British Journal of Radiology (2004) 77, 86-87
© 2004 British Institute of Radiology
doi: 10.1259/bjr/25924254
Geometric uncertainties in radiotherapy
J M Wilkinson, MSc, FIPEM
North Western Medical Physics, Christie Hospital NHS Trust, Wilmslow Road, Withington, Manchester M20 4BX, UK
 |
Introduction
|
|---|
Over the last 25 years, a succession of committees reporting to the International Commission on Radiation Units and Measurements (ICRU) have developed a philosophy of radiotherapy treatment planning in which the target volumes take account of geometric uncertainties such as the extent of spread of subclinical disease, expected movements, expected variations in shape, and inaccuracies or variations in treatment set-up [1–3]. In the more recent reports, the extent of subclinical disease is accounted for by adding an appropriate margin to expand the gross tumour volume (GTV) into a clinical target volume (CTV), which, in turn, is converted into a planning target volume (PTV) by the addition of further margins. The PTV is used in the definition of field sizes and shapes as well as in choosing beam arrangements. The ICRU philosophy is widely accepted and, in the UK, adherence to this approach has effectively become mandatory as it is now included in the Manual of Cancer Service Standards [4].
Following publication of the most recent ICRU report in this series, the British Institute of Radiology (BIR) commissioned a Working Party with the remit:
- to review the sources of geometric uncertainty in radiotherapy and to show how to generate margins around target volumes and organs-at-risk that are large enough to accommodate the combined uncertainties; and
- to describe sources of uncertainty, including volume delineation, that must be considered in particular cancer sites, to indicate the sizes of these uncertainties where known, and to give practical advice that is either not available in the literature or is particular to protocols used in individual centres on how to determine those uncertainties.
The report of this Working Party has now been published [5] and the Honorary Editors of the British Journal of Radiology are of the opinion that the information contained in it is of sufficient importance to warrant this commentary. Continuity with the ICRU is ensured by the fact that Torsten Landberg, who chaired the three ICRU committees, and Jane Dobbs, who was involved with the last two, also served as members of the BIR Working Party, and they have written the first chapter giving a clinical overview of the issues involved.
|
Clinical overview
|
|---|
For the foreseeable future, both determining the size, shape and location of the GTV, and choosing the margins by which this will be expanded into a CTV, will remain clinical responsibilities. The method used for evaluation of GTV will be of critical importance, so careful consideration needs to be given as to what is best for any particular disease site. Increasingly for modern radiotherapy, this evaluation relies on some form of imaging and as the imaging armamentarium is continually developing and expanding, expert advice and guidance are required to optimally exploit what is available. Even with an optimal technique, interoperator variations in interpretation of the images will be inevitable, but the associated uncertainties may be reduced in magnitude by ensuring good liaison with the diagnostic radiology experts and ensuring good and detailed training.
Expanding the GTV into the CTV is, in the opinion of Dobbs and Landberg, the biggest source of geometric uncertainty in the whole process. Certainly there is currently very little useful information in the literature, and frequently it is necessary to rely on what is very loosely described as "clinical experience". There is scope here for future systematic study.
The next step in the process is expanding the CTV into the PTV and although this is more concerned with technical issues, the clinician must remain closely involved, not only because many of the movement problems are patient related, but also because there may have to be compromises to balance local control probability with the risk of treatment complications. Successively adding an adequate margin for each uncertainty independently will ensure, barring mistakes, that the whole of the CTV is, and remains, enclosed within the treated volume, but in the majority of cases this will also lead to a PTV that is too large to be compatible with radical radiotherapy. A more acceptable size of PTV is achieved, without significantly prejudicing disease control probability, by first adding the margin that covers the best estimate of subclinical disease extent, and then adding the margins for all other uncertainties in combination such that there is a high probability of the CTV remaining enclosed within the PTV. This approach recognizes that it is highly unlikely that all errors will act in the same direction, and it is highly likely that some errors will, in a particular case, be smaller than the magnitude of their uncertainties. However, even this approach will not necessarily give an acceptable PTV in every case, as the proximity of some critical structure may impose a need for a combined margin that is less than ideal, and in extreme cases even the first expansion of the GTV into the CTV may have to be compromised. Good clinical judgment will always be required in deciding whether or not to compromise, or how big a compromise is necessary.
