Skeletal Radiol (1992) 21:375-379

Skeletal Radiology

Contrast-enhanced fat saturation magnetic resonance imaging for studying the pathophysiology of osteonecrosis of the hips King C.P. Li, M.D., F.R.C.P.(C.)* and Paul Hiette, M.D. Department of Radiology, St. Joseph's Hospital and Medical Center, Phoenix, Arizona, USA

Abstract. We imaged 75 hips in 40 patients using fat saturation technique before a n d after intravenous injection o f c o n t r a s t (0.1 m m o l / k g gadolinium diethylene triamine pcnta-acetic acid, G d - D T P A ) . Eighteen hips in 1 patients were determined to be osteonecrotic, either by p a t h o l o g i c or clinical examination. In the osteonecrotic hips, three distinct patterns o f e n h a n c e m e n t were f o u n d : (I) focal area outlined by brightly enhancing rim (7 hips); (II) diffuse e n h a n c e m e n t in the femoral h e a d and neck extending into the femoral shaft (3 hips); a n d (III) a c o m b i n a t i o n o f patterns I and II (8 hips). O u r data s u p p o r t the hypothesis that early n o n t r a u m a t i c osteonecrosis is associated with h y p e r e m i a a n d / o r an increase in capillary permeability rather than acute devascularization, and that diffuse m a r r o w edema is the initial finding in early n o n t r a u m a t i c osteonecrosis.

Key words: H i p -

O s t e o n e c r o s i s - Magnetic resonance imaging ( M R I ) - C o n t r a s t e n h a n c e m e n t - Fat suppression

A l t h o u g h m a g n e t i c resonance imaging ( M R I ) has been shown to be a sensitive means o f detecting osteonecrosis o f the hip [3, 5, 10], the p a t h o p h y s i o l o g i c a l basis o f the various a b n o r m a l M R I findings in osteonecrotic hips is not well u n d e r s t o o d [9]. T h e use o f g a d o l i n i u m diethylene triamine penta-acetic acid ( G d - D T P A ) for contraste n h a n c e d M R I o f lrmsculoskeletal disorders has been investigated in recent years [2], a n d G d - e n h a n c e d M R I examination with a gradient-recalled pulse sequence has been s h o w n to be useful in evaluating b o n e m a r r o w perfusion in a d o g m o d e l [4]. D u e to the n o r m a l high intensity o f m a r r o w fat, c o n t r a s t - e n h a n c e d M R I m a y decrease the visualization o f m a r r o w abnormalities, and the use o f fat suppression imaging techniques, such as Correspondence to: K.C.P. Li, *present address." Department of Radiology, Stanford University School of Medicine, Room S078, Stanford, CA 94305, USA

chemical shift imaging or selective presaturation, has been a d v o c a t e d as a solution to this p r o b l e m [2, 13]. The p u r p o s e o f this study is to investigate the usefulness o f the c o n t r a s t - e n h a n c e d fat saturation technique [12] in determining thc p a t h o p h y s i o l o g y o f osteonecrosis o f the hips.

Materials and methods We imaged 75 hips in 40 patients (25 F, 15 M, mean age 48.5 years, range 7-85 years) with hip pain using the contrast-enhanced fat saturation technique. All MR examinations were performed with a 1.5-T superconducting whole-body imager (Signa, General Electric, Milwaukee, Wis.). Before contrast injection, all patients were imaged in the coronal and axial planes with a Tl-weighted spin echo pulse sequence (TR 600-850 ms/echo time TE 10 20 ms). All patients but one were imaged using the coronal Tl-weighted spin echo fat saturation technique (TR 600-800 ms/TE 16-20 ms). T2-weighted spin echo pulse sequences (TR 1800-2500 ms/TE 8090 ms) were also obtained in 14 patients. All sequences used a matrix size of 256 x 192, two excitations, and a 5-mm slice thickness with an interslice gap of 1 mm. All patients received intravenous injections of 0.1 mmol/kg Gd-DTPA. Contrast-enhanced fat saturation images were obtained within 5 min of contrast injection with parameters idcntical to those of the precontrast fat saturation scans. The investigators carefully studied all images, paying special attention to the various enhancement patterns in the osteonecrotic hips. The details of the fat saturation technique have been described previously [6, 12]. Briefly, this tcchnique suppresses fat signal by transmitting a low amplitude radlofrequency (RF) pulse of long duration centered on the frequency of lipid resonance, followed by a "homospoil" gradient immediately before the beginning of each excitation pulse sequence. The excitation pulse sequence is then carried out in the usual fashion with signal reception centered on the frequency of water.

