Functional Perfusion MRI Predicts Later Occurrence of Steroid-Associated Osteonecrosis: An Experimental Study in Rabbits Hui Sheng, 1 Ge Zhang, 1 Yi-Xiang Wang, 2 David Ka-Wai Yeung, 2 James Francis Griffith, 2 Kwok-Sui Leung, 1 Ling Qin 1 1 Musculoskeletal Research Laboratory of the Department of Orthopaedics and Traumatology, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, New Territories, Hong Kong SAR, China, 2 Department of Diagnostic Radiology and Organ Imaging, The Chinese University of Hong Kong, Hong Kong SAR, China Received 2 May 2008; accepted 24 July 2008 Published online 24 November 2008 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jor.20765 ABSTRACT: Ischemia is the defined pathway leading to steroid-associated osteonecrosis (ON). Early detection of ischemic condition may help predict later ON occurrence. Bone marrow perfusion function evaluation by perfusion magnetic resonance imaging (MRI) may be a unique modality for this application. Twenty-five adult male New Zealand white rabbits were used in this study. Lipopolysaccharide (LPS) and methylprednisolone (MPS) were administrated for ON induction based on a published protocol. T1-weighted and fat suppression T2-weighted MR imaging (conventional MRI) were performed for ON lesion detection based on the abnormal signal in the proximal femora at week 0 as the baseline (before LPS injection), and week 1 and week 2 after MPS injection. At the same time, the blood perfusion function in the proximal femora was measured by perfusion MRI. Maximum enhancement (ME)—an index of MRI perfusion function was analyzed. After MRI scanning, the proximal femora were prepared histopathologically for ON lesion analysis. The rabbit with bilateral histopathological ON lesions was defined as an ONþ rabbit and included in the ONþ group evaluated at week 1 and week 2, respectively, and the rabbit without ON lesions in bilateral femora was classified into the ON group. For the underlying mechanism of perfusion change, the extravascular marrow fat cells were measured and the intravascular endothelium inflammation injury indicator of tissue factor (TF) expression and thrombus formation were detected. In ONþ group, ME in perfusion MRI showed a significant decrease at week 1 and week 2 as compared with the baseline (p < 0.01). There was a more than 50% decrease in ME at week 1 in ONþ group; whereas there were no detectable ON lesions by conventional MRI at week 1, though 93% (14/15) rabbits could be detected at week 2 in ONþ group. In ON group, ME showed a slight decrease at week 1 (less than 30%), and nearly recovered to normal at week 2 as compared with the baseline. Histological results showed a much larger average marrow fat area and more severe marrow blood sinusoids compression from surrounding crowded fat cells, and stronger positive TF expression in marrow endothelium and more thrombus formation in ONþ rabbits than ON rabbits. This study demonstrated that functional perfusion MRI could predict development of steroid-associated ON. Our experimental data suggested that perfusion MRI might be a sensitive noninvasive modality for monitoring steroid-associated ON in patients. ß 2008 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 27:742– 747, 2009 Keywords: steroid-associated osteonecrosis; perfusion MRI; marrow fat deposition; endothelial cells injury Steroid-associated osteonecrosis (ON) is an intractable and disabling disease in orthopedics, accounting for more than 50% in nontraumatic ON and showing poor prognosis especially in its advanced stage. 1–3 Early diagnosis or detection of ON is very important as its prognosis is stage-dependent. 4,5 Up to date, the most sensitive and commonly used modality for early ON detection relies on the abnormal signal of conventional magnetic resonance imaging (MRI) (T1-weighted and T2-weighted MR imaging), which matches the necrotic and repairing reaction in bone and marrow. 6–8 Yet more than 80% of ON cases found in this stage would progress into symptomatic ON requiring surgical interven- tions. 9,10 More sensitive modality is desirable for earlier diagnosis or prediction of ON development. The deeper understanding of the pathogenesis of steroid-associated ON may help provide new ideas for early ON detection. Many theories about the patho- genesis of steroid-associated ON have been proposed, including enhanced adipogenesis, endothelial cell injury, and fat embolism formation. 11–15 However, the current consensus is that ischemia is the final and common pathway leading to the formation of ON lesions. 9,16 As ischemia proceeds ON lesion formation, the function change of blood vessel perfusion must be an earlier event than the structural alteration of bone and its marrow. This implies that evaluation of blood perfusion function may provide an early diagnosis or predict ON development. Perfusion MRI is a sensitive noninvasive tool for evaluation of blood vessel perfusion function and has been used for evaluation of vasculature function in traumatic bone disorders and osteoporosis. 17,18 Accord- ingly, we hypothesized that functional perfusion MRI might effectively predict steroid-associated ON develop- ment. In order to prove this hypothesis, the present study was designed to investigate blood perfusion function and its correlation with ON development using an established steroid-associated ON rabbit model. 16,19 The underlying pathogenesis of ischemia was also investigated. MATERIALS AND METHODS Animals and ON Induction Twenty-five 28–30-week-old male New Zealand white rabbits weighing 3.5–4 kg were used. The ON induction procedure was adopted from our published protocol. 16 Briefly, the rabbits were intravenously injected with 10 mg/kg body weight of lipopolysaccharide (LPS) (Escherichia coli 0111:B4, Sigma- Aldrich, Inc., St. Louis, MO). Twenty-four hours later, three injections of 20 mg/kg body weight of methylprednisolone (MPS) (Pharmacia & Upjohn, Somerset County, NJ) were given intramuscularly at a time interval of 24 h. The rabbits were kept in cages and received a standard laboratory chow and water ad libitum. All animal experimental procedures described below 742 JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2009 Correspondence to: Ling Qin (T: 00852-2632-3701; F: 00852-2632- 4618; E-mail: lingqin@cuhk.edu.hk) ß 2008 Orthopaedic Research Society. Published by Wiley Periodicals, Inc.