Article Text


Original research
Increase in fluoroscopic radiation dose in successive sessions of multistage Onyx embolization of brain arteriovenous malformations compared with the first session
  1. Jae Jon Sheen1,
  2. Yuan Yuan Jiang2,
  3. Young Eun Kim1,
  4. Jun Young Maeng1,
  5. Tae-Il Kim1,
  6. Deok Hee Lee1
  1. 1 Department of Radiology, Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea (the Republic of)
  2. 2 Department of Biotechnology, Dongguk University, Ilsan, Korea (the Republic of)
  1. Correspondence to Dr Deok Hee Lee, Department of Radiology, Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul 05505, Korea (the Republic of); dhlee{at}


Background and purpose Onyx embolization is a treatment for brain arteriovenous malformations (AVMs). However, multistage embolization usually involves the presence of radiodense Onyx cast from the previous sessions, which may influence the fluoroscopic radiation dose. We compared the fluoroscopic dose between the initial and final embolization sessions.

Materials and method From January 2014 to September 2016, 18 patients underwent multistage Onyx embolization (more than twice) for brain AVMs. The total fluoroscopic duration (minutes), dose–area product (DAP, Gy×cm2), and cumulative air kerma (CAK, mGy) of both the frontal and lateral planes were obtained. We compared the frontal and lateral fluoroscopic dose rates (dose/time) of the final embolization session with those of the initial session. The relationship between the injected Onyx volume and radiation dose was tested.

Results The initial and final procedures on the frontal plane showed significantly different fluoroscopic dose rates (DAP: initial 0.668 Gy×cm2/min, final 0.848 Gy×cm2/min, P=0.02; CAK: initial 12.7 mGy/min, final 23.1 mGy/min, P=0.007). Those on the lateral plane also showed a similar pattern (DAP: initial 0.365 Gy×cm2/min, final 0.519 Gy×cm2/min, P=0.03; CAK: initial 6.2 mGy/min, final 12.9 mGy/min, P=0.01). The correlation between the cumulative Onyx volume (vials) and radiation dose ratio of both planes showed an increasing trend (rho 0.4325–0.7053; P=0.0011–0.0730).

Conclusion Owing to the automatic exposure control function during fluoroscopy, successive Onyx embolization procedures increase the fluoroscopic radiation dose in multistage brain AVM embolization because of the presence of radiodense Onyx mass.

  • angiography
  • arteriovenous malformation
  • liquid embolic material
  • embolic
  • intervention

Statistics from


The recent advent of flat panel detector systems in DSA for various neurointerventional procedures has helped reduce the total radiation dose and improved the quality of images. However, concerns about radiation safety remain because the overall radiation burden has been increasing, mostly owing to the expanded indications of endovascular procedures and increased chances of multiple procedures in various situations. One of the tasks as neurointerventionists is to reduce or maintain the radiation dose even with increased use of technologies.1–5

Endovascular treatment of brain arteriovenous malformations (AVMs) using Onyx (Medtronic, Irvine, California, USA) requires a long procedure time owing to the nature of the Onyx embolization technique, and is a situation that requires our attention to the issue of radiation dose. Moreover, the procedure usually requires multiple sessions of embolization.6 7 Frequent users of Onyx are already aware of the problem and have been trying to reduce the radiation burden to patients.8 As most of the radiation burden occurs during the Onyx injection period under fluoroscopy, some neurointerventionists, including us, set the pulse rate of the fluoroscopic imaging system following the ‘as low as reasonably acceptable’ (ALARA) principle. Currently, we set the fluoroscopic rate at 4 frames/s, which is helpful in reducing the total radiation burden without causing difficulty in monitoring the Onyx injection flow.

Although we were satisfied with the protocol change (reduced fluoroscopic frame rates), we found that the radiation dose tended to be higher in successive embolization procedures than in the first session. We hypothesized that the presence of a radiodense Onyx cast in the angiographic and fluoroscopic fields in subsequent sessions could affect the acquisition setting that is automatically controlled by the automatic exposure control (AEC) system of the angiographic machine (figure 1).

The purpose of this retrospective study was to compare the radiation doses between the first embolization session, which was done without any pre-existing Onyx cast, and the last embolization session, which involves a certain volume of Onyx cast from the previous session(s), focusing on the fluoroscopic doses.

