|
|
||||||||
Original Research |
1 All authors: Department of Radiology, University Hospital Basel, Petersgraben 4, 4031 Basel, Switzerland.
Received November 12, 2007;
accepted after revision February 23, 2008.
This work was financially supported by the EC-EURATOM 6 Framework Programme
(2002–2006) and forms part of the CT safety and efficacy project:
"Safety and Efficacy of Computed Tomography (CT): A Broad
Perspective" (contract no. FP/002388).
Abstract
|
|
|---|
MATERIALS AND METHODS. The medical literature from 1995 to 2007 was searched using PubMed, Medline, and Cochrane Library databases for articles on studies that used low-dose CT (< 3 mSv dose applied for the entire CT examination) as a diagnostic test for the detection of urolithiasis. Prospective and retrospective studies were included if they separately reported the rate of true-positive, true-negative, false-positive, and false-negative diagnoses of urolithiasis from low-dose CT compared with the positive and negative rates of normal-dose CT or a combination of diagnostic tests. Two readers assessed the quality of the studies.
RESULTS. The pooled sensitivity and specificity of low-dose CT for the diagnosis of urolithiasis were 0.966 (95% CI, 0.950–0.978) and 0.949 (95% CI, 0.920–0.970), respectively.
CONCLUSION. The results of this meta-analysis suggest that a low-dose CT protocol can be used as the initial imaging technique in the workup of patients with suspected urolithiasis.
Keywords: CT genitourinary imaging kidney stones low-dose CT urinary tract urolithiasis
|
|
|---|
The clinical manifestation in terms of, for example, acute flank pain depends on the location and size of the stones. In the literature, it remains unclear whether nonobstructing nephrolithiasis can cause symptoms [4]. In addition to locating the stone, another important part of the diagnostics is to determine why the stone formed and develop treatment to prevent recurrence.
Before the introduction of MDCT, excretory urography was the declared reference standard for evaluation of acute flank colic and diagnosis of ureteral concrements; however, even in patients with complete ureteral obstruction the cause cannot be detected in all cases [5]. Besides this limitation, pain induced by diuresis of contrast medium is a common problem in excretory urography. The detection rate for concrements in excretory urography is described in the literature as approximately 70–90% [6]. For conventional radiography, a sensitivity of 57% and specificity of 71%, as well as a poor detection rate of 50–70%, are reported [6, 7]. Sonography as an imaging alternative is mainly used for imaging the kidneys and the proximal parts of the ureters. The quality of the sonography images depends on many factors that cannot be influenced by the radiologist (e.g., intestinal gas, obesity), and sonography detects only 50–60% of ureteral calculi [8]. The sensitivity of CT for ureteral calculi—especially in the distal parts of the ureters—is higher than that of sonography [9]. In normal-dose CT, the sensitivity and specificity are 94–100% and 97%, respectively [10–13]. Therefore, CT is now recommended by many authors as the initial diagnostic imaging technique in patients with suspected renal or ureteral concrements [14, 15].
Recent radiologic CT protocols for the detection of urolithiasis show a large variation in effective doses administered. Depending on the protocol used, the effective dose ranges from 8 to 16 mSv [16, 17]. The higher dose of unenhanced helical CT compared with excretory urography is of particular concern because of repeated stone formation in young patients and, therefore, repeated diagnostic tests. By the introduction of MDCT, the high radiation dose has not been altered significantly [18]. Many studies focus on the reduction of the applied mAs to reduce the effective dose. In recently published articles, investigators describe a reduction of the effective dose administered to 0.7–2.3 mSv [13, 19–21].
CT image noise varies proportional to the reciprocal value of the square root of the milliampere product [22]. A main concern is that higher noise and therefore lower image quality resulting from low-dose CT studies possibly decreases accuracy in detecting both ureteral concrements and alternative diagnoses, which is one of the advantages of CT compared with other imaging techniques [23]. The most important alternative diagnoses are cholecystolithiasis; diverticulitis; appendicitis; tumors located in the pelvis; and aortic aneurysms [9], which might be visible even in dose-reduced CT studies [24].
