Genitourinary Imaging Original Research

Transcription

1 Genitourinary Imaging Original Research Genitourinary Imaging Original Research Koen Mermuys 1 Frank De Geeter 2 Klaus Bacher 3 Kris Van De Moortele 1 Kenneth Coenegrachts 1 Luc Steyaert 1 Jan W. Casselman 1 Mermuys K, De Geeter F, Bacher K, et al. Keywords: calculi, CT, digital radiography, digital tomosynthesis, kidney, radiation dose, ureter DOI: /AJR Received May 20, 2009; accepted after revision December 16, Department of Radiology, A.Z. St.-Jan Hospital, Ruddershove 10, 8000 Brugge, Belgium. Address correspondence to K. Mermuys (koen.mermuys@azbrugge.be). 2 Department of Nuclear Medicine, A. Z. St.-Jan Hospital, Bruges, Belgium. 3 Department of Medical Physics and Radiation Protection, Ghent University, Ghent, Belgium. AJR 2010; 195: X/10/ American Roentgen Ray Society Tomosynthesis in the Detection of Urolithiasis: Diagnostic Performance and Dosimetry Compared With Radiography With MDCT as the Reference Standard OBJECTIVE. The purpose of this study was to evaluate the diagnostic performance of digital tomosynthesis in comparison with digital radiography in the detection of urinary stones with MDCT as the reference standard. SUBJECTS AND METHODS. Fifty consecutively enrolled patients (32 men, 18 women; mean age, 51.5 years; range, years) referred for unenhanced MDCT of the abdomen with suspicion of urinary stones also underwent digital tomosynthesis and digital radiography (anteroposterior and bladder inlet views). Images from all examinations were randomly read by three blinded radiologists. The mean effective doses for digital tomosynthesis, digital radiography, and low- and high-dose MDCT were measured on a male phantom. Free-response receiver operating characteristics and receiver operating characteristics analyses were used to compare the diagnostic performance of digital radiography with that of digital tomosynthesis. RESULTS. Both types of analysis showed significantly better performance of tomosynthesis over digital radiography for all urinary stones (p < 0.05). No such improvement was found for ureteral stones. The gain in sensitivity with tomosynthesis was largest for stones between 2 and 5 mm in diameter. The mean effective dose was 0.5 msv for digital radiography, 0.85 msv for tomosynthesis, 2.5 msv for low-dose MDCT, and 12.6 msv for high-dose MDCT. CONCLUSION. Use of digital tomosynthesis of the abdomen results in improved detection of urinary stones in general over digital radiography with only a slight increase in effective dose. Use of tomosynthesis, however, was not associated with major improvement in the diagnosis of ureteral stones. The technique has potential as an alternative imaging technique in the detection and follow-up of urinary stones. I n digital tomosynthesis, a series of very-low-dose projection images are acquired with a digital detector during a single sweep of the x-ray tube over a limited angle. The data are reconstructed with a filtered back-projection algorithm to generate a set of images at a slice interval and start and end height defined by the user. The advantage of this new technique is that digital tomosynthesis removes overlying structures and provides depth information about the structure of interest. Limitations are that only slices parallel to the detector plane can be obtained and the loss of overview with blurring of the tissues outside the region of interest [1, 2]. Areas that have received the most interest are breast and chest imaging [1, 3 5]. Urinary tract stone disease frequently affects younger patients and has a high risk of recurrence. The acute symptoms of renal colic and the need for repeated examinations, either in the follow-up of known stones under treatment or when recurrence is suspected, necessitate an imaging technique that involves little risk while being highly sensitive and specific for detecting calculi. Unenhanced MDCT is widely accepted as the most sensitive imaging technique for the diagnosis of urolithiasis in patients with acute flank pain; the reported sensitivity is as high as 97% [6 12]. However, use of unenhanced MDCT for screening or followup raises concern about radiation exposure, although proposed low-dose CT protocols have had good results [13 16]. Ultrasound and radiography have only limited value. The reported sensitivities in the detection of urinary stones, even in retrospective review of radiographs, are only between 58% and 62% [17 19]. These techniques are limited mainly by bowel superimposition, although a phantom study [20] has shown that digital radiography should AJR:195, July