Figure 1

A middle-aged patient underwent cerebral angiography for evaluation of refractory seizure due to a large left parietal cortical arteriovenous malformation. (A) Three-dimensional angiogram. After a discussion on management planning, a combination of multistage embolization to reduce the nidus volume and stereotactic radiosurgery for the residual lesion were planned. (B) The first session of Onyx embolization was performed successfully, and the total fluoroscopic dose–area product (DAP) rate of the procedure was 0.44 Gy×cm2/min. The total fluoroscopic time and the DAP rate of the lateral plane were 41.2 min and 0.34 Gy×cm2/min, respectively. (C) The last (fifth) session of Onyx embolization was performed. More than half of the lateral fluoroscopic image field area was occupied by a very radiopaque Onyx cast. The total fluoroscopic DAP rate of the session was 0.83 Gy×cm2/min. The fluoroscopic time and the DAP rate of the lateral plane were 44.2 min and 0.84 Gy×cm2/min, respectively. The rates were more than double those in the first session. (D, E) A dense Onyx cast was noted in the lateral skull radiograph obtained for stereotactic radiosurgery of the residual lesion.



This retrospective study was approved by our institutional review board. From January 2014 to September 2016, 18 patients underwent multistage Onyx embolization (more than twice) for brain AVMs, mostly for volume reduction before gamma knife radiosurgery. The number of sessions ranged from 2 to 7 (two in 10 patients, three in five patients, five in two patients, and seven in one patient).


Under general anesthesia, we performed Onyx embolization with the use of a biplane neuroangiography system (Axiom Artis Zee; Siemens Healthcare, Erlangen, Germany). After placing a 6 F guiding catheter (Envoy; Codman Neurovascular, Raynham, Massachusetts, USA), a detachable tip microcatheter (Apollo; Medtronic) was inserted into the target portion of the nidus through the feeder. Under subtracted fluoroscopic monitoring, the liquid embolic material (Onyx 18; Medtronic) was injected using a standard injection technique after plug formation. The basic radiographic acquisition setting was 73 kV, pulse width was 40 ms, and a small focal spot was used with a dose setting of 1820 µGy/frame. The fluoroscopic imaging setting was 73 kV with a pulse width of 25 ms (45 nGy/pulse). The pulse rate of the fluoroscopy unit was set at 4 pulses/s for monitoring of the Onyx injection status. We used single plane monitoring (either frontal or lateral plane) during most of the Onyx injections. Biplane monitoring was also used intermittently. The magnification view was set to obtain optimal coverage of the field and visibility of the embolic material according to the operator’s discretion. Nidus occlusion status was checked intermittently either with a DSA run or brief manual contrast injection under a blank road map. We limited the number of Onyx vials to three in a single session to minimize potential complications from massive embolization.

Radiation dose data and statistical analysis

From the vendor provided radiation dose chart of each procedure, we collected several parameters related to both frontal and lateral plane fluoroscopic examinations. We obtained the total fluoroscopic duration (min), dose–area product (DAP, Gy×cm2), and cumulative air kerma (CAK, mGy) on both planes. For the dose comparison, unit values were produced by dividing those values by fluoroscopic time. We calculated the DAP rate (Gy×cm2/min) and CAK rate (mGy/min) of all fluoroscopic exposures required for both catheterization and monitoring of Onyx embolization status.

To investigate the relationship between the metal burden of Onyx (cumulative total number of vials injected until the last session) and radiation dose, we devised new values (DAP ratio and CAK ratio) by dividing the DAP and CAK values of the last session by those of the first session.

We compared the values between the initial and final sessions of the multistage embolization using a paired t test, and tested the relationship of DAP and CAK ratios with the cumulative volume (vials) of Onyx using the Spearman correlation (Stata/IC 13.0 for Mac; StataCorp LLC, Lakeway, Texas, USA).


There were 12 men and 6 women with multiple sessions of Onyx embolization for brain AVMs. Mean patient age was 36 (18–51) years. The mean largest diameter of the AVM nidus was 39 (14–72) mm. We used a total mean cumulative 6.5 (2–17) vials of Onyx during the embolization procedure. The mean fluoroscopic times for frontal and lateral projections were 31.6 and 33.5 min, respectively. Brain AVMs were distributed in the order of frontal lobe, fronto-parietal area, and parieto-occipital area. Patient baseline characteristics are summarized in table 1.

Table 1

Baseline characteristics of the patients

Compared with the initial fluoroscopic values of the DAP and CAK rates, significantly higher values were observed during the final sessions of embolization on both the frontal and lateral planes (table 2figure 2).

