGHD and the degree of growth hormone defi ciency are
associated with nonregional perfusion delay in morpho-logically normal adenohypophyses. The lack of lateraliza-tion of perfusion delay may suggest that microvascular structural abnormalities play a role in IGHD.
q RSNA, 2010
1 From the Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan (C.Y.W., H.W.C., H.S.L., C.J.J.); Department of Radiology, Tri-Service General Hospital, 325 Cheng-Kung Road, NeiHu, Taipei, Taiwan 114, Republic of China (C.Y.W., H.W.C., N.Y.C., H.S.L., H.W.K., C.J.J., G.S.H., C.Y.C.); Institute of Biomedical Engineering, Yang-Ming University, Taipei, Taiwan (N.Y.C.); Department of Medical Imaging and Radiological Sciences, Kaohsiung Medical University, Kaohsiung, Taiwan (M.C.C.); and School of Public Health, National Defense Medical Center, Taipei, Taiwan (M.S.L.). From the 2003 RSNA Annual Meeting. Re-ceived March 18, 2010; revision requested May 4; revision received June 3; accepted June 22; fi nal version accepted July 21. Supported in part by grant TSGH-C93-21, Medical Research Council under civil service of Tri-Service General Hospital. C.Y.C. supported in part by the National Science Council (grants NSC 98-2221-E-002-095-MY3 and NSC 97-2314-B-016-028-MY3). Address correspondence to C.Y.C. (e-mail: [email protected] .tw ).
On the other hand, the topographic distribution of the somatotroph cells, a functional organization of pituitary cells, has been shown to be in the lateral portions of the anterior pituitary gland in an immunochemical study ( 20 ). It would be interesting to know if growth hormone defi ciency is also topograph-ically dependent in a similar way in terms of perfusion alterations. There-fore, the purpose of our study was to in-vestigate quantitatively the topographic perfusion characteristics of the adeno-hypophysis by using dynamic contrast-enhanced MR imaging in a subgroup of patients with IGHD and with normal-appearing pituitary morphology on MR images.
Materials and Methods
Subjects
The prospective study protocol was approved by the National Defense Med-ical Center and Tri-Service General Hos-pital institutional review board, and informed consent was obtained from all children’s parents. This study was Health Insurance Portability and Ac-countability Act compliant. Between October 2002 and August 2009, 25 chil-dren with isolated IGHD (16 male pa-tients, nine female patients; mean age, 10.6 years 6 3.3 [standard deviation]; age range, 6–17 years) were enrolled age, these tests do not refl ect the
un-derlying physiologic abnormality of the pituitary gland. Physiologic factors such as perfusion to the pituitary gland may be an important factor infl uencing hor-mone secretion ( 12–14 ). The complex hypophyseal-portal vasculature system consists of numerous small branches of indirect supplying vessels from the su-perior and inferior hypophyseal arteries and has a feature of interarterial anasto-moses which may render adenohypoph-ysis vulnerable to ischemia or pituitary apoplexy ( 15,16 ). Moreover, because the blood-brain barrier is absent in the pituitary gland, perfusion as measured with T1-weighted MR imaging with in-travenous contrast medium is theoreti-cally feasible. Recent advances in MR technique have made it possible to as-sess the pituitary gland at higher spatial and temporal resolution and speed than previously reported in the literature. Dynamic contrast material–enhanced T1-weighted MR imaging can help exploit tissue perfusion properties by using ap-propriate tracer kinetic models ( 17–19 ). Because the pituitary gland is a small, pea-shaped structure contained within a sellar fossa abutting the air-containing sphenoid sinus and clivus that is rich with marrow, the surrounding air and bone could cause susceptibility artifacts, making dynamic contrast-enhanced MR imaging study of the gland more diffi cult. Hence, early studies ( 12–14 ) using one section with 3- to 4-mm thickness on a sagittal plane had major limitations in the observation of the regional pituitary perfusion properties.
I
diopathic growth hormone defi ciency(IGHD) is the most common type among children with growth hormone defi ciency. Although the pathogenesis of IGHD remains unclear, most inves-tigators have suggested that perinatal injuries to the hypothalamic-pituitary axis, such as breech presentation and neonatal hypoglycemia, could be the possible causes ( 1,2 ). These hypoth-eses are likely based on the observa-tions of morphologic alternaobserva-tions of the hypothalamus, pituitary stalk, and the pituitary gland in IGHD and in multiple pituitary hormone defi ciency ( 3,4 ), in which typical magnetic resonance (MR) imaging fi ndings often include a small pituitary gland and/or ectopic posterior lobe with or without a truncated stalk ( 5–7 ). However, morphologic changes in the pituitary gland may not be the sole determinant for inadequate secre-tion of growth hormone ( 8–10 ). Other underlying causes such as mutation of GH1 or impaired perfusion may play a role ( 11,12 ).
