Impact of data-driven respiratory gating on pre-therapeutic Tc99-MAA-based SIRT dose estimations

Objectives Tc99m-MAA SPECT/CT scans are an essential part of SIRT therapy planning. In addition to extrahepatic shunting, they may be used to estimate dose to healthy and disease tissue. However, respiratory motion imposes quantitative bias on SPECT liver images that may hinder this process. The aim of this study was to evaluate the extent to which respiratory gating of the planning scan might affect pre-therapy dose calculations.

Methods 13 datasets from 11 patients injected intraarterially with 97.8±29.3 MBq Tc99m-MAA were acquired with a Symbia T2 SPECT/CT (Siemens Medical Solutions). All patients granted informed consent. 60 views were acquired at 15 sec each, and list-mode data was also available from the system. In addition to ungated projections, fully-automated data-driven respiratory gating was applied to the list-mode data to subdivide it into 5 gates [1]. These 6 datasets were reconstructed using a prototype version of the xSPECT Quant software (20 iterations/1 subset, attenuation and scatter correction). Non-necrotic, hyper-vascularized lesions were located in coregistered MR images, and Volumes of Interest (VOIs) were demarcated on SPECT images using isocontours centered on each focus. Isocontour thresholds were chosen such that the volume of each VOI in the nuclear image matched its corresponding lesion in the MR image. Prospective tumor and liver doses (TD, LD) were evaluated for each patient using a partition model [2], and quantities from the reconstructed gate with the maximum TD/LD ratio were then compared to the ungated values.

Results Figure 1 shows coronal slices from one representative patient. Relative to the ungated image (left), clear head/foot motion is visible between gated images at beginning- (center) and end-inspiration (right). There was no significant difference between LD in gated and ungated reconstructions (p=0.774). However, TD was significantly higher in gated images (p=0.003) by an average of 13±8%, leading to a significant increase in TD/LD ratio by 15±9% (p<0.001). Increases after respiratory gating ranged from 0 to 32%.

Conclusions We found a significant increase in the TD/LD ratio in our gated reconstructions by up to 32%. This indicates that data-driven respiratory gating could allow physicians to more accurately estimate tumor dose while planning a SIRT intervention. Further work must be done to account for necrotic lesions and lung shunting to fully evaluate the effect this information will have on patient management.

• Download figure
• Open in new tab
• Download powerpoint