2 have better results than computed radiography and at a lower patient exposure [20]. The rate of detection of renal stones with ultrasound alone is only 24 60% [21, 22]. Excretory urography has the highest reported sensitivity, as high as 90% for suspected ureterolithiasis, but introduces the need for IV contrast administration and the risk of adverse reactions [12, 23]. Conventional tomography has been found to improve detection of urolithiasis owing to reduced bowel superimposition [24]. Therefore, digital tomosynthesis may be promising for this purpose. To our knowledge, however, no data are available on the use of digital tomosynthesis in the detection of urolithiasis. We therefore performed a prospective comparative study of the diagnostic performance of digital tomosynthesis and conventional digital radiography of the abdomen. Unenhanced MDCT was used as the reference standard. We also used an anthropometric phantom to measure the mean effective doses of digital tomosynthesis, digital radiography, and low- and high-dose MDCT. Subjects and Methods Study Protocol Institutional review board approval was obtained for this prospective study, and informed consent was obtained from all subjects. Between November 2007 and June 2008, 50 consecutively enrolled patients (32 men, 18 women; mean age, 51.5 years; median, 52.5 years; range, years) referred for unenhanced MDCT of the abdomen with suspicion of urinary stones underwent additional digital tomosynthesis and digital radiography (anteroposterior view and bladder inlet view). All images were read independently by three skilled radiologists (12, 10, and 26 years of experience). The readers were informed that urinary stones were suspected. First all radiographs and then all digital tomosynthesis images were read at random with the readers blinded to patient name to avoid linking the digital radiographs and digital tomosynthesis images of the same patient. All readers recorded any location they considered possibly corresponding to urinary stones and rated these locations on a 3-point scale according to the level of confidence in the presence of a urinary calculus: 1, possible but unlikely; 2, probable; 3, definite. The urinary stones found at MDCT were used as the reference against which the readers locations found with both digital radiography and digital tomosynthesis were classified as true- or false-positive. In case of disagreement between the three readers on the MDCT result, a fourth reader (7 years of experience) decided the final MDCT result to be used as the reference standard. For the analysis, unmarked lesions were rated 0. Stones were measured on MDCT images in the axial plane with standard body imaging windows. To study the effect of stone location and size on sensitivity and reproducibility, stones were classified according to their location in the kidney or ureter. Kidney stones were further subdivided in those smaller than 2 mm in diameter, those 2 5 mm, and those larger than 5 mm, according to the largest dimension measured on MDCT images. Imaging Technique MDCT was performed with a 16-MDCT scanner (LightSpeed, GE Healthcare) with high-dose and low-dose protocols. In the high-dose protocol, MDCT was performed in helical mode at 120 kvp; 280 mas; pitch, 0.8; and slice thickness, 1.25 mm. For the low-dose protocol, helical scanning mode was used with a slice thickness of 1.25 mm and a pitch of 0.8, but the tube voltage was reduced to 100 kvp and the tube current time setting to 120 mas. For digital tomosynthesis, a multipurpose digital tomosynthesis system (VolumeRAD option of Definium 8000, GE Healthcare) was used. For digital tomosynthesis examinations of the abdomen, the x-ray tube performs a continuous vertical or horizontal movement for 10 seconds, collecting 60 low-dose projection images within ± 15 around the required radiographic view. A constant tube load is used for each exposure, determined