A patient with liver-dominant metastases is mapped for Y-90 microsphere therapy. The Tc-99m MAA study is injected from a common hepatic artery catheter, but the planned treatment will infuse from a more distal position after a variant artery is embolized, and the shunt fraction is calculated from lung and liver counts on planar images. Which approach makes the lung shunt fraction representative of the planned microsphere distribution?

The pretherapy MAA study earns its value by acting as a miniature duplicate of the therapy, so the particles have to enter at the same arterial point, with the same vessels open, that the microspheres will see. A proximal injection lets activity spread into branches the planned treatment never reaches, which shifts the split between liver uptake and shunted lung activity and makes the calculated fraction a poor predictor of what the spheres will do, because the entry point still decides which territories receive particles. Injecting at the planned site before the variant artery is embolized fails by concept because the mapping then reflects flow through a vessel that will be closed on treatment day, so the distribution no longer matches the plan. Chest-only imaging fails because the fraction is a ratio of lung counts to lung-plus-liver counts, and without the liver denominator there is no fraction at all. Exam caveat: shunt thresholds that trigger activity reduction or exclusion vary by protocol and regulator, so remember the method rather than a cutoff. Practical check: confirm the entry point and the vessel state both match the treatment plan, then confirm images cover lungs and liver before counts are summed.