Abstract
BACKGROUND: (90)Y microsphere selective internal radiation therapy ((90)Y-SIRT) has been widely used for the treatment of unresectable primary or metastatic liver tumors. (99m)Tc-macroaggregated albumin single-photon emission computed tomography-computed tomography ((99m)Tc-MAA SPECT/CT) and (90)Y positron emission tomography-computed tomography ((90)Y PET/CT) are performed before and after (90)Y-SIRT, respectively, for pretreatment dose distribution assessment and posttreatment verification of microsphere distribution. The purpose of this study was to evaluate the predictive value of (99m)Tc-MAA SPECT/CT for dose distribution assessment and analyze its potential impact on the calculation of absorbed dose to normal liver tissue. To achieve this, the differences in tumor imaging, liver tumor-to-normal ratio (TNR), and normal liver tissue absorbed dose (Dliver) between (99m)Tc-MAA SPECT/CT and (90)Y PET/CT were analyzed. METHODS: Thirty-eight patients with primary or metastatic liver cancer who underwent (90)Y-SIRT were retrospectively enrolled. Their (99m)Tc-MAA SPECT/CT and (90)Y PET/CT images were analyzed, and the TNR measured by (99m)Tc-MAA SPECT/CT (TNR(SPECT)) and (90)Y PET/CT (TNR(PET)) was compared. The absolute differences in TRN (ΔTNR) and percent differences in TNR (%TNR) between the two values were also calculated. Correlation analysis was conducted with patient's age, body mass index, total liver volume, and target tumor volume. The absolute differences in D(liver) (ΔD(liver)) and percent differences in D(liver) (%D(liver)) as calculated by TNR(SPECT) and TNR(PET) were respectively compared. RESULTS: TNR(SPECT) and TNR(PET) were not significantly different across imaging modes (P=0.118) and were highly correlated (r=0.854; P<0.01). When stratification by tumor type (primary: n=27; metastatic: n=37) and location (left lobe: n=20; right lobe: n=44) was applied, significant TNR differences emerged only in the left-lobe group (TNR(SPECT) 5.9 vs. TNR(PET) 4.7; P=0.015). Bland-Altman analysis demonstrated good agreement in both total and subgroup TNR measurements. The mean ΔTNR was 0.4±1.7, with %TNR ≤92.1%. Dliver calculations showed no significant intermodal difference (P=0.198). Analyses of ΔD(liver) and %D(liver) revealed maximum overestimations of 37.3 Gy (ΔD(liver) >0) and underestimations of 44.5 Gy (ΔD(liver) <0), with peak %D(liver) reaching 92.0%. CONCLUSIONS: Comparison with (90)Y PET/CT imaging indicated that (99m)Tc-MAA SPECT/CT could accurately reflect TNR and D(liver) in (90)Y-SIRT. However, TNR(SPECT) may have the tendency to overestimate the true TNR, and appropriate dose adjustments should be considered during treatment planning. The differences in TNR observed in the left-lobe group and the potential differences in the primary group should be further analyzed in larger-sample studies.