Uncertainty analysis for absorbed dose of a brain receptor imaging agent
Radiation absorbed dose is calculated in order to predict the risk in a nuclear medicine procedure. However, it is not possible to calculate the true absorbed dose a patient would receive as a result of the exposure to ionizing radiations. Although absorbed dose estimates are known to contain uncertainty, no rigorous work which investigated the uncertainty associated with absorbed dose estimates existed before this study.
A method of uncertainty analysis for absorbed dose calculations was developed and calculations were performed for a brain imaging agent 123I-IPT. The two major sources of uncertainty considered in this research were the inter-patient organ mass variabilities and uncertainty associated with the determination of residence times. The latter included the uncertainty associated with inter-patient biokinetic variabilities and the uncertainty associated with the least squares parameter estimation procedure. In this research the classical approach of statistics was employed.
Absorbed dose uncertainties for a seven patient population and for each patient sepaerately were determined by the use of a "Latin Hypercube Sampling" method. The upper bound of the 95% confidence interval of the dose distribution for an individual was found to be 2.5 times larger than the estimated mean absorbed dose. Neglecting the variability in organ masses, the upper bound of the 95% confidence interval of the dose distribution was about 45% more than the estimated population mean. Incorporation of the inter-patient organ mass variabilities, without including correlations to absorbed fraction, into uncertainty calculations almost doubled the range of the 95% confidence intervals of the population dose distributions. For instance, the 95% confidence interval of the population liver dose distribution was between 5.6E-05 rad/μCi and 1.7E-04 rad/&me;Ci with a mean of 9.3E-05 rad/μCi. This study concluded that a new patient who is similar to the study group might receive an absorbed dose of twice as large as the estimated population mean due to the considered uncertainties.
Although there are several cases where diagnostic imaging may involve risk, in general, it is believed that diagnostic imaging entails little or no risk. Therefore, the determination of uncertainty is potentially more important in therapy applications of radiopharmaceuticals due to the much larger absorbed dose. Uncertainty information for a patient population can be used to predict the maximum dose (with the desired confidence such as 90% or 95%) a new patient who is similar to the study group might receive. The physician can then include the uncertainty information in his risk/benefit evaluation in radiation therapy. The physician might be aided in estimating potential patient radiation doses by the construction of absorbed dose data bases which also include absorbed dose uncertainties for the frequently used therapeutic agents in nuclear medicine. The method established in this work can be used in order to construct such absorbed dose data bases. Future studies should include correlation of absorbed dose with mass variabilities.
Thesis96.A95.pdf
13.19 MB
Adobe PDF
7eb2f5f932fcbd9c9b6990276707959c