Department of Biomedical Engineering & Medical Physics, North Staffordshire Hospital, Princes Road,. Hartshill, Stoke-on-Trent, Staffordshire ST4 7LN, UK.
T he British Journal of Radiology, 70 (1997 ), 299–302
© 1997 The British Institute of Radiology
Short communication
A survey of current in vivo radiotherapy dosimetry practice C R EDWARDS, BSc, MPhil, M H GRIEVESON, BSc, MIPEMB, P J MOUNTFORD, PhD, FIPEMB and P ROLFE, PhD, FIEE Department of Biomedical Engineering & Medical Physics, North Staffordshire Hospital, Princes Road, Hartshill, Stoke-on-Trent, Staffordshire ST4 7LN, UK Abstract. A questionnaire was sent out to 57 radiotherapy physics departments in the United Kingdom to determine the type of dosemeters used for in vivo measurements inside and outside X-ray treatment fields, and whether any correction is made for energy dependence when the dose to critical organs outside the main beam is estimated. 44 responses were received. 11 centres used a semi-conductor for central axis dosimetry compared with only two centres which used thermoluminescent dosimetry (TLD). 37 centres carried out dosimetry measurements outside the main beam; 25 centres used TLD and 12 centres used a semi-conductor detector. Of the 16 centres measuring the dose at both sites, 11 used a semi-conductor for the central axis measurement, but only four of those 11 changed to TLD for critical organ dosimetry despite the latter’s lower variation in energy response. None of the centres stated that they made a correction for the variation in detector energy response when making measurements outside the main beam, indicating a need for a more detailed evaluation of the energy response of these detectors and the energy spectra outside the main beam.
In vivo radiotherapy dosimetry (i.e. the monitoring of the actual dose received during treatment) provides a quality assurance check of the accuracy of the tumour dose, and a means of determining the radiation dose to a critical organ outside the main beam, such as the lens of the eye or the spinal cord. P-type semi-conductor diodes and thermoluminescent dosimetry (TLD) are most commonly used for in vivo dosimetry. Both devices have advantages and disadvantages when used to measure a radiation dose inside or outside the main beam (Table 1 ) [1–8 ]. An important disadvantage for both detectors is the energy dependence of their response, particularly at low photon energies. This can have considerable significance when measuring critical organ doses in regions of scattered radiation outside the treatment area. Further inherent errors affecting the measurement of delivered dose are associated with the use of surface detectors to infer dose at the depth required for maximum dose. The clinical value of in vivo dosimetry is well recognized [ 9–11 ] and both types of detector are commercially available for routine use. However, there is little literature describing their application to in vivo dosimetry or the corrections to be made for non-uniform energy response. A survey was Received 21 October 1996 and in revised form 14 November 1996, accepted 5 December 1996. T he British Journal of Radiology, March 1997
therefore carried out to assess the current in vivo dosimetry practice within the United Kingdom.
Survey design The aims of the survey were: 1. To identify how many radiotherapy departments within the United Kingdom perform routine in vivo dosimetry measurements, and the type of detector used at each centre. 2. To identify how many radiotherapy departments correct for the energy dependence of the detectors when measuring the dose outside the main beam, and the method used for this correction.
Results Of the 57 questionnaires sent out, 44 were returned during the period October 1994 to January 1995, equivalent to a return rate of 77%. Table 2 summarises the questions used in the survey and the replies received. From question 1 it appears that the majority of radiotherapy centres (63%) do not routinely carry out in vivo measurements of the central axis dose, despite its value in identifying errors such as inaccurate patient positioning, use of wrong accessories (e.g. wedge filter, compensator) or incorrect calculation of monitor units. The response to question 2 indicates that 299
C R Edwards, M H Grieveson, P J Mountford and P Rolfe Table 1. Summary of the characteristics of semi-conductors and TLDs when used for in vivo dosimetry measurements [1–8] Detector type
Response characteristics Advantages
Disadvantages
P-type semi-conductor diode
Instantaneous [1] Small sensitive volume [ 1] Independent of pressure [ 1]
Energy dependent [2] Temperature dependent [3 ] Directionally dependent [1] Susceptible to radiation damage [ 4] Dose rate dependent without pre-irradiation [4 ] Connecting cables Inherent build-up prevents skin dose measurement
Thermoluminescent dosimetry
Dose rate independent [1] Directionally independent [5 ] Small sensitive volume [ 5] Independent of pressure [ 1] No build-up present
Energy dependent [6] Superlinear [7 ] Not instantaneous [7] Susceptible to fading effects [ 8]
Table 2. Survey questions and responses Question number
Question
Response
1
Does your department routinely measure given doses on the central axis during external beam radiotherapy treatments? (n=44) On what treatment units do you carry out these measurements? What type of dosimetry system do you use when making central axis measurements? (n=17) Does your department measure critical organ doses? (n=44) Are the same dosemeters used for critical organ measurements as those stated earlier for question 3?
