Article Text
PDF
Abstract
Background One goal of increasing awareness of radiation dose is to encourage personal and technical modifications in order to reduce the radiation exposure of patients and staff.
Objective To analyze the radiation doses incurred during diagnostic cerebral angiography and the angiographic techniques practiced over a 4-year period, in order to demonstrate the effectiveness of implementing radiation dose reduction strategies.
Methods A retrospective review of the first 50 consecutive adult and pediatric patients undergoing diagnostic cerebral angiography each year from 2010 to 2013 was performed. Angiograms and procedure examination protocols were reviewed for patient age, gender, diagnosis, angiography techniques, fluoroscopy time, reference point air kerma (Ka,r in mGy), and kerma-area product (PKA in μGym2).
Results From January 2010 to June 2013, a total of 231 diagnostic cerebral angiograms were reviewed (200 adults, 31 children). Adult patients were aged from 19 to 94 years and included 77 men and 123 women. Pediatric patients were aged from 2 to 18 years and comprised 11 boys and 20 girls. Median Ka,r and PKA significantly decreased from 2010 to 2013 in adults (1867 mGy; 21 231 µGym2 vs 653 mGy; 7860 µGym2) and children (644 mGy; 6495 µGym2 vs138 mGy; 1465 µGym2), (p<0.001).
Conclusions Increased awareness and implementation of dose reduction strategies resulted in decreased radiation doses for diagnostic cerebral angiography both in adult and pediatric patients. The use of lower and variable digital subtraction angiography frame rates and tailored examinations contributed significantly to the reduced radiation doses observed during diagnostic cerebral angiography.
- Angiography
- Technique
Statistics from Altmetric.com
Introduction
Despite advances in non-invasive imaging, cerebral digital subtraction angiography (DSA) remains the ‘gold standard’ for evaluation of many cerebrovascular disorders. Beyond diagnostic purposes, radiographically guided neurointerventional procedures are increasingly used for the treatment of intracranial aneurysms and vascular malformations1 and are of great benefit as they are minimally invasive. The exposure of patients and staff to ionizing radiation, however, has become an increasing concern1 ,2 because of deterministic effects such as skin epilation, erythema, and desquamation,3 ,4 as well as other long-term risks, including cancer.5 ,6 Therefore, much effort has been devoted to physician education and to identifying methods for minimizing these risks.7 ,8
One fundamental component for reducing radiation exposure is adequate recording of radiation parameters for each procedure.9 A number of metrics have been developed for the measurement of radiation dose incurred during fluoroscopic procedures, including peak skin dose, reference air kerma (Ka,r), kerma-area product (PKA), and fluoroscopy time.9 At our institution, we now record the patient dose indicators Ka,r (in mGy), PKA (in μGym2), and fluoroscopy time (in minutes) in each procedure report. While we improved our documentation of these parameters, we simultaneously implemented specific techniques for dose reduction in 2010 (previously published),8 and began to observe a decreasing trend in radiation doses. This prompted us to conduct a more systematic review of the radiation doses incurred and angiographic techniques practiced during diagnostic cerebral angiography from 2010 to 2013 in order to demonstrate the effectiveness of implementing radiation dose reduction techniques.
Materials and methods
This study was approved by our institutional review board. A retrospective review of a single operator's prospectively collected database of diagnostic cerebral angiograms from 2010 to 2013 was carried out. Inclusion criteria included consecutive patients undergoing diagnostic cerebral angiography from 2010 to 2013, during the months of January to June each year. Children, categorized as ≤18 years, were considered as a distinct group. Patients who had undergone embolization procedures were excluded and also patients for whom the radiation dose reports were unavailable. A maximum number of 50 cases a year were considered for each group. All angiograms were obtained using a biplane angiography system (Artis Zee, Siemens, Germany).
Cerebral angiography and procedure examination protocols were reviewed for patient age, gender, diagnosis, angiography techniques (femoral artery evaluation, DSA frame rate, number of exposures, number of vessels catheterized, inclusion of aortic arch evaluation, 3D DSA, or DynaCT), fluoroscopy time, and patient radiation dose indicators: Ka,r (in mGy) and PKA (in μGym2).
Statistical analysis
The Department of Biostatistics of the Johns Hopkins Bloomberg School of Public Health assisted with the statistical evaluation. Differences in radiation doses and the total fluoroscopy time were calculated with the Kruskal–Wallis test. For the isolated comparison of adults younger and older than 50 years of age, a two-way analysis of variance was performed. For both tests, a p value <0.05 was considered significant.
Results
Patient demographics
From January 2010 to June 2013, a total of 231 diagnostic cerebral angiograms were reviewed including 200 adults and 31 children. Adult patients were aged from 19 to 94 years (median 54 years) and included 77 men and 123 women (table 1).
