Comparative Analysis of Radiation Dose and Image Quality Between Thyroid Shielding and Unshielding During CT Examination of the Neck

Transcription

1 Medical Physics and Informatics Original Research Medical Physics and Informatics Original Research Young Hen Lee 1 Eun-tae Park 1 Pyong Kon Cho 1 Hyung Suk Seo 1 Bo-Kyung Je 1 Sang-il Suh 2 Kyung-Sook Yang 3 Lee YH, Park ET, Cho PK, et al. Keywords: bismuth shielding, CT, radiation dose DOI: /AJR Received May 11, 2010; accepted after revision June 28, This work was supported by a grant from Korea University, Seoul, Republic of Korea. 1 Department of Radiology, Ansan Hospital, Korea University College of Medicine, 516 Gojan 1-Dong, Danwon-Gu, Ansan-Si, Gyeonggi-Do , Republic of Korea. Address correspondence to Y. H. Lee (younghen@korea.ac.kr). 2 Department of Radiology, Guro Hospital, Korea University College of Medicine, Seoul, Republic of Korea. 3 Department of Biostatistics, Korea University College of Medicine, Seoul, Republic of Korea. AJR 2011; 196: X/11/ American Roentgen Ray Society Comparative Analysis of Radiation Dose and Image Quality Between Thyroid Shielding and Unshielding During CT Examination of the Neck OBJECTIVE. The purpose of this study is to evaluate the feasibility of thyroid shielding by measuring radiation dose, CT attenuation, and noise of superficial neck structures during CT examination. SUBJECTS AND METHODS. We divided 84 patients without abnormalities seen on CT into two groups depending on whether shielding with a cotton spacer was applied over the thyroid. On CT images, we measured the CT attenuation and noises in the strap and sternocleidomastoid (SCM) muscles. The superficial radiation dose was measured using a head CT dose phantom containing ionization chambers located at the 3, 6, 9, and 12 o clock positions. RESULTS. With thyroid shielding, the CT attenuation was significantly increased (shielded strap and SCM muscles, ± 19.2 HU and ± 31.8 HU, respectively; unshielded strap and SCM muscles, 84.1 ± 12.2 HU and 78.4 ± 10.1 HU, respectively; p < 0.05), whereas noise was unaffected (shielded strap and SCM muscles, 7.2 ± 4.2 HU and 10.8 ± 4.9 HU, respectively; unshielded strap and SCM muscles, 8.6 ± 4.9 HU and 10.7 ± 6.6 HU, respectively; p > 0.05). On the phantom study, the shield significantly reduced the superficial unshielded dose at the 12 o clock position only (27.5% reduction; p < 0.01). CONCLUSION. Below the shielded surface, thyroid shielding significantly reduced the superficial radiation dose of the neck without a remarkable noise increase while increasing CT attenuation. I n recent years, the greatly increased use of CT and the accompanying radiation exposure have raised concerns about potential radiation-induced malignancies [1]. According to a 2004 report, approximately 260 million CT examinations are performed annually worldwide, and CT accounts for up to 67% of medical radiation exposure [2]. Mazonakis et al. [3] reported that primary irradiation of the thyroid gland during a neck CT examination results in an absorbed dose in the range of mgy and an increased risk for the development of thyroid malignancies (up to 390 per million patients) [3]. Therefore, several radiologists and CT manufacturers have proposed strategies to reduce radiation exposure to patients undergoing CT, particularly infants and children [4]. Among the trials conducted, selective inplane shielding has been shown to effectively reduce the radiation doses to the lens, thyroid, and breast, which are susceptible to radiation exposure, by attenuating photons [5 13]. Previous studies have shown that the use of shielding optimizes the radiation dose without compromising image quality for CT of the brain and chest [5, 6, 9, 12, 13]. However, in most studies, imaging quality has been evaluated only using a phantom. Moreover, several others have reported that inplane shields impair imaging quality by increasing the noise or streaky artifacts near the shielded surface [8, 10, 11]. Although a reduction in radiation dose is an important issue, each CT examination should be individualized on the basis of clinical history, suspected disease, or patient size to provide an accurate and definitive diagnosis. To our knowledge, there are no published studies that have addressed the effect of thyroid shielding on image quality of neck CT. In the current study, we assessed the effectiveness of thyroid shielding by analysis of CT attenuation and noise, as well as radiation dose reduction by dose phantom measurement. AJR:196, March

