- Email: [email protected]
Comment on “Effective dose range for dental cone beam computed tomography scanners”
European Journal of Radiology 81 (2012) 4219–4220
Contents lists available at SciVerse ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Letter to the Editor Comment on “Effective dose range for dental cone beam computed tomography scanners” To the Editor, I wish to congratulate Dr. Pauwels and his co-authors on their manuscript, “Effective dose range for dental cone beam computed tomography scanners,” which is currently in press in the journal [1]. The authors have undertaken an impressive survey of the dosimetry of many of the currently available CBCT models for dentistry. Because this will be read by many who are interested in a better understanding of the risk of CBCT examinations, it is an important contribution to the literature. To place this work in the context of previously published studies it is important to understand differences in the methods that were used and the potential impact of the authors’ approach on the internal validity of the data and its extensibility to previous and future studies. A large number of dosimeters (147 and 152) were employed for each of the 2 phantoms used in this study. The rationale for this was to insure an even spread over each of the radiosensitive organs in the head and neck area that are utilized in the 2007 ICRP calculation of effective dose. The intent was to produce dose estimates that were as accurate as possible. While this is a laudable goal, it is important to note that sampling strategy is just one of a number of factors that are important in measuring dose. Just as important are the actual location and extent of the radiosensitive organs that are being measured and the location of the field of view when exposing the phantom. Within patient populations, the location and size of tissues of interest can vary significantly. Positioning of patients for imaging is also subject to considerable variation depending on the operator and the positioning aids that are available with different units. While the current study attempted to position phantoms as closely as possible to a typical patient, use of different “local radiographic staff” for the different units may have resulted in inconsistent phantom positioning and important differences in dose measurements in some units. This is in fact suggested by an examination of the data for large FOV units. The effective doses reported for most of the large FOV units in the current study are in line with previously reported studies and unpublished data from our group; however, results for the NewTom
VGi are almost double that of our recent findings [2]. A potential source for this variation is the high thyroid dose seen in the current study. It is useful to compare and contrast thyroid and salivary gland doses as both are important components of the effective dose calculation. With a properly positioned patient in a large FOV, the salivary glands are directly exposed by the primary beam. In contrast, the thyroid gland is below the field of direct exposure. For a given field size the amount of scatter radiation is directly proportional to the amount of primary radiation. As mAs is increased or reduced, the ratio of scattered to primary radiation will be a constant. For a RANDO phantom the ratio of salivary gland dose to thyroid dose will remain constant as mAs varies. Changes in kVp, filtration, and shape of the FOV for CBCT units with similarly sized large FOVs have a relatively small influence on this ratio. For the most part the foregoing theoretical discussion is confirmed by data from the current study. Examination of the equivalent doses for the organs in the following table is revealing. A mean salivary gland/thyroid dose ratio of 4.5 is seen for the 8 units that were measured in this study (Table 1). Although ratios for 7 of the 8 units tested are within a standard deviation of that mean, the ratio for the NewTom VGi is 2.3 standard deviations below the mean. The thyroid dose (2045 Sv) is about 3 times the average thyroid dose for all machines. It is also more than 4 times greater than the dose found in my evaluation of the unit seen in the first column of data. This data is based on imaging geometry seen in Fig. 1. A likely reason for the difference in the estimate of dose between the two studies is a more caudal position of the FOV in the Pauwels et al. study. It is also likely that the FOV is lower for the VGi than for the other units in the current study. Lower positioning of the FOV may result in direct exposure of the thyroid gland. Because this gland has a weighting factor of 0.04, it makes a significant contribution to the total effective dose calculation. Indeed, dividing the reported salivary gland dose by the mean salivary gland/thyroid ratio, we might estimate an expected thyroid dose of 640 Gy. The difference in expected and reported doses accounts for approximately 56 Sv of excess effective dose that is due to direct exposure of the thyroid in this case. This excess represents 29% of the reported total effective dose. While higher patient positioning in the FOV can occur during
Table 1 Equivalent dose and effective dose comparisons for large FOV examinations. Measurement
Thyroid dose (Gy) Salivary gland dose (Gy) Salivary gland/thyroid ratio Number of std. dev. from mean: thyroid Effective dose (Sv)
Study Ludlow NewTom VGi
Pauwels NewTom VGi
Galileos
iCAT NG
Iluma
Kodak 9500
NewTom VGi
Scanora 3D
Sky View
Mean (SD)
447 1891 4.2 −0.4 100
2045 2855 1.4 2.3 194
380 2104 5.5 −0.5 84
355 1830 5.2 −0.6 83
1230 7225 5.9 0.9 368
585 2676 4.6 −0.2 136
354 1690 4.8 −0.6 83
296 1568 5.3 −0.7 68
474 1582 3.3 −0.4 87
685 (583) 2602 (1793) 4.5 (1.4)
The value in bold is the focus of arguments within the manuscript. 0720-048X/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2012.05.027
4220
Letter to the Editor / European Journal of Radiology 81 (2012) 4219–4220
Fig. 1. Location of RANDO phantom for dosimetry of NewTom VGi in Ludlow study. Position of FOV includes minimum amount of anatomy below soft tissue contour of chin. Location of thyroid dosimeters is below the lowest directly exposed level of the phantom.
