Doses from pediatric CT examinations in Norway Are pediatric scan protocols developed and in daily use?

Transcription

1 Doses from pediatric CT examinations in Norway Are pediatric scan protocols developed and in daily use? Eva Godske Friberg * Norwegian Radiation Protection Authority, P.O. Box, Østerås, Norway Abstract. Doses to pediatric patients from CT examinations are known to be unnecessarily high if scan protocols developed for adult patients are adopted. This overexposure is most often not recognized by the operating radiographer, due to the digital behavior of the imaging system. Use of optimized size-specific pediatric scan protocols is therefore essential to keep the doses at an appropriate level. The aim of this study was to investigate the doses to pediatric patients from CT examinations and to evaluate the level of optimization of the scan protocols. Patient data, applied scan parameters together with the dose parameters volume computed tomography dose index (CTDI vol ) and dose length product (DLP) for examinations of the head, chest and abdomen were collected by means of a questionnaire from five university hospitals. The effective dose was estimated from the total DLP by use of region-specific conversion coefficients (E DLP ). Totally 6, 8 and 8 questionnaires were received for examinations of the head, chest and abdomen, respectively. Large variations in patient doses between the hospitals were observed, addressing the need for optimization of the scan protocols in general. Most of the hospitals applied successive lower mas with decreasing patient age for all scan areas, while the use of lower tube voltage for small patients and a higher tube voltage for large patients were more rarely. This indicates the presence, to a certain level, of size specific scan protocols at some Norwegian hospitals. Focus on developing sizespecific scan protocols for pediatric patients are important to reduce the doses and risks associated with pediatric CT examinations. KEYWORDS: Pediatric, CT examination, optimization, scan protocol, effective dose, dose length product (DLP).. Introduction Doses to pediatric patients from CT examinations are known to be up to three times higher if scan protocols developed for adult patients are adopted without adjusting the scan parameters according to the patient s size [-]. This overexposure is most often not recognized by the operating radiographer, due to the digital behavior of the imaging system. CT examinations are further recognized as a relative high dose examination compared to conventional X-ray examinations and CT are more and more taking over for conventional X-ray []. Pediatrics is also exposed to a greater risk of developing stochastic late effects like cancer due to their increased radiosensitivity compared to adults []. Practicing justification and development of optimized size-specific pediatric scan protocols is therefore essential to keep the doses at an appropriate level and to reduce the risks associated with the CT examination. The aim of this study was to estimate doses to pediatric patients from CT examinations of the head, chest and abdomen and to evaluate the level of optimization of the scan protocols for pediatric patients at Norwegian university hospitals.. Material and method Six university hospitals carrying out pediatric CT examinations were invited to participate in this study. The dosimetric quantities volume computed tomography dose index (CTDI vol ), dose length product (DLP) and effective dose (E) were used as dose indicators. Key data on local CT practice were collected by means of a questionnaire for examinations of the head, chest and abdomen for ideally patients within each scan area. Information about the patient (sex, age, weight and height), clinical indication, applied scan parameters together with the available dose parameters displayed on the scanner console, CTDI vol and DLP, were among the collected data. * Presenting author, eva.friberg@nrpa.no

