Author: Farhad Jafari, Brandon Hai Nguyen 👨🔬
Affiliation: Department of Radiation Oncology, University of Minnesota Medical School, Department of Radiology, University of Minnesota 🌍
Purpose: To validate a bottom-up approach to dosimetry in nuclear medicine.
Methods: A pharmacokinetic model is developed to represent the flow of activity between different compartments of the body. Using a Python script, synthetic activities are generated for multiple compartments over a short time using random rate constants. An optimization function is then applied to the synthetic activity data to reconstruct the original rate constants, showing the retrieval of the rate constants from brief patient activity measurements. These rate constants are then used to determine the exact longitudinal activity curves for each organ and each injected radionuclide. The least-squares errors are calculated for the reconstructed rate constants and activity.
Results: In over a hundred simulations of activities within a ten-minute time span and reconstructing of the rate constants, only five simulations had a least-squares error for activity greater than 10-8 . Removing these outliers give an average least-squares error of 2.68*10-10. Furthermore, the algorithm was successful in fitting activity values over uniform and nonuniform time intervals from 5 minutes up to 200 minutes long.
Conclusion: With the proliferation of theranostics in nuclear medicine, the development of an exact method to determine organ dose has become essential to patient treatment. This bottom-up approach will provide improved dose estimations. The accuracy and precision of the developed method makes it clinically practical.