Quantitative Magnetic Resonance-Derived Biomarkers as Predictors of Function and Histotype in Adenohypophyseal Tumours.

INTRODUCTION: Magnetic resonance imaging (MRI) is the main modality to diagnose adenohypophyseal tumours, while biochemical assessment of pituitary hormones allows for their functional classification. In this retrospective exploratory cohort study, we investigated if quantitative differences in tumour MR signal intensity (SI) could be utilized to predict the function and histotype. METHODS: Clinically acquired pretreatment MRI images were retrospectively analysed in 67 clinically non-functioning gonadotropinomas (NFG), 38 somatotropinomas, and 16 medically treated giant macroprolactinomas. Mean T1- and T2-weighted SI values were determined for each tumour and normalized against either centrum semiovale white matter or CSF to derive relative T1W and T2W SI values and the relative tumour T2/T1 SI ratio. Inter-group differences in quantitative MR parameters were compared, and the power of each parameter to discriminate tumour type and subtype was assessed using the area under the receiver operator characteristic curve (AUROC). In resected somatotropinomas, the relationship between tumour granulation status, relative MR SI values, and biochemical data was also compared. RESULTS: Compared to somatotropinomas, NFG and macroprolactinomas displayed higher relative T2W SI (p < 0.001) and higher relative tumour T2/T1 SI ratio values (p < 0.001, ANOVA). Compared to intermediate/densely granulated tumours, sparsely granulated somatotropinomas were larger (p = 0.006, Mann-Whitney U test), had higher relative T2W SI (p ≤ 0.005), and higher relative tumour T2/T1 SI ratios (p ≤ 0.001, 2-tailed t test). Relative tumour T2W SI values and relative tumour T2/T1 ratio values demonstrated good discriminatory power in differentiating NFG from somatotropinoma (AUROC = 0.87-0.94) and predicting somatotropinoma subtypes (AUROC = 0.87-0.95). CONCLUSION: Quantitative SI-based MR parameters derived using clinical acquisition MRI protocols may help non-invasively discriminate the functional status of adenohypophyseal tumours and the histological subtype of somatotropinomas.