Results Eighteen hips in 11 patients (6 F, 6 M, m e a n age 41.7 years, range 9 - 7 4 years) were determined to be osteonecrotic. The diagnosis was confirmed by surgical p a t h o l o gy in 7 patients and by clinical and imaging follow-up 9 1992 International Skeletal Society

376

K.C.P. Li and P. Hiette: MRI of hip osteonecrosis

Fig. 1A, B. Normal hip with little red marrow. Normal right hip in a 49-year-old man: A SE 700/20 and B SE 800/20 fat saturation contrast-enhanced images. Note that little enhancement of the yellow marrow and trabecular bone is seen on the postcontrast image Fig. 2A, B. Normal hip with a large amount of red marrow. Normal right hip in a 45year-old woman: A SE 817/20 and B SE 567,/20 fat saturation contrast-enhanced images. Note the marked difference in the enhancement pattern of the red and yellow marrow

in 4. Loss of osteocytes as evidenced by the presence of empty osteocyte lacunae in addition to marrow necrosis were used as the minimum requirements for the histologic diagnosis of osteonecrosis. Progressive collapse o f the femoral head as evidenced by radiographs, and the presence of predisposing factors for osteonecrosis constituted the imaging and clinical criteria for osteonecrosis. MRI was never the only evidence of osteonecrosis in our series. Etiologies of the osteonecrosis included trauma (2 patients, 1 after hip pinning for fracture), steroid use (6 patients), and Legg-Perthes disease (3 patients). Red marrow was found to enhance more than yellow marrow on the contrast-enhanced fat saturation images of the normal hips (Figs. 1, 2). Three distinct patterns of enhancement were found in the osteonecrotic hips: (I) focal area of relatively low enhancement outlined by brightly enhancing rim (7 hips in 5 patients: 1 posttraumatic, 3 steroid-induced, 3 Legg-Perthes) (Fig. 3); (lI) diffuse enhancement in the femoral head and neck extending into the femoral shaft (3 hips, all steroid-in-

duced in 2 patients) (Fig. 4); and (III) a combination of patterns I and II (8 hips in 6 patients: 1 postsurgical, 6 steroid-induced, I Legg-Perthes) (Figs. 5, 6). One patient had a type I pattern in one hip and a type III pattern in the other (Figs. 3, 5). Another patient had type II pattern in one hip and type III pattern in the other (Figs. 4, 6). Both of these patients had steroidinduced osteonecrosis.

Discussion Although the pathogenesis of nontraumatic osteonecrosis is not well understood, it is believed that early causes include a decrease in venous drainage resulting in bone marrow edema, an increase in intramedullary pressure, and a decrease in marrow perfusion. Turner et al. [14] reported that a diffuse edema pattern with low signal intensity on Tl-weighted images and high signal intensity on T2-weighted images could be seen in the femoral

K.C.P. Li and P. Hiette: MRI of hip osteonecrosis

377

Fig. 3. Osteonecrosis with type I enhancement pattern. Right hip of a 68-year-old woman with pathologically proven bilateral, steroid-induced osteonecrosis of the hips: SE 800/20 fat saturation contrast-enhanced image. Note the well-defined, relatively poorly enhancing area outlined by a markedly enhancing rim with little associated enhancement in the rest of the femoral head and neck Fig. 4. Osteonecrosis with type II enhancement pattern. Right hip of a 27-year-old woman with pathologically proven, bilateral, steroid-induced osteonecrosis of the hips: SE 800/20 fat saturation contrast-enhanced image. Notice the extensive enhancement of the femoral head and neck without a poorly enhancing focal area Fig. 5. Osteonecrosis with type III enhancement pattern. Left hip of the patient illustrated in Fig. 3: SE 800/20 fat saturation contrast-enhanced image. Note the poorly enhancing area outlined by a brightly enhancing rim and the enhancement of the marrow in the femoral head and neck Fig. 6. Osteonecrosis with type III enhancement pattern. Left hip of the patient illustrated in Fig. 4:800/20 fat saturation contrast-enhanced image. Note the small crescentic area of poorly enhancing tissue in the subcortical area of the femoral head surrounded by brightly enhancing marrow in the femoral head and neck

