This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Struelens, L Articles by Van Loon, R Search for Related Content PubMed PubMed Citation Articles by Struelens, L Articles by Van Loon, R

Effective doses in angiography and interventional radiology: calculation of conversion coefficients for angiography of the lower limbs

¹ SCK-CEN, Belgian Nuclear Research Centre, Boeretang 200, 2400 Mol, ² Department of Radiology, University Hospitals of the KU Leuven, Herestraat 49, 3000 Leuven and ³ Free University of Brussels, Pleinlaan 2, 1050 Brussels, Belgium

	Abstract

Top Abstract Introduction Material and methods Results Discussion Conclusions References

 
The present study reports on investigations that we have performed to allow the calculation of effective doses (E) in interventional radiology. The use of published conversion tables might not allow sufficient guidance for the establishment of optimization strategies for procedures in interventional radiology. With the Monte Carlo N-Particle transport code (MCNP4B), conversion coefficients, linking dose–area product (DAP) measurements with E, are calculated for angiography of the lower limbs in six hospitals. The influence of various parameters on the calculation of these conversion coefficients is studied in a systematic way using the 2ⁿ factorial design. In this design the effect of different parameters and their pair-wise interactions on a certain variable is explored. In our study, the relevant parameters are tube potential, total filtration and field size and position. We concluded that the influence of radiation spectrum (kVp + filtration) is large and that the effect of field position and size is moderate, except when differences are observed in respect of the gonads. In that case, the variation in conversion coefficients is large. The results of this statistical analysis are then applied to the differences observed between the conversion coefficients, calculated for angiography of the lower limbs in the six hospitals. Recommendations for optimization of patient doses are given.

	Introduction

Top Abstract Introduction Material and methods Results Discussion Conclusions References

 
Interventional radiological procedures aim to diagnose or treat percutaneously particular diseases of the vasculature. In recent years there has been an increase in the number, type and complexity of these procedures. These radiological examinations are characterized by extended fluoroscopy times and they require many radiographic images of different areas of the body. As a consequence, the X-ray doses to patients may be high and both stochastic and deterministic effects have to be considered [1]. The Euratom 97/43 directive on health protection of individuals against the dangers of ionizing radiation in relation to medical exposure, requires special attention for these examinations. The directive, as implemented recently in the Belgian national legislation, lists general measures to ensure the application of the ALARA principle, i.e. the use of a dose as low as reasonably achievable, for these types of investigations. This study was performed within the framework of the preparation of an auditing structure, as foreseen in the legislation, for dose and image quality specifically for angiographic procedures. We report on a series of investigations we have performed to allow the calculation of effective doses from a particular interventional procedure. We focused on the diagnostic procedure "angiography of the lower limbs" as this is the most frequently performed interventional procedure in X-ray departments.

The effective dose E, a measure for stochastic risk, cannot be measured directly. A practical approach starts from entrance surface dose (ESD) or dose–area product (DAP) measurements and uses dedicated conversion coefficients.

Different organizations determined conversion coefficients that link ESD or DAP measurements with E. Previously, DAP values for successive projections of angiographic procedures of the lower limbs were multiplied by published conversion coefficients for simple radiographs of corresponding anatomical regions [2, 3]. The following limitations were observed:

• Hart et al [2] did not establish conversion coefficients for the pelvis right-posterior-oblique (RPO) and left-posterior-oblique (LPO) projections and for exposures of the legs;
  
• There is a large difference in field size used by Hart et al [2] and Drexler et al [3] in comparison with the field sizes used for angiography;
  
• Drexler et al [3] tabulated mainly conversion coefficients for anterior–posterior (AP) projections, while it is common practice to use also posterior–anterior (PA) projections;
  
• We observed average differences of the order of 20% between the results using the ESD-conversion factors in the two publications. Also large differences, up to 35%, were found using on the one hand the DAP-related and on the other hand the ESD-related conversion factors in the work of Hart et al [2].
  

We concluded that the use of these conversion tables might not provide sufficient guidance for establishing optimization strategies for angiographic procedures. As the quantity effective dose (E) is likely to be a useful parameter for the optimization of a technique, more dedicated conversion coefficients seemed necessary.

