Pituitary R2 values at 3T to assess risk of iron-mediated hypogonadal hypogonadism.
Andrew L Cheng1, Thomas D Coates1, and John C Wood2

1Pediatrics, Children's Hospital Los Angeles, Los Angeles, CA, United States, 2Pediatrics and Radiology, Children's Hospital Los Angeles, Los Angeles, CA, United States


Pituitary R2 at 1.5 Tesla has been validated as a sensitive marker of pituitary siderosis and risk of clinical hypogonadism. We cross-validated pituitary R2 measurements at 3T and 1.5T in 26 patients with iron overload syndromes. Pituitary R2 scaled linearly across field strength with a relative enhancement of 42%, consistent with previous liver R2 cross-field validations. When 3T pituitary values were transformed into equivalent 1.5T R2 values, the resulting Z-score estimates were unbiased with native 1.5T R2 estimates. Thus it is not necessary to acquire normative R2 data at 3 Tesla in order to interpret 3T pituitary R2 values.


Patients with transfusional siderosis developed iron deposition in the endocrine glands once transferrin binding capacity is overwhelmed. The pituitary gland is one of the earliest sites of iron deposition and the least able to tolerate the resulting oxidative stress. In fact, 50% of patients with thalassemia in the United States suffer from hypogonadal hypogonadism resulting from pituitary iron deposition1. We previously demonstrated that pituitary R2 at 1.5T identifies patients at risk for clinical hypogonadism before irreversible volume loss of the gland occurs2. We also defined age and sex specific nomograms for pituitary R2 and pituitary volume at 1.5T3. Since most neuroimaging is performed at 3T and the majority of newer magnet installations are also 3T, it is critical to cross calibrate pituitary R2 values at 3T and 1.5T. We hypothesized that R2 at 3T would have a linear relationship with R2 at 1.5T with a relaxivity enhancement coefficient close to the values observed for 3T liver R2 measurements 4.


From 5/1/14 – 2/9/18, 26 patients with transfusional siderosis (8 with sickle cell disease, 11 with thalassemia major, four with thalassemia intermedia, one with leukemia, one with Blackfan Diamond syndrome and one with hereditary hemochromatosis) underwent head MRI examinations at 1.5T and 3T within one month of each other; five patients underwent two examinations spaced more than one year apart. Study was approved by the Committee on Clinical Investigation (CCI#2014-00034) and all patients provided informed consent. The study population was composed of six Caucasians, six Africans, four Chinese, two Vietnamese, two Indians, two Mediterraneans, and two subjects of mixed race. Seven were ethnically Hispanic and 19 were non Hispanic. There were 19 females and 7 males with ages of 22.0 ± 8.7 [range 5.3 – 46.0]. Both the 1.5 and 3.0 Tesla examinations were performed using an 8-element head coil on a Philips Achieva running system 5.1.9. Pituitary R2 was performed using a five slice, standard mulitiple echo, spin echo examination in the sagittal plane. Field of view was 20.8 cm, voxel size 1 x 1 x 3 mm, five slices with no gap, TR 875 ms, TE 15, 30, 45, 60, 75, 90, 105, 120 ms, bandwidth 218 Hz/pixel. Images were fit pixelwise to an exponential plus a constant as previously described2,3. Values from the central three slices were averaged. We calculated the relaxivity enhancement(RE) between 3T and 1.5T as follows:

R23T-R23Tref = RE x (R21.5T-R21.5Tref) [1]

where R23T and R21.5T represent the measured R2 values at the respective field strength and R23Tref and R21.5Tref were their corresponding reference values, representing a common point on the 3T and 1.5T calibration curves. Suitable reference points could include the intrinsic R2 when no iron is present or population norms4. We exploited the linear relationship between R2 and age at 1.5T to set R21.5Tref to 10.7 ms, the expected pituitary R2 value at birth2. Thus equation [1] could be recast as follows:

R23T = RE x (R21.5T-10.7) + R23Tref [2]

where RE and R23Tref could be interpreted as simply the slope and intercept of the regression equation between R23T and the difference of R21.5T and 10.7 ms. To correct for multiple measurements in five patients, we weighted each R2 value by 0.5 in patients having two examinations. Once RE and R23Tref were estimated, R23T values were transformed into its R21.5T-equivalent, and the corresponding Z-score, using the following relationships4:

R21.5Tequivalent = (R23T-R23Tref)/RE + 10.7 Hz [3]

Z-score = (R23T-R23Tref)/RE – 0.331xAge)/0.88 [4]


Figure 1 demonstrates that R23T rises linearly with R21.5T-R21.5Tref (r2=0.91, p<0.0001). Relaxivity enhancement is estimated to be 1.42 ± 0.08 and R23Tref to be 12.0 ± 0.47 Hz. The residuals of Figure 1 were independent of age and sex. Figure 2 demonstrates the Bland Altman agreement between the pituitary R2 Z-scores calculated from the 3T (using equations 3 and 4) and 1.5 T values. There was no bias and the limits of agreement were approximately 2.1.


In this paper, we demonstrate that pituitary R2 values scale predictably with field strength and can be linearly transformed into 1.5T equivalents, similar to liver iron4. Using this relationship, we can exploit previous 1.5T normative R2 values derived from 100 healthy subjects rather than re-deriving those norms at 3 Tesla3. Although the 95% limits of agreement between the two measurements was roughly two standard deviations, this is small compared with the effect size associated with clinical hypogonadism (Z score > 5)3. Thus pituitary R2 values at 3T can be used with confidence to screen patients for clinically relevant pituitary iron burdens.


Acknowledgements: This work was supported by the National Institute of Kidney, Digestive and Kidney Disease (1R01DK097115-01A1) and by support in kind from Philips Healthcare. The authors are grateful for the contributions of Bertin Valdez, Thomas Hofstra, Jackie Baskim, Sue Carson, Trish Peterson and Debbie Harris for patient recruitment.


1. Vogiatzi MG, Macklin EA, Trachtenberg FL, Fung EB, Cheung AM, Vichinsky E, Olivieri N, Kirby M, Kwiatkowski JL, Cunningham M, Holm IA, Fleisher M, Grady RW, Peterson CM, Giardina PJ. Differences in the prevalence of growth, endocrine and vitamin D abnormalities among the various thalassaemia syndromes in North America. Br J Haematol. 2009;146(5):546-556.

2. Noetzli LJ, Panigrahy A, Mittelman SD, Hyderi A, Dongelyan A, Coates TD, Wood JC. Pituitary iron and volume predict hypogonadism in transfusional iron overload. Am J Hematol. 2012;87(2):167-171.

3. Noetzli LJ, Panigrahy A, Hyderi A, Dongelyan A, Coates TD, Wood JC. Pituitary iron and volume imaging in healthy controls. AJNR Am J Neuroradiol. 2012;33(2):259-265.

4. Ghugre NR, Doyle EK, Storey P, Wood JC. Relaxivity-iron calibration in hepatic iron overload: Predictions of a Monte Carlo model. Magn Reson Med. 2015;74(3):879-883.


Figure 1: Pituitary R2 at 3T rises linearly with the difference of the corresponding R2 at 1.5T and the 1.5T reference value (pituitary R2 of a newborn baby). One outlier (highest R2 value) was excluded because of excessive signal loss.

Figure 2: Bland Altman agreement between the pituitary Z-score predicted from 3T R2 versus 1.5T R2 measurements. 3T and 1.5T Z-scores are unbiased with one another and have 95% confidence intervals of +/- 2 Z-scores.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)