Objective: We wished to analyse the relative value of computerized tomography and magnetic resonance in patients referred for evaluation of pituitary and parasellar lesions.
Design: We performed a separate evaluation by two independent neuroradiologists of computerized tomography and magnetic resonance images ordered numerically and anonymously, with no clinical data available.
Patients: We studied 40 patients submitted for hypothalamic-pituitary study; 31 were carried out preoperatively, of which histological confirmation later became available in 14. The remaining nine patients were evaluated postoperatively.
Measurements: Over 40 parameters relating to the bony margins, cavernous sinuses, carotid arteries, optic chiasm, suprasellar cisterns, pituitary, pituitary stalk and extension of the lesion were evaluated. These reports were compared with the initial ones offered when the scans were ordered, and with the final diagnosis.
Results: Concordance between initial computerized tomography and magnetic resonance was observed in 27 cases (67.5%); among the discordant cases computerized tomography showed the lesion in two, magnetic resonance in 10, while in the remaining case reported to harbour a microadenoma on computerized tomography the differential diagnosis between a true TSH-secreting microadenoma and pituitary resistance to thyroid hormones is still unclear. Both neuroradiologists coincided in their reports in 32 patients (80%); when the initial report was compared with those of the neuroradiologists, concordance was observed with at least one of them in 34 instances (85%). Discordant results were observed principally in microadenomas secreting ACTH or PRL and in delayed puberty. In the eight patients with Cushing's disease (histologically confirmed in six) magnetic resonance was positive in five and computerized tomography in two; the abnormal image correctly identified the side of the lesion at surgery.
Conclusions: In patients referred for evaluation of Cushing's syndrome or hyperprolactinaemia (due to microadenomas) or after surgery, magnetic resonance is clearly preferable to computerized tomography. In macroadenomas both scans are equally diagnostic but magnetic resonance offers more information on pituitary morphology and neighbouring structures. Nevertheless, there are cases in which the results of computerized tomography and magnetic resonance will complement each other, since different parameters are analysed with each examination and discordant results are encountered.