Avascular Necrosis of Bone: MRI Evaluation

David A. Bluemke, M.D., Ph.D.

Investigators at the Johns Hopkins Hospital have conducted extensive research in the diagnosis of avascular necrosis using MRI. MRI is the most sensitive noninvasive examination for detecting avascular necrosis. Ongoing research efforts and opportunities for patient participation are discussed below.

AVN Research Protocol

Individuals at risk for AVN of the hip are eligible candidates for MR imaging and spectroscopy evaluation of bone marrow perfusion and composition. Eligible patients include:

age 18 - 65
new treatment with corticosteroids, OR
starting bolus treatment with corticosteroids (prednisone)
not pregnant
no history of MRI incompatible internal devices (ie pacemakers, aneurysm clips, etc)


Enrolled individuals will receive free MRI evaluation of the hips. Candidates should be evaluated BEFORE starting corticosteroids, or starting new corticosteroid dose. Examinations last approximately 1 hour. Patients who wish to be enrolled should contact:

Dr. David A. Bluemke
dbluemke@rad.jhu.edu

Note: we are unable to accept direct phone calls because of the large number of patients.

However, we are happy to answer all patients inquires by email. Thank you for your consideration.

Introduction and Definition
Avascular necrosis (AVN) of bone is a process that is characterized pathologically by bone marrow ischemia and eventual death of trabecular bone. Ischemic necrosis, osteonecrosis, and aseptic necrosis are synonyms for the same disease process. Radiologic manifestations of AVN occur in the late stages of the disease, as the bone attempts to repair itself. As bone repair occurs, weight bearing bone becomes mechanically weakened and flattened, and may eventually collapse. Secondarily, this leads to debilitating pain and osteoarthritis.

Early diagnosis of AVN using MRI is important, since the disease occurs in relatively young individuals (average age 20-50 for idiopathic forms) and since treatment options for more advanced disease are frequently unsuccessful. In this article, the pathophysiology of AVN will be considered, and the use of MRI in diagnosing early AVN and differentiating it from other marrow disorders of the hip will be reviewed.

Pathogenesis and treatment
Several conditions are clearly related to AVN of the hip; one of these is interruption of the arterial supply of the femoral head, either through trauma or vascular disease (including hypertension, sickle cell disease, caisson disease or radiation-induced arteritis). Other conditions are variably associated with AVN in a more complex manner, including corticosteroid therapy, connective tissue disease, alcohol abuse, marrow storage disease (Gaucher's disease) and dyslipoproteinemia. A proposed mechanism linking these conditions as predisposing factors for AVN is that bone functions as a closed compartment. (1) . Under certain pathologic conditions, intraosseous bone marrow pressure increases. Elevation of intraosseous pressure is transmitted to small venules and capillaries within the bone, causing a decrease in blood flow to the bone. Rapid, or uncompensated, increases in intraosseous pressure are thought to result in irreversible circulatory disturbances and subsequent tissue damage. Tissue damage causes edema, which further elevates pressures in the closed compartment.

Early decompression of bone prior to irreversible damage may break the cycle of ischemia and increased marrow pressures. Bone decompression, or coring, is controversial; the success rate variably ranges from 40 to 90% (2) . The success of decompression appears to be directly related to stage of disease: hips with no radiologic evidence or only 25% involvement of the femoral head had no collapse of the femoral head, while 73% of cases with more advanced disease had femoral head collapse . Thus, early diagnosis of AVN with MRI is critical for patient management and successful therapeutic outcome. Other treatment options include muscle pedicle graft, rotational osteotomy, and joint fusion.

MRI diagnosis of AVN
MRI is the most sensitive noninvasive method for diagnosis of AVN. Diagnosis involves detection of marrow foci of decreased signal on T1-weighted images and the characteristic double line sign (discussed further below) on T2-weighted images (3) .

Imaging Protocol. For the hip, the most important imaging planes are coronal followed by sagittal acquisitions when necessary. Because both hips are frequently involved, it is necessary to use the body coil to image both hips, not just the symptomatic hip. Following body coil images, if a question as to the diagnosis or extent of disease remains, additional images with a surface coil over the affected hip should be obtained.