2 Subjects and Methods Patients, Shielding, and CT Protocols The current study was conducted according to the guidelines of the institutional review board of the hospital, and written informed consent was received from each patient. This was a nonrandomized prospective interventional study. The participating patients were recruited from our hospital. During August and September 2007, 112 adult patients underwent contrast-enhanced neck CT examinations. In August 2007, we applied a shield over the anterior neck of all patients during routine neck CT examination to protect the thyroid gland. After a 1-month trial of thyroid shielding, the other patients underwent neck CT examinations without shielding. For evaluation of image quality, we ultimately enrolled 84 patients who did not have any abnormalities seen on CT that might alter the CT attenuation of surrounding tissues around the level of the thyroid gland by thorough reviews of clinical data and CT images. Twenty-eight patients were excluded from the study for the following reasons: acute inflammatory neck swelling, history of previous head and neck surgery, or radiation therapy. There were 45 men and 39 women enrolled in the study, with a mean age of 42.3 years (range, years). We divided the subjects into two groups according to whether a thyroid shield was applied to the anterior neck (44 patients, shielded group; 40 patients, unshielded group). There were no significant demographic differences between the two groups. A commercially available bismuth thyroid shield was used for the current study. The shield was fabricated from a double layer of bismuthimpregnated latex (0.060 mm Pb equivalent; AttenuRad, F & L Medical Products) that was placed with a 2.0-cm-thick radiolucent cotton spacer to decrease streak artifacts occurring near the contact surface with the shield [9]. Contrast-enhanced neck CT scans were performed using a 64-MDCT scanner (Brilliance 64, Philips Healthcare). An 80-mL bolus of nonionic contrast material (iohexol, 350 mg/ml iodine; Omnipaque, GE Healthcare) was administered into an antecubital vein with a power injector (MCT, Medrad). CT was initiated 40 seconds after the start of the injection. All studies were performed using the fixed-tube-current helical technique with constant settings (120 kvp, 250 mas, and mm detector configuration using a pitch of 0.987), and all images were reconstructed every 5 mm. CT Image Quality Assessment Retrospective image analysis was performed on a PACS (Piview, Infinitt) with the same softtissue algorithm display (window width, 300 HU; window level, 60 HU). A neuroradiologist with 8 years of experience evaluated the differences in CT attenuation values, which were quantified as the mean of attenuation values (expressed in Hounsfield units) and image noise as the SD of the attenuation value on the CT image, according to the presence or absence of a thyroid shield, using a previously reported method [10, 14]. To assess the effect of thyroid shielding on CT image quality, the strap and sternocleidomastoid (SCM) muscles were used for quantification of each mean CT attenuation and noise, on the basis of the rationale that the strap and SCM muscles are closely located to the thyroid shield and are constantly recognized with homogeneous attenuation on neck CT, even after the administration of contrast material. On the midaxial CT levels of the thyroid gland, homogeneous-appearing regions of the strap and SCM muscles were bilaterally chosen for a circular region of interest (ROI). The size of the ROI varied depending on the region measured, but was generally a 30-mm 2 circular area (Fig. 1). Phantom and Dose Measurement To determine the absorbed superficial radiation dose, we used a 16-cm diameter CT head dose phantom (model , Nuclear Associates) containing five holes for insertion of an ionization chamber (one in the center and four approximately 90 apart, at the 3, 6, 9, and 12 o clock positions) and 6 mm from the surface; Fig. 2). In addition, an accurate ionization chamber (model CT; Radical) and a radiation monitor controller (model 9095, Radical) were used for dose measurement. The physicist measured the individual absorbed dose of the four peripheral holes (3, 6, 9, and 12 o clock) in the CT phantom using settings of 120 kvp, 250 mas, mm detector configuration, and a 5-mm slice thickness, first without the shield and subsequently with the shield. In this dose phantom study, CT was performed with the same CT scanner (Brilliance 64, Philips Healthcare) used in the neck CT examination. The shield