Comparison of Two Novel Fully-Predictive Approaches to Compute NTCP Curves for Radiation-Induced Myelopathy In Rats Exposed to H, He, C, and O Ions
Abstract
Purpose
Toxicities to organs remain a critical limitation to dose escalation in radiotherapy. Ion radiotherapy reduces the absorbed dose to normal tissue but is characterized by a higher relative biological effectiveness (RBE). Existing clinically relevant methodologies have limitations in the computation of normal tissue complication probability (NTCP) dose-response curves for ion beams and frequently require ion-derived data as input. We present and compare two fully predictive approaches to compute complete NTCP curves for serial organs after ion exposure.
Methods
We applied the two approaches to the largest available single-endpoint in vivo dataset of rat spinal cord myelopathy, simulating exposures at multiple depths along spread-out Bragg peaks (SOBPs) of protons, helium, carbon, and oxygen ions. Radiation transport simulations provided microdosimetric descriptors for subcellular energy deposition. The first approach integrates the linear-quadratic model (LQM), the critical-volume NTCP model, histological data, a novel two-stage analysis of fractionated photon NTCP data, and a microdosimetry-based RBE model. The NTCP calculations considered glial and endothelial cells as targets. The second, simpler approach does not include an NTCP model nor histological information, but combines the in vivo photon LQM dose–response with ion- and position-specific, dose-dependent RBE values to derive photon-equivalent dose. Importantly, both approaches are fully predictive in that the ion NTCP curves were computed without ion data calibration, but solely analyzing photon data.
Results
The fit of the photon data produced α/β = 2.3±0.6 Gy for 15 MV X-rays. Without any ion-based calibration, the predicted NTCP curves and RBE values showed overall good agreement with published animal data across fractionation regimens, ion species, and SOBP positions.
Conclusion
This fully predictive framework, encompassing two different mathematical approaches, enables the computation of radiation-induced NTCP curves for myelopathy in rats from photon data alone. Future work will focus on validation for additional in vivo endpoints and clinical data.