|
Technical overview
|
|---|
The second chapter of the BIR Working Party report gives a technical overview of geometric uncertainties in radiotherapy. As working definitions, one may regard "uncertainty" as the number that comes after the ± sign in the quantitative description of some parameter or variable, and "error" as the difference between the actual value and what is assumed to be the true value. There should then be a known probability of the error lying somewhere within the range of uncertainty. When considering geometric uncertainties in radiotherapy, it is important to distinguish between systematic errors, which, for any particular patient, will become fixed at some stage in treatment preparation and then remain fixed, and treatment execution errors, specifically those due to daily set-up errors and random interfractional internal anatomical movement, which will vary. Breathing is not random, so breathing positional errors are better considered separately along with the systematic errors.
The next important distinction to make is between errors that occur in what can be approximated to a Gaussian or normal distribution, and those that cannot. Breathing positional errors fall into the latter category, as would beam algorithm errors from the treatment planning system. The uncertainties associated with Gaussian distributions may be expressed as standard deviations, added in quadrature and then multiplied by a weighting factor to obtain any given probability of encompassing the CTV in the resulting margin. For treatment preparation errors the weighting factor is a fixed value for any required probability, but for treatment execution errors the weighting factor varies with treatment technique. Values of these weighting factors for 90% probability are given in the Working Party report.
|
Practicalities and best practice
|
|---|
At first sight this approach may seem very complex, but it is also very logical, and the remaining chapters of the BIR report illustrate how it may be implemented in practice for a series of different disease sites and treatment techniques. As we strive to improve both the accuracy and precision of radiation therapy, and to exploit to the full new imaging modalities, three-dimensional (3D) treatment planning and developments such as intensity-modulated radiotherapy, we must ensure that we are working with the best possible input data for treatment planning, and specifically the best possible definition of the planning volumes. The process of reaching a decision on the appropriate size, shape and location of the GTV and determining the magnitudes of the various margins is not, and probably never will be, an exact science, but it can still be approached systematically and optimized for individual sites, with a justifiable expectation that, as a result, standards of radiotherapy will improve. Following publication of the most recent of the ICRU reports [3] there was a need for objective and practical advice on how to recognize the sources of geometric uncertainties and how to assess their magnitudes in different clinical contexts, how to reduce or eliminate them if at all possible and how to allow for the residue in the treatment planning process. This is exactly the advice that the BIR Working Party report provides.
Received for publication December 15, 2003.
Accepted for publication December 15, 2003.
|
References
|
|---|
- International Commission on Radiation Units and Measurements. ICRU 29. Dose specification for reporting external beam therapy with photons and electrons. Washington, DC: ICRU, 1978.
- International Commission on Radiation Units and Measurements. Prescribing, recording, and reporting photon beam therapy. ICRU Report 50, Bethesda, MD: ICRU, 1993.
- International Commission on Radiation Units and Measurements. ICRU Report 62, Prescribing, recording and reporting photon beam therapy (Supplement to ICRU Report 50). Bethesda, MD: ICRU, 1999.
- NHS Executive. Manual of cancer service standards. London, UK: Health Services Directorate, 2000:158.
- Bidmead M, Coffey M, Crellin A, Dobbs J, Driver D, Greener A, et al. Geometric uncertainties in radiotherapy: defining the target volume. London, UK: British Institute of Radiology, 2003.
This article has been cited by other articles:
|
|
|
|
 
BJR Review of the Year - 2004
Br. J. Radiol.,
March 1, 2005;
78(927):
181 - 185.
[Full Text]
[PDF]
|
|
|
Top
Introduction
Clinical overview
