Abstract
PURPOSE To document the accuracy of CT and MR of the sella turcica for detecting adrenocorticotropic hormone-secreting adenomas in Cushing disease.
METHODS The radiologic findings of the sella turcica prior to transsphenoidal surgery are reviewed in 141 patients who had biochemical evidence of pituitary-dependent Cushing disease. Axial thin-collimation CT scans with sagittal and coronal reformations before and after contrast enhancement were obtained in 125 patients. Seventy-eight patients had MR examinations with a 1.5-T superconducting magnet. In 11 of the patients gadolinium-enhanced MR scans were also obtained. The preoperative interpretation of the imaging studies was correlated with the surgical findings and patients follow-up.
RESULTS The sella turcica was enlarged in 43 cases (30%). In 125 patients reformatted or direct coronal thin-collimation CT scans were available. Seventy-eight of the patients had MR. In the 12 patients with pituitary macroadenomas, the accuracy of CT (n = 10) and MR (n = 10) in respect to detection of the lesion was 100%. Of the 98 microadenomas assessed by CT, 47 (48%) were directly depicted as distinct hypodense lesions. In only 31 of 73 cases (42%), however, could CT predict the precise anatomic location and extent of the lesions. Only patients in whom the hypercortisolism was corrected by later surgery were considered for the correlation analysis. Of the 52 microadenomas assessed by MR, 28 (53%) were directly depicted as distinct lesions of reduced signal intensity on T1-weighted images, and in only 21 of 41 cases (52%) did MR show good correlation to the surgical findings. Some degree of partially empty sella was found in 22% of the patients.
CONCLUSIONS Although both the sensitivity and the diagnostic accuracy of imaging methods of the sella turcica have been considerably improved in comparison with previous reports, they still provide only a minor contribution to the diagnosis and differential diagnosis of Cushing syndrome.
- Copyright © American Society of Neuroradiology