The author studied the feasibility of suing gadolinium-enhanced dynamic magnetic resonance (MR) imaging to noninvasively monitor the bone marrow perfusion of the femoral head and neck in human subjects. With a spoiled gradient-recalled acquisition, sequential images of 16 hips in 8 normal human subjects were obtained for 4 minutes after an intravenous bolus injection of 0.2m mole of gadolinium per kilogram. The average of peak percentage of enhancement (SI max %) of both sides femoral heads or necks was 23.2% without significant difference (P=0.8144 by ANOVE test). Another six femoral heads of five patients with early stage avascular necrosis received core decompression procedure. Follow-up dynamic contrast-enhanced MR was performed from 2days to 9 months after the operation. The SI max% varied form 17% to 110% of each femoral head. It was highest at 2 to 3 months postoperatively and then decreased to nearly normal at 9 months after the operation. We assumed this cange of blood perfusion reflected the postoperative fibrinoid reacthion of the core channel within the femoral head and neck, not the restoration of blood flow after core decompression. Thus, dynamic contrast- enhanced MR could be used as a non-invasive method to evaluated the bone marrw blood perfusion in acute or post-therapeutic change of the femoral head.