with a scout image collected before the digital tomosynthesis acquisition. The scout image is a conventional projection of the anatomic structure of interest. The automatic exposure control determined tube load used for the scout image is multiplied by a user-adjustable dose factor and distributed over the low-dose projections. The resulting exposure for each projection is then set to the closest tube current time setting possible within the constraint of a minimum tube current time product per projection. The low-dose projections are then combined to reconstruct section images of the abdomen, as in coronal reconstructions in CT. Reconstructed slice spacing can be selected between 1 and 50 mm with a step of 1 mm. In this study the patient was placed in supine position. Twenty-five acquisitions were made in the anteroposterior view at 80 kvp, 500 ma, 1.25 mas, 2.5 milliseconds. A total of 78 slices with a 3-mm slice interval (start height, 20 mm; end height, 250 mm) were obtained. If necessary, for example in imaging of large patients, additional reconstructions were performed to visualize the anteriormost portion of the ureters. The parameters used for digital radiography (Definium 8000 system, GE Healthcare) were anteroposterior view (85 kvp, 630 ma, mas, 52.4 milliseconds) and a 20 bladder inlet view (85 kvp, 630 ma, 64 mas, 101 milliseconds). Radiation Dose Measurements To compare the radiation burden of the selected imaging techniques, an anthropomorphic (average man) phantom (Rando Man, The Phantom Laboratory) experiment was set up. The phantom was filled with 156 thermoluminescent dosimeters (TLDs) (magnesium, titanium, sintered, natural abundance; MTS-N, TLD Poland) in positions representative of the radiosensitive organs and tissues included in the concept of effective dose [25]. The choice of the TLD locations was based on a complete CT scan of the phantom. Distribution of bone marrow over the body was adopted from Cristy [26]. An extra set of 10 TLDs was used to assess the background radiation. All TLDs were individually calibrated at the Standard Dosimetry Laboratory in Ghent, Belgium, in x-ray beam qualities that corresponded to those used in our selected protocols. radiography, digital tomosynthesis, and low-dose and high-dose MDCT of the phantom were performed with our selected protocols. Positioning and collimation of the x-ray beam were performed in a manner similar to that for patients. To obtain dose measurements well above the background radiation, eight exposures were performed. Effective Dose Calculation The effective dose (E), which is representative of the risk of late radiation-induced effects such as malignant disease, is defined with the following equation [25]: E = Σ w T H T T where H T is the equivalent dose to tissue T, and w T is the weighting factor representing the relative radiation sensitivity of tissue T. The organ weighting factors were adopted from the 2008 recommendations of the International Commission on Radiological Protection [25]. After irradiation of the phantom, TLDs were read with a TLD reader (Harshaw 3500, Thermo Scientific). The background-corrected phantom TLD readings were converted into absorbed dose values after application of the corresponding calibration factor. On the basis of the latter dose values, the mean equivalent organ doses per acquisition were calculated as described by Huda and Sandison [27]. These results were combined with the corresponding tissue-weighting factor (w T ) to calculate the effective dose in microsieverts. Statistical Evaluation With the multireader multicase data sets, the diagnostic performance of digital radiography and digital tomosynthesis was evaluated with both free-response receiver operating character- 162 AJR:195, July 2010

