Assessment of volumetric absorbed dose for mobile fluoroscopic 3D image acquisition

Abstract Mobile fluoroscopy (c‐arm) units offering 3D image reconstruction are becoming more common in surgical settings. Although these images are “CT‐like” and sometimes replace the postoperative CT, the acquisition is technically very different from a traditional CT acquisition. Dose assessment is complicated by a large beam width, automatic exposure rate control, and a rotation of less than 360°. The purpose of this work was to explore the impact of these factors on the volumetric dose calculation and to provide practical recommendations for clinical physicists assessing dose from these units using commonly available equipment. CTDIW was calculated using the IAEA method for dosimetry of wide beams and compared to scans of the 32‐cm CTDI phantom using the full beam width and a 20‐mm collimated beam width. The impact of the partial rotation on the CTDIW calculation was assessed by acquiring measurements at four and twelve positions on the phantom periphery. For the system tested, the CTDIW was calculated to be 16.1 mGy using the IAEA method with default clinical protocol. Results showed that measuring CTDIW with the full beam width or a collimated beam width alone resulted in CTDI values of 19.0 mGy and 19.5 mGy, respectively. Using four peripheral measurements instead of 12 resulted in a difference of 4% for a collimated beam and 6% for an open beam. Variations in positioning on the order of a few centimeters resulted in a variation of only 4% with an open beam. The excellent reproducibility of the measurements using the full beam width suggests that this simple method is adequate for year‐to‐year comparisons. In contrast, the IAEA method is difficult to employ, particularly with 180° acquisitions. Use of peripheral measurements in excess of the usual four is time‐consuming and not necessary for most applications obtained with the geometry specific to this system.

a 180°rather than a 360°scan, and the data are acquired with a complicated tube motion consisting of two linear 7.5°translations and a 165°rotation. 2 The geometry of the system and a detailed description of the tube motion are well-described in other publications. 1,3 CTDI W is traditionally calculated using the weighted ratio of dose measurements made at the center and periphery of a standardized acrylic phantom, as follows: where CTDI 100 is the measurement acquired with a 100-mm-long pencil ion chamber. The value used for the periphery is either the average of measurements acquired at the four cardinal positions (3 o'clock, 6 o'clock, 9 o'clock, and 12 o'clock), or at the 12 o'clock position alone. For a rotation of less than 360°, however, it is questionable whether four positions are sufficient, since either one or two of the measurement points may be outside the beam depending on the relative positions of the x-ray tube arc and dosimeter. Previous studies have employed the use of four 4,5 and eight measurement points, 6-8 but they did not compare the accuracy of differing numbers of points. A study using a fixed c-arm with 200°rotation found a variation of up to 10% in the CTDI W calculated when measurements were acquired at the four cardinal positions and when they were rotated by 45°. 7 The use of the 100-mm pencil chamber was developed for narrow-beam CT scanners, but a number of modern CT units now have beam widths that exceed 100 mm. The traditional pencil chamber is insufficient to measure the entire beam in these scanners. Although alternate measurement techniques have been proposed to better measure wide beams, 6,[8][9][10][11][12][13][14][15][16]

2.A | Assessment of CTDI W
For this assessment, the IAEA method was used to calculate the CTDI W . The IAEA method requires collimation of the beam to a width as close as possible, but not more than, 20 mm. The Ziehm c-arm does not allow collimation during the 3D acquisition, so lead pieces were taped to the exit of the tube housing to collimate to a beam width of 19.6 mm at isocenter. The lead pieces were approximately 3-mm thick and of an appropriate size to cover the tube window. Based on the system geometry, the distance between the lead pieces required to achieve the required beam width was 8 mm at the tube housing (Fig. 3). The lead pieces were carefully adjusted to ensure an 8-mm gap across the entire length, and were not moved After acquiring the measurements in the phantom, the free-inair measurements were obtained following the IAEA instructions.
The detector was covered with a lead apron to increase the mA and reduce the risk of ghosting. Only one apron was used due to the difficulty of securely taping it to the detector during the rotation and the need to remove it during each preacquisition "collision check." Because the lead apron was not sufficient to drive the AERC to maximum, reproducibility was assessed again with three measurements, with a coefficient of variation of less than 1%. The mA displayed by the system was identical for all three measurements. For the free-in-air measurement with the uncollimated beam, the ion chamber was translated twice (for a total of three acquisitions) to provide complete coverage of the beam without overlapping.
For the collimated free-in-air measurement, the tube current was   Although a reasonable attempt was made to position everything accurately using the alignment lasers, less exacting attention was paid to perfect positioning during these trials. Because the purpose of this study is to determine factors important to the clinical physicist who may be testing such a unit, the effect of phantom positioning is an important factor to consider when comparing year-to-year results from annual testing.

2.C | Number of measurements required
For the purpose of routine evaluation, it is preferable to limit the number of measurements that need to be acquired. The data from the four cardinal angles acquired under each condition were used to calculate the CTDI W for comparison to the 12-angle scenario. This test was done both for the collimated and uncollimated measurements.

| RESULTS AND DISCUSSION
Using the IAEA method, the CTDI W-I was calculated to be 16.1 mGy ( commonly used for standard CT examinations because some organs within the scan volume will be exposed to far more radiation than others. In this study, measurements among the 12 different angles varied by over a factor of 50 (maximum at 90°and minimum at 300°), as compared to about a factor of 2 commonly seen in conventional CT. The specific way in which the dose distribution overlays the organs, and the radiosensitivity of these organs, must be consid-  only the dose at isocenter of a 16-cm phantom, but they found that doses of 3-10 mGy using a 100 kV beam provided adequate image quality depending on the imaging task. 22 None of these studies used the Ziehm system, but comparison to these results indicates that the Ziehm mobile c-arm performs within the range of dose levels observed from other comparable systems.
Only one other study comparing the IAEA method to phantomonly measurements was found: Gancheva et al. used a traditional (360°scan) CT unit having a 16-cm beam width and found that CTDI W-c was 11.3% higher than CTDI W-I and CTDI W-o was 41% lower than CTDI W-I. 23 It is initially surprising that CTDI W-o was found to be higher than CTDI W-I in the current study, as numerous studies have indicated that CTDI W-o for a wide beam is expected to be artificially low when using a 100-mm ion chamber and 150-mm acrylic phantom. 5,7,[9][10][11][12][14][15][16]23 Rather than contradicting the existing literature, this is most likely an indication that the IAEA method is not ideal for use with 180°acquisitions. The study by Gancheva et al.   The use of greater than four peripheral measurements in the phantom is time-consuming and unnecessary with the geometry of this unit.

ACKNOWLEDGMENTS
The author would like to thank Justin B. Faul, RT, and Joshua L.
Gary, MD. Their assistance with this project was invaluable. The author is also thankful for the use of clinical imaging equipment at Memorial Hermann Hospital -Texas Medical Center in Houston, TX.
Publication of this article was funded in part by the University of Florida Open Access Publishing Fund.

CONF LICTS OF INTEREST
The author has no conflicts of interest to disclose.