Adult patient demographic data
Pediatric patients were aged from 2 to 18 years (median 10 years) and included 11 boys and 20 girls (table 2). A wide range in the average and median age between years is noted for the pediatric population owing to the smaller sample size.
Pediatric patient demographic data
The most common diagnoses in adult patients were unruptured intracranial aneurysm (36%) and subarachnoid hemorrhage (18.5%, including aneurysmal and non-aneurysmal subarachnoid hemorrhage), whereas in children, moyamoya vasculopathy (35.5%) and arteriovenous malformation (AVM) (22.6%) predominated. The overall average dose per diagnosis in adult patients was highest for dural arteriovenous fistula (1531.7 mGy, 16 060.3 µGym2), followed by aneurysm (1208 mGy, 14 181.5 µGym2), and AVM (1115.5 mGy, 16 060.3 µGym2).
Angiography techniques
The specific procedural techniques listed below were analyzed by reviewing the examination protocols and the individual images for each angiogram on a dedicated radiology workstation.
Femoral artery access
Femoral arterial access was evaluated with a saved contrast injection performed under fluoroscopy in 93% of adults (n=172) and in 100% of children (n=26) rather than with DSA (n=13 in adults). DSA evaluation of femoral access was not performed after 2012.
DSA frame rate
The average DSA frame rate significantly decreased each year in both adults from 3.94 f/s (2010) to 2.04 f/s (2013) (p<0.001) and in children 4.22 f/s (2010) to 2.07 f/s (2013) (p<0.001) (figure 1). A combination of fixed and variable frame rates (VFRs) was used in all studies, the proportion of which depended on the pathology studied. The vast majority of exposures, however, were obtained using a VFR in both adults and children: adults: 95.8% (2010), 90.1% (2011), 93.5% (2012), 98.7% (2013); children: 93.8% (2010), 82.6% (2011), 71.9% (2012), 94.6% (2013). The default frame rates used in the VFR setting also decreased from 4 f/s VFR (4 f/s during the arterial phase, 2 f/s during the venous phase) in 2010 and 2011 to 2 f/s VFR (2 f/s during the arterial phase, 1 f/s during the venous phase).
Graph of average digital subtraction angiography (DSA) frame rate per year for adult and pediatric patients. The line graphs demonstrate a significant decrease in the average DSA frame rate from 2010 to 2013 in both adults and children.
Aortic arch imaging
Arch aortograms were obtained in 75 adult patients with a median age of 60 years, adding an average of 31.8 mGy (median 26 mGy; range 3.7–121 mGy) and 542.6 µGym2 (median 405.4 µGym2; range 87–2616 µGym2) per examination. None were obtained in the pediatric group.
3D DSA and DynaCT
A total of 103 3D DSA procedures were performed in adult patients from 2010 to 2013, range 20–33 per year. The overwhelmingly majority of 3D DSAs were performed for the evaluation of intracranial aneurysms (69.9%; unruptured n=62, ruptured n=10) followed by cases of non-aneurysmal subarachnoid hemorrhage (n=8). Less commonly, 3D DSA was performed for the following diagnoses: AVM (n=3), carotid stenosis (n=2), intraparenchymal hemorrhage (n=2), dural arteriovenous fistula (n=2), intracranial tumor (n=1), stroke (n=1), and sickle cell disease (n=1). In children, 3D DSA was less often used, performed twice from 2010 to 2013 for the evaluation of moyamoya vasculopathy (n=1) and AVM (n=1). The average dose of a 3D DSA procedure in adult patients was 93.1 mGy (median 91.0 mGy; range 54.2–163.7 mGy), 1933.2 µGym2 (median 1829.3 µGym2, range 566.1–3553 µGym2). In children, the average 3D DSA dose was 76.4 mGy (median 76.4 mGy; range 61.9–90.8 mGy), 1523.5 µGym2 (median 1523.5 µGym2, range 542.3–2504.4 µGym2).
A total number of 36 DynaCTs in adult patients were performed from 2010 to 2013, which tended to decrease in number from 15 in 2015 to seven in 2013. DynaCT was most commonly performed in patients with an AVM (27.8%, n=10), dural arteriovenous fistula (19.4%; n=7), or intracranial aneurysm (13.9%; n=5). In children, a total of six DynaCTs were performed, 50% of which were for the evaluation of an AVM. The average dose of a DynaCT in adult patients was 312.3 mGy (median 311.0; range 261–349 mGy), 7959.6 µGym2 (median 7906.4 µGym2; range 5013.3–9885.4 µGym2). In children, the average DynaCT dose was 249.4 mGy (median 274.5 mGy; range 40.5–329 mGy), 6714.4 µGym2 (median 7190.1 µGym2; range 1092.8–3088.2 µGym2).