Table 2

Paired t test comparison of the initial and final sessions in the same patient on both frontal and lateral planes

Figure 2

Box plots of the paired t test results of the doses at the initial session and the final session in the same patient on both frontal and lateral planes. (A) Comparison of the fluoroscopic dose–area product (DAP) rate (Gy×cm2/min) on the frontal plane. (B) Comparison of the fluoroscopic DAP rate (Gy×cm2/min) on the lateral plane. (C) Comparison of the fluoroscopic cumulative air kerma (CAK) rate (mGy/min) on the frontal plane. (D) Comparison of the fluoroscopic CAK rate (mGy/min) on the lateral plane.

The correlation between the cumulative volume of Onyx (in vials) and the radiation dose ratios was significant, except for the CAK ratio on the frontal plane (table 3, figure 3).

Table 3

Correlation between cumulative Onyx vials and radiation dose ratio (last session dose to first session dose)

Figure 3

Scatterplots between the cumulative volume of Onyx (in vials) and radiation ratio. (A) Cumulative Onyx volume and dose–area product (DAP) ratio on the frontal plane. (B) Cumulative Onyx volume and DAP ratio on the lateral plane. (C) Cumulative Onyx volume and cumulative air kerma (CAK) ratio on the frontal plane. (D) Cumulative Onyx volume and CAK ratio on the lateral plane.


As embolization of brain AVMs or dural arteriovenous fistulas (DAVFs) by using a liquid embolic material usually requires an extended procedural duration, potential adverse effects of radiation could occur, such as hair loss, although such problems are usually temporary. Our current observation of increased radiation dose, particularly from the fluoroscopic examination, demands attention and vigilance, especially in multiple sessions of embolization, which are common in the endovascular treatment of brain AVMs and DAVFs.

As part of the radiation dose reduction efforts based on the ALARA principle, we lowered the fluoroscopic pulse rate to 4 pulses/s, in addition to diligent application of collimation with limited use of image magnification. Even after applying these schemes, the dose from the embolization procedures with the liquid embolic material remains relatively higher than that from other procedures, such as aneurysm embolization or stenting. While reviewing the dose charts of Onyx embolization cases to seek additional areas of further dose reduction, we found that the doses from subsequent embolizations with Onyx tended to be higher than the dose at the initial session, which prompted us to perform a statistical comparison in our patient cohort.

The presence of a high density residue in the fluoroscopic examination field is considered one of the major disadvantages of the use of radiopaque embolic materials, as a pre-existing radiopaque mass hinders the proper monitoring of the procedure and serves as a source of artifacts in follow-up imaging. Different from other radiopaque embolic materials, Onyx has demonstrated the additional problem of increasing the radiation dose owing to its tendency to occupy a larger volume than coil mass, and because of its higher density than iodine based liquid embolic materials.

As we usually use the AEC setting of the fluoroscopic machine in most of our procedures, the presence of attenuating material can increase the dose production setting to maintain the image brightness controlled by automatic exposure algorithms.9 A general analogy for this problem is overexposure of a hip radiograph in a patient with a hip prosthesis. Sometimes the presence of contrast can affect the AEC. A Japanese study in 2004 showed that the skin radiation dose increased under fluoroscopy and angiography when the bladder was filled with excreted contrast material. In that study, the authors recommended the removal of urine through urethral catheterization during the procedure.10 The presence of a radiopaque Onyx cast under angiography and fluoroscopy would affect the acquisition setting. As a result, we found that multistage Onyx embolization of brain AVMs increases the radiation dose.

Although the AEC system plays a negative role in this situation, optimization of the AEC system may ultimately lead to a more effective dose reduction. AEC decreases the radiation dose exposure and optimizes the image in the angiography suite, allowing a constant image quality at the lowest dose.11 12 For the acquisition setting, it is highly likely that the AEC system of the angiography machine is controlled automatically. During an interventional procedure, the fluoroscopic dose rate is generally controlled by the AEC of the angiographic machine, and should be considered in relation to an increase in the dose rate under various conditions. Narrow collimation reduces the scatter in the image, which results in reduced image brightness and occurs at higher doses. The machine typically increases the dose rate as a compensatory mechanism. The system will assume that the image brightness is decreasing and will also increase the dose rate when the field is collimated to block part of the area that is used to control image brightness.13

As repeat or multistage embolization with a liquid embolic material has become an important treatment modality for otherwise difficult to treat complex cerebrovascular lesions, a solution other than basic radiation dose reduction needs to be found. The simplest way would be to minimize the inclusion of the radiopaque portion in the fluoroscopic field as much as possible. The disappearance of the ‘burning effect’ in the changed field would be a good sign of an appropriate exposure setting. This might avoid the overexposure problem while obtaining better fluoroscopic image quality with less fluoroscopic dose. This is possible because the proposed modification places the Onyx mass outside of the measurement field of AEC monitoring.

