A Predictive Score Incorporating Clinical, Radiologic, and Hormonal Parameters to Discriminate Lymphocytic Hypophysitis from Non-Functioning Pituitary Macroadenomas.

Non-functional pituitary macroadenomas (NFPMA) are uncommon pituitary lesions that do not cause hormonal hypersecretion and are most often discovered at the macroadenoma stage. Consequently, they are more challenging to diagnose, often mimicking other non-secreting sellar masses, among which hypophysitis should be carefully considered. This study aimed to differentiate between non-functioning pituitary macroadenomas (NFPMA) and hypophysitis, two distinct sellar pathologies with overlapping MRI features, by developing a diagnostic score based on clinical, biological, and radiological criteria. We conducted a prospective study, including 56 patients with NFPMA and 16 patients with hypophysitis primarily of the lymphocytic subtype. A total of 31 clinical, biological, and radiological variables were analyzed using univariate and multivariate statistical methods to identify significant predictors and to establish a diagnostic score. Nine significant criteria were identified: female sex, headaches, visual disturbances, corticotropic insufficiency, pituitary volume ≤ 7 cm, loss of the posterior pituitary bright spot, cavernous sinus invasion, optic pathway compression, and pituitary stalk thickening. The established score demonstrated significant performance in predicting the diagnosis of hypophysitis ( < 0.001; Area Under the Curve = 0.967; 95% CI = 0.926-1). The sensitivity and specificity of this score were 93.8% and 87.5%, respectively, using a threshold ≥0.5. The median score was -2 (interquartile range = [-3.5; 0.5]), with extremes ranging from -6.5 to 9. Among these, pituitary stalk thickening emerged as a key diagnostic indicator. This simple and effective multi-parametric score enables rapid and accurate differentiation of hypophysitis from NFPMA, helping to avoid unnecessary surgical interventions and to improve the management of pituitary insufficiencies and may be especially valuable in settings when biopsy is unavailable or risky.