This paper deals with the calculation of the conversion coefficients and the influence of different parameters on these coefficients.

	Material and methods

Top Abstract Introduction Material and methods Results Discussion Conclusions References

 
Characteristics of equipment and examination protocols
During diagnostic angiographic procedures of the lower limbs, particular arteries are visualized by the local injection of a contrast agent using a catheter. The injection of the contrast agent allows high resolution visualization and documentation of the vessels under study. For angiography of the lower limbs, we considered six adjoining standard regions: abdomen, pelvis, upper legs, knees, lower legs and feet.

In six different hospitals, the following parameters were registered for this type of examination: radiation quality, anatomical area irradiated, beam projection and the use of extra absorption filters.

Two hospitals (C and D) have an overcouch configuration, i.e. the X-ray tube is above the patient support table. The four other hospitals (A, B, E and F) have an undercouch configuration.

Questionnaires regarding working procedure were completed by all the radiologists in all hospitals. For every exposure, the projections of the beam and the exposed area of the body were clearly documented (Table 1). In Figure 1 the field positions for exposure of the abdomen and pelvis are shown for hospitals A–F, respectively. However, it is very probable that the exact field position will vary not only between the centres, but also between the patients from the same centre. In hospital C, the AP projection of the pelvis (Figure 1c, bold line) does not adjoin the AP projection of the abdomen. It starts much lower, at the bladder height. The adjoining pelvis projections are the RPO and LPO projections. Therefore, in this hospital the AP projection only partly overlaps the LPO and RPO projections of the pelvis. Consequently, the exposure of the upper legs, adjoining the pelvis AP projection, also starts lower compared with the other centres. In hospital D, there is no exposure of the abdomen.

View larger version (41K): [in this window] [in a new window]   Figure 1. Radiation fields for exposure of the abdomen, pelvis and upper legs in the six hospitals. The anteroposterior (AP)-projection of the pelvis in hospital C is illustrated by the bold rectangle. The field for exposure of the upper legs, adjoins the AP projection of the pelvis.

The radiation spectra were characterized by the total filtration and the tube potential value. With a PMX-3 solid state dosemeter (RTI Electronics, Sweden) the half value layer (HVL) was determined for every tube in the different hospitals, by measuring dose rate for different thicknesses of aluminium. From these measurements, the total filtration is determined using the conversion tables in the application note of RTI electronics [4]. These filtrations, in terms of equivalent aluminium, comprising X-ray tube and patient support table for the undercouch configurations, are shown in Table 2, together with possible additional copper filtration and the average peak voltage used for every projection. With the "Catalogue of Diagnostic X-ray Spectra and other Data" of IPEM [5] the different spectra in the hospitals were determined for every projection separately, based on the total filtration of the beam and the tube potential.

The phantom's sex and age has to be chosen, as well as the organs to be included in the model. For both sexes, the adult phantom is chosen (21 years for male and 15 years for female model).

For one specific spectrum, projection and exposed body area, MCNP calculates doses in every organ of the phantom. With these organ doses and the tissue weighting factors specified by ICRP [2], effective doses were calculated. The theoretical DAP values associated with a given number of simulated X-ray photons were calculated from the average fluence through a surface at 20 cm from the source. Dividing the calculated effective dose by the theoretical DAP value gives the conversion coefficient for one specific exposure.

Calculation and comparative study of the conversion coefficients
Conversion coefficients were calculated for every specific projection in each hospital and for male and female patients separately. For every hospital the radiation fields are chosen as illustrated in Figure 1 and radiation quality chosen as in Table 2. The difference between the coefficients will be studied in more detail. The mean difference between conversion coefficients for men and women is calculated for each anatomical region. The factorial experiment is used to explain the differences in conversion coefficients between the centres, caused by various parameters.