Although the severity of IGHD can be assessed by using endocrinologic tests or conventional radiography for bone
Implications for Patient Care
The knowledge that those with n
IGHD can have a normal-appearing adenohypophysis but abnormal pituitary perfusion may help radi-ologists better understand the limitations of morphologic MR imaging.
Model-based dynamic contrast-n
enhanced MR imaging may help when growth hormone defi ciency is suspected and when results of morphologic imaging of the pitu-itary gland are normal.
Advances in Knowledge
Characterization of topographic n
perfusion characteristics of the adenohypophysis by using dynamic contrast-enhanced MR imaging in patients with idiopathic growth hormone defi -ciency (IGHD) and with normal pituitary morphology is techni-cally feasible.
The time to peak (TTP) for the n
patients with IGHD was signifi -cantly prolonged compared with that for control subjects ( P , .005).
Prolonged TTP in the IGHD n
group was found to be diffuse. The levels of growth hormone n
defi ciency were negatively corre-lated with the peak enhancement and slope of the wash-in phase, which suggests increased blood volume in the IGHD pituitary gland ( P , .05).
Published online before print 10.1148/radiol.10100504
Radiology 2011; 258:213–221
Abbreviations:
IGHD = idiopathic growth hormone defi ciency ROI = region of interest
SNR = signal-to-noise ratio TTP = time to peak
Author contributions:
Guarantors of integrity of entire study, C.Y.W., C.Y.C.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript fi nal version approval, all authors; literature research, C.Y.W., H.W.C., M.C.C., C.Y.C.; clinical studies, C.Y.W., H.W.C., N.Y.C., H.S.L., M.C.C., H.W.K., C.J.J., G.S.H., C.Y.C.; statistical analysis, C.Y.W., N.Y.C., M.S.L.; and manuscript editing, C.Y.W., H.W.C., H.S.L., G.S.H., C.Y.C. Authors stated no fi nancial relationship to disclose.
T1-weighted sections for the assess-ment of the presence or absence of stalk abnormalities, the size of the anterior lobe, and the positions of the posterior lobe of the pituitary gland.
For dynamic contrast-enhanced MR imaging, a total of 36 images (three sec-tions, 12 frames each) were transferred to a personal computer workstation and processed by using in-house software developed with Matlab (MathWorks, Natick, Mass). Ten regions of inter-est (ROIs) (Table 1 ) with uniform shape and size (16–20 pixels, 5.76–7.2 mm 2 ),
including the infundibulum of the stalk and the central and bilateral parame-dian portions of the anterior pituitary gland in three continuous sections ( Fig 1 ), were chosen from the dynamic contrast-enhanced data sets by two experienced neuroradiologists (C.Y.C. and H.W.K.). To ensure reliability of the dynamic contrast-enhanced MR imaging data, quantitative analysis and error estima-tion were undertaken and are addressed in a later section. Quantitative perfu-sion parameters were analyzed sepa-rately on each ROI as follows: First, the signal intensity–time data were normal-ized and converted to relative tration-time data ( 17,18 ). The concen-tration-time data ( C t ) were then fi tted,
by using a nonlinear least-square curve fi tting algorithm, to the Brix pharma-cokinetic model as follows ( 19 ):
el 21 , 21 el K t k t A e e t k K § · ¨ ¸ © ¹ C
performed. Dynamic contrast-enhanced MR imaging was performed by employ-ing a fast spin-echo technique with the following parameters: repetition time msec/echo time msec, 450/15; num-ber of signals acquired, one; echo train length, seven; matrix size, 256 3 256; fi eld of view, 140 3 140 mm; and section thickness, 2 mm. Three contiguous coro-nal sections without intersection gap were allocated on a midsagittal T1-weighted fast spin-echo image covering the anterior pituitary gland. Twelve serial dynamic images from each of the three coronal sections were acquired with an 18-second interval after manual bolus injection of 0.1 mmol gadopentetate dimeglumine (Magnevist; Bayer Schering, Berlin, Germany) per kilogram of body weight into the antecubital vein. The contrast medium injection was immediately fol-lowed by 20 mL of normal saline solu-tion to fl ush the vein.
