Endocrine Abstracts (2017) 49 EP1054 | DOI: 10.1530/endoabs.49.EP1054

Quantitative analysis of the pituitary gland at magnetic resonance imaging in obese patients

Giulia Puliani1, Carla Lubrano1, Marco Fiorelli2, Emilia Sbardella1, Chiara Graziadio1, Daniele Gianfrilli1, Andrea Lenzi1 & Andrea M Isidori1


1Department of Experimental Medicine – Sapienza University of Rome, Rome, Italy; 2Department of Neurology and Psychiatry – Sapienza University of Rome, Rome, Italy.


Background: Hypothalamic-pituitary function and adipose tissue are deeply interconnected. Obesity has been linked to morpho-functional changes in several endocrine glands (thyroid, adrenal, gonads). No study has quantified the appearance of the pituitary gland in obese subjects.

Aim: To measure pituitary signal intensity, homogeneity and texture of the pituitary gland at magnetic resonance (MR) in obese subjects free of focal pituitary disease.

Materials and Methods: Sixty-four overweight and obese patients were prospectively enrolled and underwent metabolic, hormonal, body composition (DEXA scan) and pituitary MR assessment. Texture of the pituitary gland was quantified recording pixel density and distribution at sagittal and coronal non-enhanced T1-weighted images using ImageJ software. Two operators independently placed the region of interest to entirely cover the pituitary gland, calculating mean intensity and its standard deviation. All analyses were normalized for both white and gray brain matter intensity. Pituitary volume (PV) was calculated using the ellipsoid formula.

Results: MR showed an empty sella in 12/64 (18.7%) patients. Mean PV was 384±147 mm3 (427±120 mm3, excluding the empty sella). An inverse correlation was found between PV and ultrasensitive C-reactive-protein (P=0.004) and a borderline association between PV and BMI (P=0.056). Pituitary intensity in T1-weighted images was negatively correlated with BMI (P=0.03) and truncal fat (P=0.04). Linear regression analysis revealed that, after adjusting for age and sex, the percentage of truncal fat and fibrinogen were significant predictors of the mean intensity of coronal and sagittal T1-weighted scans (P=0.001). The model explained up to 29% of variance of pituitary signal intensity, and in a step-wise R2 analysis fibrinogen itself accounted for 10% of the variance. Moreover, T1-weighted coronal scans were inversely correlated with VES (P=0.01).

Conclusions: This study describes a reduction of pituitary volume and quantitative T1-weighted intensity in obese patients, that seems related to a low-grade inflammation. Data could be explained by a relative change in pituitary stromal tissue in this cohort of patients.

Article tools

My recent searches

No recent searches.