Hepatic abscesses. Magnetic resonance imaging findings using gadolinium-BOPTA

Gadolinium (Gd)-BOPTA was evaluated in a rabbit liver abscess model and compared with Gd-HP-DO3A, examining lesion conspicuity and characterization.

Five New Zealand White rabbits with a liver abscess were studied on a 1.5-tesla Siemens Vision magnetic resonance unit. The disease model was created by surgically implanting a gel capsule filled with fusobacterium into the central or left lobe of the liver. For imaging, the animals were ventilated using a Harvard pump. Pancuronium bromide (0.12 mg/kg) was administered to allow acquisition of breath-hold scans. Magnetic resonance scans were obtained in each animal on days 2 and 3 after surgery. Every animal was studied twice, once after intravenous injection of 0.3 mmol/kg Gd-HP-DO3A (gadoteridol; ProHance) and once after intravenous injection of 0.1 mmol/kg Gd-BOPTA (gadobenate dimeglumine; MultiHance). The order of injection for the two agents was randomized with the two studies in each animal, separated by 24 hours to permit clearance. Image acquisition was performed in each instance with respiration suspended. Baseline two-dimensional spin-echo T1-weighted and fast spin-echo T2-weighted breath-hold scans were obtained first. The voxel dimensions were 5 x 0.8 x 0.8 mm3. Imaging times were 23 seconds for the T1-weighted scan and 26 seconds for the T2-weighted scan. Postcontrast scans, using spin-echo T1-weighted technique, were obtained at 1, 3, 5, and 15 minutes after contrast injection, whether Gd-HP-DO3A or Gd-BOPTA was used. Additional scans were obtained at 30, 45, and 60 minutes after Gd-BOPTA administration. At the completion of imaging on day 3, each animal was killed and the liver was removed and taken to a veterinary pathologist at the University's animal disease diagnostic lab for gross and histologic examination.

The enhancement of normal liver parenchyma, assessed by region of interest measurement and specifically as (SI(t) - SI0)/SI0 x 100, peaked at 119 +/- 37% 1 minute after injection of 0.3 mmol/kg Gd-HP-DO3A and at 126 +/- 30% 30 minutes after injection of 0.1 mmol/kg Gd-BOPTA. The difference in enhancement achieved, comparing results at each time point, was statistically significant only at 1 and 3 minutes postcontrast (P = 0.003 and 0.03). Lesion conspicuity, specifically (SIliver - SIlesion/noise), increased from 272 +/- 29 precontrast to a maximum of 639 +/- 73 at 30 minutes postcontrast using a dose of 0.1 mmol/kg Gd-BOPTA, with the improvement statistically significant (P = 0.0003). Lesion conspicuity on the T2-weighted scan was 137 +/- , with the Gd-BOPTA scan markedly superior (P = 0.00004). On scans at 45 and 60 minutes after Gd-BOPTA administration, a progressive increase in signal intensity in the central necrotic portion of the lesion was observed. This was most consistent with gradual diffusion of the agent from the adjacent liver into the lesion. Using Gd-HP-DO3A at 0.3 mmol/kg (three times the dose for Gd-BOPTA), lesion conspicuity increase from 305 +/- 37 precontrast to a maximum of 701 +/- 92 at 1 minute postcontrast, with this difference also statistically significant (P = 0.0004). The abscess rim exhibited moderate contrast enhancement, greater than that of normal liver parenchyma, on early postcontrast images with Gd-HP-DO3A.

The conspicuity of an early liver abscess is improved markedly on delayed imaging after administration of 0.1 mmol/kg Gd-BOPTA. Although a similar magnitude of parenchymal enhancement can be obtained after the administration of an extracellular agent, such as Gd-HP-DO3A, high-contrast dose (0.3 mmol/kg) and early dynamic imaging are required. The appearance of a liver abscess on late scans (45 to 60 minutes) after Gd-BOPTA injection is distinct from that of nonnecrotic metastases, with diffusion of the agent into the lesion noted.