Data Analysis
Two neuroradiologists (C.Y.C. and H.W.K., with 20 and 5 years of experience, re-spectively), who were blinded to the subjects, evaluated the standard MR im-ages of the IGHD and control groups at different sessions. Conventional images of the wrist for bone age assessment were also reviewed in the same way but only for the IGHD group. When there was a difference in reading results, a consen-sus was made to complete the rating. Each rater measured the heights of the pituitary glands according to the method proposed previously ( 21 ). Special atten-tion was given to the midline sagittal for dynamic contrast-enhanced MR
im-aging study. The inclusion criteria in-cluded (a) short stature, less than the third percentile or below 2 standard deviations of mean age-matched popu-lation height, (b) less than 2 years of
skeleton maturation, (c) isolated growth hormone defi ciency, less than 10 ng/mL peak serum growth hormone level with at least two provocative stimulations, and (d) normal pituitary glandular
height, according to the reference values for those younger than 20 years from Argyropoulou et al ( 21 ). All children with IGHD were screened with a complete hypothalamic-pituitary function test to exclude the possibility of multiple pitu-itary hormone defi ciency. In addition, children with stalk abnormality, includ-ing absence, thinninclud-ing or transection, and/or ectopic posterior lobe of the pituitary gland, were excluded from the analysis. Five children were excluded according to the clinical criteria men-tioned above. Three additional subjects were excluded because of poor image quality caused by motion artifacts. Al-together, eight patients were excluded, leaving 25 in the IGHD group for anal-ysis. Another 16 children (eight male patients, eight female patients; mean age, 10.8 years 6 5.5; age range, 2–18 years), who had no clinical or laboratory evi-dence of pituitary dysfunction and were initially referred for MR study for reasons other than developmental delay such as headache, sinusitis, or orbital lesions, were included as the control group. The body heights and weights of the control group were within the normal ranges of the age-matched population. All control subjects had normal intracranial MR fi ndings.
Image Acquisition
MR studies were performed with a 1.5-T superconductive system (Magnetom Vi-sion Plus; Siemens Medical System, Er-langen, Germany) by using a standard birdcage head coil. Patients fi rst under-went standard axial fluid-attenuated in version recovery and T2-weighted imaging and precontrast axial, coronal, and sagittal T1-weighted imaging. Dy-namic contrast-enhanced imaging and subsequent postcontrast imaging were
Table 1
Ten ROIs and Corresponding Abbreviations
Abbreviation ROI
IS Infundibulum stalk
AR Right paramedian portion of the anterior section
AM Middle portion of the anterior section
AL Left paramedian portion of the anterior section
MR Right paramedian portion of the middle section
MM Middle portion of the middle section
ML Left paramedian portion of the middle section
PR Right paramedian portion of the posterior section
PM Middle portion of the posterior section
6.0 mm 6 1.2 and 5.2 mm 6 0.9, respec-tively, which shows no signifi cant differ-ence ( P . .05). In addition, the glandular heights were positively correlated with subject age ( R 2 = 0.55) ( Fig 2 ).
Fitting Error Assessment
Good curve fi tting of the concentration-time data by using the nonlinear least-square algorithm ( R 2 = 0.922 6 0.049
in the control group and 0.966 6 0.023 in the IGHD group) was obtained for our subjects. The residual errors in the signal enhancement of our data after nonlinear fi tting were small (9.16% 6 4.82 and 6.32% 6 2.40 in the control and IGHD groups, respectively), which was consistent with the high R 2 values
( Fig 3 ). Monte Carlo simulation showed that, at the SNR of 20, which was about the SNR level of all our dynamic contrast-enhanced data, the errors in A , k 21 , and K el were 33.5%, 30.1%, 32.8%, respec-tively. Because of high imprecision, these parameters were not used in subsequent comparisons. On the other hand, errors in C max , TTP, and slope were 8.22%, 7.76%, and 14.7%, respectively, which shows better estimation reliability. Regional Perfusion of Anterior Pituitary Gland between Groups
The 10 designated ROIs are illustrated on the coronal dynamic contrast-enhanced MR images shown in Figure 3a . Two rep-resentative concentration-time curves sampled from the posterior midline (or PM) ROIs of one IGHD and one control subject are shown in Figure 3b . Table 3 lists the perfusion parametric values derived from the 10 ROIs, as well as the statistical comparison results. Between-group comparison of the perfusion pa-ram eters with all ROIs averaged is shown in Figure 4 . TTP and wash-in slope showed signifi cantly higher and lower values, respectively, in the IGHD group than in the control group ( P , .0005 for TTP, P , .005 for slope). Peak enhance-ment, on the other hand, showed no significant difference ( P . .05)
be-tween the two groups. Regional analy-sis showed that TTP of the IGHD group was signifi cantly delayed compared with that of the control group for all 10 ROIs ( P , .005) ( Fig 5 ).
18 seconds to simulate the image acquisi-tion procedure. Gaussian white noise was subsequently added to the concentration-time data at 10 SNR levels ranging from 5 to 50, with each SNR level containing 1000 sets of random noise. Nonlinear least-square error fitting to the Brix model was again performed on these simulated data to derive the three per-fusion parameters ( A, k 21 , K el ) and was compared with the original values, with errors expressed as percentage devia-tion from the true values. Similarly, the percentage errors in peak enhancement, TTP, and slope of the wash-in phase were assessed.
Statistical Analysis
Sex and age distributions were exam-ined by using the x 2 test. The Student
t test was used to examine the signifi -cant difference in perfusion parameters between the control and isolated IGHD groups. Correlation of the perfusion pa-rameters with the growth hormone de-fi ciency levels was made for the IGHD group by using the Pearson correlation coeffi cient. A P value less than .05 was considered to indicate a statistically sig-nifi cant difference.
