Dose computations for three-dimensional radiation treatment planning

The knowledge of radiation dose at all points of interest within a patient's body is essential for the evaluation of treatment plans and for the judicious selection of the best one from among a set of competing plans. This paper discusses the physical and geometric aspects of radiation dose distribution calculations necessary for three-dimensional treatment planning. A practical method of obtaining dose distributions for a given arrangement of beams and beam modifiers is with the aid of computer calculations based on mathematical formalisms derived from the principles of physics. The most accurate method of computing dose is by Monte Carlo stimulation of radiation transport. However, this method is extremely CPU time intensive and too complicated to implement for routine radiation treatment planning. Approximate semi-empirical methods have to be employed to compute dose distributions within reasonably short times. Conventional approximate methods do not yield results of adequate accuracy. Newer methods that are better able to take into account the three-dimensional nature of radiation transport have been introduced but require considerably greater computing capacity. Furthermore, computing power needs for dose calculations for three-dimensional treatment planning are one to two orders of magnitude greater than the corresponding needs of two-dimensional treatment planning. However, the concern for greater computing power requirements is being lessened somewhat with the advent of high speed computers and specialized hardware, and with the development of clever algorithms for speeding up dose calculations. For three-dimensional treatment planning some geometric issues, such as the calculation of dose at points in three-dimensional space, the incorporation of three-dimensional patient and beam geometry into a system for dose calculations, the choice of optimum grid spacing, radiological pathlength calculations, and algorithms for speeding up dose calculation, take on added importance compared to the corresponding issues for conventional two-dimensional planning. These issues are also discussed here briefly.