imaging, this is not an intrinsic property of the VGi unit which allows independent control of chin rest position and FOV location when positioning patients. The issue of susceptibility of measurement of thyroid dose to even subtle differences of phantom positioning has been discussed previously where it was shown that a rotation of even 10◦ of a RANDO phantom position can result in a 92% difference in thyroid dose for a large FOV [3]. Because the current study seeks to report representative doses for the different units under comparison, it seems unfair to penalize an individual unit with a large thyroid exposure that is based on significantly different FOV location than other units. In the current study a recalculation of organ doses using a subset of 24 dosimeters positioned similarly to protocols followed in our studies [4], demonstrated deviations in organ dose estimations by 18–28% with differences up to 80% in comparison to the full complement of dosimeters. The authors suggest that this finding of variability indicates that a large number of TLDs is needed for accurate effective dose estimation. However, the authors do not provide specific results of calculations of either equivalent doses or effective dose for the 24-dosimeter subset evaluations, so it is not possible to explore potential reasons for the observed variation. From the discussion of the above data, it is quite possible that some or even a large proportion of the variation is related to phantom position within the FOV rather than limited numbers of dosimeters located in organs and tissues of interest. While use of additional dosimeters may increase the precision of calculation of organ dose, it does not guarantee an increase in the accuracy of calculation of effective dose. Having made this argument I would like to suggest that future research in this field should be focused on the development of phantoms that can provide reasonable indicators of effective dose with fewer, rather than more dosimeters. Given the diversity of human craniofacial anatomy and the uncertainty of risk estimation of low dose exposures, the value of increased accuracy for a single anatomic representative must be considered in the context of the cost of time and resources needed to obtain a marginal increase in accuracy. Variation of around 25% between large and small numbers of dosimeters may be quite acceptable as long as reproducibility of phantom positioning within and across units is
high. A simplified phantom should be easily manufactured and affordable allowing any researcher or clinician to readily quantitate doses associated with different imaging alternatives. Furthermore, as Dr. Pauwels and his co-authors suggest, while comparison of doses following “standard” exposure protocols is useful, dosimetry is more valuable when acquired in the context of an indication of image quality. Low dose imaging of insufficient diagnostic quality for the intended diagnostic task must be avoided just as excessively high dose imaging should be avoided. Calculation of effective dose is a best estimate of risk; however, risk estimation must also be applied over a ‘level playing field’ if it is to be of value in clinical decision-making. Conflict of interest The author warrants that he has no conflicts of interest. References [1] Pauwels R, Beinsberger J, Collaert B, et al. Effective dose range for dental cone beam computed tomography scanners. European Journal of Radiology [Epub ahead of print; 2010 Dec. 31]. [2] Ludlow JB. Effective doses of NewTom VGi variable volume dental CBCT unit. Abstract #32 presented at the annual meeting of the American Association of Dental Research (AADR), 2012; http://iadr.confex.com/ iadr/2012tampa/webprogram/Paper157080.html, accessed on June 16, 2012. [3] Ludlow JB. Dose and risk in dental diagnostic imaging: with emphasis on dosimetry of CBCT. Korean Journal of Oral and Maxillofacial Radiology 2009;39(4):175–84. [4] Ludlow JB, Ivanovic M. Comparative dosimetry of dental CBCT devices and 64slice CT for oral and maxillofacial radiology. Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontology 2008;106(1):106–14.