Results
The demographic data of the IGHD group, including sex, age, bone age, body height and weight, and growth hormone level, are listed in Table 2 . There was no signifi cant difference in the distribu-tion of age ( P = .17) and sex ( P . .99) between the control and IGHD groups. The mean heights of the pituitary glands in the control and IGHD groups were where A is the amplitude scaling
con-stant for the concentration-time curve of the plasma determined by factors such as the blood volume of the pitu-itary gland and so forth ( 17,18 ), K el is the elimination transfer rate constant that describes the excretion of contrast agent through the kidneys, and k 21 is the transfer rate constant describing the return of contrast agent from the extracellular extravascular space of the pituitary tissue to the plasma com-partment. Goodness of fi t was assessed by R 2 values as used in nonlinear curve
fi tting ( 22 ). After curve fi tting, the ana-lytic concentration-time data were used to obtain A, K el , and k 21 , followed by the derivation of three wash-in indexes: The peak enhancement ( C max ) was de-fi ned as the maximal value of contrast agent concentration, the time to peak (TTP) ( T max ) was defi ned as the time duration to reach C max , and the slope of the wash-in phase was defi ned as C max / T max .
Error Assessment
Monte Carlo simulation at different levels of additive noise was used to estimate possible errors in the perfusion param-eters derived from data sets acquired with fi nite temporal resolution and signal-to-noise ratio (SNR). A typical concentration-time curve obtained from one subject was taken fi rst, along with the values of the perfusion parameters being recorded. An analytic concentration-time curve was generated from the perfusion parameters and treated as the noise-free reference curve. The reference curve was then sampled at an interval of
Figure 1
Figure 1: Schematic draw-ing of 10 ROIs selected for regional perfusion analysis from dynamic contrast-enhanced MR imaging acquisitions. N/A = not applicable.
Correlation of Growth Hormone Levels with Perfusion Parameters
The levels of growth hormone showed negative correlation with the wash-in slopes and peak enhancements of the entire anterior pituitary gland ( r =
–0.430, P = .036 and r = –0.493, P = .014, respectively) in the IGHD group ( Fig 6 ). No signifi cant correlation was found between the levels of growth hor-mone and TTP. When examining the in-dividual ROIs, correlations existed only in the posterior right, posterior middle, posterior left, and middle right regions ( P , .05).
Discussion
Our results were in concert with previ-ous works in that the prolonged TTP enhancement and decreased wash-in slopes in the anterior pituitary lobe suggest insuffi cient perfusion, accom-panied by a compensatory increase in blood volume, which is particularly true in those who had lower serum growth hormone level ( Fig 6 ). The defected hy-pophyseal-portal system could compen-sate with higher perfusion effi ciency, as we have observed signifi cantly higher slope which was dominant in the poste-rior (posteposte-rior right, posteposte-rior middle, and posterior left) and the middle (mid-dle right) sections of the anterior pitu-itary gland. The higher peak enhance-ment as seen in the posterior (right, middle, and left) and the middle (mid-dle right) sections of anterior pituitary gland would suggest increased blood volume (a vasodilatory response) of the gland when cell apoptosis progresses.
Given the fact that most IGHD and multiple pituitary hormone defi ciency cases commonly harbor structural abnor-malities in the hypothalamic-pituitary axis, a subgroup of children with IGHD like our cases with normal morphology of the pituitary gland is considered un-common, and the underlying pathogen-esis becomes less consistent with those exhibiting a truncated stalk and a his-tory of perinatal injury ( 23,24 ). From the diagnostic MR imaging perspective, functional rather than anatomic infor-mation may be needed to exploit this subgroup of patients when conventional
Table 2
Clinical Data of IGHD Group
Patient No./Sex/Age (y)
Bone Age ( , 2 years) * Body Height (percentile) Body Weight (percentile) Growth Hormone Level (ng/mL) 1/F/13 11 , 3rd 6th 3.2 2/M/12 8 , 3rd 60th 5.7 3/F/11 9 , 3rd , 3rd 8.6 4/F/11 8 15th 15th 7.1 5/M/13 11 , 3rd 15th 5.3 6/F/13 11 3rd . 97th 1.4 7/M/5 2 , 3rd , 3rd 3.9 8/F/6 4 , 3rd , 3rd 6.7 9/M/7 4 4th 25th 9.6 10/M/11 8 8th 20th 9.1 11/M/6 4 10th 8th 8.6 12/M/12 5 , 3rd 90th 0.1 13/M/8 5 , 3rd , 50th 5.5 14/M/10 5 , 3rd 75th 4.3 15/M/10 7 , 3rd 13th 6.2 16/F/13 11 , 3rd , 3rd 2.2 17/M/7 4 , 3rd , 3rd 1.2 18/M/7 3 5th 10th 9.5 19/M/13 10 , 3rd 20th 3.5 20/F/18 13 , 3rd , 3rd 2.3 21/M/16 14 , 3rd 9th 8.7 22/F/11 9 5th 50th 5.2 23/M/7 4 , 3rd 4th 2.5 24/M/16 11 , 3rd , 3rd 6.1 25/F/12 10 , 3rd , 3rd 7.5
Note.—Average patient age was 10.6 years 6 3.3. Average serum growth hormone level was 5.36 ng/mL 6 2.8. * Consists of less than 2 years of skeleton maturation.