A protocol for routine imaging of the hips for AVN is shown in Table 1. Thin coronal images should be obtained directly over the femoral heads based on an axial localizer. Although the coronal T1 images alone are diagnostic of AVN in 95% of cases, it is recommended that axial images of the pelvis from the top of the sacrum to the femoral neck be obtained using a fat-suppressed T2-weighted sequence or inversion recovery sequence with TI time set to null fat. Axial images allow detection of other abnormalities that frequently mimic the clinical presentation of AVN. For example, patients receiving corticosteroid therapy are frequently osteopenic and have rapid weight gain, placing them at risk for insufficiency or stress fractures of the sacrum or pelvic bones. Muscle strains and septic arthritis of the sacroiliac joints are also readily detected on the axial fat-suppressed images.

At our institution, we have implemented a less expensive "screening MRI" examination for AVN that consists only of a 4 minute coronal T1-weighted acquisition, with 6 mm thick sections and a gap of 2 mm. This is done without a localizer series, and with properly instructed technologists, the femoral heads can be appropriately imaged in all cases. The charge for this examination is comparable to that of a radiographic plain film series of the hip. Although the method is highly accurate in detecting AVN, it is much less sensitive for detecting other hip or pelvic abnormalities that clinically mimic the disease. Therefore, we restrict the use of this "AVN protocol" to rheumatologists and orthopedic surgeons who specialize in managing high risk patient populations for AVN.

MR findings. AVN is diagnosed when a peripheral band of low signal intensity is present on all imaging sequences, typically in the superior portion of the femoral head, outlining a central area of marrow. This peripheral band is most apparent on T1-weighted sequences (Figure 1). The central area of marrow contained within the dark line may have widely varying signal intensity on various imaging sequences (see below). Rarely, bone with histologically proved AVN can appear normal by MRI.

On conventional T2 sequences, the inner border of the peripheral band shows high signal in 80% of cases (Figure 2). This is called the "double - line" sign of avascular necrosis, and is considered to be pathognomonic. Various reports state that the inner "bright" signal is due to the reactive interface, or granulation tissue, between infarcted and normal marrow. Other authors have shown that by changing the phase and frequency direction, the position of the inner "bright" signal changes in some cases, so that the etiology is that of a chemical shift artifact. Regardless of the etiology in specific cases, recognition of the double line sign is useful, since it is frequently characteristic of AVN.

On fast (or turbo) T2-weighted sequences, the double line sign usually is not well seen. This is because fat has increased signal intensity on fast spin echo sequences, thus obscuring the bright inner line. Because of this bright fat signal, edema can be obscured, so that a frequency selective pulse is frequently added to suppress signal from fat. If fat suppression is used, it is the dark, peripheral band of AVN that is not seen in contrast to the inner high signal band of AVN. Nevertheless, FSE T2 images with fat saturation are useful in demonstrating the extent of marrow edema associated with the infarct.

T2 sequences and inversion recovery sequences frequently demonstrate associated hip effusions. Increased joint fluid is commonly associated with AVN, and its presence does not indicate a septic joint effusion. The frequent presence of joint effusions has led to the hypothesis that patients are presenting with pain due to their effusion, rather than the long-standing process of AVN. Pain in association with joint effusions may be due to distention of the joint capsule by fluid.

Atypical findings of AVN. Diffuse areas of low signal in the femoral head on T1 images with high signal on T2 ("bone marrow edema pattern") may occasionally be present without a peripheral dark band of AVN on T1 sequences (4) . These areas of edema may be extensive, reaching into the femoral neck or trochanteric regions (Figure 3). On bone biopsy, AVN may be diagnosed, but the MR appearance is not specific for this condition. A primary differential diagnostic consideration is transient osteoporosis of the hip (TOH). There are no certain features to differentiate TOH from AVN by MRI, except that atypical AVN eventually progresses to MR imaging features of typical AVN, while TOH is a self-limited condition that resolves over 4 to 10 months. The MR bone marrow edema pattern is discussed further below.

