Lung shunt fraction calculations before Y-90 transarterial radioembolization: Comparison of accuracy and clinical significance of planar scintigraphy and SPECT/CT.

Abstract

To determine the accuracy and clinical significance of planar scintigraphy lung shunt fraction (PLSF) and single-photon emission computerized tomography (SPECT) computed tomography (CT) lung shunt fraction (SLSF) before Y-90 transarterial radioembolization. Seventy patients (46 men, 24 women; mean age, 64±9.5 [SD] years) who underwent 83 treatments with Y-90 transarterial radioembolization for primary or secondary malignancies of the liver with a PLSF ≥ 7.5% were retrospectively evaluated. The patients mapping technetium 99m (Tc-99m) macroaggregated albumin (MAA) PLSF and SLSF were calculated and compared to the post Y-90 delivery SLSF. A model using modern dose thresholds was created to identify patients who would require dose reduction due to a lung dose ≥ 30Gy, with patients who required >50% dose reduction considered to be delivery cancelations. A significant difference was found between mean PLSF (14.7±11.6 [SD]%; range: 7.5-84.1%) and mean SLSF (8.7±8.5 [SD]%; range: 1.7-73.5) (P < 0.001). The mean realized LSF (7.1±3 [SD]%; range:1.5-17.6) was significantly less than the PLSF (P <0.001) but not the SLSF (P=0.07). PLSF significantly overestimated the realized LSF by more than the SLSF (8.5±5.3 [SD] % [range: -0.1-21.7] vs. 0.8±3.6 [SD] % [range: -5-13.2], respectively) (P < 0.001). Based on the clinical significance model, 20 patients (20/83, 24.1%) would have required dose reduction or cancelation when using PLSF but would not require even a dose reduction when using the SLSF. Significantly more deliveries would have been be canceled if PLSF was used as compared to SLSF (22/83 [26.5%] vs. 6/83 [7.2%], respectively) (P < 0.001). SLSF is significantly more accurate at predicting realized LSF than PLSF and this difference is of clinical significance in a number of patients with a PLSF ≥ 7.5%.