| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Full paper |
1 Regional Medical Physics Department, Newcastle General Hospital, Newcastle upon Tyne NE4 6BE , 2 Regional Medical Physics Department, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4PP, UK
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Numerous workers have reported estimates of second cancer incidence following radiotherapy [3]. This is likely to remain an important issue as, following improvements in cancer treatment, more patients may survive longer than the latent period for expression of the second cancer. In addition, several new developments in radiotherapy may lead to higher doses to organs and tissues throughout the body. Intensity-modulated radiotherapy (IMRT) may lead to substantial increases in whole body doses, because of the longer beam-on times and the consequent increased exposure to leakage radiation. The use of an increased number of fields results in a larger volume of tissue exposed to relatively low doses, although the target volume may be smaller. Compared with conventional radiotherapy, Hall and Wuu have estimated that IMRT may almost double the incidence of second cancers [4]. New imaging techniques are also contributing to verification of treatment field positions during the course of radiotherapy. Although megavoltage portal imaging is a long-established technique, image-guided radiotherapy (IGRT), involving CT studies at many fractions [5], cone beam CT using megavoltage and kilovoltage X-ray energies [6, 7] and integrated systems such as tomotherapy [8], are also emerging as viable techniques for improving the accuracy of dose delivery to the target volume and may increase organ doses and subsequent risks. In contrast, the potential for reduction of second cancers by using proton radiotherapy has been investigated by Miralbell et al [9].
Previous estimates of second cancer risk have necessarily made simplifying assumptions about doses and risks. Followill et al [10] estimated a single whole body dose at a point distant from the treatment field and applied a risk factor for the general population. Mutic and Low [11] measured doses outside the target volume using thermoluminescent dosimetry (TLD) and a block phantom. Verellen et al [12] used personal dosemeters to measure the personal dose equivalent Hp(10) for six patient treatments, related this to the effective dose and used a nominal risk factor for the general population. More recently, Kry et al [13] have used NCRP risk factors [14] to compare several IMRT treatments of the prostate. The purpose of these papers was to compare different irradiation techniques (e.g. conventional radiotherapy versus IMRT), thus placing the emphasis on relative, rather than absolute estimates of risk.
In other fields, diverse exposures and irradiation patterns are usually combined using the quantity effective dose. However, there are several problems associated with using this concept in the radiotherapy context [15]. ICRP risk and tissue weighting factors refer to the general population, whereas the age distribution of radiotherapy patients will be skewed towards the higher ages, implying lower risk estimates. Conversely, tissue weighting factors refer to low doses associated with occupational exposure and a dose and dose rate effectiveness factor (DDREF) of 2 has been applied. It is not clear what modifications to the DDREF should be made for the effects of radiotherapy fractionation schedules and cell kill on induced cancer probability at high doses.
Thus, the estimation of second cancer induction probability from a combination of weighted organ doses is not trivial and no attempt has been made in this paper to combine individual organ doses, which include a radiotherapy component, to form an overall measure of risk. Nevertheless, as a first step, organ doses measured in several critical organs in a realistic anthropomorphic phantom may be useful for justification purposes and will help to place concomitant doses in perspective, particularly for developing and justifying techniques where more extensive imaging is used to improve the accuracy of dose delivery.
This paper describes the experimental simulation of radical prostate radiotherapy together with associated CT scanning and portal imaging, with doses measured using lithium fluoride TLD chips loaded in a male anthropomorphic phantom. In anticipation of eventually calculating second cancer induction probabilities, doses were measured in organs (apart from the skin) to which ICRP has ascribed tissue weighting factors [1]. Skin was omitted because of the uncertainties arising from the dose gradient at the surface due to build up effects at megavoltage energies, the identification of the effective point of measurement of a TLD chip under these conditions and the large number of surface measurements which would in any case be necessary in order to form an unambiguous estimate of total skin dose. Furthermore, the skin has the lowest fatal cancer probability coefficient of all the main organs identified by ICRP and is not as critical as other tissues when considering second cancers following radiotherapy.
