Evaluating the accuracy of SPECT/CT LSF estimations in SIRT therapies using Monte Carlo simulations with virtual 4D...

Recent publications advocate the use of SPECT/CT for LSF estimation for more accurate lung dose estimations in Selective Internal Radiation Therapy (SIRT) treatments. These studies have been limited to patient studies where the ground truth is not known or to a small number of phantom studies of a single body size. This research aimed to investigate the accuracy of LSF estimation derived using a Monte Carlo SPECT/CT model incorporating 4D virtual phantoms of varying BMIs.

Six 4D virtual XCAT phantoms were used to model patients of varying BMIs. SPECT acquisitions were simulated and assessed using the SIMIND Monte Carlo modelling framework. LSF estimates derived from SPECT imaging were evaluated and subsequently compared to corrected planar LSF estimates.

SPECT LSF estimates which incorporated scatter correction and with a CT based attenuation correction were shown to be the most accurate. Moreover, the SPECT LSF estimates were shown to be comparable with planar LSF estimation with incorporated scatter correction.

Given that the LSF estimate is a critical input in calculating the absorbed dose to the target volume in SIRT treatments, errors in its estimation can propagate through the dosimetric workflow. This work showed that both the SPECT/CT and the planar LSF estimation with scatter correction methods can significantly improve the LSF accuracy, over the current geometric mean method, and are easily implementable into the clinical workflow.