head, neck, and intertrochanteric region in patients with early osteonecrosis w i t h o u t focal abnormalities in the femoral head. K r a m e r et al. [8] histologically examined b o n e cores f r o m patients with the typical M R pattern o f b o n e m a r r o w e d e m a a n d f o u n d changes o f early osteonecrosis. The histologic changes observed included an increase in intrasinusoidal fluid, an a c c u m u l a t i o n o f interstitial fluids in the m a r r o w , and extended fatty marr o w degeneration alternating with fibrovascular regeneration o f the m a r r o w . O u r d a t a d e m o n s t r a t e d that m a r r o w areas with a diffuse e d e m a pattern (types I1 and III) showed m a r k e d e n h a n c e m e n t following intravenous G d - D T P A administration. This finding correlates well with the increase in radionuclide u p t a k e in these areas reported by previous investigators [8, 14]. Since G d - D T P A is freely dis-

tributed t h r o u g h o u t the extracellular space, the increase in e n h a n c e m e n t could be due to an increase in b l o o d flow, increase in capillary permeability, or a c o m b i n a tion o f the two. In the future, use o f c o n t r a s t - e n h a n c e d d y n a m i c M R scans [4] or new intravascular contrast agents [1, 11, 15] m a y be helpful in differentiating the effect o f an increase in b l o o d flow f r o m that o f an increase in capillary permeability. Diffuse m a r r o w e n h a n c e m e n t with or w i t h o u t focal lesions was c o m m o n in our series (10/16 n o n t r a u m a t i c osteonecrotic hips). In one patient with bilateral osteonecrosis, one hip showed only diffuse m a r r o w enhancement, and the contralateral hip showed diffuse m a r r o w e n h a n c e m e n t with a focal area o f p o o r enhancement. In a n o t h e r patient with bilateral osteonecrosis, one hip showed only diffuse m a r r o w enhancement, and the con-

378 tralateral hip had a focal area of poor enhancement outlined by a brightly enhancing rim. These observations offer indirect evidence supporting the hypothesis that diffuse marrow edema is the initial finding in early nontraumatic osteonecrosis which, if left untreated, may progress to focal areas of abnormality demonstrable by MRI. In the later stages, the marrow edema may subside, leaving the focal abnormalities as the only findings on M R examination. The diffuse marrow edema pattern of early nontraumatic osteonecrosis seen on MRI is indistinguishable from that seen in transient osteoporosis of the hip [8, 9, 14]. That is why we used stringent criteria to exclude cases of transient osteoporosis of the hip from our study. It is also possible that the diffuse marrow edema pattern seen on MRI can be caused by various pathologic processes including infection, neoplasm, and trauma and is not specific to transient osteoporosis of the hip or osteonecrosis. In our series, we did not observe diffuse marrow edema or focal abnormalities suggestive of osteonecrosis in the patients who were determined to be normal by clinical and imaging follow-up. All the patients with documented osteonecrosis had M R abnormalities. Other investigators have reported a spontaneous regression of M R findings of osteonecrosis [7]; conversely, patients with documented osteonecrosis may not demonstrate any abnormal M R findings [5]. Most investigators agree, however, that the accuracy of M R in diagnosing osteonecrosis is very high [3, 5, 10]. The high accuracy in our series probably relates to the highly selected nature of our population. Patients with hip pain with or without predisposing factors for osteonecrosis were included in the study, making the prevalence of disease lower than some of the reported series and increasing the true negative rate. All patients with osteonecrosis in this study were symptomatic; therefore, early asymptomatic osteonecrosis that may be missed by M R I was not included. Both of these factors probably contributed to the high accuracy o f this study. The fat saturation technique has several advantages over other fat suppression techniques [6, 12]. It does not require superimposition of two imaging matrices for subtraction, it is compatible with any regular spin echo sequence, and it does not require special postprocessing of the imaging data. There are, however, some disadvantages of this technique [6, 12]. First, it depends on the homogeneity of the main magnetic field for signal selection. Fat suppression is usually less homogeneous at the periphery of an image obtained with a large field-of-view than it is at the center of the image. In our experience, however, fat suppression in the region of the femoral heads was almost always good. Another minor problem is that, due to the additional R F saturation pulse, spin echo sequences obtained with the fat saturation technique yield 15% fewer slices than sequences without this pulse. To cover the same area, sequences with longer T R must be used. Lastly, for accurate fat suppression, the spectral peak of lipids needs to be identified so that the chemically selective pulse can be centered properly. With minimal training, our technologists were able to

