Abstracts 2006
Abstract Detail


« back


2nd ISRTRD - Physics 1: Quantitative Analysis & Treatment Planning
Sunday  October 1, 2006  11h - 13h
Room: MC 2

96	Development and validation of an organ residence time estimation method for high dose Y-90 ibritumomab tiuxetan therapy E. C. Frey, B. He, G. Sgouros, I. W. Flinn, R. L. Wahl; Johns Hopkins University, Baltimore, MD, UNITED STATES.

Aim: We are currently engaged in a dose escalation trial to find the maximum tolerated dose (MTD) for a myeloablative Y-90 ibritumomab tiuxetan therapy regimen with autologous stem cell transplant. In this study, the escalation variable is dose to critical organs. Thus, accurate estimation of organ residence times is essential. The goal of this work was to develop and validate a method for estimating organ residence time of the In-111 labeled antibody for use in this study.Materials and Methods: In the method, anterior/posterior planar scans are acquired at 5 time points plus abdominal and thoracic SPECT/CT acquisitions at 24 hrs post injection. SPECT images were reconstructed using a quantitative SPECT (Q-SPECT) reconstruction method based on iterative reconstruction including compensation for attenuation, scatter, and the collimator-detector response. Planar scans were processed using scatter subtraction and geometric mean methods. Organ volumes-of-interest were defined using the registered CT and SPECT images and regions of interest were manually defined on the planar scans. Organ time-activity curves (TACs) were estimated from the planar scans and, for the hybrid method, rescaled to pass through the organ activity estimated from the Q-SPECT images. The residence times were estimated by integrating exponential functions fitted to the TACs. The method was evaluated using physical phantom experiments and Monte Carlo simulation studies by comparing accuracies of estimated residence times for the liver, the dose-limiting organ in all patients to-date. The planar and hybrid methods were also compared using data from 12 patient studies.Results: The errors in the 24 hr estimates of liver activity were 1.3% and 13% for the Q-SPECT and planar methods, respectively. The planar method resulted in a 19% underestimation in the liver compared to 8% for the hybrid method. In the patient studies, the difference between the planar and hybrid estimates of the residence times was 11% +/- 14%. Of importance, the variation in the accuracy over the patient population was larger than the average difference. Large variations in accuracy of dose estimates could result in large variations in the administered dose compared to the true dose, and thus errors in estimating the MTD.Conclusions: We conclude that, the hybrid planar/SPECT estimation method provided a substantial improvement compared to residence time estimates based on planar images. This methodology has been applied in the ongoing dose escalation study where a 28 Gy dose to the liver has been administered to three patients without adverse reactions.

« back