John B. Ludlow ∗ North Carolina Oral Health Institute, Koury Oral Health Sciences, Room 5411-K, 385 South Columbia Street, Chapel Hill, NC 27599-7455, United States ∗ Tel.: +919 537 3141. E-mail address: [email protected]
18 October 2011
Contents lists available at SciVerse ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Letter to the Editor Comment on “Effective dose range for dental cone beam computed tomography scanners” To the Editor, I wish to congratulate Dr. Pauwels and his co-authors on their manuscript, “Effective dose range for dental cone beam computed tomography scanners,” which is currently in press in the journal [1]. The authors have undertaken an impressive survey of the dosimetry of many of the currently available CBCT models for dentistry. Because this will be read by many who are interested in a better understanding of the risk of CBCT examinations, it is an important contribution to the literature. To place this work in the context of previously published studies it is important to understand differences in the methods that were used and the potential impact of the authors’ approach on the internal validity of the data and its extensibility to previous and future studies. A large number of dosimeters (147 and 152) were employed for each of the 2 phantoms used in this study. The rationale for this was to insure an even spread over each of the radiosensitive organs in the head and neck area that are utilized in the 2007 ICRP calculation of effective dose. The intent was to produce dose estimates that were as accurate as possible. While this is a laudable goal, it is important to note that sampling strategy is just one of a number of factors that are important in measuring dose. Just as important are the actual location and extent of the radiosensitive organs that are being measured and the location of the field of view when exposing the phantom. Within patient populations, the location and size of tissues of interest can vary significantly. Positioning of patients for imaging is also subject to considerable variation depending on the operator and the positioning aids that are available with different units. While the current study attempted to position phantoms as closely as possible to a typical patient, use of different “local radiographic staff” for the different units may have resulted in inconsistent phantom positioning and important differences in dose measurements in some units. This is in fact suggested by an examination of the data for large FOV units. The effective doses reported for most of the large FOV units in the current study are in line with previously reported studies and unpublished data from our group; however, results for the NewTom
VGi are almost double that of our recent findings [2]. A potential source for this variation is the high thyroid dose seen in the current study. It is useful to compare and contrast thyroid and salivary gland doses as both are important components of the effective dose calculation. With a properly positioned patient in a large FOV, the salivary glands are directly exposed by the primary beam. In contrast, the thyroid gland is below the field of direct exposure. For a given field size the amount of scatter radiation is directly proportional to the amount of primary radiation. As mAs is increased or reduced, the ratio of scattered to primary radiation will be a constant. For a RANDO phantom the ratio of salivary gland dose to thyroid dose will remain constant as mAs varies. Changes in kVp, filtration, and shape of the FOV for CBCT units with similarly sized large FOVs have a relatively small influence on this ratio. For the most part the foregoing theoretical discussion is confirmed by data from the current study. Examination of the equivalent doses for the organs in the following table is revealing. A mean salivary gland/thyroid dose ratio of 4.5 is seen for the 8 units that were measured in this study (Table 1). Although ratios for 7 of the 8 units tested are within a standard deviation of that mean, the ratio for the NewTom VGi is 2.3 standard deviations below the mean. The thyroid dose (2045 Sv) is about 3 times the average thyroid dose for all machines. It is also more than 4 times greater than the dose found in my evaluation of the unit seen in the first column of data. This data is based on imaging geometry seen in Fig. 1. A likely reason for the difference in the estimate of dose between the two studies is a more caudal position of the FOV in the Pauwels et al. study. It is also likely that the FOV is lower for the VGi than for the other units in the current study. Lower positioning of the FOV may result in direct exposure of the thyroid gland. Because this gland has a weighting factor of 0.04, it makes a significant contribution to the total effective dose calculation. Indeed, dividing the reported salivary gland dose by the mean salivary gland/thyroid ratio, we might estimate an expected thyroid dose of 640 Gy. The difference in expected and reported doses accounts for approximately 56 Sv of excess effective dose that is due to direct exposure of the thyroid in this case. This excess represents 29% of the reported total effective dose. While higher patient positioning in the FOV can occur during
Table 1 Equivalent dose and effective dose comparisons for large FOV examinations. Measurement
Thyroid dose (Gy) Salivary gland dose (Gy) Salivary gland/thyroid ratio Number of std. dev. from mean: thyroid Effective dose (Sv)
Study Ludlow NewTom VGi
Pauwels NewTom VGi
Galileos
iCAT NG
Iluma
Kodak 9500
NewTom VGi
Scanora 3D
Sky View
Mean (SD)
447 1891 4.2 −0.4 100
2045 2855 1.4 2.3 194
380 2104 5.5 −0.5 84
355 1830 5.2 −0.6 83
1230 7225 5.9 0.9 368
585 2676 4.6 −0.2 136
354 1690 4.8 −0.6 83
296 1568 5.3 −0.7 68
474 1582 3.3 −0.4 87
685 (583) 2602 (1793) 4.5 (1.4)
The value in bold is the focus of arguments within the manuscript. 0720-048X/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2012.05.027
4220
Letter to the Editor / European Journal of Radiology 81 (2012) 4219–4220
Fig. 1. Location of RANDO phantom for dosimetry of NewTom VGi in Ludlow study. Position of FOV includes minimum amount of anatomy below soft tissue contour of chin. Location of thyroid dosimeters is below the lowest directly exposed level of the phantom.
