Theranostic Digital Twins Framework for Virtual Theranostic Trials
Abstract
Purpose
Theranostic digital twins (TDTs) are virtual representations of individual patients, that integrate clinical patient history, imaging, and physiologically based pharmacokinetic (PBPK) modeling to simulate radiopharmaceutical therapies (RPTs). Their validation is challenging due to limited datasets containing pre- and post-treatment imaging, and outcomes. To address this, we developed and validated an automated pipeline that generates synthetic theranostic imaging data from CTs of real patients.
Methods
A cohort of 10 patients with prostate cancer was used in this preliminary study. For each patient CT, organs were automatically segmented using Total-Segmentator. A prostate specific membrane antigen (PSMA) PBPK model generated organ-specific biodistribution time courses, which were mapped to voxelized activity distributions on each segmented organ across multiple frames. SPECT projections were simulated using SIMIND Monte Carlo with clinically relevant acquisition settings and subsequently reconstructed using PyTomography. Organ uptake from reconstructed images was compared with PBPK generated time activity curves (TACs) andreconstructed TACs were additionally fit with a bi-exponential model to assess kinetic behavior.
Results
Reconstructed SPECT organ activities were consistent with simulated ground truth (errors −43% to +23%), within expected clinical uncertainties. Derived pharmacokinetic parameters such as organ washout decay rateagreed well with ground truth model, supporting the use of simulated data to optimize time–activity curve fitting and imaging schedules for dosimetry.
Conclusion
We presented a CT-driven TDT pipeline that generates realistic SPECT time series from patient anatomy and PBPK-modeled kinetics, enabling quantitative validation through organ TAC agreement and kinetic-parameter comparison. Ongoing work will incorporate synthetic lesions, expanded validation and dosimetry to enable scalable virtual theranostic trials.