was situated from the anterior surface of the phantom by an identical thickness (2 cm) of the cotton spacer used during neck CT examinations. We measured radiation doses five times and then obtained the mean value of radiation doses for the shielded and unshielded phantom over the identical position. Statistical Analysis We used Student s t test to compare the mean CT attenuations and image noises of the superficial anterior neck muscles for the shielded group versus the unshielded group. Analysis of variance was performed to determine where the measured radiation doses were significantly reduced among the different locations in the CT phantom with the shield application. Multiple post hoc comparisons between the different groups were also performed using Dunnett s T3 test. All statistical analyses were performed with a standard PC software package (SPSS version 12 for Windows, SPSS). A Fig. 1 Measurement of mean CT attenuation and noise in unshielded (left) and shielded (right) neck CT images through level of thyroid. Regions of interest (circles) measuring approximately 30 mm 2 are drawn in indicated areas of strap (squares) and sternocleidomastoid (asterisks) muscles at same axial level in range of shield. 612 AJR:196, March 2011

3 Fig. 2 Photograph of CT radiation dose phantom measurement. After applying thyroid shield over anterior surface of phantom with 2-cm-thick cotton spacer, dose measurements are repeatedly performed using ionization chamber of four peripherally located holes with 3, 6, 9, and 12 o clock positions. p value less than 0.05 was considered to indicate statistical significance. Results Mean CT Attenuation Values and Noise Measured on Neck CT Images Table 1 summarizes the effects of the use of the thyroid shield on the mean CT attenuations and image noises of the superficial neck muscles. Using the shield application while keeping all other CT parameters constant, the measured mean CT attenuation values determined by ROIs drawn in the strap and SCM muscles were ± 19.2 HU and ± 31.8 HU with shielding, and 84.1 ± 12.2 HU and 78.4 ± 10.1 HU without shielding, respectively. These differences between the shielded and unshielded neck CT examinations were statistically significant (p < 0.01), which implies that the use of the shield resulted in increased CT attenuation of the superficial neck muscles compared with the unshielded CT examination. However, the noises for the strap muscle and SCM muscles were 7.2 ± 4.2 HU and 10.8 ± 4.9 HU with shielding and 8.6 ± 4.9 HU and 10.7 ± 6.6 HU without shielding, respectively (Table 1). These differences in noises measured in the superficial neck muscles between shielded and unshielded CT examinations were not statistically significant (p = and p = 0.953, respectively). Examples of the measured values of the mean CT attenuation and noise on shielded and unshielded neck CT examinations are shown in Figure 3. Superficial Radiation Doses Measured With a CT Phantom The effects of the use of the shield on the superficial radiation dose are summarized in Table 2. The mean radiation dose measured with the use of a CT phantom was cgy (unshielded) to cgy (shielded) for the 12 o clock position, cgy (unshielded) to cgy (shielded) for the 3 o clock position, cgy (unshielded) to cgy (shielded) for the 6 o clock position, and cgy (unshielded) to cgy (shielded) for the 9 o clock position. To determine the dose reduction effect by bismuth shielding, the following formula was used: [(unshielded dose shielded dose) / unshielded dose] 100. These results indicate that the average CT dose index reduction resulting from bismuth thyroid shielding was 5.0%, 3.4%, 7.7%, and 27.5% at the 3, 6, 9, and 12 o clock positions, respectively. The phantom study showed that the use of the shield significantly reduced the superficial unshielded dose at the 12 o clock position compared with the other directions (3, 6, and 9 o clock positions, Table 3; p < 0.05). Discussion Selective shielding of radiation-sensitive organs during a CT examination was proposed by Hopper et al. in 1997 [15] and has been applied in several clinical situations [5 13]. Shields are made of thin sheets of flexible latex impregnated with bismuth and shaped to cover the lens, thyroid, and breast during brain, cervical spine, and chest CT examinations. These shields have been shown to reduce the radiation dose to the lens, thyroid, and breast by as much as 65.4% [6, 8], 47% [8, 9, 11], and 32% [5, 8, 9], respectively, without compromising image quality. However, there is a tendency to overestimate