Staging. There are 2 staging classifications of AVN, one based on radiographs (Table 2) (Ficat and Arlet (5) ) and the other based on MR signal intensities (Table 3) (Mitchell et al. (3) ) The accuracy of radiographic staging may be improved using CT to detect a subchondral lucency indicating advanced, or Stage III disease. Note, however, that CT does not depict the earliest marrow abnormalities resulting in osteonecrosis.

MR staging of AVN is based on the signal intensity of the center of the marrow inside the dark line of necrosis (Table 3). Radiographically occult AVN will generally be depicted on MRI as any of classes A to C. The MR classification implies that the infarcted bone progresses in an orderly manner through the various classes. This, however, is not necessarily the case, since often several "classes" of signal intensity are present within the infarcted marrow. Further, unlike radiographic staging, MR classes have little predictive value regarding the prognosis for collapse of the femoral head. However, the MRI size and position of the AVN lesion is related to prognosis, as discussed below.

Relationship of MR findings to prognosis. The extent of AVN has been related to favorable outcome (pain relief) versus poor outcome (permanent disability) (6) . AVN that was entirely circumscribed, and that did not extend cranially to the cortical subchondral margin, had a good outcome, independent of the overall size of the AVN lesion. The percentage of the weight bearing surface (Figure 1) occupied by the AVN lesion was the most reliable predictor for predicting outcome. The overall percentage of the femoral head occupied by the AVN lesion was least reliable in predicting outcome.

Role of contrast enhancement. No role for routine gadolinium administration has been
demonstrated for detection or diagnosis of nontraumatic AVN. Dynamic imaging evaluating the time course of perfusion of the femoral head is currently being investigated to determine if patients with AVN show different rates of perfusion than those without AVN (7) .

There is a high risk of AVN following fracture of the femoral neck. Bone marrow enhancement after Gd-DTPA administration has been shown to correlate with preservation of blood flow to the hip on angiography. Long-term follow-up, however, has not been performed to determine the prognostic significance of these findings.

Bone marrow edema pattern. Occasionally, AVN may manifest as a diffuse area of decreased signal on T1-weighted images and increased signal on T2-weighted images involving the femoral head, neck, and occasionally the intertrochanteric femur. This has been termed the "bone marrow edema" pattern on MR imaging, since the signal intensities are compatible with increased free-water content. Although pathologic proof is frequently lacking in reports of AVN presenting with this pattern, follow-up MR examinations or radiographs demonstrate that the bone marrow edema pattern can evolve into focal patterns entirely characteristic of AVN.

The MR pattern of bone marrow edema is not specific for AVN, however, and the differential diagnosis includes transient osteoporosis, bone bruise, infiltrative disease, and transient bone marrow edema syndrome. Although the clinical history can be helpful in distinguishing between these entities (e.g., a history of trauma as the etiology of a bone bruise), in other cases a definite diagnosis can only be made based on the time course of the imaging and clinical findings.

Differential Diagnosis
Transient osteoporosis of the hip. Transient osteoporosis of the hip is a self-limiting cause of hip pain described in middle-aged men or in women in their third trimester of pregnancy. Patients present with hip pain and limp in the absence of trauma or infection. The etiology of this condition is unknown, but a neurogenic origin has been proposed, similar to reflex sympathetic dystrophy. Transient osteoporosis resolves spontaneously over a period of 4-10 months. Osteopenia of the subchondral cortex is evident on radiographs and is a useful feature for making a specific diagnosis of transient osteoporosis. Bone scans demonstrate diffuse increased radiotracer uptake in the femoral head and frequently the neck. The MR findings of diffuse bone marrow edema (Figure 11), which may also involve the acetabulum, and associated hip joint effusion, may precede radiographic evidence of osteopenia by several weeks. There is no evidence of a double-line sign on MR images, as is frequently present in AVN. The MR findings are not characteristic, and the primary diagnostic consideration is AVN.

Transient bone marrow edema syndrome. This entity is similar to transient osteoporosis in that the condition is a self-limited cause of hip pain that resolves over several months. Since the MR findings and clinical presentation are similar to transient osteoporosis (Figure 12), Hayes et al. have proposed that transient bone marrow edema syndrome be reserved for those cases in which there is no radiographic evidence of osteopenia (8) . Again, the etiology of transient bone marrow edema syndrome is unknown.

