| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
3 cm) Renal Masses with MDCT: Benefits of Thin Overlapping Reconstructions
1 Department of Radiology, Brigham and Women's Hospital, 75 Francis St., Boston,
MA 02115.
2 Present address: Department of Diagnostic Radiology, Keio University School of
Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan.
Received August 14, 2003;
accepted after revision January 21, 2004.
Address correspondence to M. Jinzaki
(jinzaki{at}sc.itc.keio.ac.jp).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Thirty-seven patients were scanned with MDCT using 2.5-mm collimation. Nephrographic phase data were reconstructed in two ways: a standard protocol (5-mm section thickness, no overlap) and an experimental protocol (3-mm section thickness, 50% overlap). Masses were detected and classified into three groups: group 1, measuring less than 20 H on both protocols (classified as cysts on both); group 2, measuring 20 H or greater on standard protocol and less than 20 H on experimental protocol (classified as cysts using experimental protocol only); and group 3, measuring 20 H or greater on both protocols (not classified as cysts using either protocol). Masses 10 mm or larger in group 3 were evaluated further for enhancement. Statistically significant differences between protocols were assessed using an analysis of counts and proportions.
RESULTS. Of 175 detected lesions, 29 (17%) were detected only with the experimental protocol; all but one were smaller than 5 mm. Using the experimental protocol, of 45 masses between 5 and 10 mm, the number of masses that could be characterized as cysts increased from 13 (29%) to 38 (84%). The overall number of indeterminate lesions was reduced from 101 (69%) of 146 lesions detected with the standard protocol to 86 (53%) of 161 lesions detected with the experimental protocol.
CONCLUSION. Using MDCT and thin overlapping reconstructions, renal cysts as small as 5 mm can be diagnosed with more confidence than is possible with standard reconstructions, and the overall number of indeterminate renal masses is reduced.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
When imaging renal masses with single-detector helical CT, most radiologists use a collimation of 5 mm, a reconstruction thickness of 5 mm, and a reconstruction interval of 5 mm [1–10]. Although renal masses 10 mm or larger can be characterized using 5-mm collimation [2], the characterization of smaller renal masses is difficult without using thinner collimation [1–3]. Thinner collimation requires an increased table translation to cover both kidneys during a single breath-hold, which could result in compromised image quality [12]. Although almost all lesions smaller than 10 mm have no CT features to suggest that they are anything but simple cysts, the lesions are too small to be diagnosed as cysts with confidence. When a collimation of 5 mm is used, attenuation measurements of these lesions are often inaccurate and imaging characteristics are nondiagnostic because of partial volume averaging.
Compared with single-detector helical CT, MDCT allows the kidneys to be
scanned with a collimation of less than 5 mm during a single breath-hold
[13]. From a single data set
obtained with thin collimation, both thin and thick sections can be
reconstructed. Consequently, no additional radiation exposure is required to
obtain the thin sections. Because reducing section thickness reduces partial
volume averaging, protocols using thinner sections would be expected to
improve the detection and characterization of small renal masses. We sought to
determine, when using MDCT and 2.5-mm collimation, if viewing images
reconstructed with a 3-mm thickness and a 50% overlap would improve the
detection and characterization of small (
3 cm) renal masses compared with
images reconstructed with a 5-mm thickness and no overlap.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The projection data from the nephrographic phase were reconstructed with 5-mm-thick sections at 5-mm increments (standard protocol) and with 3-mm-thick sections at 1.5-mm increments (experimental protocol). We overlapped the reconstructions in the experimental protocol to optimize the detection of renal masses, the characterization of renal masses, and the quality of multiplanar reformatted images. Because the two reconstruction protocols used the same projection data, the two image sets were obtained with no increased radiation.
Image Analysis
In our study, only unenhanced and nephrographic phase scans were analyzed.
The urographic phase was not included in the analysis because the
nephrographic phase has been shown to be the optimal phase to image renal
masses for enhancement [5,
9]. Evaluating renal masses
with the urographic phase has not been shown to be of additional value when
the protocol includes both unenhanced and nephrographic phases. Furthermore,
the purpose of this study was to analyze how to optimize the nephrographic
phase scans. Both the unenhanced and the nephrographic phase images were
interpreted using a commercially available workstation (Agfa Medical Imaging).
Two radiologists independently determined the number of lesions detected on
each series without knowledge of prior imaging or clinical or histologic data.
Each observer independently reviewed the entire set of standard protocol
images first and then the experimental protocol set 3–5 days later.