Figure 2
Figure 2: Graph shows distribution of anterior pituitary height for control (6.0 mm 6 1.2) and IGHD (5.2 mm 6 0.9) subjects, both with gland size falling within the normal range suggested in the literature as a function of age. Solid line = mean. Dotted lines = upper and lower bounds. Linear regression showed positive correlation of gland height with age ( R 2 = 0.55), which was not signifi cantly different compared with control subject data reported in the literature ( P = .384).
system. While it is understandable that delayed enhancement of the anterior pituitary lobe may occur in growth hor-mone defi ciency or multiple pituitary hormone defi ciency with stalk abnor-malities because the main blood sup-ply to the adenohypophysis is from the portal vasculature, it may be debatable when this occurs in patients with intact stalk morphology. In the Kornreich et al study ( 28 ), MR imaging fi ndings of pi-tuitary structural abnormality have con-tributed to the prediction of patterns and severity of hypopituitarism. It may ap-pear that both types of IGHD (ie, types with and types without normal pituitary morphology) sustain the common feature of perfusion defi cits, and the severity of perfusion impairments goes along with the levels of growth hormone defi ciency as seen in our cases.
The functional cytomorphology of adenohypophysis with immunochemistry techniques is well documented ( 29 ). Growth hormone (somatotroph) cells account for approximately 50% of the pituitary cell mass and are anatomically distributed in the lateral and postero-lateral portions of the adenohypophysis. Although the one-cell, one-hormone the-ory is no longer tenable, given that pluri-potential cells are detected with more advanced techniques ( 30 ), somatotroph cells remain the main growth hormone secretary cells. The topographic later-alization of the somatotroph cells in ad-enohypophysis suggests that the dimin-ished growth hormone secreting activity may be accompanied by decreased vas-cular perfusion in the capillary bed. To prove this point, a multiregional com-parison of the perfusion properties in the entire anterior pituitary lobe with respect to that of a control is required. In our study, we conducted multisection and multiregional perfusion analysis for this purpose, and the results revealed region-independent perfusion delay in the IGHD group. This may explain in part that the pathogenesis of IGHD is multifactorial, and a single hormone defi ciency could accompany global per-fusion delay in the context of compli-cated vascular anatomy involved in the anterior pituitary gland. On the other hand, the fi ndings of the overwhelming Impaired blood perfusion to the
anom-alous hypothalamic-pituitary system in children with IGHD or multiple pituitary hormone defi ciency has been implicated in two early studies with single-section dynamic contrast-enhanced MR imaging at low temporal and spatial resolution ( 12,14 ). Both studies showed delayed enhancement in growth hormone defi -ciency and multiple pituitary hormone defi ciency groups, which suggests that abnormality of vasculature exists in hy-pophyseal arteries and the portal venous MR imaging results are normal. On the
other hand, the growth responses of an individual patient to growth hormone treatment are variable. Growth hormone treatment may induce adverse events such as leukemia, benign intracranial hypertension, type II diabetes mellitus, and even death ( 25–27 ). A functional im-aging tool for pituitary response moni-toring may help to individualize patient treatments, especially in cases with pro-gression from IGHD to multiple pitu-itary hormone defi ciency.
Figure 3
Figure 3: (a) Coronal contrast-enhanced T1-weighted (450/15) MR images in 12-year-old female control subject show typical selection of ROIs (16–20 pixels each, 5.76–7.2 mm 2 ), according to the guidance illustrated in Figure 1. (b) Typical relative concentration-time curves generated from the posterior mid-line (PM) region, along with their fi tting results. Note the delayed TTP for the IGHD enhance-ment compared with control enhancement.
diagnosis, as opposed to the use of group average for the concentration-time curves ( 12 ). Furthermore, quantitative derivation of the perfusion parameters allowed an objective exploration of the phenomena rather than merely descrip-tive investigations ( 14 ). Missing any el-ements from above might risk sub-stantial lowering of the measurement accuracy.
There were some limitations to our study. First, the case numbers were small because of the relatively uncom-mon clinical entity and strict inclusion criteria. This, however, can be partially compensated by rigorous data collec-tion and analysis of the dynamic MR data sets. That is why we tested the reliability supply to the hypophyseal-portal system
in dwarfi sm ( 12,14,32,33 ), several techni-cal features used in our study provided key improvements toward successful quantifi cation of the pituitary perfusion. Fast specho acquisition was used in-stead of rapid gradient-echo acquisition ( 34,35 ) for dynamic imaging; thus, the possible susceptibility-induced signal loss from the air-tissue interface was alleviated. An appropriate pharmacoki-netic model was used for curve fi tting, with validating simulation performed; thus, the identifi cation of peak enhanc-ement was relatively immune to noise effects ( 36 ). With successful curve fi t-ting, data analysis can be performed on a case-by-case basis as in routine incidence of congenital anomalies in
children with IGHD from the study by Garel and Léger ( 31 ) suggest that un-enhanced MR imaging is appropriate for regular morphologic diagnosis. Our study further implied that, in a sub-group of children with IGHD and with normal pituitary morphology, dynamic contrast-enhanced MR imaging may help to delineate the intrinsic defi cits such as delayed perfusion and increased blood volume in the adenohypophysis.
