PROJECT SUMMARY
The effort of the proposed project is directed towards (1) resolving a controversy about two main approaches
for optimizing intensity-modulated proton therapy (IMPT), and (2) testing the validity of assumptions and approximations in the proton biological effect computation models. The current approach, used almost universally in the clinics, is to assume that relative biological effectiveness (RBE) of protons to be 1.1. In reality, RBE
is a complex variable function of dose, linear energy transfer (LET), tissue and cell type, endpoint and other
parameters. This fact is now being realized increasingly and there have been efforts to define IMPT optimization criteria in terms of variable RBE-weighted dose. RBE may be computed using one of many models, which
make questionable approximations and assumptions and the parameters of tissue response and their dependence on LET are uncertain. Consequently, the computed RBE and the results of IMPT optimization based on
RBE-weighted dose have corresponding uncertainties. Therefore, many researchers and clinical practitioners
are advocating an alternative approach in which IMPT optimization and evaluation are based on the criteria
that attempt to minimize LET in normal tissues and/or maximize it in the target volume while maintaining the
physical (or the constant RBE=1.1 weighted) dose distributions to be the same as, or similar to, the current
clinical approach. However, among the limitations of the LET-based approach is that biological effect does not
depend upon LET alone, and increasing or decreasing LET in a tumor or normal tissue would not correctly reflect its clinical consequences. While it is commonly acknowledged that current clinical approach is suboptimal, it is unclear which of the two alternative approaches would lead to safer and more effective IMPT treatments. In addition to this controversy, both approaches, explicitly or implicitly, employ average of LET contributions from all protons of highly disparate energies. We assert that this approximation tends to underestimate
the biological effect. Our hypothesis is that criteria for optimizing IMPT must incorporate variable RBE, taking
into account the non-linear dependence of RBE on LET spectra and the physical and biological uncertainties to
produce biologically effective dose distributions that would be safer and more effective compared to the LET-
based approach and, especially, the current clinical approach of using RBE of 1.1. We propose three specific
aims to test this hypothesis: (1) Investigate the pros and cons of optimization criteria based on LET vs. biologically effective dose computed using current models. (2) Explore the potential of biological effect optimization
(instead of optimization of biologically-effective dose) based on new or reparametrized models, including those
that eliminate LET averaging. (3) Investigate the impact of physical uncertainties on biological effects through
robust optimization incorporating physical uncertainties. The innovative optimization approaches and findings
from this project are aimed at immediate translation into clinical practice of IMPT to improve therapeutic ratio
and to facilitate clinical trials and research.