External beam radiotherapy of the prostate was chosen for study because it is the most common cancer in UK men and advances in treatment have shown an increase in 10 year survival from 20% in 1971 to 50% in 2001 [16]. Brenner et al [17], in a comparison of second cancers in patients treated for prostate carcinoma by radiotherapy compared with surgery, have shown that there is a small but significant increase in solid tumours from 5 years following radiotherapy, principally showing as carcinomas of the bladder, rectum and lung, and sarcomas in or near the treatment field. A further study has also shown a significant increase in second rectal cancers [18].
Measurement of organ doses from CT scanning and portal imaging allows doses to be calculated for any number and combination of these imaging techniques, thus facilitating the estimation of the organ dose contribution from image-intensive IGRT.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
The simulated prostate treatment consists of two phases. In phase 1, 64 Gy is delivered in 32 fractions to the planning target volume (PTV), followed in phase 2 by 10 Gy in 5 fractions to a reduced target volume. The prostate, bladder and rectum were identified on slabs 32, 33 and 34 of the phantom, following consideration of a series of patient CT scans, and a three-field plan for both phases developed as shown in Figure 1
. Field sizes and gantry angles for both treatment phases are given in Table 1. Anteroposterior (AP) and lateral digitally reconstructed radiographs were also generated (Figure 2).
|
|
|
shows the numbers of TLD chips contributing to the dose measurement for each organ or tissue.
|
Neutron dose contribution
At 15 MV, photoneutron production in the linear accelerator head will contribute to the organ and tissue doses. Measurement of individual organ doses due to neutrons was not attempted, but instead calculated from previously reported measurements of neutron fluence. Expressions for direct, thermal and scattered neutrons from medical linear accelerators, derived from Monte Carlo calculations by McCall [19, 20] were used to calculate the neutron fluence at the centre point of each slice of the RANDO phantom. They are reproduced in Equation (1) where the three terms give direct, scattered and thermal neutron components, respectively. The total neutron fluence per photon dose at the isocentre (
), is given by:
|
|
where a is the transmission factor for neutrons in head shielding (taken as 0.85); Q is the neutron source strength per unit photon dose at the isocentre for a Siemens PRIMUS accelerator (taken as 0.21x1012 neutrons Gy–1) [21]; d is the distance from the source; and S is the treatment room surface area (180 m2).
Conversion of to equivalent dose per Gy at the isocentre used the fluence-to-dose conversion factors given by NCRP [14], assuming a mean neutron energy of 0.5 MeV for an accelerator running at 15 MV [22, 23] and taking values of radiation weighting factor wR of 20 and 5 for fast and thermal neutrons, respectively. Although the ionization suffered by lithium fluoride TLD will include contributions from (n,p) and (n,
) reactions, this has been ignored in practice because the wR values ascribed to the neutron absorbed dose are significantly higher than unity.
CT scanning for localization
Using the same TLD loading scheme as for the radiotherapy treatment, the phantom was scanned using the standard departmental protocol for prostate localization, using a Siemens Emotion Duo scanner (Siemens Medical Solutions, Erlangen, Germany). The irradiation consisted of an AP topogram (130 kV; 30 mAs; topogram length 512 mm) followed by an axial scan (130 kV; 105 mAs; slice width 5 mm; pitch 1.5; 33 slices). Ten identical consecutive scans (topogram and axial) of the phantom were carried out in order to ensure that the doses at the TLD measurement positions were well above the minimum detectable value. Measured doses were subsequently scaled to give the doses received for a single localization scan. (Doses from conventional simulation were not measured, since a simulator is not routinely used in this centre for prostate localization).
Portal imaging
Using the same TLD loading scheme as for the radiotherapy treatment, the phantom was irradiated by a 15 cmx15 cm AP portal image exposure, as used in conjunction with the Siemens portalvision electronic portal imaging system (Siemens Medical Solutions, Erlangen, Germany) mounted on the same linear accelerator as used for treatment, but running at 6 MV. Typically, a clinical portal image would be acquired using 3 monitor units (MU). For simulation, 200 MU were used in order to ensure a measurable dose to distant organs and the subsequent doses scaled to give the commensurate clinical exposure for a single portal image. A 10 cmx10 cm lateral portal image exposure was also simulated in the same way for comparison.