organ dose reduction by the use of a shield [4], and the effectiveness for image quality has been questioned because of image noise and streak artifacts [8]. Additionally, no prior study, to our knowledge, has quantitatively evaluated how the use of a thyroid shield affects image quality when located in an area under diagnostic investigation. The findings of the current study are consistent with those of prior studies showing that in-plane shields substantially decrease radiation dose [8, 9, 11]. Furthermore, we noted that the dose-reducing effect by the shield was particularly pronounced at the superficial surface just below the shield (i.e., the 12 o clock position) compared with the TABLE 1: Comparison of Mean CT Attenuation Values and Mean Noise Values Measured for Neck CT Images With or Without Thyroid Shield CT Attenuation CT Noise Location Unshielded (n = 40) Shielded (n = 44) p Unshielded (n = 40) Shielded (n = 44) p Strap muscle 84.1 ( ) ( ) < ( ) 7.2 ( ) 0.20 Sternocleiodomastoid muscle 78.4 ( ) ( ) < ( ) 10.8 ( ) 0.96 Note Data in parentheses are minimum maximum value in Hounsfield units in each muscle by drawing region of interest (~30 mm 2 ). AJR:196, March

4 Fig. 3 Examples of measured values of mean CT attenuation and noise on neck CT images. Compared with unshielded CT image (left), mean attenuations of both superficial neck muscles (strap muscle [squares] and sternocleidomastoid [asterisks]) are higher in shielded image (right), whereas image noises do not show significant difference. TABLE 2: Measurement of Mean Radiation Doses by Use of 16-cm Diameter CT Head Dose Phantom With or Without Shield Location Unshielded Dose (cgy) a Shielded Dose (cgy) a Dose Reduction (%) b 3 o clock o clock o clock o clock a Based on dose length product. b Defined as [(unshielded dose shielded dose) / unshielded dose] 100. TABLE 3: Comparison of Mean Differences of Reduced Doses at Different Ionization Chambers of CT Phantom Location Mean Difference of Reduced Doses (95% Confidence Interval) a Standard Error p 12 o clock and 3 o clock ( ) o clock and 6 o clock ( ) o clock and 9 o clock ( ) a Based on dose length product. other surfaces, such as lateral or deep to the shield (i.e., the 3, 6, and 9 o clock positions). However, the shield, as a radiation dose reduction tool, resulted in increased CT attenuation near the shield during neck CT examinations, while image noise was maintained. Similarly, these results are consistent with the findings reported by Kalra et al. [10], who reported that the shield application increased CT attenuation for the shielded surface in an anthropomorphic chest phantom study, which was likely related to a metal artifact caused by bismuth implanted within the shield. When metal implants are in the CT scan field, they lead to a degraded CT image quality as the result of a combination of beam hardening, photon starvation, scattering, and edge gradient effects [16]. The artifacts occur both near and far from the implants, because the density of the metal is beyond the normal range that can be handled by the computer, resulting in incomplete attenuation profiles [17] (Fig. 3). One phantom study by Hohl et al. [9] showed that an increase in image noise by the use of shields could be limited by applying a 1-cm spacer between the shield and the skin. Our results with neck CT image assessment showed that CT attenuation values were inevitably increased in superficial anterior neck muscles below the shielded surface, although a 2-cm cotton spacer successfully prevented an increase in noise associated with shield application. These results might provoke difficulties in neck CT image interpretation when a thyroid shield is present, especially for disease involving the superficial soft-tissue structures adjacent to the thyroid shield, such as cellulitis. Thus, the use of the shield should be carefully considered in CT examinations if the suspected abnormalities are located near the shield. In such instances, other methods should be used for the purpose of dose reduction, such as reducing the tube current, and direct contact of the shield over the skin should be avoided in all situations. With respect to several investigations that consider both CT image quality and dose, tube current modulation or simply reducing tube current can accomplish greater dose reductions than shield application for the resulting image noise [8, 11]. There are several limitations to our study. First, this study did not involve pediatric patients, who are the main concern for dose-reduction