The need for radiological investigation of the lumbosacral spine is based on the results of a thorough clinical examination. A useful and basic preliminary step, which will avoid unnecessary investigations, is to determine whether the predominant symptom is back pain or leg pain. Leg pain extending to the foot is indicative of nerve root compression and imaging needs to be directed towards the demonstration of a compressive lesion, typically disc prolapse. This is most commonly seen at the L4/5 or L5/S1 levels (90–95%), and MRI should be employed as the primary mode of imaging. If the predominant symptom is back pain, with or without proximal lower limb radiation, then invasive techniques may be required, including discography and facet joint arthrography. The presence of degenerative disc and facet disease demonstrated by plain films, CT or MRI has no direct correlation with the incidence of clinical symptomatology. The annulus fibrosus of the intervertebral disc and the facet joints are richly innervated, and only direct injection can assess them as a potential pain source. However, unless there are therapeutic implications, there is no indication to go to these lengths, as many patients can be managed by physiotherapy and mild analgesics.

Conventional radiography

Routine radiographic evaluation at the initial assessment of a patient with acute low back pain does not usually provide clinically useful information. Eighty-five per cent of such patients will return to work within 2 months having received only conservative therapy, indicating the potential for non-contributory imaging. Despite the known limitations of radiographs, it is often helpful to obtain routine radiographs of the lumbar spine before another investigation is requested. The role of conventional radiographs can be summarized in the following points:

1. They assist in the diagnosis of conditions that can mimic mechanical or discogenic pain, e.g. infection, spondylolysis, ankylosing spondylitis and bone tumours, though in most circumstances ^99mTc scintigraphy, CT and MRI are more sensitive.

2. They serve as a technical aid to survey the vertebral column and spinal canal prior to myelography, CT or MRI, particularly in the sense of providing basic anatomical data regarding segmentation. Failure to do this may lead to errors in interpreting correctly the vertebral level of abnormalities prior to surgery.

3. Correlation of CT or MRI data with radiographic appearances is often helpful in interpretation.

Computed tomography and magnetic resonance imaging of the spine

CT and MRI have replaced myelography as the primary method for investigating suspected disc prolapse. High-quality axial imaging by CT is an accurate means of demonstrating disc herniation but, in practice, many studies are less than optimal due to obesity, scoliosis and beam-hardening effects due to dense bone sclerosis. For these reasons, and because of better contrast resolution, MRI is the preferred technique and CT is only employed when MRI cannot be used. MRI alone has the capacity to show the morphology of the intervertebral disc, and can show ageing changes, typically dehydration, in the nucleus pulposus. It provides sagittal sections, which have major advantages for the demonstration of the spinal cord and cauda equina, vertebral alignment, stenosis of the spinal canal, and for showing the neural foramina. Far lateral disc herniation cannot be shown by myelography, but is readily demonstrated by CT or MRI. CT may be preferred to MRI where there is a suspected spinal injury, in the assessment of primary spinal tumours of bony origin, and in the study of spondylolysis and Paget’s disease. MRI in spinal stenosis provides all the required information showing all the relevant levels on a single image, the degree of narrowing at each level and the secondary effects such as the distension of the vertebral venous plexus. The relative contributions of bone, osteophyte, ligament or disc, while better evaluated by CT, are relatively unimportant in the management decisions. Furthermore, MRI will show conditions which may mimic spinal stenosis such as prolapsed thoracic disc, ependymoma of the conus medullaris and dural arteriovenous fistula.

In addition to the diagnosis of prolapsed intervertebral disc, CT and MRI differentiate the contained disc, where the herniated portion remains in continuity with the main body of the disc, from the sequestrated disc, where there is a free migratory disc fragment. This distinction may be crucial in the choice of conservative or surgical therapy, and of percutaneous rather than open surgical techniques. MRI studies have shown that even massive extruded disc lesions can resolve naturally with time, without intervention. Despite the presence of nerve root compression, a disc prolapse can be entirely asymptomatic. Gadolinium enhancement of compressed lumbar nerve roots is seen in symptomatic disc prolapse with a specificity of 95.9%.¹

Finally, in the decision as to whether to choose CT or MRI it should be remembered that lumbar spine CT delivers a substantial radiation dose, which is important, particularly in younger patients.

The main remaining uses of myelography are in patients with claustrophobia or who are otherwise not suitable for MRI. There are advocates for the use of CT myelography in the investigation of MRI-negative cervical radiculopathy. Myelography also allows a dynamic assessment of the spinal canal in instances of spinal stenosis and instability. The use of a special MR-compatible spinal harness that provides axial loading, and the availability of open and upright MR scanners also provide non-invasive dynamic MR imaging capability.

The problems of the ‘post-laminectomy’ patient or ‘failed back surgery syndrome’ are well known. Accurate pre-operative assessment should limit the number of cases resulting from inappropriate surgery and surgery at the wrong level. The investigation of the postoperative lumbar spine is difficult and re-operation has a poor outcome in many cases. Although the investigation of the postoperative lumbar spine is difficult, it is vital to make the distinction between residual or recurrent disc prolapse at the operated level and epidural fibrosis in order to minimize the risk of a negative re-exploration. The best available technique is gadolinium-enhanced MRI.