TLD calibration and dose measurement
Following irradiation, chips were left for a fixed time of 48 h to allow the same fading time for all irradiations and read out in a Harshaw 5500 TLD reader (Thermo Electron Corporation, Solon, OH).
Individual chip calibration factors were established based on irradiation of each chip to a known dose using a Cs-137 calibration source. These factors provide a correction for interchip variability within a batch. In addition, batches of chips were irradiated within the range of doses and energies used in the simulations in order to generate a calibration curve. Doses were measured with ionization chambers and electrometers whose calibrations were traceable to national standards.
Dose calculation
Raw data from each chip readout were modified by the individual chip calibration factor and the dose calibration factor for the appropriate photon energy, giving the result as absorbed dose to water at the chip position. Organ doses were then calculated as weighted averages over all the TLD measurement points, using the mass fraction of each organ within each slab [24]. For lung, bone surface and bone marrow, the mass fractions given by Huda [25] were used. Doses to bone marrow and bone surface were calculated from the average of measurements to the positions of these organs within each slab and also, for comparison, from the average dose over the whole slab, in both cases using the same mass fractions. The total doses to these organs were obtained by summing the dose contributions for each slab. All organs, except for bone surfaces, were assumed to be water-equivalent. Doses to bone were derived from doses to water by application of mass stopping power ratios (bone/water) of 0.91 and 0.93 for 6 MV and 15 MV, respectively, and for CT energies a mass energy absorption coefficient ratio (bone/water) of 3.1 was used. Chips which had received doses >2000 mGy were not re-used.
Calculated neutron doses for each organ were added to the photon contributions for the simulation of treatment at 15 MV.
The current protocol in use in this centre defines the maximum number of concomitant images which can be acquired under a single authorization by the Practitioner, as defined under UK legislation [2]. It allows for the following: (i) three localization episodes on the CT scanner; (ii) three pre-treatment verification episodes, at least one on the CT scanner and one on the treatment unit; and (iii) eight verification episodes during the course of treatment. A localization or verification episode on a CT scanner consists of a maximum of three topograms, two CT scans up to 120 cm in length and three single axial slices. A pre-treatment verification episode on the treatment unit consists of a maximum of four portal images. A verification episode during treatment delivery consists of a pre-treatment verification episode and a localization or verification episode on the CT scanner. Several combinations of CT scans and portal images are possible within this framework and organ doses for two possibilities were calculated: (i) 10 CT scans and 36 portal images, where portal imaging is the primary method of verification; and (ii) 26 CT scans and 4 portal images, where CT scanning is the primary method of verification.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Neutron results
Figure 3 shows the neutron dose (in mSv Gy–1 at the isocentre) as a function of distance from the isocentre normal to the central axis, i.e. along the long axis of the phantom. This enabled a representative neutron dose to be calculated for each slab. The decrease of dose with distance is simply a consequence of an inverse square term acting on the direct neutron component and involving distance from the X-ray target.
|
gives organ and tissue doses for the complete radiotherapy treatment, single CT localization scan and topogram, and single AP portal image. Organ doses resulting from a lateral 10 cmx10 cm portal exposure were less than those arising from the 15 cmx15 cm AP exposure and the latter have been adopted as representative of prostate portal images. Doses are given for all organs for which tissue weighting factors have been identified [1], apart from the skin, and include remainder organs. Apart from muscle, bone marrow and bone surface, ranges of doses are also given for organs containing more than 5 TLD measurement points, or where mean organ doses are >0.1 mSv. Although the dose to the rectum is included within the estimate for the colon, the rectal dose has been calculated separately, because of its proximity to the prostate, and is included in the footnote to Table 3. In fact, part of the rectum and bladder were included in the target volume and received measured doses in excess of 60 Gy and 73 Gy, respectively. The neutron component of organ doses ranges from 19–50 mSv for the whole treatment.
