Dose Reduction in Tomosynthesis

Transcription

1 Musculoskeletal Imaging Original Research Becker et al. Dose Reduction in Tomosynthesis of the Wrist Musculoskeletal Imaging Original Research Anton S. Becker 1 Katharina Martini Kai Higashigaito Roman Guggenberger Gustav Andreisek Thomas Frauenfelder Becker AS, Martini K, Higashigaito K, Guggenberger R, Andreisek G, Frauenfelder T Keywords: radiography, radiometry, tomosynthesis, wrist joint, x-ray DOI: /AJR Received May 10, 2016; accepted after revision July 26, A. S. Becker and K. Martini contributed equally to this work. 1 All authors: Department of Radiology, Institute of Diagnostic and Interventional Radiology, University Hospital Zurich, University of Zurich, Raemistrasse 100, 8091 Zurich, Switzerland. Address correspondence to A. S. Becker (anton.becker@usz.ch). AJR 2017; 208: X/17/ American Roentgen Ray Society Dose Reduction in Tomosynthesis of the Wrist OBJECTIVE. The purpose of this study was to quantitatively and qualitatively determine the impact of radiation dose reduction on the image noise and quality of tomosynthesis studies of the wrist. MATERIALS AND METHODS. Imaging of six cadaver wrists was performed with tomosynthesis in anteroposterior position at a tube voltage of 60 kv and tube current of 80 ma and subsequently at 60 or 50 kv with different tube currents of 80, 40, or 32 ma. Dose-area products (DAP) were obtained from the electronically logged protocol. Image noise was measured with an ROI. Two independent and blinded readers evaluated all images. Interreader agreement was measured with a Cohen kappa. Readers assessed overall quality and delineation of soft tissue, cortical bone, and trabecular bone on a 4-point Likert scale. RESULTS. The highest DAP (3.892 ± Gy cm 2 ) was recorded for images obtained with 60 kv and 80 ma; the lowest (0.857 ± Gy cm 2 ) was recorded for images obtained with 50 kv and 32 ma. Noise was highest when a combination of 50 kv and 32 ma (389 ± 26.6) was used and lowest when a combination of 60 kv and 80 ma (218 ± 12.3) was used. The amount of noise on images acquired using 60 kv and 80 ma was statistically significantly different from the amount measured on all other images (p < ). Interreader agreement was excellent (κ = 0.93). Delineation of anatomy and overall quality were scored best on images obtained with 60 kv and 80 ma and worst on images obtained with 50 kv and 32 ma. The difference in delineation and quality on images obtained using 50 kv and 40 ma was not statistically significantly different compared with images obtained using 60 kv and 80 ma. CONCLUSION. Significant dose reduction for tomosynthesis of the wrist is possible while image quality and delineation of anatomic structures remain preserved. T omosynthesis is based on geometric tomography, which consists of reconstructing single planes from a set of projection images. The basic principle has been known for almost a century [1], although the term tomosynthesis was coined by Grant in 1972 [2]. Because of the isometric dataset in CT, tomosynthesis was fully replaced by the former in the 1980s. Recently, with increasing radiation dose awareness in CT, tomosynthesis has been rediscovered as a promising modality with higher diagnostic sensitivity than radiography at a comparable radiation dose [3]. Tomosynthesis is already widely used clinically in breast imaging, where it has been shown to reduce the recall rate in screening examinations [4], and it is an active field of research in thoracic and musculoskeletal radiology. In thoracic imaging, for example, tomosynthesis is considered to be a reasonable alternative to CT for serial follow-up imaging in patients with cystic fibrosis [5], because it provides similar information and its radiation dose is significantly lower than CT, almost equal to that from radiography [3]. Because cystic fibrosis mainly affects younger patients, a low or reduced radiation dose is preferable. Similarly, a low or reduced radiation dose is also favored in musculoskeletal radiology in settings such as wrist imaging in patients with inflammatory arthropathies like rheumatoid arthritis, psoriasis arthritis, or inflammatory bowel disease related arthritis, which may affect adolescents or young adults under 30 years and may require serial follow-up imaging for erosive subchondral bone destruction. In early stages of inflammatory arthropathies in which soft-tissue changes precede the osseous reaction, any modality relying on x-rays is suboptimal because of the inherently poor soft-tissue contrast. Instead, ultrasound or MRI is usually AJR:208, January

