PCXMC – A Monte Carlo program for calculating patient doses in medical x-ray examinations (2nd Ed.)
Tapiovaara M and Siiskonen T
STUK – Radiation and Nuclear Safety Authority
Jokiniemenkuja 1, 01370 Vantaa, Finland
This report (STUK-A231) has been published as the second edition of ”Tapiovaara M, Lakkisto M and Servomaa A. PCXMC: A PC-based Monte Carlo program for calculating patient doses in medical x-ray examinations. Report STUK-A139, Helsinki: Finnish Centre for Radiation and Nuclear Safety, (1997)”.
X-ray diagnostics is a significant source of radiation exposure among the population. Therefore, it is important that x-ray examinations are conducted using techniques that keep the patients' exposure as low as possible but still compatible with the medical purposes of the examinations (ICRP 1996). In order to achieve this, it is necessary to understand the factors that affect the exposure and to be able to assess the patients' doses.
Patient dose has often been described by the patient’s entrance surface dose, which is measured on the patient’s skin at the centre of the x-ray beam. An alternative to this is to make the measurement free-in-air, without the contribution of the radiation that is backscattered from the patient, and express the result in terms of air kerma (incident air kerma, ICRU 2005). In some cases such simple measurements may be sufficient. This is the case, for example, in quality control measurements which concern the stability of equipment, and where the same x-ray exposure conditions are used in each measurement. However, the entrance surface dose is not sufficient for comparison or assessment of patients’ doses if the irradiation conditions (the size of the patient, the radiation quality, the exposed body-part, or other factors) are changed. In such cases, the patient dose needs to be characterised by quantities that are more directly related to the detriment caused by radiation (ICRU 2005). Optimisation of imaging techniques is an example of a case where the incident air kerma or the entrance surface dose are not sufficient for quantifying the patient dose, unless the local skin dose –not cancer induction– is the primary concern.
Presently, stochastic harm to humans from ionising radiation is assessed by the mean absorbed doses (or equivalent doses) in different organs or tissues in the body (ICRP 2007). For some purposes, the detriment can be assessed – and reported more simply – by referring to the effective dose: this enables one to express the dose as a single number, instead of a long list of doses in various organs and tissues. However, the use of the effective dose in describing the radiation exposure of patients is sometimes criticized and it is suggested that appropriate risk values for the individual tissues at risk should be used instead (ICRP 2007). Therefore, also the ability of assessing the risk of exposure-induced cancer has been built to PCXMC 2.0. The risk assessment is done according to the model of BEIR VII Committee (BEIR 2006).
Organ doses and the effective dose cannot be measured directly in patients undergoing x-ray examinations, and they are difficult and time consuming to obtain by experimental measurements using physical phantoms. However, they can be calculated to a reasonable approximation, provided that sufficient data on the x-ray examination technique are available. Today, such calculations are most often made using the Monte Carlo calculation method, where random numbers are used for simulating the transport of radiation in a complex medium, in this case the human body (for references on the Monte Carlo method in medical applications, see, for example, Andreo 1991). The physical interactions between radiation and matter are reasonably well-known, and the accuracy of the calculation is limited mainly by the accuracy of the anatomical model used to describe actual patients and by the characterisation of the applied radiation field (Zankl et al. 1989, Jones and Wall 1985).
Monte Carlo data on organ doses and the effective dose in general projection radiography of adults have been presented in tabular form in Rosenstein (1976a and 1976b), Rosenstein et al (1992), Jones and Wall (1985), Drexler et al. (1990), Hart et al. (1994a and 1994b) and Stern et al. (1995). Similar data for children have been presented in Rosenstein et al. (1979), Zankl et al. (1989) and Hart et al. (1996a and 1996b). Such data and the methods for obtaining them have been reviewed in ICRU Report 74 (2005). In addition to these tabulated conversion factors, several publications consider special cases of x-ray examinations and give dose conversion factors for them. In spite of the extensive tabulation of organ dose conversion factors in the above-mentioned references, not all x-ray projections or x-ray spectra of interest are covered and the data apply only for those individuals whose size and body-build correspond to the phantoms used in deriving the data.
STUK (Radiation and Nuclear Safety Authority, Finland) first published PCXMC (PC program for X-ray Monte Carlo) in 1997 (Tapiovaara et al 1997). This program allowed computation of organ doses for patients of different ages and sizes in freely adjustable x-ray projections and other examination conditions that are used in projection radiography and fluoroscopy. Since 1997, the program has been improved in several occasions. PCXMC versions 1.0–1.5 used slightly modified mathematical phantom models of Cristy (1980). In the present version (PCXMC 2.0) the phantom is still basically the same, but has been updated to the phantom models of Cristy and Eckerman (1987) with some further modifications (modification of the head, correction of some apparent errors in the data of Cristy and Eckerman and inclusion of some new organs: extrathoracic airways, oral mucosa, prostate and salivary glands). These modifications of the phantom enable the calculation of the effective dose using the tissue weighting factors introduced in ICRP Publication 103 (2007). The program is now also able to assess age- and sex-dependent radiation risks.
The Monte Carlo simulation time depends on the desired accuracy and on the speed of the PC, but takes typically from a few seconds to a few minutes with a PC with a 1.8 GHz processor. After having made the Monte Carlo calculation once, the user can calculate the organ doses for various x-ray spectra with a minimal computation time.