|
Table 4 gives the total organ and tissue doses for a complete course of radiotherapy involving a total of 10 CT scans and 36 portal images. The data are also shown graphically as total doses (Figure 4a) and as percentages of the total doses from radiotherapy, CT scanning and portal imaging for each organ (Figure 4b). Table 5 gives organ doses for 26 CT scans and 4 portal images. Figure 5 shows total doses and doses as percentages of the total organ dose, respectively. A subset of these data have previously been reported in preliminary form [26].
|
|
|
|
, which shows a plot of mean dose for each slab.
|
shows the difference between the delivered and measured doses at 11 points as a fraction of the delivered dose.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
. In addition, there are uncertainties associated with the location of organs and the choice of sampling points within them. These are difficult to quantify. Finally, measurements on a single size and shape of anthropomorphic phantom will not represent the interpatient variation seen in practice. Estimating doses to bone surfaces and red bone marrow is particularly difficult because of the extended and complex distribution of these tissues throughout the body. Although dosemeters were placed close to selected bone locations, the uncertainties in dose estimation for marrow and surfaces are likely to be greater than for more well-defined organs. As pointed out by Golikov and Nikitin [24], the mass fractions for bone given by Huda and Sandison [25] and used here, refer to the mass distribution of the whole skeleton, which may differ from the corresponding distribution of bone surface cells. Doses to bone surfaces and marrow are also some of the higher organ doses measured. This emphasises the potential difficulty of using this technique as a reliable basis for absolute risk assessment and suggests that the technique may be more appropriate for the calculation of relative risks, for example between two different treatment modalities or between treatment and imaging exposures, where dose measurement locations and assumptions about organ position and extent are invariant.
For regions in or near the target volume, TLD measurement is likely to be considerably less accurate than calculation using the well-developed algorithms used in current treatment planning systems in combination with accurate ionization chamber measurements of dose distributions in water. Hence, for organs and tissues close to the target volume, use of the treatment planning system is probably the best method for organ dose calculation. However, planning systems are not designed for calculation of doses to distant organs and do not, in any case, include neutron dose estimates.
Calculated organ neutron doses are subject to several assumptions inherent in the values given for the terms in Equation (1). Furthermore, it has been assumed that the estimated dose to a slab is appropriate for all organs within the slab, whereas, in practice, this assumption will overestimate the neutron dose to centrally placed organs. Whilst negligible in comparison with doses to the target and its immediate surroundings, the neutron contribution forms a substantial fraction of the total dose to remote organs and tissues, although this would only apply for treatments at energies in excess of approximately 10 MV.
Dose and risk estimates
The bladder and rectum, as expected, receive the highest dose, since part of the bladder wall and rectum will be contained within the target volume for this treatment. Considerable variation in bladder and rectal volumes, both within and between patients, have been reported [27] and this will lead to corresponding variations and uncertainties in bladder and rectal doses – and risks – in practice. The observation of second bladder and rectal cancers in patients who have previously received radiotherapy for prostate cancer is consistent with these high doses [17], since it implies that parts of each organ may receive doses near the peak of the dose–response curve.
In Figures 4 and 5, the numbers of concomitant exposures have deliberately been set to correspond to the upper limits of what is expected in this centre and it is seen that only for bone surfaces do the concomitant exposures exceed 10% of the total organ dose for both cases illustrated, with red bone marrow contributing >10% for the case of treatments in which portal imaging is the verification technique used.
To place concomitant doses in perspective, we may consider the excess relative risk (ERR) associated with a population receiving radiotherapy and associated concomitant exposures, compared with a hypothetical population receiving radiotherapy only, assuming that the probability of cancer induction per unit dose is constant. The ERR is given approximately by:
|
|
where: DT,radio and DT,concom are the doses to organ T from radiotherapy only and concomitant only exposures, respectively.
Taking the two examples of concomitant exposures (10 CT scans + 36 portal images and 26 CT scans + 4 portal images), in both cases, ERRs for most organs are <0.1, with bone surfaces, small intestine and muscle <0.3. A single CT scan or portal image will give an ERR <0.01 for all organs.