2 Fig. 1 Screen shot of noise measurement in tomosynthesis image. Screen shot of ROI placement for noise measurement in one specimen in upper right corner (top) for background noise and 3.5 cm below radiocarpal joint into soft tissue between radius and ulna (bottom). Perim = perimeter, ERMF = Estimated Radiographic Magnification Factor, avg = average, Dev = SD. preferred. Both of these modalities have other drawbacks: Ultrasound is heavily operator dependent, and MRI is associated with high costs and is not readily available in many parts of the world. The use of CT for follow-up series in young patients has been discouraged because of high radiation doses, although CT would depict erosions with a much higher sensitivity than radiography [6, 7]. In recent studies, standard tomosynthesis acquisition parameters were used that ranged from 50 kv to 60 kv and ma with radiation doses lower than CT but still slightly higher than standard radiography [6 10]. This decrease is already favorable for clinical applications such as replacing CT in some cases, but any further dose reduction would be of great benefit for the further acceptance of tomosynthesis as a clinical tool and alternative to radiography. Our hypothesis was that the radiation dose in wrist tomosynthesis can be further reduced without hampering overall image quality. Thus, the aim of our study was to evaluate different tomosynthesis acquisition parameters with reduced radiation doses to quantitatively and qualitatively determine the impact of the dose reduction on the image noise and quality. Becker et al. Materials and Methods Cadaveric Wrists Ethics committee approval was granted for imaging of six complete cadaveric wrists (three left, three right). The specimens originated from four subjects (two women, two men; age range, years; mean age, 81.5 years) who had donated their bodies to the Institute of Anatomy at the University of Zurich for medical research. The use of these specimens complied with local and federal legislation. All specimens had been fixated in Thiel solution (five specimens for 4 years, one specimen for 6 years), which maintains the normal physiologic soft-tissue density. Imaging Technique Tomosynthesis series were performed using a commercially available tomosynthesis unit (FDR AcSelerate, Fujifilm Europe Medical Systems). All specimens were fixed on the table in a posterior-anterior position. The position of the detector was fixed under the table, and no grid was used. A constant collimation of the irradiation field was set. Special metal markers were placed to the left and the right of the specimens in all images to computationally compensate vibrations and subsequent motion artifacts caused by the engine. The x-ray anode moved at constant speed from 20 to 20 (tube angle, 40 ) above the table and the specimen, acquiring images in dorsopalmar projection. The scan duration was 5 s. The number of exposures was 30; one exposure lasts 12.5 ms. Tube voltage was set to 50 or 60 kv and tube current to 80, 40, or 32 ma, resulting in a total of six different acquisitions, each with 36 images (2- mm increment) per specimen. The corresponding dose-area products (DAP) were obtained from the automatic electronically logged protocol. The radiation dose was calculated using the adapted formula by Noël et al. [8]. All images were transferred to the hospital PACS (IMPAX, AGFA Healthcare) and stored for further image analysis. Fig. 2 Bar graphs show dose-area product produced with tube voltage of 50 kv (left) and 60 kv (right) and varying tube current. Dose reduction by 71.3% was measured changing from standard 60 kv and 80 ma to 50 kv and 40 ma combination. Tube current reduction to 32 ma had only minimal impact on radiation dose (additional 6.7%). Dose-Area Product (Gy cm 2 ) Image Evaluation Quantitative evaluation For quantitative image evaluation, noise levels were measured by one author with 15 years of experience with quantitative image analysis. On the image focused on the midcarpal