Abstract
OBJECTIVE: To compare computed tomography and magnetic resonance imaging in investigating patients suspected of having a lesion in the posterior cranial fossa. DESIGN: Randomised allocation of newly referred patients to undergo either computed tomography or magnetic resonance imaging; the alternative investigation was performed subsequently only in response to a request from the referring doctor. SETTING: A regional neuroscience centre serving 2.7 million. PATIENTS: 1020 Patients recruited between April 1986 and December 1987, all suspected by neurologists, neurosurgeons, or other specialists of having a lesion in the posterior fossa and referred for neuroradiology. The groups allocated to undergo computed tomography or magnetic resonance imaging were well matched in distributions of age, sex, specialty of referring doctor, investigation as an inpatient or an outpatient, suspected site of lesion, and presumed disease process; the referring doctor's confidence in the initial clinical diagnosis was also similar. INTERVENTIONS: After the patients had been imaged by either computed tomography or magnetic resonance (using a resistive magnet of 0.15 T) doctors were given the radiologist's report and a form asking if they considered that imaging with the alternative technique was necessary and, if so, why; it also asked for their current diagnoses and their confidence in them. MAIN OUTCOME MEASURES: Number of requests for the alternative method of investigation. Assessment of characteristics of patients for whom further imaging was requested and lesions that were suspected initially and how the results of the second imaging affected clinicians' and radiologists' opinions. RESULTS: Ninety three of the 501 patients who initially underwent computed tomography were referred subsequently for magnetic resonance imaging whereas only 28 of the 493 patients who initially underwent magnetic resonance imaging were referred subsequently for computed tomography. Over the study the number of patients referred for magnetic resonance imaging after computed tomography increased but requests for computed tomography after magnetic resonance imaging decreased. The reason that clinicians gave most commonly for requesting further imaging by magnetic resonance was that the results of the initial computed tomography failed to exclude their suspected diagnosis (64 patients). This was less common in patients investigated initially by magnetic resonance imaging (eight patients). Management of 28 patients (6%) imaged initially with computed tomography and 12 patients (2%) imaged initially with magnetic resonance was changed on the basis of the results of the alternative imaging. CONCLUSIONS: Magnetic resonance imaging provided doctors with the information required to manage patients suspected of having a lesion in the posterior fossa more commonly than computed tomography, but computed tomography alone was satisfactory in 80% of cases...