imaging, this is not an intrinsic property of the VGi unit which allows independent control of chin rest position and FOV location when positioning patients. The issue of susceptibility of measurement of thyroid dose to even subtle differences of phantom positioning has been discussed previously where it was shown that a rotation of even 10◦ of a RANDO phantom position can result in a 92% difference in thyroid dose for a large FOV [3]. Because the current study seeks to report representative doses for the different units under comparison, it seems unfair to penalize an individual unit with a large thyroid exposure that is based on significantly different FOV location than other units. In the current study a recalculation of organ doses using a subset of 24 dosimeters positioned similarly to protocols followed in our studies [4], demonstrated deviations in organ dose estimations by 18–28% with differences up to 80% in comparison to the full complement of dosimeters. The authors suggest that this finding of variability indicates that a large number of TLDs is needed for accurate effective dose estimation. However, the authors do not provide specific results of calculations of either equivalent doses or effective dose for the 24-dosimeter subset evaluations, so it is not possible to explore potential reasons for the observed variation. From the discussion of the above data, it is quite possible that some or even a large proportion of the variation is related to phantom position within the FOV rather than limited numbers of dosimeters located in organs and tissues of interest. While use of additional dosimeters may increase the precision of calculation of organ dose, it does not guarantee an increase in the accuracy of calculation of effective dose. Having made this argument I would like to suggest that future research in this field should be focused on the development of phantoms that can provide reasonable indicators of effective dose with fewer, rather than more dosimeters. Given the diversity of human craniofacial anatomy and the uncertainty of risk estimation of low dose exposures, the value of increased accuracy for a single anatomic representative must be considered in the context of the cost of time and resources needed to obtain a marginal increase in accuracy. Variation of around 25% between large and small numbers of dosimeters may be quite acceptable as long as reproducibility of phantom positioning within and across units is
high. A simplified phantom should be easily manufactured and affordable allowing any researcher or clinician to readily quantitate doses associated with different imaging alternatives. Furthermore, as Dr. Pauwels and his co-authors suggest, while comparison of doses following “standard” exposure protocols is useful, dosimetry is more valuable when acquired in the context of an indication of image quality. Low dose imaging of insufficient diagnostic quality for the intended diagnostic task must be avoided just as excessively high dose imaging should be avoided. Calculation of effective dose is a best estimate of risk; however, risk estimation must also be applied over a ‘level playing field’ if it is to be of value in clinical decision-making. Conflict of interest The author warrants that he has no conflicts of interest. References [1] Pauwels R, Beinsberger J, Collaert B, et al. Effective dose range for dental cone beam computed tomography scanners. European Journal of Radiology [Epub ahead of print; 2010 Dec. 31]. [2] Ludlow JB. Effective doses of NewTom VGi variable volume dental CBCT unit. Abstract #32 presented at the annual meeting of the American Association of Dental Research (AADR), 2012; http://iadr.confex.com/ iadr/2012tampa/webprogram/Paper157080.html, accessed on June 16, 2012. [3] Ludlow JB. Dose and risk in dental diagnostic imaging: with emphasis on dosimetry of CBCT. Korean Journal of Oral and Maxillofacial Radiology 2009;39(4):175–84. [4] Ludlow JB, Ivanovic M. Comparative dosimetry of dental CBCT devices and 64slice CT for oral and maxillofacial radiology. Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontology 2008;106(1):106–14.
John B. Ludlow ∗ North Carolina Oral Health Institute, Koury Oral Health Sciences, Room 5411-K, 385 South Columbia Street, Chapel Hill, NC 27599-7455, United States ∗ Tel.: +919 537 3141. E-mail address: [email protected]
18 October 2011