3 istics (FROC) and receiver operating characteristics (ROC) analyses in implementations available at The current version of this software is JAFROC_V3d_BETA. The methods are described in detail in the literature [28, 29]. Briefly, whereas ROC analysis is used to assess the true-positive rate versus the false-positive rate on a per patient basis, FROC is used to analyze the true-positive rate on a per lesion basis versus the false-positive rate on a per patient basis (false-positive findings on abnormal images are ignored). From such curves, figures of merit are calculated that correspond to the nonparametrically estimated area under the curve. The significance of differences observed between techniques is determined by jackknifing cases and analyzing the pseudovalue by analysis of variance. Because only three readers were involved in our study, analysis of variance was performed for random cases but not for random readers, allowing the results to be generalized for all cases. The analyses were performed on all stones and on ureteral stones separately. Interobserver reproducibility of readings was assessed with the intraclass correlation coefficient. These values were calculated with software (SPSS for Windows version 17.0, SPSS). Two-way analysis of variance and the absolute agreement definition for single measures were used. To investigate the effects of stone size on the detection rate, observer rates were dichotomized to yes-or-no responses. Responses were considered positive for confidence ratings of 2 or 3 (i.e., readings rated at least probable). The effects of size and location were subsequently assessed with Fisher s exact test (two-sided for size effects, one-sided for location effects). The McNemar test was used to assess differences between the sensitivity values of digital radiography and digital tomosynthesis. Chisquare tests were used on average results. Values of p < 0.05 were considered significant. TABLE 1: Figures of Merit From Receiver Operating Characteristics and Free-Response Receiver Operating Characteristics Analyses of the Diagnostic Performance of Planar Radiography and Tomosynthesis Reader Receiver Operating Characteristics Radiography Tomosynthesis p a Radiography Free-Response Receiver Operating Characteristics Tomosynthesis All stones Average Ureteral stones Average NS NS Note NS = not significant. a radiography versus digital tomosynthesis. TABLE 2: Location Dependency of Sensitivity Reader radiography Stone Location Kidney (n = 63) Ureter (n = 15) Results In the 50 patients in the study, a total of 78 urinary stones were found in 31 patients, 63 of which were in the kidney (19 at most 2 mm in diameter, 30 between 2 and 5 mm, and 14 larger than 5 mm). The other 15 stones were in the ureter (four at most 2 mm in diameter, seven between 2 and 5 mm, and four larger than 5 mm). Renal stones were present in 27 patients and ureteral stones in 13. No bladder stones were found. For the diagnosis of the presence of urinary stones (regardless of location), the results of FROC and ROC analyses pointed toward better diagnostic performance of digital tomosynthesis than of digital radiography. For ureteral stones, however, no significant difference was observed between these two techniques (Table 1). This finding was also reflected in the dichotomous analysis in which all locations scored as at least probable were considered positive (Table 2). Whereas with digital radiography no significant differences in sensitivity were observed for renal versus ureteral stones, with digital tomosynthesis the sensitivity was significantly worse for ureteral stones than for those located in kidneys. With digital tomosynthesis, a significant gain in sensitivity was accomplished for renal stones but not for those located in the ureters NS NS NS Average (NS) tomosynthesis NS Average (0.009) p b NS NS NS Note Values are sensitivity. a For location dependency of sensitivity according to Fisher s exact test or according to chi-square test (for average results, indicated by parentheses). b For digital radiography versus digital tomosynthesis, according to McNemar test. p a p a AJR:195, July

