A Quantitative Comparison of PET/CT and SPECT/CT for post Y-90 SIRT Therapy


Objectives: The two main aims of this project were as follows, firstly to optimise PET/CT acquisition length and reconstruction for the purposes of quantitative image of 90Y post SIRT. Secondly, to compare this optimised PET/CT protocol with a previously optimised SPECT/CT reconstruction for quantitative imaging of 90Y post-SIRT. This was done with a view to carry out individualised patient dosimetry. Methods: A NEMA IQ phantom with an 8:1 sphere-to-background ratio was scanned overnight on PET/CT. This overnight scan was rebinned into varying acquisition lengths and reconstructed using a BPL algorithm to determine both the optimum scan time and reconstruction for quantification. Contrast Recovery (CR), background variability (BV) and recovered activity percentage (RAP) were used as a basis for optimisation. The same phantom was scanned on both SPECT/CT and PET/CT in order to compare the quantification of the two modalities. SPECT/CT datasets were reconstructed using an optimised OSEM algorithm (5 iterations, 15 subsets, Monte Carlo collimator modelling) using Hermes Hybrid Recon (Hermes Medical Solutions AB). PET/CT datasets were reconstructed using the optimised reconstruction and acquisition time that had been identified. This investigation was repeated using an anthropomorphic liver phantom (AbdoMan™) with a 4:1 ratio. Five patients that had had both SPECT/CT and PET/CT post-SIRT were retrospectively reconstructed using these optimised algorithms. Voxel-based dosimetry was carried out using Hermes Internal Radionuclide Dosimetry (HIRD) to compare differences in liver doses for both modalities. Results: A beta value of 1000 produced the highest CR and RAP (76% & 73%, 37mm sphere). There was no statistically significant increase (P<0.05) in either the CR or RAP for scan times of >15 minutes. There were significant increases in CR & RAP values for all analysed spheres in the NEMA phantom (P<0.05) for PET/CT over SPECT/CT. For all spheres in the AbdoMan Phantom, bar the 40mm, there was a significant increase in CR & RAP for PET/CT (P<0.05). For patient dosimetry, PET/CT produced, on average, a 7% higher absorbed liver dose compared to SPECT/CT. Conclusions: Results indicate that an acquisition length of 15 minutes and beta value of 1000 (when using Q.Clear reconstruction) is optimum for quantitative 90Y PET/CT imaging. Optimised PET/CT appears to be superior to SPECT/CT for the purposes of quantitative imaging of 90Y