In vivo dosimetry: implementation issues

While the value of IVD as a routine clinical procedure has been the subject of much debate and many meetings, far less attention has been paid to another key issue: the physical limitations of the detectors used for this dosimetry. Writing in the British Journal of Radiology, radiotherapy physicist Craig Edwards from the UK's University Hospital of North Staffordshire takes a closer look at the characteristics of such detectors and how these may affect the implementation of IVD (Br. J. Radiol. doi: 10.1259/bjr/32563777).¹

"I would say that the number of centres doing IVD hasn't really changed much since a survey paper that I wrote in 2007," said Edwards.² "But to be honest, I can't really see an argument for not doing IVD. In France, they recently issued a report making IVD mandatory for every patient. It will be interesting to see how they get on in the next few years."

Caution required

Edwards notes that in the UK, just under half of all radiotherapy centres perform central-axis IVD. Two thirds of centres, however, measure critical-organ dose outside the main treatment field, where the X-ray spectrum differs from that at the (calibrated) central-axis position. Unfortunately, the detectors most commonly employed for IVD – semiconductor diodes and thermoluminescent dosimeters (TLDs) – both exhibit a non-uniform energy response. As such, it's crucial that detectors are also evaluated outside the main treatment field to enable correction for this energy dependence.

A previous study by Edwards and his colleagues revealed that, compared with its central-axis response, a diode over-estimated the X-ray dose by 10–70% at distances of 1–10 cm from the edge of an X-ray field (ranging in size from 4 x 4 to 15 x 15 cm). The TLD, on the other hand, produced more reliable dose measurements outside the main field, with a response ratio between central-axis and outlying distances of around unity.³

Based solely upon its photon energy characteristics, the TLD thus appears the obvious choice for critical-organ IVD outside the main field. Unfortunately, says Edwards, things aren't that simple. Interactions between the X-ray beam and components in the linac head generate contaminating electrons, which also contribute to the overall dose. The spectra incident on any detector will thus be a mixture of X-ray photons and electrons, the relative contributions of which differ inside and outside the treatment field.

For a 6 MV X-ray beam, for example, the electron contribution to total diode response at the central-axis is estimated as about 7–8% (for field sizes from 4 x 4 to 15 x 15 cm). Outside the main field, this contribution increases with distance from the field edge and decreasing field size, reaching 58% at 10 cm from the edge of a 4 x 4 cm X-ray field.⁴ Such results reveal the large impact that electrons can have upon detector response. Dose measurements outside the field must therefore be interpreted with great caution.

The TLD has a similar response variation to electron energy as the diode, so the contaminating electron dose should be comparable. Edwards notes, however, that as the TLD response is less dependent on incident photon energy, thus reducing the photon dose component of its overall response, it could be more affected by contaminating electrons than the diode – possibly even under-responding outside of the main field. "Repeating the diode experiments with TLDs is something that I'm hoping to finish off soon," he told medicalphysicsweb.

For central-axis IVD, Edwards recommends use of a semiconductor diode, since it has no real energy-dependence problems and can provide instant dose measurements, while accurate TLD measurements take some time to produce. "But I remain on the fence with regard to recommending one system over the other for critical-organ IVD outside the main field," he said.

Take it forward

Edwards and colleagues conclude that there's a definite need for further studies of the effects of X-ray energy and electron contamination on detector response outside the main field. Such work could lead to the introduction of guidelines as to the relative X-ray and electron contribution to the total detector response and dose.

"There is one final piece of the jigsaw that hasn't been put in place yet," he added. "That is the creation of a position-dependent, calibration matrix that can be used to correct an IVD system for spectra variations when measuring critical-organ dose outside the main field."

He also points out the importance of developing similar guidelines for advanced delivery techniques like intensity-modulated radiotherapy (IMRT), which will alter the central-axis spectra. The team is planning to investigate the impact of IMRT on the response of an IVD system, using Monte Carlo methods. Subsequently, if required, the researchers hope to develop methods to improve the accuracy of IVD in such situations.

"On a national level, I think it may be time to be more explicit in stating that all departments should be doing IVD and put a time frame on its implementation," said Edwards. "While it's true that the number of mistakes that this type of system will be picking up is, or should be, getting less and less with the advent of automated record-and-verify systems etc, people seem to forget that we are here for the patients. I feel that IVD systems provide a type of comfort blanket for the patient – and you can't really put a price on peace-of-mind."