Yes: 17 centres (37%) No: 27 centres (63%)
2 3 4 5
6
When measuring critical organ doses, do you allow for any change of sensitivity in your dosimetry system with energy? (n=38)
Megavoltage machines (inc. 60Co): 17 centres ( 100%) Diode: 12 centres ( 70%) TLD: 2 centres (12%) Both: 3 centres (18%) Yes: 38 centres (86%) No: 6 centres (14%) (a) Centres also measuring central axis dose (n=16)a Diode: 8 centres ( 50%) TLD: 8 centres (50%) Both: 0 centre (0%) (b) Centres not measuring central axis dose (n=22) Diode: 4 centres ( 18%) TLD: 17 centres (77%) Both: 1 centre (5%) Yes: 6 centres (16%) No: 32 centres (84%)
a Does not include the centre which recorded central axis dose but not critical organ dose.
these measurements were only carried out on megavoltage treatment equipment (including 60Co units). This is probably because radical treatments requiring a high level of accuracy are almost universally carried out on megavoltage equipment, while orthovoltage units are used for palliative treatments. The semi-conductor diode was preferred to TLD for central axis measurements by a ratio of approximately 651 (question 3 ). A possible explanation is that the instantaneous response of the diode system was more important than the 300
advantages offered by the TLD for this measurement (Table 1 ). The number of centres monitoring critical organ dose was twice the number who recorded central axis dose (question 4 ). One centre recorded central axis dose but not critical organ dose. For those centres performing central axis and critical organ dose measurements, there was an equal division between the use of the diode and TLD systems for the latter measurements (question 5 ). For those centres who did not record central axis dose but T he British Journal of Radiology, March 1997
Short communication: In vivo radiotherapy dosimetry practice
did record critical organ dose, the TLD was the preferred dosimetry system by a ratio of over 451 compared with the use of semi-conductors. Only six centres (four using semi-conductors, two using TLDs) indicated that they corrected for the nonuniformity in energy response (question 6). However, in each case corrections were derived for the conditions of total body irradiation and therefore the detector response was calibrated only for the energy spectrum of the primary beam. No centres indicated that corrections of critical organ dose allowed for changes in the energy spectrum outside the main beam.
Discussion Despite the obvious benefits of in vivo dosimetry, less than half of the United Kingdom radiotherapy departments which responded to this survey routinely carried out central axis in vivo dosimetry with the majority of these preferring to use a semiconductor diode system. However, the use of the word ‘‘routinely’’ in question 1 gave rise to some uncertainty since some departments measured the given dose on the first treatment fraction for every patient and other centres measured the first fraction less often or only on certain categories of patients. All centres who questioned whether their frequency could be defined as ‘‘routine’’ were included in the ‘‘yes’’ response group. More than twice as many respondents carry out critical organ dose measurements outside the treatment field compared with the number carrying out central axis dosimetry, indicating the perceived importance of these measurements, with the TLD being the preferred dosimetry system. For the 16 centres performing dosimetry at both sites, Table 3 gives a more detailed breakdown of the choice of detector for critical organ dosimetry according to the type of detector used for central axis dosimetry. 11 of these centres used the diode for central axis dosimetry, but only four of these 11 used TLDs to record the critical organ dose. This is despite the smaller variation in TLD response for low photon energies which is a particularly important characteristic if the critical organ is outside the main
Table 3. Comparison of detection systems between central axis and critical organ dosimetry for the 16 centres performing dosimetry at both sites Critical organ dosemeter
Central axis dosemeter Diode
TLD
Both
Diode TLD Both
7 4 0
0 2 0
1 2 0
Total
11
2
3
T he British Journal of Radiology, March 1997
beam. The potential importance of this characteristic was recognized by two of the three centres who used both systems to record central axis dose, but preferred to use the TLD to record critical organ dose. Although Table 3 also shows that both centres which used TLD to record central axis dose remained with that type of detector for critical organ dose measurement, this choice may have been dictated by the lack of a diode system. It was evident from the survey that none of the centres who carry out critical organ dose measurements outside the main beam made any correction based on the energy response of the detector and the energy spectrum of the radiation at the critical organ. The pragmatic approach reported by four centres is that the greater sensitivity of both types of dosemeter to lower energy scattered radiation helps to protect critical organs by overestimating their dose.
