Limited magnetic resonance imaging in low back pain instead of plain radiographs: experience with first 1000 cases.

AIM: We report our experience with the first 1000 patients with non-traumatic low back pain (LBP) without radiculopathy undergoing limited sequence magnetic resonance imaging (MRI) instead of plain radiography.
METHODS: Between January 1996 and December 1998, 1042 patients with low back pain unresponsive to conservative treatment were examined using a limited MR protocol comprising sagittal T1-weighted and STIR imaging. Plain radiographs were not performed.
RESULTS: Malignancy, infection, vertebral fracture, spondylitis, pars defects and cord tumours were detected in 20%. Of the 82 osteoporotic vertebral fractures detected, 51 (62%) were recent and 31 had normal marrow signal indicating that they were old. Eighty pars defects were identified, 45(56%) had spondylolisthesis, 29(37%) were undisplaced and 6 (7%) had pars oedema only. Neoplastic disease was found in 17(8%) of which none was suspected before imaging. Benign neoplastic diseases such as vertebral AVM/haemangiomata were excluded. Twenty-one patients had a variety of disorders including ankylosing spondylitis (7), large vessel aneurysm (3), discitis (2), ovarian cyst (2), sequestered disc (2), sacral insufficiency fracture (2) and one patient each with burst fracture, retroperitoneal haematoma and a previously unsuspected horseshoe kidney.
CONCLUSION: The majority of patients with LBP are best assessed clinically and imaging is usually not required. In patients with worrying symptoms, MRI with a limited protocol detects a greater number of abnormalities than previously reported studies using plain radiographs and has replaced plain radiography in our hospital. We report our experience with the first 1000 patients and highlight issues such as protocols, detection rates and communication issues. Copyright 2001 The Royal College of Radiologists.