strategies during CT examinations; thus, the importance of the clinical implications might be lessened. Second, we attempted to measure the CT attenuation values of muscle, but not of any other structures, with and without thyroid shielding. Image contrast in CT is fundamentally determined by differential attenuation in different types of tissues, so it is unclear whether the measured changes in a single tissue type significantly affect the image contrast. Therefore, further investigation between two different tissue types (such as muscle and fat) is needed to determine whether these higher values merely represented a linear shift of all tissues (in which case, relative contrast differences would be preserved) or a more significant change in relative differences (e.g., perhaps the difference between fat and muscle is less in the images with the shield, causing a loss of contrast resolution). Third, the effects of the combined use of the bismuth shield and automatic tube current modulation must be further assessed regarding dose reduction. In conclusion, both phantom and neck CT image assessment studies have shown that shield application substantially reduces the 614 AJR:196, March 2011

5 radiation dose below the shielded surface without an increase in image noise. However, erroneously increased CT attenuation might be possible. Therefore, care should be exercised for shield application during neck CT examinations, particularly for abnormalities that might be near the shielded surface. References 1. Brenner DJ, Hall EJ. Computed tomography an increasing source of radiation exposure. N Engl J Med 2007; 357: Frush DP, Applegate K. Computed tomography and radiation: understanding the issues. J Am Coll Radiol 2004; 1: Mazonakis M, Tzedakis A, Damilakis J, Gourtsoyiannis N. Thyroid dose from common head and neck CT examinations in children: is there an excess risk for thyroid cancer induction? Eur Radiol 2007; 17: McCollough CH, Primak AN, Braun N, Kofler J, Yu L, Christner J. Strategies for reducing radiation dose in CT. Radiol Clin North Am 2009; 47: Fricke BL, Donnelly LF, Frush DP, et al. In-plane bismuth breast shields for pediatric CT: effects on FOR YOUR INFORMATION radiation dose and image quality using experimental and clinical data. AJR 2003; 180: Hopper KD, Neuman JD, King SH, Kunselman AR. Radioprotection to the eye during CT scanning. AJNR 2001; 22: Hopper KD. Orbital, thyroid, and breast superficial radiation shielding for patients undergoing diagnostic CT. Semin Ultrasound CT MR 2002; 23: Geleijns J, Salvado Artells M, Veldkamp WJ, Lopez Tortosa M, Calzado Cantera A. Quantitative assessment of selective in-plane shielding of tissues in computed tomography through evaluation of absorbed dose and image quality. Eur Radiol 2006; 16: Hohl C, Wildberger JE, Suss C, et al. Radiation dose reduction to breast and thyroid during MDCT: effectiveness of an in-plane bismuth shield. Acta Radiol 2006; 47: Kalra MK, Dang P, Singh S, Saini S, Shepard JA. In-plane shielding for CT: effect of off-centering, automatic exposure control and shield-to-surface distance. Korean J Radiol 2009; 10: Leswick DA, Hunt MM, Webster ST, Fladeland DA. Thyroid shields versus z-axis automatic tube current modulation for dose reduction at neck CT. Radiology 2008; 249: Mukundan S Jr, Wang PI, Frush DP, et al. MOS- FET dosimetry for radiation dose assessment of bismuth shielding of the eye in children. AJR 2007; 188: Yilmaz MH, Albayram S, Yasar D, et al. Female breast radiation exposure during thorax multidetector computed tomography and the effectiveness of bismuth breast shield to reduce breast radiation dose. J Comput Assist Tomogr 2007; 31: Wintermark M, Maeder P, Verdun FR, et al. Using 80 kvp versus 120 kvp in perfusion CT measurement of regional cerebral blood flow. AJNR 2000; 21: Hopper KD, King SH, Lobell ME, TenHave TR, Weaver JS. The breast: in-plane x-ray protection during diagnostic thoracic CT shielding with bismuth radioprotective garments. Radiology 1997; 205: Liu PT, Pavlicek WP, Peter MB, Spangehl MJ, Roberts CC, Paden RG. Metal artifact reduction image reconstruction algorithm for CT of implanted metal orthopedic devices: a work in progress. Skeletal Radiol 2009; 38: Barrett JF, Keat N. Artifacts in CT: recognition and avoidance. RadioGraphics 2004; 24: Unique customized medical search engine service from ARRS! ARRS GoldMiner is a keyword- and concept-driven search engine that provides instant access to radiologic images published in peer-reviewed journals. For more information, visit AJR:196, March