4 TABLE 3: Size Dependency of Sensitivity Reader radiography Stone Size (mm) 2 (n = 23) > 2 and 5 (n = 37) > 5 (n = 18) Average (0.001) Tomosynthesis NS Average (0.02) p b NS NS Note Values are sensitivity. a For location dependency of sensitivity according to Fisher s exact test or according to chi-square test (for average results, indicated by parentheses). b For digital radiography versus digital tomosynthesis, according to McNemar test. TABLE 4: Intraclass Correlation Coefficients Patient Level Radiography Tomosynthesis All stones ( ) ( ) Ureteral stones ( ) ( ) Stone All stones ( ) ( ) Ureteral stones ( ) ( ) Stones > 5 mm ( ) ( ) Stones between 2 and 5 mm ( ) ( ) Stones 2 mm ( ) ( ) Note Values in parentheses are 95% CI. A p a For all readers, the detection rate increased with increasing stone size for both digital radiography and digital tomosynthesis (Table 3). For stones up to 5 mm, digital tomosynthesis was significantly more sensitive than digital radiography, and the gain in sensitivity with digital tomosynthesis was most marked for stones between 2 and 5 mm in diameter. Interobserver agreement as measured by intraclass correlation coefficient proved to be better with digital tomosynthesis than with digital radiography, although the differences did not reach statistical significance (Table 4). This result was consistent when data were analyzed at the patient level (presence or absence of stones in a patient) and at the stone level (presence or absence of a stone in a stone localization). A typical example of the improved detection with digital tomosynthesis is shown in Figure 1. The effective dose value was 0.85 msv for digital tomosynthesis. radiography (anteroposterior and bladder inlet views) had an effective dose of 0.5 msv, and MDCT had much higher mean effective radiation dose values of 2.5 msv and 12.6 msv for the low- and high-dose protocols (Fig. 2). The mean effective dose of digital tomosynthesis of the abdomen was 66% lower than that for low-dose MDCT and 93% lower than that for high-dose MDCT but still 70% higher than that for two-view digital radiography. Discussion In this preliminary study, we found that digital tomosynthesis had significantly better diagnostic performance than digital radiography in the diagnosis of urinary stones in general. The radiation dose was only slightly increased in comparison with that for digital radiography and remained substantially lower than the dose in high- and even low-dose Fig year-old woman with suspected urinary stones. A, anteroposterior radiograph shows three possible stones (arrows) in left kidney. Right kidney is not delineated owing to superposition of feces and air. B and C, Contiguous digital tomosynthesis images reconstructed with spacing of 3 mm shows numerous stones in both kidneys. Calculi and contour of right kidney are much clearer than in A. B C 164 AJR:195, July 2010

5 MDCT protocols. In contrast, in the small sample studied, digital tomosynthesis afforded no major improvement in the diagnosis of ureteral stones. The overall sensitivity of digital tomosynthesis for urinary stones on a per stone basis was only 59% in the current study. This result was partly due to the relatively large contribution of stones smaller than 2 mm (23 of the 78 stones, 29%), for which the sensitivity was only a disappointing 36%, decreasing the overall sensitivity for digital tomosynthesis. Noise on the images makes it difficult to appreciate these small stones (Fig. 3). Sensitivity increases substantially to 64% and 76% for stones between 2 and 5 mm and those larger than 5 mm. Low-density stones (uric acid stones) remain invisible, and intrarenal vascular and cyst calcifications can give false-positive results. A second cause of the relatively low overall sensitivity for urinary stones is the contribution of stones located in the ureters, for which a sensitivity of only 29% was obtained. One of the causes of the poor result in the detection of ureteral stones is that the readers were blinded to information on whether the patient had renal colic, whether hydronephrosis was found with ultrasound, and the side of the symptoms. Because the wall of the ureter is not visible on tomographic images, a large number of ureteral stones were misinterpreted as phleboliths and mesenteric calcifications and vice versa (Fig. 4). We believe that the sensitivity for ureterolithiasis would increase substantially if digital tomosynthesis were combined with clinical and ultrasound information. It also appears from this study that digital tomosynthesis is more reproducible between readers than is digital radiography. Interobserver agreement always was better with digital tomosynthesis than with digital radiography. A Fig. 2 Graph shows total effective dose measured on anthropomorphic phantom for four imaging techniques. Effective Dose (msv) In a 2009 