2 All pediatric doses (CTDI vol, DLP and E) presented in this study are related to dose measurements performed in the standard head dosimetry phantom (6 cm in diameter). For examinations of the pediatric trunk, most CT scanners provide dose data related to the standard body dosimetry phantom ( cm in diameter). These doses were converted to the head phantom by multiplying with the ratio between normalized weighted CTDI (CTDI w ) measured in the head and body phantoms. The ratio between normalized CTDI measurements free in air (CTDI air ) for body and head scan field of view (SFOV) was also applied to account for potential differences in filtration between these two SFOVs. Scanner-specific normalized CTDI w and CTDI air measurements provided by ImPACT were used in this conversion []. Effective doses to pediatric patients were estimated from the total DLP by use of region-specific conversion coefficients (E DLP ) published by Chapple et al [6]. These conversion coefficients were expressed as a function of patient size, making them applicable to all patient sizes. If the patient s weight and height were missing in the questionnaire, average weight and height for the given age and gender, were used in the calculation of E DLP. These average growth parameters were taken from height and weight curves provided by the Norwegian maternal and child health centre.. Results Five of the six invited hospitals approved to participate in this study. Totally 6, 8 and 8 questionnaires were received for pediatric CT examinations of the head, chest and abdomen, respectively. Chest examinations were further divided into high resolution chest examinations (HRCT) (n=) and ordinary chest examinations (n=8) due to fundamental differences in the scan protocols. Ten of the questionnaires were filled out incorrectly or incompletely and were excluded from the survey. The total number of questionnaires included for the different scan areas from each hospital are given in Table, while the total number of questionnaires within specified age groups for the same scan areas are given in Table. As can be seen, none of the hospitals managed to collect data for patients within each scan area. The response rate was highest for head examinations and more modest for trunk examinations for all hospitals. Two of the major CT vendors were represented in this study, covering totally 8 different CT scanner models (all helical). All the scanners were multi slice scanners giving (n=), (n=), 8 (n=), 6 (n=) and 6 (n=) simultaneous slices per rotation. Table : Number of questionnaires included for examinations of the different scan areas from the five hospitals. Scan area Total Head 6 Chest 6 8 Chest HRCT 6 Abdomen Table : Total number of questionnaires for the different scan areas within each specified age group. Age group/scan area Head Chest Chest HRCT Abdomen s 6 s 7 s 7 6 s s The mean and corresponding ranges of the dose parameters CTDI vol, DLP and effective dose for all pediatric patients for the different scan areas are summarized in Table. Analysis of the dose parameters with respect to patient age and different hospitals are shown in Figure to. Large variations in CTDI vol, DLP and effective dose within each scan area and age group were observed

3 between the hospitals. Large variations in individual patient doses were also observed for some hospitals, illustrated by the large error bars representing the standard deviation. Despite these huge variations some general trends in the different dose levels with respect to patient age could be pointed out. A reduction of CTDI vol with decreasing age was observed for all scan areas with some exceptions for the youngest patients. Only head examinations at hospital deviated significantly from this general trend with increasing CTDI vol with decreasing patient age. A similar trend of reduction in DLP with decreasing age was observed for all scan areas for all patient ages, with some deviations for head examinations at hospital and having a lower DLP for the oldest patents. When considering the effective dose, a decrease was observed with decreasing patient age until the age of two followed by an increase in dose by a factor of up to. for the youngest patients. Mentionable is the significant increase in effective dose with decreasing patient age over the whole age range observed for head examinations at hospital. In this particular case, babies received an effective dose almost a factor higher then for 6 years old patients. Typical effective doses for pediatric patients were.8 msv (.-.6),. msv (.-.),.6 msv (.-.) and. msv (.-.8) for examinations of the head, chest HRCT, chest, and abdomen, respectively. Examinations of the trunk were associated with the highest effective doses. Table : Mean CTDI vol, DLP and effective dose for all pediatric patients for the different scan areas. The corresponding ranges are given in brackets. Scan area (all ages) CTDI vol [mgy] DLP [mgycm] E [msv] Head (-96) (-8).8 (.-.) Chest. (.-7) (6-69).6 (.-.9) Chest HRCT. (.7-) 7 (9-7). (.-.) Abdomen 8. (.-9) 8 (8-). (.-) Figure : Mean CTDI vol for the different scan areas divided on specific age groups for the five hospitals. Error bars represents the standard deviation. CTDIvol [mgy] Head CTDIvol [mgy] Abdomen CTDIvol [mgy] Chest CTDIvol [mgy] Chest HRCT

4 Figure : Mean DLP for the different scan areas divided on specific age groups for the five hospitals. Error bars represents the standard deviation. DLP [mgycm] 8 6 Head DLP [mgycm] Abdomen DLP [mgycm] 6 Chest DLP [mgycm] 6 Chest HRCT Figure : Mean effective dose for the different scan areas divided on specific age groups for the five hospitals. Error bars represents the standard deviation. Effective dose [msv],,,,,, Head Effective dose [msv] 6,,,, 8, 6,,,, Abdomen Effective dose [msv] 6,,,,,, Chest Effective dose [msv] 6,,,,,, Chest HRCT,,

