Clinical, hormonal and magnetic resonance imaging (MRI) predictors of transsphenoidal surgery outcome in acromegaly

Abstract

OBJECTIVE: Progress in the treatment of acromegaly with drugs is making it necessary to improve the prediction of the outcome of transsphenoidal surgery. DESIGN: We evaluated clinical, hormonal and radiologic predictors based on magnetic resonance imaging (MRI) of surgical outcome in patients with acromegaly. METHODS: This retrospective analysis included 125 consecutive patients investigated for acromegaly in a single endocrine unit since the use of MRI imaging began (1988). Eighty-three of these patients (50 women) underwent transsphenoidal surgery and were investigated before and after surgery in our department. A neuroradiologist unaware of the surgical outcome analyzed the results of pituitary gland MRI investigations. RESULTS: Surgical remission rates were 44%, 43%, 61% and 59% based on mean basal GH concentration under 2.5 microg/l, GH/oral glucose tolerance test (OGTT) of <1 microg/l, GH/OGTT of <2 microg/l or IGF-I concentration normal for age and sex respectively. In univariate logistic regression analysis with IGF-I concentration used as the criterion for cure, young age (P<0.001), high IGF-I concentration before surgery (P<0.01), high basal GH concentration before surgery (P<0.02), and high nadir GH/OGTT before surgery (P=0.03) were predictors of poor outcome. The following results in standardized MRI analysis were associated with a higher probability of not being cured: adenoma greater than 15 mm in diameter (P<0.02), infrasellar extension (P=0.04), suprasellar extension (P<0.005) and invasive adenoma (0.02) according to MRI staging. MRI analysis of the intracavernous extension showed that stages above B2 (possible sinus extension with sign of invasion of the space below the carotid artery) were associated with a lower probability of postoperative normal GH plasma levels (P=0.01). In multivariate analysis, age, preoperative hormonal levels and adenoma size remained the major predictors of surgical outcome. CONCLUSIONS: This report provides the first evidence that detailed MRI analysis of adenoma size, location and potential invasion, together with preoperative clinical and hormonal parameters, can be used for the prediction of hormonal outcome after transsphenoidal surgery for acromegaly.