New automated methods for integrated quality control of 18F-FDG

1836 Objectives: Quality Control is the most labor-intensive part of the radiopharmaceutical production. Highly skilled operators perform multiple tests for every batch of radiopharmaceutical produced, some of these tests are highly subjective. Production facilities allocate considerable resources to remain compliant with cGMP guidelines. Here we present a set of tests that can be automatically performed on a standard laboratory hardware, with no operator intervention. Specifically, we focus on the incorporation of the bacterial endotoxin assay into the QC protocol. Methods: A multimode plate reader integrated with a pipetting robot was used as a hardware platform. The analysis was performed using modified 384 well plate and the reactions were performed in 96 well plates located on the deck of the pipetting robot. Bacterial endotoxin was analyzed using a stabilized LAL extract co-lyophilized with chromogenic reagents. Results: Endotoxin assay needed to be incorporated into the suite of test comprised of visual assessment, analysis of residual kryptofix, ethanol and acetonitrile as well as TLC analysis, half-life measurement and radioactivity concentration measurements. Since these assays are performed on the same platform they compete for both sensing and pipetting hardware, and cannot be performed at the optimal timing. For that reason, a compromise in individual tests performances had to be reached to unify all of them into one procedure. Specifically, Endotoxin assay had to be performed at suboptimal time and temperature, that led to loss of linearity. To circumvent these limitations, we have developed a statistical algorithm that allows for robust quantification of Endotoxin in the presence of these confounding factors. Conclusions: We successfully integrated the Bacterial Endotoxin assay into the comprehensive suite of QC test that are required for release of clinical doses of 18F-FDG. A mathematical algorithm, robust in presence of random light scattering events, temperature variability and pipetting imprecision, was developed to convert the experimental data into endotoxin concentrations. This allowed for Endotoxin assay performance to improve under suboptimal conditions established by the requirements of other assays in the suite.