Septic arthritis. Septic arthritis may occur from hematogenous spread of an infectious agent or by contiguous spread. On MR images, a joint effusion is present that is bright on T2 images and is nonspecific in appearance. If septic arthritis is suspected, immediate joint aspiration must be performed in order to obtain cultures to determine the infectious agent. Underlying bone changes are not typically present, although if the condition is prolonged, evidence of marrow edema may be present (increased signal on T2 images).

Stress fracture. Patients with stress fractures of the femoral neck may have a similar clinical presentation to patients suspected of AVN. They occur in young patients, resulting from overuse and repeated stress with underlying normal bone, such as in runners and military recruits. Insufficiency fractures occur in osteoporotic women in whom activity levels are seemingly normal (Figure 5). MR findings include a diffuse area of increased signal on T2 images in the area of the fracture, typically in the femoral neck. This corresponds to edema, and is of intermediate signal of T1 images. In addition, the band of edema frequently has a linear component, which may be more obvious on T1 images.
Table 1: AVN screening protocol. (body coil, both hips)
· Axial STIR or T2 with fat saturation
· Coronal T1, 256x256, 2 NEX, TR 400-500, TE minimum, centered on the femoral heads. (Add surface coil imaging to either hip, if necessary, sagittal T1 sequence).
· Coronal T2 (FSE or conventional), 256x192-256, 1-2 NEX, TR 3000-400, TE 80-100, (fat suppression if FSE)

Table 2: Ficat and Arlet Staging of AVN: (Radiographic staging)
Stage Findings
0 Diagnosis by MR or bone scan. No radiographic findings.
1 Slight osteoporosis on plain films. No sclerosis.
2 Diffuses osteoporosis and sclerosis on plain films. A reactive shell of bone delimits the infarct. Spherical femoral head.
3 Crescent sign (radiolucency) under the subchondral bone representing a fracture. Joint space preserved.
4 Femoral head collapse. Joint space narrowing.

Table 3: MRI staging of AVN (3)

Class	 T1	                T2              Definition
A       bright            intermediate    "fat" signal.
B       bright            bright          "blood" signal.
C       intermediate      bright          "fluid" or "edema" signal.
D       dark              dark            "fibrosis" signal.

References
1. Hungerford DS, Zizic TM: Pathogenesis of ischemic necrosis of the femoral head. Hip 249-262, 1983.
2. Hopson CN, Siverhus SW: Ischemic necrosis of the femoral head: treatment by core decompression. J Bone Joint Surg [Am] 70A:1048-1051, 1988.
3. Mitchell DG, Rao VM, Dalinka MK, et al.: Femoral head avascular necrosis: correlation of MR imaging, radiographic staging, radionuclide imaging, and clinical findings. Radiology 162:709-15, 1987.
4. Turner DA, Templeton AC, Selzer PM, Rosenberg AG, Petasnick JP: Femoral capital osteonecrosis: MR finding of diffuse marrow abnormalities without focal lesions. Radiology 171:135-40, 1989.
5. Ficat RP, Arlet J. Necrosis of the femoral head. In: D. S. Hungerford, ed. Ischemia and necross of bone. Baltimore: Williams and Wilkins, 1980: 171-182.
6. Lafforgue P, Dahan E, Chagnaud C, Schiano A, Kasbarian M, Acquaviva PC: Early-stage avascular necrosis of the femoral head: MR imaging for prognosis in 31 cases with at least 2 years of follow-up. Radiology 187:199-204, 1993.
7. Bluemke DA, Petri M, Zerhouni EA: Femoral head perfusion and composition: MRI and MRS evaluation in patients at risk for avascular necrosis. Radiology 197:433-438, 1995.
8. Hayes CW, Conway WF, Daniel WW: MR imaging of bone marrow edema pattern: transient osteoporosis, transient bone marrow edema syndrome or osteonecrosis. Radiographics 13:1001-1011, 1993.

[Go Back | Spiral CT Liver | MRI Liver | MRI lymph nodes | Cardiac MRI]