Vessels catheterized
On average, the number of vessels catheterized steadily decreased from 2010 to 2013: 5.08 (2010); 4.86 (2011); 4.70 (2012); and 4.28 (2013) (p=0.119) in adult patients. No significant difference (p=0.961) was noted in the average number of vessels catheterized in children: 5.00 (2010); 5.20 (2011); 4.57 (2012); 4.78 (2013).
Number of exposures
Excluding 3D DSA and DynaCT, the average number of exposures for each adult patient decreased from 2010 to 2013: 11.4 (2010); 10.9 (2011); 10.4 (2012); 9.0 (2013).
In children, the same decreasing trend was observed: 9.6 (2010); 9.2 (2011); 9.1 (2012); 6.2 (2013).
Radiation dose indicators
Fluoroscopy time
In 2010, the median total fluoroscopy time for adults was 8.6 min per patient compared with 6.6 min in 2013, corresponding to a 23.3% reduction. Patient age affected total fluoroscopy times with an overall median fluoroscopy time of 5.8 min for adults aged <50 years versus 9.5 min for patients aged ≥50 years (p<0.001). The median total fluoroscopy time for children steadily dropped from 6.0 min per procedure in 2010 to 2.5 min in 2013 (p=0.012).
Ka,r (air kerma) and PKA (kerma-area product)
The median Ka,r and PKA for each adult patient significantly decreased from 2010 (1867 mGy, 21 231 µGym2) to 2013 (653 mGy, 7860 µGym2), (p<0.001) (figure 2). The median Ka,r and PKA per exposure per adult patient also decreased from 2010 to 2013: 87 mGy, 987 µGym2 (2010); 77 mGy, 905 µGym2 (2011); 44 mGy, 524 µGym2 (2012); 33 mGy, 393 µGym2 (2013). With respect to age groups (<50 and ≥50 years), no difference between younger and older patients was identified (p=0.070).
Box plots of reference air kerma (Ka,r) and kerma-area product (PKA) values in adult patients from 2010 to 2013. The box plots of the radiation metrics Ka,r (left image) and PKA (right image) per adult patient from 2010 to 2013 show a 65% and 63% dose reduction, respectively by 2013.
The same trend was seen in pediatric patients with the largest median Ka,r and PKA values in 2010 (644 mGy, 6495 µGym2) and the smallest in 2013 (138 mGy, 1465 µGym2) (figure 3). The radiation metrics fluctuate during this period, reflecting our small sample size and differences in patient age and weight. When adjusting the children's Ka,r and PKA by weight, both radiation metrics decreased from 2010 to 2013: 17 mGy/kg (Ka,r), 192 µGym2 (PKA) in 2010; 17 mGy/kg (Ka,r), 189 µGym2 (PKA) in 2011; 10 mGy/kg (Ka,r), 155 µGym2 PKA in 2012; and 5 mGy/kg (Ka,r), 68 µGym2 (PKA) in 2013, p<0.001 (Ka,r) p=0.008 (PKA), respectively.
Box plots of reference air kerma (Ka,r) and kerma-area product (PKA) values in pediatric patients from 2010 to 2013. The box plots of the radiation metrics Ka,r (left image) and PKA (right image) per pediatric patient from 2010 to 2013 show a 78.6% and 77.4% dose reduction, respectively by 2013.
Discussion
In this study we critically analyzed 200 adult diagnostic cerebral angiograms and observed a 65% (Ka,r) and 63.0% (PKA) decrease in radiation doses incurred during a 4-year period of increased radiation dose awareness beginning in 2010. In the 31 pediatric cerebral angiograms reviewed during the same period, the reduction was even more substantial: 78.6% (Ka,r) and 77.4% (PKA). These documented reductions in radiation dose confirm our anecdotal observations of a trend toward lower doses throughout this period and support the notion that increased radiation awareness and implementation of radiation dose reduction techniques are effective in reducing dose.
During this time, our documentation of radiation exposure improved by recording the patient dose indicators (Ka,r and PKA) in addition to fluoroscopy time in each patient's dictated report. Before 2010, we recorded only fluoroscopy time, which is an insufficient indicator of radiation dose. Furthermore, we began systematically reviewing the manufacturer-generated examination protocol after each case and uploading that data to our picture archiving and communication system (PACS) system. These improvements in data collection allowed for the analysis of our radiation doses after each procedure and enabled us to identify areas that would benefit from modifying operator-related techniques and equipment-related settings. As our cognizance of radiation dose grew, we began to monitor the Ka,r and PKA real time during each case on the display monitors in the angiography suite.
A number of factors can be attributed to the observed decrease in radiation doses achieved during diagnostic cerebral angiography from 2010 to 2013. The first is the result of a generalized increase in radiation consciousness and the Hawthorne effect, which has been shown to shorten fluoroscopy times10 and radiation doses.11 As our radiation awareness increased, we actively implemented radiation dose reduction strategies (previously published)8 in 2010 and, importantly, continued to analyze our doses and adjust our technique and equipment-related settings.7 We encourage readers to examine their radiation parameters (Ka,r, PKA, and fluoroscopy time) real time during each case and document them in their procedural reports as a first step toward analysis of their own practices.