Conclusions Three main conclusions were drawn from this survey. Firstly, central axis in vivo dosimetry was carried out much less frequently than critical organ dosimetry. Secondly, there was a lack of consensus over the type of dosemeter to be used for either purpose. Thirdly, no corrections were made for energy response when critical organ dose was being determined. Compared to the TLD, the main advantages of the semi-conductor for critical organ dosimetry are the instantaneous response and its smaller sensitive volume giving a better spatial resolution. Although both detectors have a non-uniform response at low energies, the greater non-uniformity of the diode remains the main disadvantage for this application. Knowledge of the energy spectrum at any point outside the main beam and the energy response of the TLD and the diode will provide a more accurate estimate of critical organ dose. Therefore, future work will concentrate on a detailed evaluation of the energy sensitivity of direct patient dosimetry systems, together with Monte Carlo calculations of the energy spectra of scattered radiation at points outside the treatment field, for various nominal beam energies and treatment situations; the objective of this work being to improve the overall accuracy of in vivo critical organ dosimetry. Publication of this data may encourage an increase in the general frequency at which this type of dosimetry is carried out.
Acknowledgments We would like to express our gratitude to all those respondents who took the time to complete and return this questionnaire; to Mr A J Moloney for his constructive comments and to the West 301
C R Edwards, M H Grieveson, P J Mountford and P Rolfe
Midlands NHS Executive for their financial support of Mr C R Edwards.
References 1. Aukett RJ. A comparison of semi-conductor and thermoluminescent dosemeters for in vivo dosimetry. Br J Radiol 1991;64:947–52. 2. Rikner G, Grusell E. Selective shielding of a p-Si detector for quality independence. Acta Radiol Oncol 1985;24:65–9. 3. Grusell E, Rikner G. Evaluation of temperature effects in p-type silicon detectors. Phys Med Biol 1986;31:527–34. 4. Rikner G, Grusell E. Effects of radiation damage on p-type silicon detectors. Phys Med Biol 1983;28:1261–7. 5. Kron T, Schneider M, Murray A, Mameghan H. Clinical thermoluminescence dosimetry: how do expectations and results compare? Radiother Oncol 1993;26:151–61.
302
6. Marshall TO. Accuracy and precision in thermoluminescence dosimetry. In: Hufton A, editor. Practical Aspects of Thermoluminescence Dosimetry. London: Hospital Physicists’ Association, 1984;19. 7. Horowitz YS. The theoretical and microdosimetric basis of thermoluminescence and applications to dosimetry. Phys Med Biol 1981;26:765–824. 8. Cameron J. Radiation dosimetry. Env Health Perspect 1991;91:45–8. 9. Leunens G, van Dam J, Dutreix A, van der Schueren E. Quality assurance in radiotherapy by in vivo dosimetry. 1. Entrance dose measurements, a reliable procedure. Radiother Oncol 1990;17:141–51. 10. Leunens G, van Dam J, Dutreix A, van der Schueren E. Quality assurance in radiotherapy by in vivo dosimetry. 2. Determination of the target absorbed dose. Radiother Oncol 1990;19:73–87. 11. Mayles WP, Heisig S, Mayles HM. Treatment verification and in vivo dosimetry. In: Williams JR, Thwaites DI, editors. Radiotherapy Physics in Practice. Oxford: Oxford University Press, 1993;227–51.
T he British Journal of Radiology, March 1997