Factorial experiments
Several parameters, such as field size, field position, patient thickness and radiation spectrum (kVp and filtration) are known to affect the conversion coefficients for estimating the effective dose. Ideally, the influence of each parameter would be independent, so that an estimate of a conversion coefficient is the product of several parameters [9]:

where is a representative value of the coefficient and _i, _j... _n are the factors which take into account the effect of the variations of the n parameters. It is important not only to determine if the parameters have an influence on the conversion coefficients, but also if there is a significant interaction between the parameters. This means that Equation (1) has to be expanded to:

where _ij... _in... _jn... are correction factors which include the additional effect of interaction between the parameters. In general, higher-order interactions do not contribute significantly to the observed variations. This model converts into a linear model if natural logarithms are taken of Equation (2) and resembles the additive effect model with interactions by Snedecor and Cochran [10]. To explore the effects of the different parameters and their interactions in a systematic and efficient way, the 2ⁿ factorial experiment is used [10]. For every parameter, one "low" and one "high" level is given, even though this notation might be arbitrary in the case of qualitative parameters. The complete factorial experiment requires that each level of every parameter occurs with each level of every other parameter, giving a total of 2ⁿ combinations, with n being the number of parameters involved in the analysis.

In our experiment, the effect of the following parameters on the calculation of conversion coefficients for the effective dose were considered:

• Tube voltage (kVp)
  
• Beam filtration
  
• Field position and size
  

The low and high level values chosen for the parameters are determined from our detailed observations for several interventional procedures in different hospitals. We believe that these ranges are representative values for the parameters considered in interventional radiology for the exposure of abdomen and pelvis. For tube voltage (65–85 kVp) and for filtration (4.1 mm Al–6.5 mm Al+0.1 mm Cu). In Figure 2, the low and high levels for the field position are illustrated.

View larger version (34K): [in this window] [in a new window]   Figure 2. Low and high level for (a) the abdomen and (b) the pelvis exposure in the factorial experiment. In practice, the fields have the same horizontal dimensions. In the figures they are offset to illustrate the different fields more clearly.

Statistical analysis is performed for each of the projections on the logarithms of (E/DAP) – conversion coefficients and the main factorial effects and all pair-wise interactions were investigated. These analyses were done using the generalized linear model of the computer package STATISTICA.

The results of the factorial experiment are applied to the differences between the conversion coefficients, calculated for angiography of the lower limbs for the six specific hospitals.

	Results

Top Abstract Introduction Material and methods Results Discussion Conclusions References

 
In Table 3 the conversion coefficients are shown that link measured DAP values to the effective dose, both for men and women and for the six hospitals. The statistical error of all these factors is less than 0.1%. This was calculated from the relative errors given on the MCNP output files for every calculated dose.

Effective doses can be calculated for the full procedure by multiplying the conversion coefficients in Table 3 by the measured DAP values per projection in all of the six hospitals: with i the different projections.

In Table 4, the average DAP data from each hospital is given for each projection. This results in average effective doses of 8.0 mSv for hospital A (15 patients), 15.5 mSv for hospital B (16 patients), 8.0 mSv for hospital C (20 patients), 3.9 mSv for hospital D (13 patients), 16.8 mSv for hospital E (18 patients) and 11.3 mSv for hospital F (12 patients).

The results of the factorial analysis are given in Table 5. The first value is a representative value for the conversion coefficient and corresponds to the geometrical mean of the eight possible combinations. The main and interaction effects are expressed as multipliers, illustrating how big the effects of the different parameters and their pair-wise interactions are on the mean value of the conversion coefficients. In Table 5, the multipliers are given for the low level of every parameter and their interactions. According to the constraints in the statistical analysis [9], the multiplier for the high level is the reciprocal of the multiplier for the low level. For multipliers <1, the conversion coefficient increases and for multipliers >1, the coefficient decreases when the parameter changes from the low to the high level. By analogy with West et al [9], variations between the low and high value of less than 2% are described as "insignificant", of 2–5% as "small", of 5–15% as "moderate" and of over 15% as "large".

Conversion coefficients for every combination of the three parameters can be calculated per projection, following Equation (2).

For example, for an abdomen PA exposure at 65 kVp and 6.5 mm Al+0.1 mm Cu with a high field position, the conversion coefficient can be calculated from Table 5 as:

CF=0.170 x 0.842 x (1/0.834) x (1/0.914) x (1/0.986) x (1/0.995) x 0.998=0.191

Since for the abdomen PA projection most interactions are insignificant, they can be omitted. For other levels of a specific parameter, the multiplier can be determined from an interpolation between the low and the high multiplier of that parameter.