Arachnoiditis is a cause of postoperative symptoms and its features are shown on myelography, CT myelography and MRI. In the past, many cases were caused by the use of myodil (Pantopaque) as a myelographic contrast medium. It is likely that the use of myodil as an intrathecal contrast agent caused arachnoiditis in most cases, but this became symptomatic in only a minority. The potentiating effects of blood in the CSF, particularly as a result of surgery, have been evident in many cases. New cases of arachnoiditis are now rarely seen, but there is residue of chronic disease still presenting from time to time.

Conclusions

MRI has revolutionized the imaging of spinal disease. Advantages include non-invasiveness, multiple imaging planes and lack of radiation exposure. Its superior soft tissue contrast enables the distinction of nucleus pulposus from annulus fibrosus of the healthy disc and enables the early diagnosis of degenerative changes. However, up to 35% of asymptomatic individuals less than 40 years of age have significant intervertebral disc disease at one or more levels on MRI images. Correlation with the clinical evidence is, therefore, essential before any relevance is attached to their presence and surgery is undertaken. As MRI is, at present, not as accurate as discography in the diagnosis and delineation of annular disease, and in diagnosing the pain source, there has been a resurgence of interest in discography. MRI should be used as a predictor of the causative levels contributing to the back pain with discography having a significant role in the investigation of discogenic pain prior to surgical fusion.²

References

1. Tyrrell, PN, Cassar-Pullicino, VN, McCall, IW. Gadolinium-DTPA enhancement of symptomatic nerve roots in MRI of the lumbar spine. Eur Radiol. 1998; 8(1):116–122.

2. Cassar-Pullicino, VN. MRI of the ageing and herniating intervertebral disc. Eur J Radiol. 1998; 27(3):214–228.

Further Reading

Boden, SD, Davis, DO, Dina, TS, et al. Abnormal magnetic resonance scans of the lumbar spine in asymptomatic subjects. J Bone Joint Surg Am. 1990; 72(3):403–408.

Butt, WP. Radiology for back pain. Clin Radiol. 1989; 40(1):6–10.

Cribb, GL, Jaffray, DC, Cassar-Pullicino, VN. Observations on the natural history of massive lumbar disc herniation. J Bone Joint Surg Br. 2007; 89(6):782–784.

du Boulay, GH, Hawkes, S, Lee, CC, et al. Comparing the cost of spinal MR with conventional myelography and radiculography. Neuroradiology. 1990; 32(2):124–136.

Horton, WC, Daftari, TK. Which disc as visualized by magnetic resonance imaging is actually a source of pain? A correlation between magnetic resonance imaging and discography. Spine. 1992; 17(Suppl 6):S164–S171.

Hueftle, MG, Modic, MT, Ross, JS, et al. Lumbar spine: post-operative MR imaging with gadolinium-DTPA. Radiology. 1988; 167(3):817–824.

Myelography and radiculography

Contrast media

Historically, the contrast media that have been used for myelography include gas (CO₂, air), lipiodol, abrodil, myodil (Pantopaque, iodophenylate), meglumine iothalamate (Conray), meglumine iocarmate (Dimer X), metrizamide (Amipaque), iopamidol (Niopam), iohexol (Omnipaque) and iotrolan (Isovist). The early oil-based media were diagnostically poor and led to arachnoiditis. The early water-soluble media were very toxic and also led to arachnoiditis, and the first intrathecal contrast medium with acceptably low toxicity was metrizamide, a non-ionic dimer. Iohexol and iopamidol are non-ionic monomers licensed for intrathecal use and the ones most commonly used currently.

Cervical myelography

This may be performed by introduction of contrast medium into the thecal sac by lumbar puncture and then run up to the cervical spine, or by direct cervical puncture at C1/2.

Lumbar radiculography

This may be performed by injection of contrast medium into the lumbar thecal sac. If for any reason lumbar puncture is not possible, e.g. because of lumbar spine deformity or arachnoiditis, it is possible to introduce the contrast medium from above by cervical injection. Dilution of the contrast as it passes downwards is a major disadvantage.

Indications

Suspected lumbar root or cauda equina compression, spinal stenosis and conus medullaris lesions in patients who are unable or unwilling to undergo MRI.

Contraindications

Lumbar puncture is potentially hazardous in the presence of raised intracranial pressure. It may be difficult to achieve satisfactory intrathecal injection in patients who have had a recent lumbar puncture, since subdural CSF accumulates temporarily, and this space may be entered rather than the subarachnoid space. Accordingly an interval of 1 week or so is advisable.

Contrast medium

The contrast medium is the same as that used for cervical myelography. The maximum dose is again the equivalent of 3 g of iodine, and may be given as 10 ml of contrast medium of 300 mg ml⁻¹ concentration, or 12.5 ml of 240 mg ml⁻¹.