Although previous studies have shown similar delayed wash-in for the blood
Table 3
Regional Difference between IGHD and Control Group according to Perfusion Parameters
Perfusion Parameter Region PR PM PL IS MR MM ML AR AM AL Peak enhancement (%) * Control group 89.6 6 43 89.2 6 43 97.9 6 35 77.3 6 32 136.6 6 33 143.9 6 59 139.8 6 27 129 6 32 133.3 6 49 124.2 6 33 IGHD group 102.4 6 51 90.6 6 46 104.6 6 48 72 6 36 134.2 6 51 130.9 6 44 124.7 6 45 114.9 6 32 128.7 6 50 117.6 6 45 TTP (sec) † Control group 50.4 6 12 31.2 6 15 57.6 6 18 28.3 6 15 62.4 6 21 39.6 6 19 64.8 6 17 65.6 6 19 48.9 6 18 64.3 6 18 IGHD group 95.8 6 27 62.6 6 17 100.8 6 29 45 6 21 96.5 6 26 68.4 6 20 102.2 6 34 111.1 6 44 86.9 6 34 121.3 6 43 Slope (percentage per
second) †
Control group 1.8 6 1 2.9 6 1 1.7 6 0 3.1 6 1 2.4 6 1 4.2 6 2 2.4 6 1 2.2 6 1 3.2 6 2 2.1 6 1
IGHD group 1.1 6 1 1.6 6 1 1.1 6 1 1.8 6 1 1.5 6 1 2.1 6 1 1.4 6 1 1.3 6 1 1.7 6 1 1.1 6 1
Note.—Data are means 6 standard deviations. * P value for all ROIs was .05.
† P value for all ROIs was .005.
Figure 4
Figure 4: Graph shows comparison of TTP, peak enhancement, and wash-in slope of the entire anterior gland between the control and IGHD groups. The units of measure for the vertical axis are percentage per second for slope, percentage for peak enhancement, and second for TTP. Error bars = standard deviations. ∗ = signifi cant difference. Data show delayed TTP and reduced wash-in slope in IGHD group at P less than .0005 and P less than .005 levels, respectively.
Figure 5
Figure 5: Graph of regional analysis of the TTP shows that IGHD group exhibits TTP values 1.5 to two times greater than that of the control group in all 10 ROIs ( P , .005). In addition, the median parts (PM, IS, MM, and AM) of the anterior hypophysis showed faster wash-in than the bilateral parts (PR, PL, MR, ML, AR, and AL) in both groups ( P , .005). ∗ = signifi cant difference.
stalk transection . J Clin Endocrinol Metab 1988 ; 67 ( 4 ): 817 – 823 .
4 . Murray PG , Hague C , Fafoula O , et al . Likelihood of persistent GH defi ciency into late adolescence: relationship to the pres-ence of an ectopic or normally sited poste-rior pituitary gland . Clin Endocrinol (Oxf) 2009 ; 71 ( 2 ): 215 – 219 .
5 . Kelly WM , Kucharczyk W , Kucharczyk J , et al . Posterior pituitary ectopia: an MR feature of pituitary dwarfi sm . AJNR Am J Neuroradiol 1988 ; 9 ( 3 ): 453 – 460 .
6 . Root AW , Martinez CR , Muroff LR . Subhy-pothalamic high-intensity signals identifi ed by magnetic resonance imaging in children with idiopathic anterior hypopituitarism: evidence suggestive of an ‘ectopic’ poste-rior pituitary gland . Am J Dis Child 1989 ; 143 ( 3 ): 366 – 367 .
7 . Fujisawa I , Kikuchi K , Nishimura K , et al . Transection of the pituitary stalk: development of an ectopic posterior lobe assessed with MR imaging . Radiology 1987 ; 165 ( 2 ): 487 – 489 . 8 . Zucchini S , di Natale B , Ambrosetto P ,
De Angelis R , Cacciari E , Chiumello G . Role of magnetic resonance imaging in hypotha-lamic-pituitary disorders . Horm Res 1995 ; 44 ( suppl 3 ): 8 – 14 .
9 . Bozzola M , Adamsbaum C , Biscaldi I , et al . Role of magnetic resonance imaging in the diagnosis and prognosis of growth hormone defi ciency . Clin Endocrinol (Oxf) 1996 ; 45 ( 1 ): 21 – 26 .