|
|
|---|
Data Extraction
A comprehensive computer literature search of abstracts about studies of
human subjects was performed to identify articles on the diagnostic
performance of low-dose CT in patients with suspected urolithiasis compared
with the diagnostic performance of normal-dose CT or the combination of other
radiologic diagnostic tests, surgical results, and follow-up as the reference
standard. The PubMed databases were used with the following search query,
which contains Medical Subject Heading (MeSH) vocabulary from the National
Library of Medicine and the "TIAB" tag directing the search of the
title and abstract fields (ptyp = publication type):
("urinary calculi"[MeSH Terms] OR "urinary calculi"[TIAB] OR urolithiasis[TIAB] OR "ureteral calculi"[MeSH Terms] OR "ureterolithiasis"[TIAB] OR "kidney calculi"[MeSH Terms] OR nephrolithiasis[TIAB]) AND ("ct" [TIAB] OR ct[TIAB]) AND "humans"[MeSH Terms] AND ("sensitivity and specificity" [MeSH Terms] OR sensitivity [TIAB]) NOT Case Reports[ptyp].
The search query was translated accordingly for MEDLINE databases and the Cochrane Library.
Review articles, case reports, letters, and comments were not selected. The list of articles was supplemented with articles found from an extensive cross-check of the references. We excluded redundant studies based on the authors' affiliation and the patient population. The search was limited to articles published after 1995 because the use of MDCT for the diagnosis of urolithiasis was first described then [3]. To check for eligibility criteria, we retrieved the study articles for more detailed investigations.
Study Selection
Two observers independently checked all retrieved articles for the
inclusion criteria. Disagreements were resolved in consensus. The inclusion
criteria were as follows: First, articles were reported in English, German, or
French language; second, low-dose CT was used (< 3 mSv effective dose
applied); third, normal-dose CT or a multitechnique strategy was used as the
reference standard; and, fourth, absolute numbers of true-positive,
false-negative, false-positive, and true-negative findings were available or
could be derived from the published data.
A study was excluded if the published data were incomplete.
Critical Appraisal of the Quality of Studies
The quality of the results was independently assessed by two radiologists
using the specified Standards for Reporting of Diagnostic Accuracy (STARD)
checklist [26]; these
assessments yielded a total score for each study evaluated. Detailed
guidelines for interpretation of the items were discussed by the two
radiologists during earlier meetings.
|
|
|
|---|
One study was excluded because of a possible selection bias of the patients [34]: All patients exceeding the weight of 90 kg were excluded. Two articles did not meet the inclusion criteria for low-dose CT [30, 35]. One study was excluded because different results were reported for three readers, and supplementary statistical analysis revealed significant heterogeneity for those readers; therefore, pooling inner specificities to include that study in the meta-analysis could not be performed [20]. Another study was excluded for reasons of insufficient data reporting [19].
Only one study had to be excluded because the mean effective dose applied of 3.5 mSv was slightly above the defined threshold of < 3 mSv [32]. The resulting seven studies represented a total of 1,061 patients.
Overview of the Studies
Of the 156 studies screened, seven met the inclusion criteria for the
meta-analysis. A summary of the results of each study and of selected study
characteristics is outlined in Table
1.
|
|
|
Four studies used helical single-detector CT for image acquisition [24, 31, 33, 36] and three studies used MDCT [21, 37, 38].
Critical Appraisal of the Quality of the Studies
The results of an assessment of the quality of the studies can be found in
Table 2. The median score
according to the STARD criteria was 0.72. The interreader agreement was
excellent at 85.5% [29]. None
of the studies yielded the maximum STARD score of 25 (100%).
|
Pooling Sensitivities and Specificities
The chi-square test for heterogeneity in sensitivity analysis revealed a
value of 2.19 and therefore indicated heterogeneity was not statistically
significant with p = 0.9. In general, heterogeneity among different
studies is caused either by differences in the ways clinicians define a test
as positive for urolithiasis or by wide variations among patients in terms of
their background. Heterogeneity analysis for pooling specificity revealed a
chi-square of 4.45—indicating that heterogeneity was not statistically
significant—with p = 0.615. Therefore, we pooled the
sensitivities and specificities of the studies selected. A pooled sensitivity
of 0.966 (95% CI, 0.950–0.978) and a pooled specificity of 0.949 (95%
CI, 0.920–0.970) were calculated (Figs.
2 and
3).