joint line, a 100-mm 2 circular ROI was placed 3.5 cm below the radiocarpal joint into the soft tissue between the radius and ulna (Fig. 1). The mean and SD of tissue density was taken for further statistical evaluation. In addition, the background noise was measured on the same plane in each corner of the image (Fig. 1). Qualitative evaluation Qualitative image evaluation was performed by two blinded and independent radiologists (12 and 15 years of experience as board-certified radiologists). The differently exposed tomosynthesis examinations of the six specimens were presented in a random order and both radiologists were additionally blinded to the radiation exposure data of the examinations. Each examination consisted of 36 individual tomographic images. Both radiologists were permitted to use diagnostic PACS workstation software features such as interactive window and level setting or magnification of images for comprehensive image evaluation. The latter was performed using 4-point Likert scales adapted from Bolte et al. [11] for the overall image quality and delineation of various anatomic structures: soft tissue (fat planes and skin), cortical surface (cortical and vascular edges) and individual trabeculae. A score of 1 (excellent) meant that the individual anatomic structure could be extremely well distinguished. A score of 2 (good) meant that the individual anatomic structure could be well distinguished. A score of 3 (sufficient) was defined as a structure that could be distinguished to a degree that still would allow a diagnostic assessment in the clinical routine. A score of 4 (poor) meant that the delineation of anatomic structures was not sufficient for a reliable diagnostic assessment [11] Tube Current (ma) 160 AJR:208, January 2017

3 Dose Reduction in Tomosynthesis of the Wrist Noise Tube Current (ma) Statistical Methods For the statistical analysis, we used R software (version 3.2.1, The R Software Foundation). Means and SDs for radiation dose, image noise, overall image quality, and delineation of Fig. 3 Box-andwhisker plots show noise levels produced with tube voltage of 50 kv (left) and 60 kv (right) and varying tube current. Noise level decreases as tube voltage and tube current increase. Differences in noise levels for all parameter combinations were statistically significant compared with clinical standard combination of 60 kv and 80 ma (far right) (p < 0.001). Line within boxes = mean, top line of box = 75th percentile, bottom line of box = 25th percentile, whiskers = 95% CI, circle = outlier. anatomic structures were computed and reported descriptively. Interreader agreement was determined using kappa2 (irr, cran.r-project.org/web/packages/irr/), which uses Cohen kappa values to account for agreements by chance. Kappa values were interpreted as follows: slight, κ < 0.20; fair, κ = ; moderate, κ = ; substantial, κ = ; or excellent agreement, κ > 0.80 [12]. Because of the advanced age of the specimen donors and the small number of specimens, we did not assume a normal distribution of our data. Hence, Likert scale ratings for the different combinations of tube voltage and current for image acquisition were compared using Mann-Whitney-Wilcoxon tests separately, as well as in a cumulated score in which the four individual scores were added up. DAPs were compared with a univariate ANOVA and t test. A Bonferroni correction was used (0.05 / 6 comparisons) and a p value of < was considered statistically significant. Bar and box plots depicting radiation dose, noise, and image quality in relation to the different tube voltage and current combinations were generated with the ggplot2 package [13]. Results Dose-Area Product, Dose, and Image Noise The mean DAP for all patients ranged between 3.89 and Gy cm 2 and the noise Fig. 4 Single plane of tomosynthesis examination with magnified in-plane structures. Magnification corresponds to actual size seen on diagnostic monitor. First row shows soft tissue, second row shows cortical bone, and third row shows trabecular bone. Image in first column was obtained using standard combination of 60 kv and 80 ma; second column, 50 kv and 32 ma; third column, 50 kv and 40 ma; and fourth column, 60 kv and 80 ma. Decreasing noise and increasing definition of anatomic boundaries can be appreciated from left to right (i.e., from lower to higher radiation dose). While there is visible increase in noise from protocol in third to second column, subjective quality was only significantly worse when compared with first column. AJR:208, January

