[en] ⁹⁰Y microspheres are used for intra-arterial treatment of liver tumours. In the patient preparation, a hepatic angiogram is performed and all arteries that could transport microspheres from the targeted liver vasculature to extrahepatic organs are blocked. ^99mTc-labelled macroaggregated albumin (MAA) is injected intra-arterially to simulate the treatment and whole-body scintigraphy and single photon emission computed tomography (SPECT) of the abdomen are performed. Various aspects of lung shunt fraction (LSF) estimation were studied: interobserver and intrapatient variability, influence of scan quality and underlying disease. Secondly, the interobserver variability in reading the MAA SPECT of the abdomen was investigated. We reviewed 90 whole-body scans and 20 SPECT scans performed at our institution. Readers were blinded to each other's findings. Scoring the scan quality was based on the visualization of tracer degradation. The mean difference in LSF between the readers was 1%. In 1 of 23 patients who underwent repeated MAA injections a marked change in LSF was observed. No significant differences in LSF were recorded for primary versus secondary liver tumours. There was a correlation between scan quality and LSF, suggesting that low scan quality leads to overestimation of the LSF. Concordant results in ruling out the presence of extrahepatic tracer deposition were reached in 17 of 20 scans (85%). Interobserver and intrapatient variability in LSF calculation was limited. LSF was clearly dependent on scan quality. The underlying disease had no significant impact on the LSF. Interobserver variability for reading the MAA SPECT scans was acceptable. (orig.)

[en] Purpose: Portal vein embolization is performed to increase the future liver remnant before liver surgery in patients with liver malignancies. This study assesses the feasibility of a transsinusoidal approach for portal vein embolization (PVE) with the ethylene vinyl alcohol copolymer, Onyx. Methods: Indirect portography through contrast injection in the cranial mesenteric artery was performed in eight healthy pigs. Onyx was slowly injected through a microcatheter from a wedged position in the hepatic vein and advanced through the liver lobules into the portal system. The progression of Onyx was followed under fluoroscopy, and the extent of embolization was monitored by indirect portography. The pigs were euthanized immediately (n = 2), at 7 days (n = 4), or at 21 days postprocedure (n = 2). All pigs underwent necropsy and the ex vivo livers were grossly and histopathologically analyzed. Results: Transsinusoidal PVE was successfully performed in five of eight pigs (63%). In 14 of 21 injections (67%), a segmental portal vein could be filled completely. A mean of 1.6 liver lobes per pig was embolized (range 1–2 lobes). There were no periprocedural adverse events. Focal capsular scarring was visible on the surface of two resected livers, yet the capsules remained intact. Histopathological examination showed no signs of recanalization or abscess formation. Mild inflammatory reaction to Onyx was observed in the perivascular parenchyma. Conclusions: The porcine portal vein can be embolized through injection of Onyx from a wedged position in the hepatic vein. Possible complications of transsinusoidal PVE and the effect on contralateral hypertrophy need further study.