5 To evaluate the level of optimization of the scan protocols according to patient size, the average applied mas-product and tube voltage from each hospital were plotted as a function of patient age for the different scan areas, see Fig.. Most of the hospitals applied successive lower mas with decreasing patient age for all scan areas, with exceptions of head examinations at hospital. Use of automatic exposure control (AEC) did not necessarily result in a mas reduction with decreasing patient age as might be expected. The absolute value of the applied mas-product within each scan area varied by maximum a factor of. to.9 between the hospitals, depending on the different age groups. The majority of the hospitals used a tube voltage of kv for examinations of all patient ages, while the use of a lower tube voltage for the smallest patients and higher tube voltage for the largest patients were more rarely. Mentionable is the opposite trend observed for head examinations at hospital and, applying the lowest tube voltage at the largest patients. The trends in reduced mas and kv with decreasing patient age observed in some hospitals indicate the presence, to some level, of optimized scan protocols with respect to patient age. Figure : Mean mas-product and tube voltage from each hospital as a function of patient age. mas-product Head Chest HRCT Chest Abdomen Tube voltage [kv] Head Chest HRCT Chest Abdomen Typical numbers of subsequent scan sequences, applied pitch and use of AEC are summarized in Table. The variation in these parameters was larger between different hospitals than between different age groups within each hospital. As can be seen, examinations of the head are associated with the lowest pitch and highest number of subsequent scan sequences and lowest use of AEC. AEC were most frequent used for examinations of the abdomen. Variations in applied collimation were also observed among the hospitals, but not analyzed any further since these data were incorrect or insufficient filled out in the questionnaires. No attempt was done to categorize the examinations after the clinical indication given in the questionnaires, because they were too vague and imprecise. Most common clinical indications were trauma, malignancy (including controls), infection and different respiratory disorders. The mean E DLP used in the effective dose estimations in this study are summarized in Table. A decrease in E DLP with increasing patient age was observed for all scan areas, as expected.

6 Table : Average number of subsequent scan sequences, applied pitch and use of AEC for the different scan areas. The corresponding ranges are given in brackets. Scan area No. of sequences Pitch Use of AEC Head.7 (-).9 (.-.7) % Chest. (-). (.7-) % Chest HRCT. (-) 8 % Abdomen. (-). (.969-) 7% Table : Mean E DLP based on actual patient height and weight used in the estimation of effective doses for the different age groups and scan areas. Scan area / age group Head Chest Chest HRCT Abdomen upper part pelvis part DISCUSSION Dose estimations from pediatric CT examinations are not an easy task and few dose surveys are carried out for pediatric patients. The main problem when estimating doses from pediatric CT examinations is the large variation in patient size. It is demonstrated that the absorbed dose to dosimetric phantoms of different size is increasing with decreasing phantom diameter [7-8]. It is therefore recommended that, irrespective of patient age and scan location, doses for all pediatric examinations should be expressed in terms of absorbed dose to the 6 cm diameter phantom (head phantom) [9]. Despite of this, most CT manufacturers still display the dose indicators (CTDI vol and DLP) related to the cm diameter phantom (body phantom) for all examinations of the pediatric trunk. Generally, CTDI w measured in the head phantom is about twice that for the body phantom, under similar exposure conditions []. For examinations of the trunk, the displayed consol CTDI vo l and DLP do not reflect the physical absorbed dose to pediatric patients. Regardless of this, they still constitute a useful tool in dose surveys and optimization of examination scan protocols. To ensure reliable dose readings, it is important that validation of the consol CTDI vol and DLP is integrated in the routine quality control of the CT scanner performed locally at the hospital by the medical physicist. The most common method for estimation of the effective dose is by use of DLP to effective dose conversion coefficients (E DLP ) for specific body regions. As already mentioned, the absorbed dose to the patient will increase with decreasing patient size. For estimates of effective dose to pediatric patients, these patient size-organ dose-effects has to be included in the E DLP. E DLP expressed as a function of patient size (e.g. equivalent diameter), will probably give the most reliable dose estimate for pediatric patients. It is crucial that the DLP used in the dose estimation is related to the same dosimetry phantom as the DLP used for deriving the E DLP. Most E DLP for pediatric patients are based on DLP related to the head phantom, and hence the consol DLP can not directly be used in effective dose estimations for examinations of the trunk. In these cases DLP has to be converted to the head phantom prior to use in effective dose estimations as done in this study. It is assumed that by using the DLP displayed on the scanner consol, the effect of additional rotations necessary for data interpolation at either side of the planned image volume in helical scanning (overscan) is included. Care has to be taken to include this effect if DLP is manually calculated from the applied scan parameters and own CTDI measurements. 6