	Discussion

Top Abstract Introduction Material and methods Results Discussion Conclusions References

 
Male versus female patients
In Hart et al [2], no difference is made between conversion coefficients for male and female patients. From Table 3, however, we see that there is a significant difference in conversion coefficients for both sexes. The position of the gonads is crucial here. For the abdomen and pelvis projection, the conversion coefficients for the female patients are higher than those for the male patients. From Figure 1, we can see that the female gonads are near or in the radiation fields for both projections. For the pelvis projection, the male gonads are also near or in the radiation field. However for the hospitals with tube-under-table configuration, the female gonads are exposed more directly than the male gonads. The conversion coefficients for the upper legs are larger for the male patients, because the male gonads are much closer to the field than the female gonads.

Overcouch versus undercouch configurations
The position of the different radiation sensitive organs in the human body determines the difference in conversion coefficients between the anterior and posterior projections. For the anterior projection of the abdomen, radiation sensitive organs like the liver, stomach and lower large intestine (LLI) have larger doses compared with the posterior projection. For the latter projection, the doses to the mineral bone surface and the red bone marrow are larger compared with the anterior projection. However, they contribute less to the effective dose, because their doses are lower compared with those of the liver, stomach and LLI.

Abdomen projection
From the statistical analysis we can see that the maximum variation in conversion coefficients for the change in field position for the AP and PA exposure of the abdomen is around 15%. The difference between the conversion coefficients in Table 3, for this projection, is therefore mainly due to differences in tube configuration and radiation spectrum. The variation for these projections due to tube potential and filtration is large. This explains why hospitals A and F have larger conversion coefficients for the abdomen exposure. From Table 2, we observed the largest filtration for hospital A and the highest tube potential value for hospital F. The conversion coefficient in hospital C is larger due to the anterior projection, as explained in the previous paragraph.

Pelvis projection
The influence of the field position is much larger for the pelvis exposure: 31% for the AP-projection and 51% for the PA-projection. As mentioned before, this is caused by the difference in exposure of the gonads. This explains the high conversion coefficient for male patients in hospital C in Table 3: the inclusion of the gonads in the radiation field and the direct exposure due to the anterior projection. For the female patients, the conversion coefficient for the pelvis exposure in hospital C is much smaller compared with the other hospitals. The ovaries are far away from the radiation field in this hospital (Figure 1). Again for the pelvis exposure, we observe from the statistical analysis that tube potential and filtration have a large influence for the PA and AP projections, respectively. This explains the difference between the conversion coefficients in the hospitals where the difference in field position with respect to the gonads is small.

Upper leg projection
For the upper leg projection, the conversion coefficient for the male patients in hospital D is much larger than in other hospitals. Again, this can be explained by the inclusion of the gonads in the radiation field and by the anterior projection. In hospital C, the conversion coefficient for the upper leg for men is lower than in any other centre. Because of the lower position of the radiation field for this projection (Figure 1c), the gonads are more distant from the field. We can also observe a difference in conversion coefficients for the hospitals with tube-under-table configuration: for hospital E (0.032 mSv Gy^–1cm^–2) where the gonads are not in the radiation field, for hospital B (0.046 mSv Gy^–1cm^–2) with gonads partly exposed and for hospital F (0.064 mSv Gy^–1cm^–2) with gonads completely exposed. The higher conversion coefficient in hospital A (0.095 mSv Gy^–1cm^–2) is mainly caused by the larger filtration.

Patient thickness
The calculation of conversion coefficients for the six hospitals is made for standard-sized patients. In practice, this will not be the case. For the same tube potential and filtration, thicker patients will have smaller conversion coefficients because the organ doses are smaller compared with the standard patient. In practice, however, thicker patients could have larger tube potential values, implying a higher conversion coefficient. However, this will not necessarily mean a larger effective dose to the patient. A higher tube potential might imply a lower mAs value, depending on the tube potential curve and the dose level set in the system. Consequently, the DAP value could decrease. The steeper the tube potential curve, the smaller will be the mAs and the DAP.