4 Becker et al. TABLE 1: Radiation Dose-Area Products (DAP) With Different Imaging Parameters Parameter Combination Mean DAP ± SD (Gy cm 2 ) between 218 and 395 units. The values for the individual cadavers varied between 0.6 Gy cm 2 and 1.01 Gy cm 2 for images obtained using 50 kv and 32 ma and between 3.03 Gy cm 2 and 4.11 Gy cm 2 for images obtained using 60 kv and 80 ma. Table 1 shows the DAP and noise values in relation to tube voltage and tube current. Reducing the tube settings to 50 kv and 32 ma resulted in a 78% reduction of the DAP and an increase in tissue noise of 81% compared with the standard setting of 60 kv and 80 ma (50 kv and 32 ma: DAP = Gy cm 2, dose = mgy, noise = 395; 60 kv and 80 ma: DAP = 3.89 Gy cm 2, dose = mgy, noise = 218). Reducing the tube current from 80 ma to 40 ma and maintaining the tube voltage halved the DAP (Fig. 2). It also increased the noise by about 20% (Fig. 3). However, reducing the tube voltage from 60 kv to 50 kv and maintaining the tube current resulted in a DAP reduction of 55 60% with an increase of image noise of about 70% (Fig. 3). Comparing the DAP of all tube voltage and current combinations to the original standard 60 kv and 80 ma presetting, all abovementioned results for the DAP were significantly smaller (p < 0.001). There was Relative DAP Reduction (%) Mean Noise ± SD Relative Noise Increase (%) Clinical standard 60 kv and 80 ma ± NA 218 ± 12.2 NA Test acquisitions 60 kv and 40 ma ± ± kv and 32 ma ± ± kv and 80 ma ± ± kv and 40 ma ± ± kv and 32 ma ± ± Note All differences from the test acquisition parameters were statistically significant compared with the clinical standard (p < 0.001). NA = not applicable. also a statistically significantly higher noise value for tube voltage and current combinations when compared with the standard presetting (p < ). TABLE 2: Visual Score and p Values From the Qualitative Evaluation Qualitative Analysis of Image Quality and Delineation of Structures Interreader agreement was excellent for all factors examined (κ = ). Visibility of soft tissue, cortical vascular edges, and individual trabeculae were scored highest (score 1) for the combination of 60 kv and 80 ma and that of 50 kv and 80 ma. Although the delineation score for the 50 kv and 32 ma protocol was still good to excellent, it did significantly differ from that of the standard protocol (overall image quality: 1.8 for reader 1 and 1.7 for reader 2; p = ). At 60 or 50 kv and 80 ma, the visibility of all structures was scored as excellent. The overall image quality as well as delineation of anatomic structures was not significantly different for the protocols using 60 kv and 40 ma, 50 kv and 80 ma, or 50 kv and 40 ma protocols when compared with the standard 60 kv and 80 ma protocol. For the remaining protocols, the scores differed significantly (Table 2). Whereas trabecular bone retained excellent delineation with 50 kv and 40 ma, soft tissue and cortical surface was scored somewhat lower with a score of good (60 kv and 40 ma, 1.4 ± 0.2; 50 kv and 40 ma, 1.6 ± 0.5). Figure 4 shows a representative example with those three anatomic regions magnified; soft tissue, cortical bone, and trabecular bone lost some delineation with a tube current of 32 ma, which led to the large drop in cumulative scores at 32 ma seen in Figure 5. For both tube voltages at a tube current of 32 ma, the visibility score for soft tissue was lower, as well as for trabecular bone only with 50-kV tube voltage, compared with images obtained with the standard protocol. Other scores did not differ significantly (Table 2). Discussion Digital radiography remains the preferred examination for osseous abnormalities in patients with osteoarthritis. However, radiography is known to be inferior to CT, which in cases of unclear findings is the method of choice [10, 14]. Digital tomosynthesis is a new technique falling in between these two methods that has not often been used for musculoskeletal imaging. Because of the low cost of tomosynthesis examinations (comparable with that of a conventional radiograph), significant