evaluation report [30] from the United Kingdom National Health Service Purchasing and Supply Agency on digital tomosynthesis for general radiography, it was calculated that in a best case approximation, a digital tomosynthesis system becomes cost-effective if 0.6 CT examinations per week are transferred to the digital tomosynthesis. Therefore, it seems likely that cost-effectiveness will be assured Radiography A B 0.85 Tomosynthesis 2.5 MDCT Low-Dose 12.6 MDCT High-Dose Fig year-old woman with renal stone. A, Enlarged digital tomosynthesis image of left kidney shows 2-mm tiny calculus (arrow) in central part of kidney. B, Enlarged MDCT image of left kidney confirms presence of tiny calculus. With these very small stones, it becomes difficult to differentiate tiny calculi from surrounding noise with digital tomosynthesis. A larger clinical study is needed to define the exact position of digital tomosynthesis in the detection and follow-up of urinary stones. It also would be worthwhile to compare the results of digital tomosynthesis with those of low-dose MDCT. The data from our preliminary study suggest that digital tomosynthesis can replace digital radiography as the first-line imaging technique Fig year-old man with right-sided renal colic. A, Axial MDCT image shows obstructive calculus in right vesicoureteral junction. B, Bladder inlet digital radiograph does not show distal ureterolithiasis owing to superposition of bowel and feces. (Fig. 4 continues on next page) B AJR:195, July

6 for patients with suspected urinary stone disease and would increase the detection rate of urinary stones, possibly even more so in combination with ultrasound. tomosynthesis also may be useful in follow-up after extracorporeal shockwave lithotripsy to look for residual stone disease, stone migration, and fragmentation of known stones. Nevertheless, MDCT will remain the reference standard for detecting urolithiasis with almost perfect sensitivity by means of direct stone visualization or visualization of secondary CT signs of urinary obstruction such as ureteral dilatation and perinephric and periureteral fat stranding, even after stone passage [6, 9 11, 31]. We found that use of digital tomosynthesis of the abdomen improves the diagnostic performance over that reached with planar digital radiography and results in better interreader reproducibility with only a limited increase in radiation dose. Therefore, digital tomosynthesis shows promise as a cost-effective alternative to MDCT in the detection of urinary stones and especially in follow-up. It has the potential to reduce the number of MDCT examinations requested and thus decrease the overall radiation dose received by patients. Acknowledgment We thank D. Chakraborty, University of Pittsburgh, for advice on ROC and FROC analysis. References 1. McAdams HP, Samei E, Dobbins J 3rd, Tourassi GD, Ravin CE. Recent advances in chest radiography. Radiology 2006; 241: 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: Dobbins JT 3rd, McAdams HP, Godfrey DJ, Li CM. tomosynthesis of the chest. J Thorac C Fig. 4 continued 49-year-old man with right-sided renal colic. C and D, tomosynthesis image (C) and magnified image of inset in C (D) distinctly show ureteral calculus (arrow). Central lucency in stone mimics phlebolith. Because they were blinded to clinical information, all readers misinterpreted image finding as phlebolith. D Imaging 2008; 23: James TD, McAdams HP, Song JW, et al. tomosynthesis of the chest for lung nodule detection: interim sensitivity results from an ongoing NIHsponsored trial. Med Phys 2008; 35: Park JM, Franken EA Jr, Garg M, Fajardo LL, Niklason LT. Breast tomosynthesis: present considerations and future applications. RadioGraphics 2007; 27[suppl 1]:S231 S Chu G, Rosenfield AT, Anderson K, Scout L, Smith RC. Sensitivity and value of digital CT scout radiography for detecting ureteral stones in patients with ureterolithiasis diagnosed on unenhanced CT. AJR 1999; 173: Remer EM, Herts BR, Streem SB, et al. Spiral noncontrast CT versus combined plain radiography and renal US after extracorporeal shock wave lithotripsy: cost-identification analysis. Radiology 1997; 204: Sheafor DH, Hertzberg BS, Freed KS, et al. Nonenhanced helical CT and US in the