Magnetic resonance imaging: when is one sequence sufficient?

We have studied 100 patients undergoing magnetic resonance imaging (MRI) in eight well-defined clinical problems. The relative values of the clinical details and the initial imaging sequence in reaching the final MR diagnoses were assessed. For each patient, two radiologists independently predicted the likely radiological findings from the clinical details. They then assessed the radiological appearances shown by the initial imaging sequence (which was chosen according to the clinical problem). Lastly, they made a final interpretation using all the available information. Prediction of likely radiological abnormalities from the clinical details proved unreliable. However, the radiological assessment of the initial imaging sequence was reliable for clinical problems related to the pituitary fossa, posterior fossa, internal auditory meatus and for suspected multiple sclerosis. In these patient groups additional sequences might be reserved for those with equivocal findings. Conversely, assessment of the initial imaging sequence proved unreliable compared with the full radiological assessment for clinical problems in the lumbar spine, the axilla and the knee. This study has led us to reduce the number of sequences performed for some clinical problems, with a commensurate increase in the throughput of patients.