cost savings may be possible if the diagnostic performance for a given disease entity was similar to CT. Several possible applications have been described in the literature. For example, some case reports have described the improved detection of clinically suspected but otherwise radiographically occult scaphoid fractures using tomosynthesis [15, 16]. A larger prospective study of wrist fractures [9] showed tomosynthesis to be more useful than radiography but less useful than CT in fracture detection. This radiation dose reduction for tomosynthesis probably comes at the cost of a lower diagnostic accuracy. However, another study found that digital tomosynthesis is almost equal to CT concerning the detection of Item Scored 60 kv and 80 ma 60 kv and 40 ma 60 kv and 32 ma 50 kv and 80 ma 50 kv and 40 ma 50 kv and 32 ma Overall quality (p = 0.405) 1.67 (p = 0.025) 1.17 (p = 0.405) 1.67 (p = 0.405) 1.8 (p = 0.008) Soft tissue (p = 0.025) 2.17 (p = 0.009) 1 (p = 1.000) 1.67 (p = 0.025) 2.2 (p = 0.002) Cortical bone 1 1 (p = 1.000) 1 (p = 0.174) 1 (p = 1.000) 1.67 (p = 0.405) 1.83 (p = 0.027) Trabecular bone (p = 0.405) 1.50 (p = 0.027) 1 (p = 1.000) 1 (p = 1.000) 2 (p = 0.001) Cumulated scores (p = 0.027) 6.34 (p = 0.003) 4.17 (p = 0.505) 6.01 (p = 0.028) 7.83 (p = 0.002) Note The tube voltage and current combinations were compared with the standard combination (60 kv and 80 ma). Boldface type indicates statistically significant differences (p < 0.008) compared with the standard combination. 162 AJR:208, January 2017

5 Dose Reduction in Tomosynthesis of the Wrist erosions in patients with rheumatoid arthritis [7], indicating that the difference in accuracy varies for different conditions. Few studies have addressed hardware imaging with tomosynthesis, and the results from those studies suggest that tomosynthesis may have fewer artifacts than CT [17 19]. Overall, the evidence for using tomosynthesis in musculoskeletal imaging is still sparse; however, an increased usage has been seen, raising concerns about radiation exposure of patients. Most work regarding radiation exposure has been published about dose reduction for tomosynthesis in breast or chest imaging [3, 20 22]. In the breast, for example, higher soft-tissue contrast is needed, leading to a substantially lower tube voltage and hence lower radiation doses. However, the experience from breast imaging is not directly relevant to musculoskeletal imaging. The latter involves anatomic structures of different densities, and tube current and voltage have to be optimized separately for each anatomic region to reach a sufficient tissue penetration using the lowest dose possible. The few reports that have examined radiation exposure with regard to musculoskeletal applications for tomosynthesis have found that the radiation dose is lower than CT [7, 8]. In our study, we evaluated different tomosynthesis acquisition parameters with reduced radiation doses and found that the tube voltage could be reduced from 60 to 50 kv and the tube current from 80 to 40 ma without significant effect on the visibility of anatomic structures. This adjustment resulted in a dose reduction of up to 70% (Table 1 and Fig. 2). At the same time, the DAP ranged between 1 and 4 mgy cm 2 for an anterior-posterior projection of tomosynthesis. Similar studies of chest tomosynthesis have shown that the dose could be substantially reduced to a certain point, but a further decrease of the tube current had a detrimental effect on image quality because of the increasing electronic noise at very low doses [23]. Our results confirm these prior observations; we found a significantly lower qualitative score and an increase in image noise of up to 80% for the lowest radiation dose setting. Therefore, the optimal tube settings in the wrist were 50 kv and 40 ma in our study, corresponding to a mean DAP of Gy cm 2 and mean dose of 0.51 mgy. Our value is within the range of a prior study from Noël et al. [8], who measured a DAP of 1.72 Gy cm 2 and a dose of 1.01 mgy in a single wrist. However, Noël et al. assessed Fig. 5 Box-andwhisker plot shows cumulative quality and delineation scores for images obtained with tube voltages of 50 kv (left) and 60 kv (right). Only small, not statistically significant differences were noted between images