Biodi: an application for storage, comparison, and dosimetry of biodistribution experiments in radiopharmaceutical development

Introduction: Preclinical animal biodistribution studies performed by animal dissections are an essential step in radiopharmaceutical development. They allow to determine radiotracer tissue uptake (e.g tumor, major organs, blood), and to predict behavior in humans. Moreover, this data can be used to perform dosimetry calculations to ensure safety and to aid in the determination of optimal injected activities, which is particularly important for therapeutic radiopharmaceuticals. Here we present Biodi, an application aiming to automate the analysis of biodistribution studies, simplify/enable dosimetry calculations, and centralize the storage of multiple experiments in a database for simpler radiopharmaceutical comparisons and statistical analyses. The application is web-based, easy to deploy, and has been made open-source for access by researchers in the nuclear medicine community.

Methods: The Biodi framework is composed of two main components: 1) an application programming interface (API), and 2) a front-end user interface (UI). The framework communicates with an external database (e.g. PostgreSQL) and, if desired, an authentication server (LDAP). The API was developed using Flask, a Python web-framework, while the UI was developed using VueJs. A standard excel template that includes the experimental information such as injected activity, injection/euthanasia times, and tissue masses is filled by researchers when performing a study. The template, in combination with tissue radioactivity measurements exported from gamma counters, is used to calculate the %ID/g of uptake in each sample. For this, the API uses custom parsers to ingest CSV and XLSX files into the database. The database stores physical parameters of radioisotopes such as half-life, emissions, and calibration factors used for decay correction and %ID/g calculations. To perform dosimetry calculations, the results from the biodistributions studies are automatically selected. The user has the option of accepting either mono- or bi-exponential fits of the data, with the options of adding a trapezoid area calculation for the first time points and to select an uptake phase in the analysis. Thus, the application calculates the number of emissions per injected activity for each tissue. Mouse data is extrapolated to human data using ratios between tissue masses and whole-body masses between the MOBY phantom and human ICRP-89 phantoms. For the last step of the calculation, Biodi generates a case file that can be easily imported to OLINDA (Hermes Medical Solutions, Sweden) to obtain dose results.

Results: Biodi is an application that enables storage, comparison, and rapid analysis of biodistribution experiments. It allows users to upload, analyse, view, and export biodistribution and dosimetry results. The isotope configuration and calibration feature provides logging and quality control of gamma counter calibrations. Biodi creates interactive plots and data tables to summarize results and provides visualizations to the calculations to the number of disintegrations used for dosimetry. Biodi allows users to dynamically build queries using a drag and drop interface; a powerful feature for researchers to compare the efficacy of their radiopharmaceuticals. Analyzed and raw data are easily downloadable in a format that is compatible with multiple statistical software packages for more complex analysis. Biodi has been containerized using docker; allowing researchers to easily deploy it in their own laboratories. The source code and documentation of Biodi has been made publicly available at https://github.com/qurit/biodi.

Conclusions: Biodi is a novel web application that may help the nuclear medicine community in the development of novel radiopharmaceuticals. It allows for storage and automated analysis of biodistribution studies. Furthermore, it simplifies dosimetry calculations in rodents with the option of extrapolation to humans.