Author: Lukas M. Carter, Jazmin Schwartz 👨🔬
Affiliation: Department of Medical Physics, Memorial Sloan Kettering Cancer Center, Memorial Sloan Kettering Cancer Center 🌍
Purpose: Accurate lung shunt fraction (LSF) estimation is essential for the safety and efficacy of 90Y-radioembolization therapy. This study compares planar gamma camera imaging and tomographic SPECT-CT imaging of pre-treatment 99mTc-microaggregated albumin (MAA) to assess how tumor location within the liver affects LSF calculation. The data will be used in future studies to explore how imaging artifacts—such as partial volume effects, organ contouring inaccuracies, and organ overlap—and catheter placement variability between MAA planning and therapy contribute to LSF discrepancies. These findings will inform future investigations into whether LSF variability arises from true physiological shunting, catheter positioning, or imaging limitations.
Methods: Lung and liver activity distributions were quantified using planar and SPECT-CT imaging in a cohort of 25 patients undergoing pre-treatment evaluation of 99mTc-MAA for 90Y-radioembolization planning. LSF values were compared between modalities, with a focus on the impact of tumor proximity to the lung-liver interface. For quantification via SPECT-CT images, the lung and liver were automatically contoured from co-registered CT images using TotalSegmentator. As a first-order indicator of the tumor/treatment site, the tumor was contoured using a 42% of maximum intensity threshold; the centroid of the tumor was then computed. Finally, the Euclidian distance from the tumor centroid to the dome of the liver was determined as a metric of tumor proximity to the lung-liver interface. For planar gamma camera images, the lung and liver were contoured manually.
Results: SPECT imaging tended to produce higher LSF estimates when tumors were located near the liver dome, while planar imaging showed no significant correlation.
Conclusion: The observed increase in LSF with SPECT imaging for tumors near the liver dome suggests potential contributions from imaging limitations. These findings provide a foundation for future studies aimed at distinguishing between physiological and imaging-related factors affecting LSF and improving the accuracy of 90Y-radioembolization planning.