Optimization guidelines
When minimizing effective doses for patients, the doses to the gonads play an important role. For procedures such as angiography of the lower limbs, exposure of the gonads is inevitable. In all hospitals the gonads for both genders are often exposed completely or partly in the pelvis projection. However, due to the possible additional oblique views, the pelvis is often more exposed than the other body regions. Therefore, it would be preferable for male patients to expose the gonads with the upper legs projection. Moreover, the beam path through the upper legs is smaller than through the pelvis. This results in lower tube potential and/or mAs values and consequently in lower DAP values. The higher conversion coefficient for the upper legs projection (due to the presence of the gonads) will contribute less to the effective dose, because of the lower associated DAP value.

For the female patients, the ovaries could be exposed with the abdomen projection to avoid the extra exposure from the oblique projections of the pelvis. The ovaries, however, are situated at that point where the abdominal artery splits into the two arteries leading to the legs. As it is mainly for this position of the arterial system that the oblique projections are performed for angiography of the lower limbs, it is not possible to include the ovaries in the abdominal exposure from a medical point of view.

Kicken [11] reported (E/DAP)-conversion coefficients for an abdomen PA projection. They were 0.102 mSv Gy^–1cm^–2 for the male patients for 80 kVp and 0.171 mSv Gy^–1cm^–2 for the female patients for 75 kVp, both with a total filtration of 6.0 mm Al_eq. They are comparable with the conversion coefficients in our study.

The conversion coefficients in NRPB-R262 [2, 12] are interpolated for 75 kVp and 5 mm Al_eq as 0.240, 0.149, 0.268 and 0.112 for the abdomen and pelvis AP and PA projections, respectively. These values, averaged over male and female patients, are comparable with the average values given in Table 5.

The conversion coefficients in this study are for the procedure of angiography of the lower limbs in several hospitals. However, they can also be used for similar projections in other procedures. For example, the conversion coefficients for the pelvis projections can be used in an iliac stenting procedure or the conversion coefficient for the abdomen irradiation can be used in a stent procedure of the kidneys or for a TIPS (transjugular portosystem shunt) procedure, provided that filtration and tube potential are known and that DAP values for each of these projections in the specific procedure are available.

One of the limitations of the study is the small number of participating hospitals. Also, an accurate calculation of effective dose requires a separate DAP value for every projection. This is not routinely available yet.

	Conclusions

Top Abstract Introduction Material and methods Results Discussion Conclusions References

 
For the six hospitals in this study, conversion coefficients were calculated for every specific field position and radiation spectrum. This is possible but not routinely very practical. We investigated which parameters have an important influence on the calculation of conversion coefficients in order to limit the set necessary to calculate effective dose in angiography and interventional radiology. From the statistical analysis and the interpretation of the conversion coefficients for the six hospitals, we can conclude that:

1. X-ray systems with a tube-over-table configuration give higher conversion coefficients and therefore higher effective doses than systems with tube-under-table configuration;
   
2. The influence of tube potential and filtration on the conversion coefficients is large (30%). It is recommended to establish a set of conversion coefficients for the entire range of tube potential and filtration in angiography and interventional radiology;
   
3. Radiation fields are small in interventional radiology, compared with other standard radiographic measurement. This means that male- or female-specific organs are not necessarily irradiated by the same radiation fields. Hence, it is recommended to calculate separate conversion coefficients for male and female patients;
   
4. The influence of field position on the calculation of conversion coefficients is large (31–51%) whenever variations are observed in respect of the gonads. For an accurate calculation of effective dose in an angio room, one should investigate if there are large deviations between the actual position of the radiation field and the reference field for which the conversion coefficients were calculated in respect to the gonads. If necessary, a new coefficient should be calculated specific for the radiation field used.
   

In order of priority, we propose to treat separately undercouch and overcouch set ups, to account then for male versus female patients, to adjust conversion coefficients in terms of radiation quality and finally, if necessary, to fine tune for differences in radiation field. If the position of the radiation field is close to or involves the gonads, greater attention should be given to the calculation.