10 . Inoue Y , Nemoto Y , Fujita K , et al . Pituitary dwarfi sm: CT evaluation of the pituitary gland . Radiology 1986 ; 159 ( 1 ): 171 – 173 .
the anterior pituitary gland, dynamic contrast-enhanced MR imaging may be potentially feasible to monitor the ther-apeutic response and to help in clinical decision making, especially for children with IGHD and normal morphology. Finally, although all control subjects had normal body heights, serum growth hormone levels and bone age informa-tion were not obtained because of the original experiment design that was partly compromised by the traditional Chinese folklore and ethical issue.
In conclusion, dynamic contrast-enhanced T1-weighted fast spin-echo MR imaging and Brix model analysis allow multisection and multiregional quantitative evaluation of the anterior pituitary gland in patients with IGHD and with normal pituitary morphology. Acknowledgment: The authors thank Shih-I Tsao, BS, for MR data retrieving.
References
1 . Craft WH , Underwoood LE , Van Wyk JJ . High incidence of perinatal insult in children with idiopathic hypopituitarism . J Pediatr 1980 ; 96 ( 3 pt 1 ): 397 – 402 .
2 . Kucharczyk W . Etiology of congenital growth hormone defi ciency . AJNR Am J Neuroradiol 2000 ; 21 ( 6 ): 999 – 1000 .
3 . Kikuchi K , Fujisawa I , Momoi T , et al . Hypothalamic-pituitary function in growth hormone-defi cient patients with pituitary and SNR dependency of the perfusion
parameter calculations. Simulation re-sults indicated that the estimation of the washout parameter K el was prone to large errors, which originated from inadequate temporal coverage of our dynamic contrast-enhanced MR imag-ing protocol spannimag-ing less than 4 min-utes. Accompanying the K el imprecision were the consequently large errors in A and k 21 . As a result, we restricted our data analysis within the wash-in phase for an accurate derivation of TTP, C max , and the wash-in slope. Second, the temporal and spatial resolution of the dynamic contrast-enhanced MR images were still limited, though having been much improved compared with earlier studies, as far as the size of the anterior pituitary gland is concerned. Further technical improvements in dynamic MR imaging with shorter imaging time and higher temporal resolution, while main-taining a suffi cient SNR level for reliable quantitative analysis, can help in this regard. Because we did not perform serial dynamic contrast-enhanced MR imaging studies individually over time, the longitudinal sensitivity of our per-fusion quantifi cation and its correlation to hormone level requires further inves-tigation. Given that a reasonable SNR can be achieved and perfusion quantifi -cation can be done in small organs like
Figure 6
Figure 6: Graphs of growth hormone level in IGHD show negative correlation with (a) wash-in slope ( r = 2 0.430, P = .036) and (b) peak enhancement ( r = 2 0.493, P = .014) for the entire anterior pituitary. No correlation was found between the growth hormone level and TTP. When investigating individual ROIs, similar correlations existed only in PR, PM, PL, and MR regions ( P , .05, data not shown). The results may suggest that the impaired perfusion in IGHD presented as delayed wash-in is accompanied by a compensating increase in blood volume, a vasodilatory response which was dominant in the posterior (PR, PM, and PL) and middle (MR) sections of the anterior pituitary gland.
11 . Binder G , Nagel BH , Ranke MB , Mullis PE . Isolated GH defi ciency (IGHD) type II: imaging of the pituitary gland by magnetic resonance reveals characteristic differences in compari-son with severe IGHD of unknown origin . Eur J Endocrinol 2002 ; 147 ( 6 ): 755 – 760 . 12 . Liu HM , Li YW , Tsai WY , Su CT . Dynamic
enhancement MRI of anterior lobe in pitu-itary dwarfi sm . Neuroradiology 1995 ; 37 ( 6 ): 486 – 490 .
13 . Tien RD . Sequence of enhancement of various portions of the pituitary gland on gadolinium-enhanced MR images: correlation with re-gional blood supply . AJR Am J Roentgenol 1992 ; 158 ( 3 ): 651 – 654 .
14 . Maghnie M , Genovese E , Aricò M , et al . Evolving pituitary hormone defi ciency is as-sociated with pituitary vasculopathy: dynamic MR study in children with hypopituitarism, diabetes insipidus, and Langerhans cell his-tiocytosis . Radiology 1994 ; 193 ( 2 ): 493 – 499 . 15 . Sheehan HL , Stanfi eld JP . The pathogen-esis of post-partum necrosis of the anterior lobe of the pituitary gland . Acta Endocrinol (Copenh) 1961 ; 37 ( 4 ): 479 – 510 .
16 . Reid RL , Quigley ME , Yen SS . Pituitary apo-plexy: a review . Arch Neurol 1985 ; 42 ( 7 ): 712 – 719 .
17 . Buckley DL , Kerslake RW , Blackband SJ , Horsman A . Quantitative analysis of multi-slice Gd-DTPA enhanced dynamic MR images using an automated simplex minimization procedure . Magn Reson Med 1994 ; 32 ( 5 ): 646 – 651 .