obtained using 60 kv and 80 ma and 50 kv and 40 ma. Large, statistically significant drops in perceived quality were seen with 32-mA combinations. Line within boxes = mean, top line of box = 75th percentile, bottom line of box = 25th percentile, whiskers = 95% CI, circle = outlier. Cumulative Quality and Visibility Scores neither image quality nor delineation of anatomic structures. Image quality, image noise, and delineation of anatomic structures are influenced by several factors, and other methods have potential to further improve them while still allowing reduction of the radiation dose. For example, the air gap is a well-known method of reducing scatter radiation in radiography and thus increasing the signal-to-noise ratio of a given protocol (i.e., at the same radiation dose level) [24]. An additional air gap might be useful to enhance signal-to-noise ratio in tomosynthesis and allow further radiation dose reduction. Grids are another useful method in conventional radiography to control scatter radiation. However, because of the changing angle of incidence of the x-rays on the detectors during the different acquisitions from projection to projection, a conventional grid would result in severe loss of the primary x-rays when the x-ray source is located at oblique angles. As a result, grids are not normally used in tomosynthesis. Because tomosynthesis combines elements of both radiography and CT, it has challenges similar to both of those two methods with regard to radiation exposure levels. However, comparing CT and tomosynthesis is inherently difficult because of the different underlying physical principles of the two methods. For example, surface entrance doses cannot be recalculated into effective doses for comparison. In addition, no conversion coefficient (k) tables for the upper extremities exist to determine the effective dose for either CT or radiography. To calculate the effective dose, additional parameters (size of penetrated body region, tissue type) Tube Current (ma) have to be taken into consideration (e.g., tomosynthesis of the femur requires higher tube voltage and current than tomosynthesis of the wrist) [25, 26]. To measure the effective doses between the different modalities, anthropomorphic phantoms have to be used. This technique appears to be the only reliable estimation of dose when comparing two different imaging techniques [27, 28]. We are planning an extended phantom and cadaver study using multiple bodies, in which we want to insert dosimeters directly into various body parts at predefined locations. However, this design is laborious and many other issues also need to be solved before the study can begin. Until then, precise dosimeter study results are not available in the literature and one can only refer to prior reports from the literature where uniformly CT was found to have higher effective doses than tomosynthesis [6 9]. A cadaver-model study such as ours has inherent limitations. First, the use of cadavers allows scanning without motion artifacts, which does not reflect clinical reality. Motion artifacts in vivo might impair image quality, because the scan takes 5 seconds. Second, all cadaveric wrists were from older subjects, which is typical for cadaver studies because younger individuals rarely donate their bodies voluntarily to research. Third, we used a small number of cadaveric wrists (n = 6). However, given the excellent interreader agreement, we do not think that a higher number of cadaveric wrists would have changed our final results and conclusion. In addition, our number of specimens is typical for such studies. Lastly, we only focused on image quality and visibility of ana- AJR:208, January