The SROC curve is presented in Figure 4. The area under the SROC curve (AUC) represents the overall diagnostic accuracy of a test. The AUC was calculated as 99.32% (standard error [SE] = 0.0016).
|
|
|
|---|
The mean effective dose for an excretory urography examination is reported as 2.6 mSv [40]—less than one third the effective dose of routinely used CT protocols for the detection of urolithiasis [16, 17], which ranges between 8 and 16 mSv. The dose-reduced studies in our meta-analysis reported applied maximum effective doses of 0.7–2.8 mSv. Only one study exceeded the effective dose for excretory urography [31], whereas the other studies ranged from 0.7 to 2.1 mSv.
For CT detection of urolithiasis, a substantial dose reduction should be possible because of the high contrast difference between most urinary tract stones and the surrounding soft tissue. In general, characterizing stone morphology and determining location remain the major advantages of CT over other imaging techniques (i.e., abdominal radiography, sonography) in the evaluation of patients with suspected renal colic [15, 41].
The effects of CT dose reduction in obese patients remain unclear. In these times of automatic tube current modulation techniques, concepts of absolute effective dose reduction are difficult to apply in all patients because of differences in body mass index (BMI) values of patients. Therefore, the main focus of dose reduction should be less the reduction of absolute tube current, but rather the increase of image noise. This need to change focus is evident because recent publications distinguish between several levels of BMI and consecutively applied tube current [21, 38].
Modern CT scanners provide possibilities for automatic patient-adapted tube current modulation and therefore a wider potential for a BMI threshold–independent relative dose reduction by relatively increasing noise [42]. Further studies are necessary to define the performance of low-dose CT in the detection of urolithiasis at different noise levels. A definition of sufficient noise levels for both the detection of concrements and the related differential diagnosis will permit a significant relative dose reduction in all patients and therefore reduce collective dose. Further research will be necessary to define these levels.
On the basis of composition, various common stone types can be differentiated including major types—such as calcium oxalate monohydrate, calcium oxalate dihydrate, brushite, struvite, cystine, and uric acid—and a few minor types. The stone composition and the related fragility might directly influence treatment outcome in shock wave lithotripsy. As has been suggested in the literature, differentiation of stone characteristics might be possible using CT attenuation values [43–45]. These characteristics directly influence the ability to predict response rates to shock wave lithotripsy [46–48]. In standard-dose CT, uric acid, cystine, calcium oxalate monohydrate, and brushite calculi can be identified with a probability of correct diagnosis exceeding 85%. Knowing the composition of stones is important because these calculi are generally refractory to extracorporeal lithotripsy and because uric acid and cystine calculi are usually treated medically [49]. In a recent study the authors clearly showed the feasibility of using helical CT to identify cystine stones that will be susceptible to shock wave lithotripsy [50]. A major limitation of that study was that it was performed as an in vitro analysis and the authors used thinner CT collimation than would be used clinically. As stated in their discussion, the extrapolation of their results to coarser slice widths is not obvious [50]. The clinical value of dose-reduced CT to predict stone composition can be doubted. However the feasibility of stone component characterization using low-dose CT still remains unclear in the literature.
In conclusion, the results of this meta-analysis suggest that a low-dose CT protocol can be used as the first-line imaging tool in the clinical workup of patients with suspected renal colic providing that clinicians and radiologists are aware of the limitations of dose-reduced fixed-tube-current protocols, especially in obese patients, compared with standard-dose CT. A low-dose protocol for the detection of urolithiasis seems to be most valuable in patients requiring follow-up scanning to ensure highest accuracy.
Further research is necessary to define noise thresholds for the detection of urolithiasis and therefore reduce the collective dose using recent techniques such as automatic tube current modulation.
|
|
|---|
This article has been cited by other articles:
![]() |
A. K. Hara, R. G. Paden, A. C. Silva, J. L. Kujak, H. J. Lawder, and W. Pavlicek Iterative Reconstruction Technique for Reducing Body Radiation Dose at CT: Feasibility Study Am. J. Roentgenol., September 1, 2009; 193(3): 764 - 771. [Abstract] [Full Text] [PDF] |
||||
![]() |
S.-M. Joo, K. H. Lee, Y. H. Kim, S. Y. Kim, K. Kim, K. J. Kim, and B. Kim Detection of the Normal Appendix with Low-Dose Unenhanced CT: Use of the Sliding Slab Averaging Technique Radiology, June 1, 2009; 251(3): 780 - 787. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Tack and P. A. Gevenois Body MDCT at 140 kV Am. J. Roentgenol., March 1, 2009; 192(3): W139 - W140. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |