Imaging Considerations (Magnetic Resonance, Computed Tomography, Myelography, Plain)

Introduction

Regardless of the presence of lumbar instrumentation, recurrent symptoms after a lumbar decompressive surgery are not uncommon; for example, after undergoing microlumbar discectomy up to 12% of patients may have a recurrent disc herniation within the first year. The incidence of continued or new pain after lumbar fusion may be as high as 30%. Further, the incidence of repeat lumbar surgery following decompression alone ranges from 1.6% to 10.8% at 1 year versus 0% to 7% following fusion. Duration of symptoms can influence imaging protocols. In the early postoperative period, magnetic resonance imaging (MRI) is notoriously difficult to interpret. Some authors recommend not pursuing MRI until 6 weeks postoperatively. Once the patient is out of the immediate postoperative period, imaging algorithms are similar to those in unoperated patients.

The onset of new symptoms prompts a new differential diagnosis. This helps guide further imaging. Common scenarios include recurrent disc herniation, recurrent stenosis, spondylitic disease at another level, spondylolisthesis (mobile or immobile), adjacent segment disease, spinal deformity, disc space infection or osteomyelitis, other infection, arachnoiditis, cerebrospinal fluid (CSF) leak pseudomeningocele, or wound healing issues. The first step in generating a differential diagnosis is a thorough history and physical exam. Although the breadth of these two activities is outside the scope of this chapter, suffice it to say that many diagnoses can be made by history alone (e.g., neurogenic claudication) and by examining the patient’s incision and posture one can discern infection and progression of spinal deformity. Imaging studies are helpful to further confirm a diagnosis.

Magnetic Resonance Imaging

MRI is the modality of choice to image the postoperative lumbar spine. However, care must be taken when interpreting images within the first 6 to 8 weeks following an operation. In the early postoperative period, characteristic changes such as T2 hyperintensity in both the soft tissues and osseous structures can occur and be considered normal. Regardless of timing, relative to the surgical procedures the MRI should be obtained with and without intravenous gadolinium. Commonly postoperative changes can mimic disc herniation or nerve root compression. In fact, within 6 weeks of surgery nearly 25% of patients can have imaging findings consistent with residual or recurrent disc herniation. Some can have severe thecal sac narrowing. In three-quarters of these patients, interval MRI at 6 months may show persistence of this mass effect.

In a study using 3T MRI to surveil 30 operated patients after lumbar discectomy, 80% of patients were found to have residual mass effect on the nerve root in the early postoperative period. Changes of high T2 signal were found to reliably resolve within 3 months. The addition of gadolinium allows for differentiation of recurrent disc herniation and epidural fibrosis ( Fig. 5.1 ). Recurrent disc herniation will demonstrate an area that does not enhance, whereas epidural fibrosis will demonstrate confluent enhancement. This allows for determining the difference between the two diagnoses in 96% of the cases.

A common imaging finding after lumbar fusion will call into question the diagnosis of disc space infection versus degenerative change. Frequently this can be seen at the terminal end of a long construct. Both entities can have Modic type I changes, which are characterized by low T1 and high T2 signal within the bone marrow of a vertebral body adjacent to the disc space. If there is high T2 signal within the disc space, most clinicians will favor infection; however, this is not always the case. In a small series, Patel et al. reported a diffusion “claw sign” that can help differentiate infection from degenerative change. This is seen on trace/combined diffusion weighted images as well-defined, linear paired regions of high signal of adjacent vertebral bodies at the border of normal and vascularized bone marrow. If this “claw sign” is present, there is a near greater than 97% chance the imaging findings represent degenerative changes. Conversely the absence of a claw sign yields a greater than 93% chance of infection.

Contrast enhancement of nerve roots is abnormal. Nerve roots have a blood-nerve barrier, up to the dorsal root ganglia, which avidly enhances. Disc herniations or other compressive lesions may cause nerve roots to enhance. In a study of 200 unoperated patients, 5% had nerve root enhancement. In 70% of these patients, this enhancement was associated with disc pathology. This enhancement may persist for months in the postoperative period.

Postoperative fluid collections are best assessed with MRI. Common fluid collections include hematoma, pseudomeningocele, seroma, and abscess. History may be important in differentiating these conditions. If a patient has recurrent or residual symptoms in the early postoperative period, a diagnosis of hematoma should be considered. Often the surgical incision will be full or bulging, although this may not be the case. Hematomas are typically extradural and demonstrate mixed signal on MR sequences owing to blood products ( Fig. 5.2 ). Patients with pseudomeningocele typically have a known dural tear at the time of index surgery. Additionally, they may have CSF leaking from the incision and postural headaches. On MRI, these collections follow CSF signal on all sequences ( Fig. 5.3 ). Seromas present in a more subacute fashion, in the days or weeks following surgery. Bone morphogenetic protein-2, which is used as a fusion material, is known to cause seroma formation. Imaging characteristics on MRI will be similar to fluid or CSF signal. Finally, abscesses develop in the weeks after surgery.

Evaluation of these patients may be complex. Commonly patients will report worsening back pain. This may be accompanied by constitutional symptoms such as fever, chills, rigors, or night sweats. However, less than 50% of patients will report fever. The surgical site is often tender, warm, erythematous, tense, and draining. Bloodwork can be helpful as well. Erythrocyte sedimentation rate (ESR) and C reactive protein (CRP) will be elevated in most infections. CRP, however, is more sensitive than the ESR. The sensitivity is 100% and the specificity is 96.8%. CRP peaks typically on postoperative day 3 and then falls. Any escalation beyond that calls into question the possibility of infection. MRI imaging of these collections will frequently show a collection that follows fluid signal and demonstrates brisk rim enhancement. Caution and clinical judgement need to be exercised, however, because seromas may demonstrate enhancement as well.

Sagittal plane malalignment may not always be evident on MRI; however, secondary indicators of this are commonly seen. Facet joint effusions seen on MRI imaging become more significant with size. An effusion greater than 1.5 mm should raise the concern for spondylolisthesis and prompt upright x-rays. Chaput et al. found for each millimeter increase in effusion the chance of finding spondylolisthesis increased 5.6 times.

Spinal implants typically degrade the MRI images. Depending on the type of metallic implant, artifact may make the MRI imaging uninterpretable ( Fig. 5.4 ). Stainless steel implants cause significant artifact. Titanium implants, although they are MRI compatible, in general cause minimal artifact. Newer generation systems that employ cobalt chromium in rods and other portions of the screw can cause significant artifact. However, if the construct is relatively short, even this artifact will permit diagnostic quality imaging.

In previously operated patients, refractory pain can be caused by arachnoiditis. This can cause significant refractory back and leg pain. Common imaging findings include nerve root clumping, “empty thecal sac” sign, CSF loculations, and intradural calcifications ( Fig. 5.5 ). This clinical entity is seldom successfully addressed with surgery.