Validation of clinical image quantification and dosimetry using 3D printed anthropomorphic phantoms

Abstract

Patient-specific image quantification and dosimetry are vital to optimise the delivery of molecular radiotherapy. This thesis describes the validation of image quantification using 3D printed patient-specific organ models and Monte Carlo simulations of radiation transport to calculate radiation dose. In order to calculate the radiation dose to an organ the curve describing the activity in an organ over time must be integrated. The impact of experimental noise on curve fitting in nuclear medicine is demonstrated using artificial sets of imaging data. A set of patient-specific 3D printed organ models, comprising the liver spleen and both kidneys, based on a diagnostic CT scan, was produced. Patient-specific calibration factors for 177Lu were calculated. These patient-specific calibration factors were compared to those previously reported for organ models based on the Cristy and Eckerman phantom series and a 113 ml sphere. Overestimations of recovered activity and hence dose of up to 135 % are shown. The calculation of calibration factors revealed a dependence on the position of the insert. The source of this position dependence was determined to result from the application attenuation correction. Users of reconstruction systems must ensure that the behaviour of attenuation correction on their system is understood. Patient-specific dosimetry calculations were compared to clinical tools for 177Lu and 131I. These calculations demonstrate the consistency of clinical methods for 177Lu if mass scaling is applied to the generic calculations. Inconsistencies in the calculations are present for 131I due to the greater contribution of gamma rays to the radiation dose. Summed exponential functions are often used to describe the activity in an organ over time. However these functions are fundamentally hard to fit to data points due to their mathematical structure. Artificial data sets representing measurements of the activity in organs over time were generated and curves fitted to them. Calculations of the area under these curves demonstrates the difficulty in fitting and calculation of uncertainty as applied to nuclear medicine. The work on anthropomorphic calibration factors is published in Physica Medica. A paper describing the examination of the positional dependence is published in Nuclear Medicine Communications.

Date of Award	31 Dec 2019
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Andrew Robinson (Supervisor) & David Cullen (Supervisor)