In addition to increased awareness, modifications of many of our angiographic techniques during this period contributed to the radiation dose reduction. The evaluation of femoral artery access changed from DSA to a saved fluoroscopic cine8 ,12 in those patients who were candidates for an arterial closure device. In children, no DSA evaluations were performed and we moved from a saved fluoroscopic cine to a single saved fluoroscopic image of the micropuncture needle and wire without contrast, as no closure devices were used in children. The numbers of exposures and vessels catheterized per patient decreased throughout this time, reflecting a trend towards a more tailored examination per patient—specifically, using available imaging data from previous studies (CT angiography, MR angiography, or DSA) and not repeating evaluation of the aortic arch or carotid bifurcations as well as interrogating only those vessels supplying a known lesion, particularly for follow-up studies.
From 2010 to 2013, the default DSA frame rates decreased from 4 f/s to 2 f/s for both fixed and VFR settings. For AVMs, we decreased our initial ‘fast’ frame rate evaluation from 7.5 f/s to 6 f/s and returned to our default 2 f/s VFR for other vessels requiring interrogation not contributing to the AVM. While we used a VFR setting in 2010, the initial frame rates were set to 4 f/s during the arterial phase followed by 2 f/s during the venous phase, but these settings are not readily apparent on the console in the angiography suite. In 2011, we changed our VFR settings to 2 f/s during the arterial phase followed by 1 f/s during the venous phase. This underscores the need to routinely examine the default manufacturer settings, particularly after a system upgrade. Similar to Kahn et al,7 we have adopted routine use of the VFR for the majority of our studies, with selective use of faster fixed frame rates for lesions requiring greater hemodynamic stratification to delineate the angioarchitecture of high-flow lesions such as AVMs or high-flow arteriovenous fistulas.
Although these basic techniques have shown to significantly reduce radiation dose during diagnostic cerebral angiography, many opportunities remain for additional radiation dose reduction. This includes the creation of pediatric protocols for 3D DSA and DynaCT acquisitions as well as lower-dose DynaCT protocols. In addition to these technical considerations, the education of our practicing and training clinicians performing these procedures is a vital component of a successful radiation reduction program.
Previous studies have offered other suggestions for reducing radiation dose. Most recently Kahn et al7 agreed that altering factory settings on the equipment allows a significant decrease in radiation exposure. Levitt et al13 claim that increased awareness of lowering radiation dose during endovascular procedures is necessary, although the image quality may be reduced. However, our group and others have presented techniques that lower the radiation dose while maintaining a diagnostic level of image quality.7 ,8 A study by Balter et al14 agreed that in order to lower radiation dose, adjusting the frame rate is essential to maintaining image quality. Ultimately, it is each practitioner's responsibility to investigate his or her own practice and to limit unnecessary radiation exposure according to the ALARA (As Low as Reasonably Achievable) principle.2
Limitations of this study include a single operator's experience and no specific data on the dose per frame, dose per pulse, or whether or not radiation scatter grids were removed for pediatric cases. In addition, although a significant decrease in radiation dose was seen in adult cases, doses were not adjusted for body mass index. Despite these limitations, we have demonstrated that significant decreases in radiation doses incurred during diagnostic cerebral angiography can be easily achieved by a combination of modifications both to operator-related techniques and equipment-related settings.
Conclusion
Increased awareness and implementation of dose reduction strategies results in significantly lower radiation doses for diagnostic cerebral angiography in both in adult and pediatric patients. The following suggestions are based on the analysis of our angiographic techniques during this time period:
Record the radiation parameters Ka,r, PKA, and fluoroscopy time in all procedure reports. Monitor these factors real time during each case and strive to reduce these numbers while maintaining image quality and patient safety.
Use imaging data from previous studies to limit aortic arch evaluations and minimize the number of DSA exposures to answer specific clinical questions.
Evaluate femoral artery access with saved fluoroscopically acquired images rather than by DSA.
Consider a 2 f/s VFR as the default DSA frame rate for routine studies. Selectively choose a fixed higher faster frame rate (4 or 6 f/s) for the arterial feeders to high-flow lesions.
Although not specifically evaluated in this study, the creation of lower-dose protocols that reduce the default manufacturer settings for dose delivered per frame and dose per pulse7 are additional opportunities for even more dose reduction.
References
Footnotes
Contributors All authors contributed to this manuscript and satisfied the criteria for authorship.
Competing interests None declared.
Ethics approval Institutional review board.
Provenance and peer review Not commissioned; externally peer reviewed.