Improving dosimetry for Selective Internal Radiotherapy with 90Y microspheres

Abstract

90Y Selective internal radiotherapy (SIRT) treatments are commonly used in the treatment of non-operable cancerous tumours in the liver. Disease is treated through both embolisation of blood vessels and the deposition of dose from the beta radiation produced in the decay of 90Y. Current best practice is to quantify the dose delivered in molecular radiotherapy (MRT) treatments. This work addresses gaps in knowledge for the dosimetry calculations of SIRT therapies in three areas: (1) The validity of Monte Carlo (MC) radiation transport codes for predicting the bremsstrahlung photon spectrum produced by the emitted beta radiation was investigated. This is essential if MC is used to make predictions about SPECT imaging of SIRT therapies. The transport and detection of this radiation was investigated using both SPECT and HPGe detector systems. It was determinedthat whilst MC codes are faithful in their description of gamma-ray transport, they areunable to accurately model bremsstrahlung photon production. A photon source was developed utilising the observed disagreement between simulated and experimental bremsstrahlung. The source proved superior for SPECT acquisition modelling than native bremsstrahlung but fell short of the accuracy observed with true gamma-ray emitters. (2) An investigation of the dose delivered by SIRsphere therapies was undertaken at a cellular length scale. A MC generator based on histological data was developed to produce realistic distributions of microspheres in different sized tumours. These were used as an input to a series of GATE simulations to quantify the dose distribution to the tumour and the healthy liver. The dose was found to strongly preferentially accumulate near the tumour surface, with a large degree of inhomogeneity throughout the tumour. (3) List mode acquisition was used to compare the quantitative accuracy of 90Y SPECT images using different energy windows. A patient-like 3D printed phantom scan was performed using a GE Discovery 670 with MEGP collimators. The data were sorted into 36 different effective emission window sets and reconstructed on a Xeleris workstation. A figure of merit (FoM) was defined to measure the quantitative accuracy of the reconstructed images. Trends across the bremsstrahlung spectrum were investigated in addition to the effect of varying the widths of window. The best quantification was found for an emission window from 50 to 60 keV.

Date of Award	31 Dec 2017
Original language	English
Awarding Institution	The University of Manchester
Supervisor	David Cullen (Supervisor) & Jill Tipping (Supervisor)