Imaging for Low Back Pain: An Evidence-Based Guide for GPs

Low back pain is one of the most common presentations in Australian general practice, and one of the most over-imaged. Australian and international guidelines - including RACGP recommendations and Choosing Wisely Australia - are consistent: for acute non-specific low back pain without red flags, imaging is not indicated in the first weeks of the episode.

The rationale is well established. Most acute episodes improve substantially within four to six weeks regardless of imaging. Degenerative findings on lumbar imaging are near-universal in asymptomatic adults, so scans frequently reveal changes that are incidental to the presentation. And there is evidence that early imaging in non-specific low back pain is associated with worse outcomes - more procedures, more cost, and no faster recovery.

The clinical task is therefore twofold: reliably identify the small group who do need imaging, and confidently manage the majority who don't.

Imaging (and often urgent referral) is indicated when features suggest a specific underlying pathology:

• Cauda equina syndrome - saddle anaesthesia, urinary retention or incontinence, faecal incontinence, bilateral leg symptoms. This is an emergency requiring same-day MRI and surgical referral, usually via the emergency department. • Suspected malignancy - history of cancer, unexplained weight loss, night pain, age-related risk, failure to improve as expected. • Suspected infection - fever, IV drug use, recent spinal procedure, immunosuppression, raised inflammatory markers. • Significant trauma - or minor trauma in patients with osteoporosis or long-term corticosteroid use, where vertebral fracture is possible. • Progressive or severe neurological deficit - worsening motor weakness or multi-level signs.

Radiculopathy alone, without these features, is generally managed conservatively first; imaging becomes relevant if symptoms persist beyond several weeks or an intervention is being contemplated.

When imaging is indicated, match the modality to the question:

• X-ray lumbar spine - useful for suspected vertebral fracture (trauma, osteoporosis) and for alignment questions. It does not show discs, nerve roots, or the cord, so it contributes little to radiculopathy or suspected malignancy workups. • MRI lumbar spine - the modality of choice for suspected cauda equina, infection, malignancy, and persistent radiculopathy. It shows discs, nerve roots, the cord, marrow, and soft tissues without ionising radiation. • CT lumbar spine - reserved for bony detail (complex fracture characterisation) or when MRI is contraindicated. It carries a meaningful radiation dose to the pelvis and is not a routine substitute for MRI.

Note that Medicare eligibility for GP-referred lumbar spine MRI is limited - check current MBS criteria - so for non-urgent cases, a specialist pathway or a well-timed re-assessment is often the practical route.

The greyest zone is the patient whose back pain has persisted beyond six weeks without red flags. Guidelines support considering imaging at this point if it will change management - for example, if radicular symptoms persist and the patient would consider an epidural injection or surgical opinion.

Before referring, ask what decision the scan will inform. If the patient would not contemplate an intervention, an MRI showing a disc protrusion changes nothing except, potentially, the patient's beliefs about their back - and there is good evidence that pathologising language on reports can entrench disability.

If you do image, prepare the patient for the likelihood of incidental degenerative findings before the scan: "Most people your age have some wear-and-tear changes on a scan, so we expect to see some - what we're checking for is anything that needs specific treatment." This single sentence prevents a great deal of downstream anxiety.

Many patients present with the expectation that a scan is the definitive test for back pain, and declining a request badly can damage the therapeutic relationship. Approaches that work in practice:

• Validate first - acknowledge that the pain is real and significant. Patients often interpret "no scan" as "you don't believe me". • Explain what a scan would and wouldn't show - that imaging findings correlate poorly with pain, and that most scans in this situation show age-typical changes. • Offer a concrete plan instead - analgesia strategy, activity advice, physiotherapy, and a defined review date. • Safety-net explicitly - describe the red-flag symptoms that would change the plan and prompt urgent review. • Frame imaging as deferred, not refused - "If this isn't clearly improving in a few weeks, imaging is one of the things we'll consider."

A planned review appointment converts a potential confrontation into a shared strategy.

When you do refer, the quality of the request shapes the quality of the report:

• State the specific question - "MRI lumbar spine - 6 weeks progressive left L5 radiculopathy, weak EHL, positive SLR; assess for disc protrusion with nerve root compression" tells the radiologist exactly what to assess. • Include red-flag context where relevant - a history of malignancy changes the radiologist's search pattern and reporting emphasis. • Mark urgency honestly - suspected cauda equina is an emergency-department presentation, not a routine referral marked "urgent". • Note prior imaging - comparison with previous studies materially improves reporting.

For broader referral technique, see our referral best practices guide. As always, this article is general professional information - consult current RACGP, RANZCR and Choosing Wisely Australia guidance for individual clinical decisions.