6 Becker et al. tomic structures. Diseases and abnormalities were not evaluated, as we were not allowed to dissect specimens for standard of reference purposes but had to return them without damage for the use in other experiments. In conclusion, significant dose reduction for tomosynthesis of the wrist is possible while image quality and visibility of anatomic structures remain preserved. Acknowledgment We thank Joerg Mueller from Fujifilm Europe Medical Systems for his help and sharing his expertise for this project. References 1. Plantes BZ. Eine neue methode zur differenzierung in der röntgenographie (planigraphie). Acta Radiol 1932; 13: Grant DG. Tomosynthesis: a three-dimensional radiographic imaging technique. IEEE Trans Biomed Eng 1972; 19: Asplund SA, Johnsson AA, Vikgren J, et al. Effect of radiation dose level on the detectability of pulmonary nodules in chest tomosynthesis. Eur Radiol 2014; 24: Durand MA, Haas BM, Yao X, et al. Early clinical experience with digital breast tomosynthesis for screening mammography. Radiology 2015; 274: Vult von Steyern K, Bjorkman-Burtscher IM, Hoglund P, Bozovic G, Wiklund M, Geijer M. Description and validation of a scoring system for tomosynthesis in pulmonary cystic fibrosis. Eur Radiol 2012; 22: Canella C, Philippe P, Pansini V, Salleron J, Flipo RM, Cotten A. Use of tomosynthesis for erosion evaluation in rheumatoid arthritic hands and wrists. Radiology 2011; 258: Simoni P, Gérard L, Kaiser MJ, et al. Use of tomosynthesis for detection of bone erosions of the foot in patients with established rheumatoid arthritis: comparison with radiography and CT. AJR 2015; 205: Noël A, Ottenin MA, Germain C, et al. Comparison of irradiation for tomosynthesis and CT of the wrist. [in French] J Radiol 2011; 92: Ottenin MA, Jacquot A, Grospretre O, et al. Evaluation of the diagnostic performance of tomosynthesis in fractures of the wrist. AJR 2012; 198: Perry D, Stewart N, Benton N, et al. Detection of erosions in the rheumatoid hand: a comparative study of multidetector computerized tomography versus magnetic resonance scanning. J Rheumatol 2005; 32: Bolte H, Sattler EM, Jahnke T, et al. Low dose MDCT of the wrist: an ex vivo approach. Eur J Radiol 2011; 77: Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33: Wickham H. ggplot2: elegant graphics for data analysis. New York, NY: Springer Science & Business Media, Dalbeth N, Gao A, Roger M, Doyle AJ, McQueen FM. Digital tomosynthesis for bone erosion scoring in gout: comparison with plain radiography and computed tomography. Rheumatology (Oxford) 2014; 53: Geijer M, Borjesson AM, Gothlin JH. Clinical utility of tomosynthesis in suspected scaphoid fracture: a pilot study. Skeletal Radiol 2011; 40: Mermuys K, Vanslambrouck K, Goubau J, Steyaert L, Casselman JW. Use of digital tomosynthesis: case report of a suspected scaphoid fracture and technique. Skeletal Radiol 2008; 37: Gazaille RE 3rd, Flynn MJ, Page W 3rd, Finley S, van Holsbeeck M. Technical innovation: digital tomosynthesis of the hip following intra-articular administration of contrast. Skeletal Radiol 2011; 40: Gomi T, Hirano H. Clinical potential of digital linear tomosynthesis imaging of total joint arthroplasty. J Digit Imaging 2008; 21: Machida H, Yuhara T, Sabol JM, Tamura M, Shimada Y, Ueno E. Postoperative follow-up of olecranon fracture by digital tomosynthesis radiography. Jpn J Radiol 2011; 29: Baptista M, Di Maria S, Barros S, et al. Dosimetric characterization and organ dose assessment in digital breast tomosynthesis: measurements and Monte Carlo simulations using voxel phantoms. Med Phys 2015; 42: Feng SS, Sechopoulos I. Clinical digital breast tomosynthesis system: dosimetric characterization. Radiology 2012; 263: Kumar SG, Garg MK, Khandelwal N, et al. Role of digital tomosynthesis and dual energy subtraction digital radiography in detecting pulmonary nodules. Eur J Radiol 2015; 84: Svalkvist A, Bath M. Simulation of dose reduction in tomosynthesis. Med Phys 2010; 37: Kottamasu S, Kuhns L. Musculoskeletal computed radiography in children: scatter reduction and improvement in bony trabecular sharpness using air gap placement of the imaging plate. Pediatr Radiol 1997; 27: Wrixon AD. New ICRP recommendations. J Radiol Prot 2008; 28: Petraszko A, Siegal D, Flynn M, Rao SD, Peterson E, van Holsbeeck M. The advantages of tomosynthesis for evaluating bisphosphonate-related atypical femur fractures compared to radiography. Skeletal Radiol 2016; 45: Chiarot CB, Siewerdsen JH, Haycocks T, Moseley DJ, Jaffray DA. An innovative phantom for quantitative and qualitative investigation of advanced x-ray imaging technologies. Phys Med Biol 2005; 50:N287 N Neisius A, Astroza GM, Wang C, et al. Digital tomosynthesis: a new technique for imaging nephrolithiasis: specific organ doses and effective doses compared with renal stone protocol noncontrast computed tomography. Urology 2014; 83: AJR:208, January 2017