Improved dosimetry in radioembolization using a dual isotope SPECT/CT protocol with ¹⁶⁶Ho-microspheres and ^99mTc-stannous phytate: a proof of concept

Objectives Improvements in radioembolization dosimetry are expected to lead to improvements in efficacy and toxicity. However, accurate tumor and healthy liver delineation and image co-registration is very challenging. A dual-isotope SPECT/CT protocol was developed to improve dosimetry, using a radiocolloid, 50 MBq ^99mTc-stannous phytate, for normal liver tissue segmentation, and a scout dose, 250 MBq ¹⁶⁶Ho microspheres, for treatment simulation. The image quality of the ¹⁶⁶Ho-dual-isotope SPECT/CT was compared with the ¹⁶⁶Ho-only SPECT/CT in a phantom study and in patients.

Methods First a phantom study was performed using an image quality phantom (adhering to the NEMA standard Nu 2-2007) and an anthropomorphic torso phantom. Initial ¹⁶⁶Ho-only SPECT/CT scans were acquired with ¹⁶⁶Ho-filled tumor-like depositions (ranging in volume from 0.5 ml to 130 ml). Subsequently, dual-isotope scans were acquired by adding ^99mTc-pertechnetate to the water-filled background compartments. Phantom images were analyzed by comparing the recovered activities on SPECT/CT using a volume-of-interest (VOI) around ¹⁶⁶Ho depositions. Thereafter, both ¹⁶⁶Ho-only scans and dual-isotope scans (DI) were acquired in 3 patients, using the same acquisition parameters, reconstruction method and SPECT/CT camera. In the analysis of patient scans, 5 clinically significant compartments were defined: total scan, whole liver, basal lungs and 2 tumors. VOI’s of the 2 tumors were manually drawn based on their corresponding contrast enhanced CT images.

Results With low background concentrations of ^99mTc phantom studies showed a good correlation between the Ho-only and DI. As expected, a high ^99mTc-concentration caused contamination in the ¹⁶⁶Ho photopeak window due to downscatter, resulting in a less reliable recovery of ¹⁶⁶Ho activity (volume dependent). Smaller depositions of 1 ml could differ up to 75%, whereas the difference in large depositions up to 130 ml was just 7%. To maintain adequate image quality the administrated activity of ^99mTc-radiocolloid to patients after a ¹⁶⁶Ho-only scan was limited to 50 MBq. Recovered activity showed a difference between ¹⁶⁶Ho-only and DI of: total scan: -3.7%, +2.6%, -1.2%; whole liver: -5.1%, -1.8%, -2.8%; 1^st tumor: -3.4%, +15.5%, -4.1%; 2^nd tumor: -15.3%, +7.2%, 0.8%; and basal lungs -102.6%, +153.6%, +44.1% (greater deviation due to very low counts) for 3 patients respectively.

Conclusions A dual-isotope SPECT/CT protocol is feasible. Diagnostic quality is maintained and activity recovery is acceptable. This protocol can be used for automated image-based dosimetry during radioembolization work-up.