In comparison with work by Waddington and McKenzie [28] the effective dose for a male AP pelvic portal image (15 cmx15 cm at 6 MV) is higher (0.78 mSv MU–1) compared with their 0.34 mSv MU–1). However, no account has been taken here of the fraction of bladder wall volume which lies within the target volume and for which the risk of radiation carcinogenesis may be reduced because of cell kill. This would have the effect of reducing the effective tissue weighting factor for the bladder and hence the effective dose. Moreover, our sampling of the relevant bladder tissue is coarse compared with dose estimations using the dose plotting facilities of a treatment planning system. Waddington and McKenzie assumed that 25% of the bladder was outside the treated volume and a similar assumption would reduce our figure of effective dose to 0.42 mSv MU–1. They also assumed that any cancer cells induced in the bladder tissue within the target would have a negligible chance of survival. Whilst simple exponential cell kill would support this assumption, there is conflicting evidence from epidemiological studies, summarized by Hall and Wuu [4]. We find an effective dose of 0.27 mSv MU–1 for the lateral portal field, compared with 0.32 mSv MU–1 [28] which possibly reflects the smaller field size. In any case, estimates of effective dose will vary with patient size and target volume, so that these results should be considered to be representative rather than definitive.
It has previously been argued [15] that for legislative purposes under the UK IR(ME) Regulations, a single justification for a pre-determined combination of radiotherapy and concomitant exposures is practically more appropriate than the justification of individual concomitant exposures, provided that the contributions of the latter are appreciated. These results support this suggestion, since even for a relatively large number of concomitant exposures, the contribution to the overall organ dose, whilst not negligible, is nevertheless small. However, it should be noted that for radiotherapy treatments carrying a good prognosis, especially for younger patients and children, the added risks of concomitant exposures may need to be considered carefully in the justification process. This is especially true for treatments which involve imaging at each fraction, either by CT or portal imaging, such as the developing techniques associated with image-guided radiotherapy.
The combination of radiotherapy and concomitant doses to give an unambiguous estimate of risk of second malignancy is not trivial and, because of the inappropriateness of the current definition of effective dose, no attempt has been made in this paper to do so. Nevertheless, the doses given here allow the total organ and tissue doses from various combinations of radiotherapy, localization and verification exposures to be estimated for this particular treatment.
|
Conclusions |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Realistic numbers of concomitant exposures give a small but significant contribution to the total dose to most organs and tissues. Generally this is in the range 1–10%, but can be as high a 20% for bone marrow and bone surfaces. At 15 MV, a significant proportion of the doses to organs distant from the target volume is due to photoneutrons produced in the treatment head.
These data may be useful for giving a realistic perspective on the contribution of concomitant doses to the overall organ dose burden for this treatment, for protection of the patient and compliance with the requirements of UK Regulations [2]. Whilst doses from similar abdominal radiotherapy treatments may be inferred from these data, further work will be necessary to establish organ and tissue doses for radiotherapy of other parts of the body, both for adults and children.
|
Acknowledgments |
|---|
Received for publication July 8, 2005. Revision received August 25, 2005. Accepted for publication September 23, 2005.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  L J SAWYER, S A WHITTLE, E S MATTHEWS, H C STARRITT, and T P JUPP Estimation of organ and effective doses resulting from cone beam CT imaging for radiotherapy treatment planning Br. J. Radiol., July 1, 2009; 82(979): 577 - 584. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R M HARRISON Doses to organs and tissues from concomitant imaging in radiotherapy: a suggested framework for clinical justification Br. J. Radiol., December 1, 2008; 81(972): 970 - 974. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R M Harrison, M Wilkinson, D J Rawlings, and M Moore Doses to critical organs following radiotherapy and concomitant imaging of the larynx and breast Br. J. Radiol., December 1, 2007; 80(960): 989 - 995. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| BJR | DMFR | IMAGING | ALL BIR JOURNALS |