Equipment

Tilting fluoroscopic table.

Patient preparation

As for cervical myelography.

Preliminary images

AP and lateral projections of the region under study are taken. Preliminary examination of radiographs is helpful to assess the anatomy of the spine, in order to facilitate the lumbar puncture, and to assist in interpretation of the images. It is important to draw the surgeon’s attention to any question of ambiguous segmentation, either lumbarization or sacralization. There is a potential danger of operating at the wrong level if this is not made absolutely clear in the report. A clear description of any anomaly is required, together with a statement of how the vertebrae have been numbered in the report.

Technique

1. The lumbar thecal sac is punctured at L2/3, L3/4 or L4/5. The higher levels tend to be away from the most common sites of disc herniation and stenosis, and puncture may, therefore, be easier.

2. Lumbar puncture can technically be performed in the lateral decubitus position, in the sitting position, or even in the prone position. In the prone position the needle is guided under fluoroscopy, usually at the L2/L3 interspinous space with the patient lying on a folded pillow. This is important, because in the prone position the spinous processes are approximated due to lordosis, rendering puncture more difficult. In addition, spinal extension produces a relatively narrow thecal sac. The sitting position allows easy lumbar puncture, but is unsatisfactory for two reasons. First, the injected contrast medium drops through a large volume of CSF to accumulate in the sacral sac and becomes diluted as it descends. Second, patients may faint in this position, a complication that can be very dangerous, since the radiologist is on the wrong side of the table to help in preventing the patient from coming to any harm.

3. If lumbar puncture is carried out in the lateral decubitus position (described below) moderate spinal flexion is desirable, but there is no need for the extreme flexion sometimes demanded of the patient. A small pillow is placed under the dependent lumbar angle to keep the spine straight. The relevant interspace is one or two spinous processes above the plane of the iliac crest (L4/L5). If the spinous process cannot be felt, lateral fluoroscopy may help.

4. In obese patients the apparent soft-tissue midline gravitates below the spinal midline. The midline position may be verified while introducing local anaesthetic (1% lidocaine) into the skin and subcutaneous tissue, though there is no need to infiltrate the interspinous ligament. A 22G spinal needle should be used. It should be introduced with only a 10–15° cranial angulation. The passage of the needle through the interspinous ligament has a very characteristic sensation of moderate smooth resistance. Lack of resistance implies passage in the fatty tissues lateral to the midline, and a gritty sensation means impingement on the bone of the spinous process.

5. There is a tendency in inexperienced operators to introduce the needle at too steep a cranial angle, which a casual inspection of the spinous process anatomy can show to be incorrect. While traversing the interspinous ligament, the bevel should be in the coronal plane to avoid deflection to one side or the other, but once well established in the ligament may be turned to penetrate the thecal sac parallel to its long axis.

6. There is a characteristic sudden loss of resistance as the needle enters the thecal sac, and at this time the stilette should be withdrawn to verify CSF flow out of the needle. In the prone position suction using a 2-ml syringe may be necessary to obtain CSF. It should then be reintroduced and the needle advanced about 2 mm to ensure that the whole of the bevel has entered the thecal sac. A flexible connector is attached, taking care not to disturb the needle position.

7. Most difficulties are technical, arising from non-midline positioning of the needle, but in the presence of canal stenosis it may be difficult to find a position among the crowded roots where good CSF flow will take place. In most cases, radiologists will prefer to observe the entry of contrast medium into the thecal sac on the fluoroscopic screen, and this is especially important if there has been any difficulty in achieving a good needle position.

8. After the contrast medium has been injected, the patient turns to lie prone, and a series of films is obtained. Before taking films ensure that the relevant segment of the spinal canal is adequately filled with contrast medium. This usually requires some degree of feet-down tilt of the table, and a footrest should be in place to support the patient.

Radiographic views

1. AP and oblique views are obtained. (About 25° of obliquity is usual, but tailored in the individual case to profile the exit sleeves of the nerve roots of the cauda equina.)

2. A lateral view with a horizontal beam is useful, but further laterals in the erect or semi-erect position on flexion and extension add a dynamic dimension to the study.

3. Some radiologists advocate lateral decubitus frontal oblique views, to allow full entry of contrast medium into the dependent root sleeves of the whole lumbar and lower thoracic thecal sac. When turning patients from one side to the other ensure that they turn through the prone position, as turning supine will produce irretrievable contrast medium dilution.

At this stage the study will be complete as far as the question of root compression is concerned. The films should be carefully reviewed at this stage to ensure that all areas are fully covered, as the next manoeuvre will make any return to this area impossible. If the study is good, the patient should be screened in the lateral decubitus position, and tilted level or slightly head down until the contrast medium flows up to lie across the thoraco-lumbar junction. The patient may then be turned to lie supine and films of the conus and lower thoracic area obtained in the AP projection.