Clinically Tenable Lung Dose Estimates in Y-90 Radioembolization from Truncated Maa-SPECT/CT with Unknown Lung Mass 📝

Author: Dan Giardina, John Karageorgiou, Chris Malone, Naganathan Mani, Allan Thomas 👨‍🔬

Affiliation: Washington University School of Medicine, Mallinckrodt Institute of Radiology, Washington University School of Medicine 🌍

Abstract:

Purpose: Relative to planar imaging, MAA-SPECT/CT offers more reliable lung shunt fraction (LSF) and lung mean dose (LMD) estimates in 90Y radioembolization. But lung truncation in SPECT/CT can limit its utility. Truncated SPECT/CT data can be used to estimate LSF/LMD, but errors are difficult to predict. Here, the effects of different assumed lung-mass values on LSF and LMD corrections for lung truncation in SPECT/CT were analyzed.
Methods: Lung truncation was simulated at 40, 60, and 80% lung coverage in 56 90Y cases with planar LSF >8%. Empirical correction methods were used to estimate lung counts at full lung-mass from truncated MAA-SPECT/CT data. LMD estimates were computed from four methods: 1) planar imaging, (LMDplanar), 2) uncorrected-truncated SPECT and lung-mass (SPECTTrunc), 3) truncation-corrected SPECT, 1-kg lung-mass (SPECT1-kg), 4) truncation-corrected SPECT, patient-specific lung-mass (SPECTFit). The LMD estimates were compared with non-truncated SPECT/CT (SPECTTrue) using Bland-Altman analysis. The methods were also applied to 31 independent cases with true lung truncation and no ground truth.
Results: The mean±95% prediction interval (PI: 1.96σ) of LMD differences for LMDplanar relative to SPECTTrue were 11.0±14.9 Gy. At 60% simulated lung coverage, they were 0.2±1.8 Gy for SPECTTrunc, 0.2±1.9 Gy for SPECT1-kg, and 0.1±1.7 Gy for SPECTFit. In the 31 truncated clinical cases, differences between LMDplanar and SPECTTrunc had mean±95% PI of 11.2±10.3 Gy. Differences between SPECTTrunc, SPECT1-kg, and SPECTFit were very small (mean<0.5 Gy, 95% PI<1.5 Gy), with minimal dependence on their differing lung-masses.
Conclusion: In lung-truncated MAA-SPECT/CT, offsetting errors from LSF and lung-mass in LMD produce an interesting result: the specific lung-mass used in LMD estimates is not significant. Instead, the most impactful difference is using SPECT/CT data over planar imaging for estimating LMD. Even when lung-mass is unknown or a reference value (1-kg) is used, truncated MAA-SPECT/CT offers more robust LMD estimates than planar imaging.