Received for publication December 22, 2003. Revision received September 2, 2004. Accepted for publication September 29, 2004.

	References

Top Abstract Introduction Material and methods Results Discussion Conclusions References

 

1. Faulkner K. Radiation protection in interventional radiology. Br J Radiol 1997;70:325–6.[Medline]
2. Hart D, Jones DG, Wall BF. Estimation of effective dose in diagnostic radiology from entrance surface dose and dose-area product measurements. NRPB-R262. Chilton: NRPB, 1994.
3. Drexler G, Panzer W, Widenmann L, Williams G, Zankl M. The calculation of dose from external photon exposures using reference human phantoms and monte carlo methods, part III: organ doses in X-ray diagnosis. GSF-Bericht 11/90, 1990.
4. HVL measurements with PMX-III and the PMX detector on units with tungsten target. Application note No. 03-008/01, RTI electronics AB, Sweden, February 1994.
5. Catalogue of Diagnostic X-Ray Spectra and other Data, IPEM-R78, York, UK 1997, ISBN 0 904181 88 X (CD-ROM).
6. Briestmeister JF. "A General Monte Carlo N-Particle Transport Code", Version 4B (LA-12625-M, March 1997).
7. Van Riper KA. Bodybuilder software for Windows, version 1.21, White Rock Science, Los Alamos, USA, 1997-2001.
8. Cristy M, Eckerman KF. Specific absorbed fractions of energy at various ages from internal photon sources. I. Methods. Oak Ridge National Laboratory Report ORNL/TM-8381/VI, 1987.
9. West KN, Sandborg M, Persliden J, Alm Carlsson G. Sensitivity of coefficients for converting entrance surface dose and kerma-area product to effective dose and energy imparted to the patient. Phys Med Biol 1999;44:1937–54.[CrossRef][Medline]
10. Snedecor GW, Cochran WG. Statistical methods, 7th edn. Iowa: Iowa State University Press, 1980.
11. Kicken PJH. Radiation dosimetry in vascular radiology – organ and effective dose to patients and staff. Chapter 8: Quantitative data about patient exposure conditions and dose estimates for selected procedures in the University Hospital Maastricht. 1996:p140 ISBN 90 5278 217 2.
12. Hart D, et al. Authors' reply. Br J Radiol 1998;71:995.

This article has been cited by other articles:

				H GEIJER, E NORRMAN, and J PERSLIDEN Optimizing the tube potential for lumbar spine radiography with a flat-panel digital detector Br. J. Radiol., January 1, 2009; 82(973): 62 - 68. [Abstract] [Full Text] [PDF]

				K. Smans, L. Struelens, M.-T. Hoornaert, F. Bleeser, N. Buls, D. Berus, P. Clerinx, F. Malchair, F. Vanhavere, and H. Bosmans A study of the correlation between dose area product and effective dose in vascular radiology Radiat Prot Dosimetry, July 1, 2008; 130(3): 300 - 308. [Abstract] [Full Text] [PDF]

				C. J. Martin RADIATION DOSIMETRY FOR DIAGNOSTIC MEDICAL EXPOSURES Radiat Prot Dosimetry, March 28, 2008; (2008) ncm495v1. [Abstract] [Full Text] [PDF]

				M. T. Bahreyni Toossi, H. Zare, S. Bayani, and S. Esmaili Organ and effective doses of patients arising from coronary angiography and percutaneous transluminal coronary angioplasty at two hospitals in Mashhad-Iran Radiat Prot Dosimetry, February 1, 2008; 128(3): 363 - 366. [Abstract] [Full Text] [PDF]

				L. Struelens, F. Vanhavere, H. Bosmans, R. Van Loon, and M. Geukens Data analysis from a multi-centre, comparative study of angiographic examinations leading to practical guidelines for the optimisation of patient doses Radiat Prot Dosimetry, December 1, 2005; 117(1-3): 87 - 92. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Struelens, L Articles by Van Loon, R Search for Related Content PubMed PubMed Citation Articles by Struelens, L Articles by Van Loon, R