7 This survey revealed large variations in all the dose parameters studied. The main reason for variations in CTDI vol were differences in the applied scan parameters like mas, kv, pitch and collimation. Physical differences between different CT scanners are also responsible for some of the observed differences in the CTDI vol values. Different CT scanners have different architecture and detector efficiency and may result in different patient doses for identical applied scan parameters by a factor of two to three. The observed variations in DLP are, in addition to the variation in applied scan parameters, explained by differences in scan length and number of subsequent scan sequences. Also the fact that the collected patient doses cover a wide range of clinical indications associated with different requirements to image quality, like contrast or spatial resolution, are also responsible for the large dose variations observed. Table 6 summarizes doses from this survey with doses obtained from a similar survey carried out in the United Kingdom (UK) []. Mean E DLP used in the effective dose estimation for the different age groups are also included. The Norwegian dose values were comparable with the doses obtained in the UK survey for head examinations being up to % higher at the most. On the other side, the UK dose values for chest examinations were a factor of. to higher than the Norwegian doses, indicating differences in local scan techniques between these to countries. The E DLP used for effective dose estimations were also comparable, being slightly higher in the UK survey for chest examinations. Differences in patient age and clinical indications between these two surveys may also explain some of the differences observed in the dose values. Table 6: Comparison of CTDI vol [mgy], total DLP [mgycm], effective dose [msv] and E DLP [msv/mgycm] from this survey (N) with those reported in the UK survey (UK). All dose parameters are related to the head phantom (6 cm in diameter). HEAD CHEST Age (mean ) (6. ) (. ) (6.8 ) (.6 ) UK N UK N UK N UK N UK N UK N E DLP CTDI vol DLP E Mean age of the patients within each age group from the Norwegian survey. The presence of size-specific scan protocols, to some level, was observed at most of the hospitals. The most significant evidence of this presence was the overall trend of mas-reduction with decreasing patient age observed in Fig.. On the other hand, the same figure illustrate that mas-adjustments are not a routine procedure for examinations of pediatric patients in a Norwegian hospital. When considering changes in tube voltage, some hospitals reduced the kv for small patients, but the majority used kv for all patient sizes. The combination of constant tube voltage and mas-reduction with decreasing patient size as a first approach to size-specific scan protocols for pediatric patients are also reported by others [, ]. The variation in absolute mas value are probably mainly due to lack of optimization of the scan protocol. On the other hand, some confusion between the physical and effective mas may have occurred in the filling out of the questionnaire. Some CT scanners provide the effective mas and some hospitals may have reported this value instead of the physical mas. The differences between these to mas-values are the pitch factor. Reporting of wrong mas could not alone explain the large variation in absolute mas-values and support the conclusion of lack of optimization of the scan protocols. Reporting of the wrong mas-value will not affect the doses analyzed in this study, since CTDI vol provided by the scanner already has taken care of the pitch correction. This survey revealed large variations in CT practice and level of size-specific scan protocols for pediatric patients among the hospitals, indicating an urgent need for optimization. Optimization of scan protocols is based on the balance between a reduction of radiation dose to the patient and maintaining 7

