The etiological diagnosis of ACTH-dependent Cushing's syndrome is often a problem. In fact, no endocrine or radiological examination can conclusively distinguish the ectopic from the pituitary source of disease. The aim of our study was to evaluate the role of stimulation and suppression endocrine tests in the diagnostic and therapeutic approach of patients with Cushing's disease (CD) and negative pituitary magnetic resonance imaging (MRI), considering their post-surgical outcome in comparison with patients with CD and positive MRI. We retrospectively analyzed 31 patients (25 women and 6 men, median age 40 +/- 15 years) with a confirmed diagnosis of CD who underwent transsphenoidal pituitary surgery by the same neurosurgeon between 2001 and 2005. Preoperative endocrine assessment included corticotropin-releasing hormone (CRH), desmopressin (dDAVP), and overnight 8 mg dexamethasone suppression tests (8-DST) in all patients. Fifteen patients had a normal pituitary MRI and sixteen had a clearly evident pituitary microadenoma. Bilateral inferior petrosal sinus sampling (BIPSS) was performed in patients with discordant biochemical results or with signs and symptoms highly suggestive of an ectopic source of ACTH. Post-surgical median follow-up was 38.4 +/- 22.0 months. Among patients with negative MRI, 60% had concordant positive endocrine tests and underwent neurosurgery without other examinations. BIPSS was performed in three other patients prompted by discordant endocrine tests (negative dDAVP) and in two patients with clinical suspicion of ectopic disease. Among patients with positive MRI, 87% underwent neurosurgery without BIPSS that was performed in two patients because of negative concomitant response to dDAVP and CRH tests. A pituitary adenoma, confirmed by pathological examination, was found in 40 and 81% of patients with negative and positive MRI respectively (P<0.05), corticotroph hyperplasia resulted more frequent in the group with negative MRI. Remission rate was not different between patients with negative and positive MRI (73 and 75% respectively; P=0.61) and between patients with negative MRI who did not undergo BIPSS and patients with positive MRI (P=0.56). The recurrence rate was also similar between groups (P=0.64), but higher, although not statistically different (P=0.07) in patients with corticotroph hyperplasia at histology. An accurate evaluation of presurgical endocrine tests results enabled us to reduce the number of BIPSS in patients with a negative MRI without any fallout on their post-surgical outcome. In the hands of an expert pituitary surgeon, the outcome after surgeryand the subsequent recurrence rate are much the same in patients with negative or positive MRI.