Understanding the details of the AEC system and its measurement field would be helpful in finding a solution to the reported problem. A highly absorbing object increases the radiation dose and reduces the image quality by affecting the AEC system. The change in the measurement field to a less radiodense area can lower the radiation dose and increase image quality. For the Siemens Artis Zee unit, the AEC systems of the frontal and lateral planes are controlled by different schemes. We noticed that the frontal plane provided multiple measurement field schemes, whereas the lateral plane provided only two options for choosing a rectangular or round measurement field at the center of the image field. Unfortunately, we prefer using the lateral plane during Onyx injection in most cases. The only way to minimize the burning effect and avoid increasing the radiation dose would be to place the radiopaque mass outside of the rectangular measurement field, which may limit the technical options during embolization. It would be helpful if the manufacturer could install similar diverse measuring field options for the lateral plane also. In addition, the newly introduced liquid embolic material PHIL (precipitating hydrophobic injectable liquid; Microvention, Tustin, California, USA) with less radiopacity may cause less radiation increase, but more studies are needed to evaluate its safety and efficacy.14

We also collected dose measurements such as DAP and CAK in angiograms and obtained the number of angiographic exposure frames. Both CAK and DAP were divided by the angiography frames for the dose comparison. However, these data could not show significant results because there are many confounding variables that need to be accounted for. The first consideration is that various neurointerventionists performed the diagnostic angiographies for AVMs. Moreover, the diagnostic angiograms were obtained using different angiographic machines. There was no unified angiogram protocol for settings such as radiation dose mode, magnification, and collimation, followed by the various neurointerventionists. Therefore, we could not find significant results on the radiation difference of the angiographic run between the first and last sessions. In contrast, only one neurointerventionist performed the embolization procedures for AVMs which affected the analysis of the fluoroscopic radiation dose. Thus, we excluded parameters related to the angiographic run in this study.

The kV value during DSA and fluoroscopy was also a key data point of interest. The kV data for DSA exposures were saved in the final reports of the angiographic machines, although those from the fluoroscopy sessions would no longer be available for mining. Unfortunately, we could not find a significant change in kVP between the initial and final embolization sessions for AVMs. We supposed that the lack of a unified protocol, for settings such as magnification change and copper filter use, led to various confounding factors.

This study has several limitations. First, data were collected retrospectively without controlling for potential sources of bias. Second, we used the vendor provided radiation exposure data from the dose chart, which only showed the exposure dose parameters. We used indirect measures of dose, such as DAP and CAK, instead of direct dose measurement. Although these data are considered to indirectly represent the real radiation burden to patients, they must be interpreted with caution, especially when estimating the clinical significance of a certain DAP or CAK value. Third, we did not perform a systematic comparison of all sources of radiation exposure for Onyx embolization. Owing to the technical difficulty of a retrospective dose comparison, we decided to limit the study to comparison of fluoroscopic doses only.


Owing to the presence of a radiopaque Onyx cast from previous injections in multistage embolization of brain AVMs and DAVFs, successive Onyx embolization procedures increase the fluoroscopic radiation dose. Preventing the burning effect in fluoroscopic images through the proper setting of the AEC system, in addition to basic radiological procedural principles, such as active use of collimation and limited application of magnification, may help reduce the inadvertent radiation exposure burden. A better understanding of available AEC schemes is required to minimize this inadvertent radiation burden.


View Abstract


  • Contributors All authors made substantial contributions to the design of the study, and analysis and interpretation of the data for the work. All authors drafted and/or revised it critically, and provided final approval of the version to be published, and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Funding This work was partly supported by a research grant (AMC 2015-9114) from Dongkuk Pharma Co Ltd Korea.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval The study was approved by the Asan Medical Center’s institutional review board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement The authors are ready to share spreadsheets from their data acquisition and experimental set-up details on request.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.