Purpose: 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.
Methods: 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.
Results: 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%).
Conclusion: 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.