Click here to


Are you sure ?

Yes, do it No, cancel

Planning Dosimetry for 90Y Therapy: Evaluating the Predictive Power of 99mTc-MAA and the Partition Model

M. Allan Thomas1*, Armeen Mahvash2, Mohamed Abdelsalam2, Ahmed Kaseb2, S. Cheenu Kappadath1, (1) Department of Imaging Physics, UT MD Anderson Cancer Center, Houston, TX, (2) Department of Interventional Radiology, UT MD Anderson Cancer Center, Houston, TX


(Wednesday, 7/15/2020) 2:00 PM - 3:00 PM [Eastern Time (GMT-4)]

Room: Track 1

Purpose: ???Tc-MAA-SPECT/CT is used in ?°Y-radioembolization treatment planning to assess perfused liver volumes and absorbed dose distribution. Variability in absorbed dose estimates arise due to differences in the dosimetry models and imaging modalities used. An important, often overlooked, aspect is differences in tumor uptake and distribution of perfused volumes between ???Tc-MAA and ?°Y-microspheres. In this work, we explore the effects of clinical variables on uncertainties in predicting ?°Y dosimetry from pre-therapy ???Tc-MAA.

Methods: Both partition (PM) and voxel dosimetry (VD) models were used to calculate tumor and normal liver (NL) mean doses using both planning ???Tc-MAA- and verification ?°Y-SPECT/CT in this retrospective analysis (NCT #01900002, N=33). Two categories of cases were stratified (matched versus unmatched) by comparing MAA and ?°Y perfused volumes. Bland-Altman analysis was used to compute the mean bias and 95% prediction intervals (PI) between various scenarios.

Results: Overall, ???Tc-MAA predictions for ?°Y doses are poor, with clear bias (-46%) and large PI (±185%). Only 17/33 of cases were deemed matched. In matched cases, the mean biases (PI) were -5% (±64%) for single tumor cases and -15% (±71%) for tumors >5 cm, representing the best scenarios for ???Tc-MAA dose predictions. NL doses were more reliably predicted by ???Tc-MAA, especially for matched cases with mean bias (PI) of 0% (±31%). PM was equivalent to VD with mean bias (PI) of -3% (±2%) for NL doses, -2% (±5%) for single tumor cases, and -6% (±26%) for tumors >5 cm.

Conclusion: The blind use of ???Tc-MAA to predict ?°Y dosimetry across all cases is not recommended due to large bias and uncertainty. Nonetheless, ???Tc-MAA can reliability predict ?°Y doses with low uncertainties for single tumors and when tumors are >5 cm. These findings stratify the predictive power of ???Tc-MAA based treatment planning for ?°Y dosimetry in improving treatment outcomes.

Funding Support, Disclosures, and Conflict of Interest: SCK serves as consultant for Boston Scientific, Sirtex, ABK Biomedical, and Varian Medical. AM serves as consultant for Boston Scientific, Sirtex, and ABK Biomedical.


SPECT, Dosimetry, Radiation Therapy


IM/TH- Radiopharmaceutical Therapy: Dose estimation: MIRD/deterministic

Contact Email