8 an image quality good enough to answer the diagnostic question. The process consist of adjusting the scan parameters according to the patient size, reducing the number of sequences and limiting the scan length to cover only the region necessary to answer the clinical question. The main parameters affecting the CTDI vol (representing the average dose in the irradiated slice) are mas, kv, pitch and collimation. In the optimization process of CTDI vol, it is important to have knowledge on how changes in the different scan parameters will affect radiation dose and image quality. Changes in mas will mainly affect image noise, while changes in kv will also affect the image contrast. Changes in pitch and collimation are even more complex, affecting spatial resolution and image artifacts. Unfortunately, there is nothing like a universal CT technique to be adopted, due to physical differences between scanners from different vendors (e.g. bow tie filters, focal spot to detector distance, detector efficiency, etc.). As mentioned earlier, the variation between different scanner models may result in different patient doses for identical scan parameters by a factor of two to three. New CT scanners usually maintain image quality at a much lower dose than old scanners, resulting in an unnecessary high dose if protocols designed for old scanner models are adapted to new scanners. Calibration of each individual CT scanners to establish the relationship between patient dose (mas) and image noise is important prior to the optimization process. It is therefore important to have a qualified medical physicist in diagnostic radiology locally at the hospital to assist in radiation output measurements and optimization procedures. The fact that X-ray attenuation decreases with decreasing patient size result in unnecessary high patient doses if the same mas-product are applied independent of patient size. For CT examinations of pediatric patients the applied mas should therefore be reduced to compensate for the increased patient dose due to a smaller body diameter compared to adults. Wong et al. has shown that the doses to pediatric patients from head CT could be reduced by as much as % relative to adults by reducing the mas and still obtain essentially the same image-to-noise ratio []. This support that there is still a large optimization potential for scan protocols of pediatric patients in Norwegian hospitals, especially for the youngest patients. CONCLUSION This survey reflects the current pediatric CT practice in Norwegian university hospitals. Large variations in scanning technique and the resulting patient doses were observed, addressing an urgent need for optimization with respect to patient size. The lack of size-specific scan protocols for pediatric patients in some hospitals give rise to concern, since pediatric patients are more radiosensitive than adults. The most effective way to reduce pediatric doses is to ensure that the CT examination is justified and to optimize the scan protocol with respect to scan parameters and reduce the scan length and number of scan sequences to a minimum. As a first step in the optimization procedure, the majority of the hospitals included in this survey can reduce their mas-product without undue loss of diagnostic information resulting in a significant doses reduction to pediatric patients. REFERENCES [] BRENNER, D.J., ELLISTON, C.D., HALL, E.J., BERDON W.E., Estimated risks of radiation-induced fatal cancer from pediatric CT, AJR 76 () 89. [] PATTERSON, A., FRUSH, D.P., DONNELLY, L.F., Helical CT of the body: are settings adjusted for pediatric patients? AJR 76 () 97. [] DONNELLY, L.F., et al., Minimizing radiation dose for pediatric body application of single detector helical CT: strategies at a large children s hospital, AJR 76 (). [] OLERUD, H.M., SAXEBØL, G., Diagnostic radiology in Norway from Examination frequency and collective effective dose to patients, Radiat. Prot. Dosim., 7 (997) 7. [] ImPACT CT patient dosimetry Excel spreadsheet (version.99x, October 6). Available from home webpage of the ImPACT (Imaging Performance and Assessment of CT scanners) evaluation centre of the DH Medicines and Healhcare products Regulatory Agency (MHRA), (.7.8). [6] CHAPPLE, C-L., WILLIS, S., FRAME, J., Effective dose in paediatric computed tomography, Phys. Med. Biol. 7 () 7. 8

9 [7] HUDA, W., ATHERTON, J.V., WARE, D.E., CUMMING, W.A., An approach for the estimation of effective radiation dose at CT in pediatric patients, Radiology (997) 7. [8] NICKOLOFF, E.L., DUTTA, A.K., ZHENG, F.L., Influence of phantom diameter, kvp and scan mode upon computed tomography dose index, Med. Phys. () 9. [9] SHRIMPTON, P.C., WALL, B.F., Reference doses for paediatric computed tomography, Radiat. Prot. Dosimetry 9 () 9. [] SHRIMPTON, P.C., LEWIS, M.A., DUNN, M., Doses from computed tomography (CT): Examinations in the UK review, NRPB W67, National Radiological Protection Board, Chilton (). [] MOSS, M., McLEAN, D., Paediatric and adult CT practice and patient dose in Australia, Australas. Radiol. (6). 9