K.C.P. Li and P. Hiette: MRI of hip osteonecrosis do this efficiently and accurately. Some institutions prefer the short tau inversion recovery (STIR) technique as the primary technique for fat suppression as it gives a more uniform fat suppression across the field-of-view. However, with an intravenous administration of GdDTPA, the T1 relaxation times of nonfat tissues containing G d - D T P A may be shortened to approximately that of fat, and visualization of these tissues may be suppressed by the STIR technique. Therefore, the fat saturation technique is preferable to the STIR technique for contrast-enhanced scans. In conclusion, we found that the contrast-enhanced fat saturation technique provided us with a new, noninvasive way of studying the pathophysiology of osteonecrosis. We found that nontraumatic osteonecrosis is often associated with a diffuse increase in contrast enhancement, which supports the hypothesis that early nontraumatic osteonecrosis is associated with hyperemia and/or an increase in capillary permeability rather than with acute devascularization. There was also indirect evidence that diffuse marrow edema is the initial finding in early nontraumatic osteonecrosis which, if left untreated, may progress to develop focal areas of abnormality demonstrable by MRI. Further longitudinal clinical and animal studies with the contrast-enhanced fat saturation technique are definitely warranted.

References 1. Armitage FE, Richardson DE, Li KCP (1990) Polymeric contrast agents for magnetic resonance imaging: synthesis and characterization of gadolinium diethylenetriaminepentaacetic acid conjugated to polysaccharides. Bioconjug Chem 1 : 365 2. Beltran J, Chandnani V, McGhee RA Jr, Kursunoglu-Brahme S (1991) Gadopentetate dimeglumine-enhanced MR imaging of the musculoskeletal system. AJR 156:457 3. Coleman BG, Kressel HY, Dalinka MK, Scheibler ML, Burk DL, Cohen EK (1988) Radiographically negative avascular necrosis: detection in high-risk patients with MR imaging. Radiology 168:525 4. Cova M, Kang YS, Tsukamoto H, et al. (1991) Bone marrow perfusion evaluated with gadolinium-enhanced dynamic fast MR imaging in a dog model. Radiology 179:535 5. Genez BM, Wilson MR, Houk RW, et al. (1988) Early osteonecrosis of the femoral head: detection in high-risk patients with MR imaging. Radiology 168:521 6. Keller PJ, Hunter WW, Schmalbrock P (1987) Multisection fat-water imaging with chemical shift selective presaturation. Radiology 164:539 7. Kopecky KK, Braunstein EM, Brandt KD, et al. (1991) Apparent avascular necrosis of the hip: appearance and spontaneous resolution of MR findings in renal allograft recipients. Radiology 179 : 523 8. Kramer J, Hoffman S Jr, Engel A Jr, Neuhold A Jr, Leder K Jr, Imhof H (1991) Bone marrow edema in transient osteoporosis : initial stage of avascular necrosis of the hip. Radiology 181(P): 136 9. Mitchell DG (1989) Using MR imaging to probe the pathophysiology of osteonecrosis. Radiology 171:25 10. Mitchell DG, Rao VM, Dalinka MD, et al. (1987) Avascular necrosis of the femoral head: correlation of MR imaging, radiographic staging, radionuclide imaging, and clinical findings. Radiology 162:709 11. Schmiedl U, Ogan M, Paajanen H, et al. (1987) Albumin labeled with Gd-DTPA as an intravascular, blood pool-enhanc-

K.C.P. Li and P. Hiette : MRI of hip osteonecrosis ing agent for MR imaging : biodistribution and imaging studies. Radiology 162: 205 12. Semelka RC, Chew W, Hricak H, Tomei E, Higgins CB (1990) Fat-saturation MR imaging of the upper abdomen. AJR 155:1111 13. Simon JH, Szumowski J (1989) Chemical shift imaging with paramagnetic contrast material enhancement for improved lesion depiction. Radiology 171:539

379 14. Turner DA, Templeton AC, Selzer PM, Rosenberg AG, Petasnick JP (1989) Femoral capital osteonecrosis: MR findings of diffuse marrow abnormalities without focal lesions. Radiology 171:135 15. Wang S, Wilkstrom MG, White DL, et al. (1990) Evaluation of Gd-DTPA-labeled dextran as an intravascular MR contrast agent: imaging characteristics in normal rat tissues. Radiology 175:483