18 . Larsson HB , Stubgaard M , Frederiksen JL , Jensen M , Henriksen O , Paulson OB . Quan-titation of blood-brain barrier defect by mag-netic resonance imaging and gadolinium-DTPA in patients with multiple sclerosis and brain tumors . Magn Reson Med 1990 ; 16 ( 1 ): 117 – 131 .
19 . Brix G , Semmler W , Port R , Schad LR , Layer G , Lorenz WJ . Pharmacokinetic parameters in CNS Gd-DTPA enhanced MR imaging . J Comput Assist Tomogr 1991 ; 15 ( 4 ): 621 – 628 . 20 . Elster AD . Modern imaging of the pituitary .
Radiology 1993 ; 187 ( 1 ): 1 – 14 .
21 . Argyropoulou M , Perignon F , Brunelle F , Brauner R , Rappaport R . Height of normal pituitary gland as a function of age evaluated by magnetic resonance imaging in children . Pediatr Radiol 1991 ; 21 ( 4 ): 247 – 249 . 22 . Bates D , Watts D . Nonlinear regression:
it-erative estimation and linear approximations. In: Nonlinear regression analysis and its appli-cations . New York, NY : Wiley , 1988 ; 33 – 66 . 23 . Kuroiwa T , Okabe Y , Hasuo K , Yasumori K ,
Mizushima A , Masuda K . MR imaging of pi-tuitary dwarfi sm . AJNR Am J Neuroradiol 1991 ; 12 ( 1 ): 161 – 164 .
24 . Tillmann V , Tang VW , Price DA , Hughes DG , Wright NB , Clayton PE . Magnetic reso-nance imaging of the hypothalamic-pituitary axis in the diagnosis of growth hormone de-fi ciency . J Pediatr Endocrinol Metab 2000 ; 13 ( 9 ): 1577 – 1583 .
25 . Blethen SL , Allen DB , Graves D , August G , Moshang T , Rosenfeld R . Safety of recombi-nant deoxyribonucleic acid-derived growth hormone: The National Cooperative Growth Study experience . J Clin Endocrinol Metab 1996 ; 81 ( 5 ): 1704 – 1710 .
26 . Taback SP , Dean HJ . Mortality in Canadian children with growth hormone (GH) de-fi ciency receiving GH therapy 1967-1992. The Canadian Growth Hormone Advisory Committee . J Clin Endocrinol Metab 1996 ; 81 ( 5 ): 1693 – 1696 .
27 . Nishi Y , Tanaka T , Takano K , et al . Recent status in the occurrence of leukemia in growth hormone-treated patients in Japan: GH Treatment Study Committee of the
Foun-dation for Growth Science, Japan . J Clin Endocrinol Metab 1999 ; 84 ( 6 ): 1961 – 1965 . 28 . Kornreich L , Horev G , Lazar L , Schwarz M ,
Sulkes J , Pertzelan A . MR fi ndings in growth hormone defi ciency: correlation with sever-ity of hypopituitarism . AJNR Am J Neurora-diol 1998 ; 19 ( 8 ): 1495 – 1499 .
29 . Horvath E , Kovacs K . Fine structural cytology of the adenohypophysis in rat and man . J Electron Microsc Tech 1988 ; 8 ( 4 ): 401 – 432 . 30 . Scheithauer BW , Horvath E , Kovacs K , Laws ER Jr , Randall RV , Ryan N . Plurihormonal pi-tuitary adenomas . Semin Diagn Pathol 1986 ; 3 ( 1 ): 69 – 82 .
31 . Garel C , Léger J . Contribution of magnetic resonance imaging in non-tumoral hypopitu-itarism in children . Horm Res 2007 ; 67 ( 4 ): 194 – 202 .
32 . Sato N , Sze G , Endo K . Hypophysitis: endo-crinologic and dynamic MR fi ndings . AJNR Am J Neuroradiol 1998 ; 19 ( 3 ): 439 – 444 . 33 . Manfré L , Midiri M , Rosato F , Janni A ,
Lagalla R . Perfusion MRI in normal and ab-normal pituitary gland: a preliminary study . Clin Imaging 1997 ; 21 ( 5 ): 311 – 318 .
34 . Sakamoto Y , Takahashi M , Korogi Y , Bussaka H , Ushio Y . Normal and abnormal pitu-itary glands: gadopentetate dimeglumine-enhanced MR imaging . Radiology 1991 ; 178 ( 2 ): 441 – 445 .
35 . Stadnik T , Stevenaert A , Beckers A , Luypaert R , Buisseret T , Osteaux M . Pituitary microad-enomas: diagnosis with two- and three-dimensional MR imaging at 1.5 T before and after injection of gadolinium . Radiology 1990 ; 176 ( 2 ): 419 – 428 .
36 . Maier C , Riedl M , Clodi M , et al . Dynamic contrast-enhanced MR imaging of the stim-ulated pituitary gland . Neuroimage 2004 ; 22 ( 1 ): 347 – 352 .