emergency evaluation of patients with renal colic: prospective comparison. Radiology 2000; 217: Smith RC, Rosenfield AT, Choe KA, et al. Acute flank pain: comparison of non-contrast-enhanced CT and intravenous urography. Radiology 1995; 194: Lumerman J, Gershbaum MD, Hines J, Nardi P, Beuchert P, Katz DS. Unenhanced helical computed tomography for the evaluation of suspected renal colic in the adolescent population: a pilot study. Urology 2001; 57: Smith RC, Verga M, McCarthy S, Rosenfield AT. Diagnosis of acute flank pain: value of unenhanced helical CT. AJR 1996; 166: Shine S. Urinary calculus: IVU vs. CT renal stone? A critically appraised topic. Abdom Imaging 2008; 33: Tack D, Sourtzis S, Delpierre I, de Maertelaer V, Gevenois PA. Low-dose unenhanced multidetector CT of patients with suspected renal colic. AJR 2003; 180: Kluner C, Hein PA, Gralla O, et al. Does ultra-lowdose CT with a radiation dose equivalent to that of KUB suffice to detect renal and ureteral calculi? J Comput Assist Tomogr 2006; 30: Poletti PA, Platon A, Rutschmann OT, Schmidlin FR, Iselin CE, Becker CD. Low-dose versus standard-dose CT protocol in patients with clinically suspected renal colic. AJR 2007; 188: Liu W, Esler SJ, Kenny BJ, Goh RH, Rainbow AJ, Stevenson GW. Low-dose nonenhanced helical CT of renal colic: assessment of ureteric stone detection and measurement of effective dose equivalent. Radiology 2000; 215: Levine JA, Neitlich J, Verga M, Dalrymple N, Smith RC. Ureteral calculi in patients with flank pain: correlation of plain radiography with unenhanced helical CT. Radiology 1997; 204: Roth CS, Bowyer BA, Berquist TH. Utility of the plain abdominal radiograph for diagnosing ureteral calculi. Ann Emerg Med 1985; 14: Mutgi A, Williams JW, Nettleman M. Renal colic. Utility of the plain abdominal roentgenogram. Arch Intern Med 1991; 151: Szucs-Farkas Z, Chakraborty DP, Thoeny HC, Loupatazis C, Vock P, Bonel HM. Detection of urinary stones at reduced radiation exposure: a phantom study comparing computed radiography and a low-dose digital radiography linear slit scanning system. AJR 2009; 192:1579; [web]w271 W Fowler KA, Locken JA, Duchesne JH, Williamson MR. US for detecting renal calculi with nonenhanced CT as a reference standard. Radiology 2002; 222: Ulusan S, Koc Z, Tokmak N. Accuracy of sonography for detecting renal stone: comparison with CT. J Clin Ultrasound 2007; 35: Pfister SA, Deckart A, Laschke S, et al. Unenhanced helical computed tomography vs intravenous urography in patients with acute flank pain: accuracy and economic impact in a randomized prospective trial. Eur Radiol 2003; 13: Schwartz G, Lipschitz S, Becker JA. Detection of renal calculi: the value of tomography. AJR 1984; 143: AJR:195, July 2010

7 25. [No authors listed]. The 2007 Recommendations of the International Commission on Radiological Protection: ICRP publication 103. Ann ICRP 2007; 37: Cristy M. Active bone marrow distribution as a function of age in humans. Phys Med Biol 1981; 26: Huda W, Sandison GA. Estimation of mean organ doses in diagnostic radiology from Rando phantom FOR YOUR INFORMATION measurements. Health Phys 1984; 47: Chakraborty DP, Berbaum KS. Observer studies involving detection and localization: modeling, analysis, and validation. Med Phys 2004; 31: Dorfman DD, Berbaum KS, Metz CE. Receiver operating characteristic rating analysis: generalization to the population of readers and patients with the jackknife method. Invest Radiol 1992; 27: Centre for Evidence-Based Purchasing. Evaluation report: tomosynthesis for general radiography. London, UK: NHS Purchasing and Supply Agency, Report CEP uk/pasaweb/nhsprocurement/cep. Published February Accessed March 29, Sourtzis S, Thibeau JF, Damry N, Raslan A, Vandendris M, Bellemans M. Radiologic investigation of renal colic: unenhanced helical CT compared with excretory urography. AJR 1999; 172: PQI Connect is the latest addition to the ARRS Website and serves as a source for information on meeting the growing demand for quality review programs in today s radiology practices and facilities. The interactive and easy-to-navigate site focuses on five critical topics that guide you through news items, relevant articles, and links to important information on each topic. AJR:195, July