Oncology Publications

Title

Experience-driven dose-volume histogram maps of NTCP risk as an aid for radiation treatment plan selection and optimization.

Document Type

Article

Publication Date

1-2008

Journal

Medical Physics

Volume

35

Issue

1

First Page

333

Last Page

343

URL with Digital Object Identifier

http://dx.doi.org/10.1118/1.2815943

Abstract

Commonly, the quality of treatment plans is judged by a dose-volume histogram (DVH) in regards to satisfying a series of dose-volume constraints. This paper presents a novel technique for mapping normal tissue complication probabilities (NTCP) onto regions of dose-volume space with statistical considerations of risk. Mapping is done for DVHs specific to one treatment technique for one disease site. In this study, the method is illustrated for simplified intensity modulated arc therapy of the prostate, and the resulting NTCP values apply to complications in the rectum. The method consists of implementing a Monte Carlo algorithm that creates a large set of DVH curves by simulating random walks through dose-volume space. The walks are guided by a base set of clinical DVHs. Grid points in the dose-volume space have an associated NTCP spectrum for curves passing above right of the grid point of interest. After a DVH is simulated and the NTCP estimate calculated using the Lyman model, dose-volume points located to the bottom left of the curve are scored with this NTCP value and contributed to the spectrum of each point. A NTCP tolerance value is then selected and the risk of violating this tolerance is identified by a gray-scale map in regions of dose-volume space. The generated maps distinguish technique-specific, high-risk regions, a feature which is advantageous over fixed single-point dose-volume constraints commonly used. The maps also provide a visualization tool to help select safe and robust treatment plans and open the possibility for improving the efficiency of biologically based plan optimization by focusing on the more critical sections of DVH curves.