Predictive value of 99mTc-MAA-based dosimetry in personalized 90Y-SIRT planning for liver malignancies

Abstract

Background: Selective internal radiation therapy with ⁹⁰Y radioembolization aims to selectively irradiate liver tumours by administering radioactive microspheres under the theragnostic assumption that the pre-therapy injection of ^99mTc labelled macroaggregated albumin (^99mTc-MAA) provides an estimation of the ⁹⁰Y microspheres biodistribution, which is not always the case. Due to the growing interest in theragnostic dosimetry for personalized radionuclide therapy, a robust relationship between the delivered and pre-treatment radiation absorbed doses is required. In this work, we aim to investigate the predictive value of absorbed dose metrics calculated from ^99mTc-MAA (simulation) compared to those obtained from ⁹⁰Y post-therapy SPECT/CT.

Results: A total of 79 patients were analysed. Pre- and post-therapy 3D-voxel dosimetry was calculated on ^99mTc-MAA and ⁹⁰Y SPECT/CT, respectively, based on Local Deposition Method. Mean absorbed dose, tumour-to-normal ratio, and absorbed dose distribution in terms of dose-volume histogram (DVH) metrics were obtained and compared for each volume of interest (VOI). Mann-Whitney U-test and Pearson's correlation coefficient were used to assess the correlation between both methods. The effect of the tumoral liver volume on the absorbed dose metrics was also investigated. Strong correlation was found between simulation and therapy mean absorbed doses for all VOIs, although simulation tended to overestimate tumour absorbed doses by 26%. DVH metrics showed good correlation too, but significant differences were found for several metrics, mostly on non-tumoral liver. It was observed that the tumoral liver volume does not significantly affect the differences between simulation and therapy absorbed dose metrics.

Conclusion: This study supports the strong correlation between absorbed dose metrics from simulation and therapy dosimetry based on ⁹⁰Y SPECT/CT, highlighting the predictive ability of ^99mTc-MAA, not only in terms of mean absorbed dose but also of the dose distribution.

Keywords: 90Y radioembolization; Dosimetry; SPECT; Selective internal radiation therapy; Theragnostics.

Fig. 1

Followed workflow. The acquired images are the pre-treatment diagnostic CT or MR and the SPECT and AC-CT from both simulation and therapy. For the post-therapeutic verification, segmentation and registration is performed. From these images, we proceed to calculate the activity map and then, the dose map based on self-calibration strategy and the local energy deposition approach. Finally, the MAD, DVH and TNR are calculated within the three VOIs

Fig. 2

An example of patient images showing agreement (top: a, b, c) and other example showing disagreement (bottom: d, e, f) between simulation SPECT with 99mTc-MAA and therapy bSPECT with 90Y. a, d Diagnostic contrast-enhanced CT. b, e SPECT/CT images of the pre-treatment simulation with 99mTc-MAA. c, f SPECT/CT images of the therapy session with 90Y. The segments corresponding to the perfused lobe and to the tumour are depicted in green and blue, respectively. Both examples correspond to HCC patients

Fig. 3

Dose-volume histograms of a patient with visual match from the SPECT (left) and mismatch (right) for all VOIs (TL, NTLt, NTLw), both patients of HCC with a single lesion

Fig. 4

Correlations plots for different dosimetry metrics derived from TL: mean dose (MAD), D50 and D70 (top) and D95, V120 and V205 (bottom). For visualization purposes, in each graph the axes are set to equal length, leaving some points out of frame. The entire images are found on Additional file 1: Fig. S1

Fig. 5

Correlations plots for different dosimetry metrics derived from NTLt: mean dose (MAD), D50 and D70 (top) and D95, V50 and V90 (bottom). For visualization purposes, the correlation for V20 is not shown, and in each graph the axes are set to equal length, leaving some points out of frame. The entire images are found on Additional file 1: Fig. S2

Fig. 6

Correlations plots for different dosimetry metrics derived from NTLw: mean absorbed dose (MAD), D50 and D70 (top) and D95, V50 and V90 (bottom). For visualization purposes, the correlation for V20 is not shown, and in each graph the axes are set to equal length, leaving some points out of frame. The entire images are found on Additional file 1: Fig. S3

Fig. 7

Correlation plots between MAD from simulation (Tc-MAA) and therapy (⁹⁰Y) for TL < 200 ml and TL > 200 ml. For visualization purposes, in each graph the axes are set to equal length, leaving some points out of frame. The entire images are found on Additional file 1: Fig. S4

Fig. 8

Correlation and Bland–Altman plots for TNR derived from NTLt and NTLw from simulation (top) and post-therapy (bottom)