Disagreements were resolved by consensus. When a consensus was not reached,
the final decision as to whether a mass was present was made by a third
radiologist.
Detected masses were characterized by a single radiologist. The size of
each mass was recorded and classified into one of three groups (< 5 mm,
5 and < 10 mm, 10 mm). All masses were assessed for
calcification, septa, cyst wall nodularity or irregularity, and attenuation
values after the administration of contrast medium. Using the Bosniak
classification
[1–3],
lesions without septa or cyst wall nodularity and attenuation of less than 20
H on contrast-enhanced images were considered simple cysts. The mean
attenuation value and standard deviation (SD) of each lesion on
contrast-enhanced images were measured with a centrally placed region of
interest of approximately half the diameter of the lesion. Masses with no
features to suggest that they were anything but simple cysts (i.e., no septa
or nodularity) were classified into one of three groups on the basis of their
attenuation. Masses in group 1 had attenuations of less than 20 H using both
the standard and experimental protocols and were classified as cysts using
both protocols. Masses in group 2 had attenuations of 20 H or higher with the
standard protocol, and less than 20 H with the experimental protocol, and were
therefore classified as cysts using the experimental protocol only. Masses in
group 3 had attenuations of 20 H or higher on both standard and experimental
protocols and therefore were not classified as cysts using attenuation alone
on either protocol. Of the masses in group 3, those 10 mm or larger were
evaluated further for enhancement. To determine the presence of enhancement,
unenhanced attenuation values and their SDs were recorded. These values were
subtracted from the attenuation values on the nephrographic phase images. A
threshold of 10 H was used to determine enhancement. Enhancement was not
evaluated in masses less than 10 mm because virtually all these masses were
too small to detect on 5-mm-thick unenhanced images. Thus, all masses smaller
than 10 mm in group 3 were considered indeterminate.
Data Analysis
To compare the standard and experimental protocols, the following
statistical methods were used. First, to evaluate the detectability of renal
masses in each reconstruction protocol, the detected masses was categorized
into one of the three size groups (< 5 mm, 5 and < 10 mm, 10
mm) and according to whether they were detected using the standard protocol,
the experimental protocol, or both protocols. Furthermore, statistical
hypothesis testing based on a z-test of proportion was conducted for
determining whether more lesions were detected with the experimental protocol
alone than with the standard protocol alone.
Similarly, to compare our ability to characterize lesions using the two protocols, the detected masses were not only stratified by size (as in the previously mentioned three categories) but also categorized into one of the three characterization groups already discussed. Data were analyzed statistically using an analysis of counts and proportions derived from contingency tables.
Estimation of Radiation Dose
The radiation doses for the MDCT protocol used in this study were based on
measured surface dose to a sample of clinical trials unit patients, organ
doses determined from tabular data
[14,
15], and organ weighting
factors [16,
17]. The average effective
doses were 6.4 and 2.5 mSv for the unenhanced phase and nephrographic phase,
respectively.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
Characterization of Renal Masses
None of the 132 masses detected on both protocols had calcification, septa,
nodularity, or irregularity. Among the 32 masses 10 mm or larger, 29 masses
(91%) were classified as cysts on both reconstruction protocols (group 1)
(Fig. 3A,
3B), whereas the remaining
three masses (9%) were hyperdense (group 3)
(Table 2). Of these three
masses, two were homogeneously hyperdense (64 and 52 H), did not enhance using
either reconstruction protocols, and were diagnosed as hyperdense cysts. The
third mass enhanced to 53 H on both protocols, was diagnosed as renal cell
carcinoma, and was confirmed at surgery (Fig.
4A,
4B).
|
|
|
|
|
Of the 44 masses larger than or equal to 5 mm and smaller than 10 mm that were detected on both protocols, 13 (30%) were classified as cysts with both protocols (group 1), 24 (55%) were characterized as cysts only with the experimental protocol (group 2) (Fig. 3A, 3B), and seven (16%) were not classifiable (group 3). In addition to these 44 masses, a 7-mm exophytic mass was detected and characterized as a cyst only with the experimental protocol. Thus, of 45 masses larger than or equal to 5 mm but smaller than 10 mm, 38 (84%) could be characterized as cysts with the experimental protocol, compared with only 13 (29%) with the standard protocol.
Among the 56 masses smaller than 5 mm, five (9%) were characterized as cysts with the experimental protocol only (group 2), all of which were between 4.5 and 5 mm. The other lesions in this group could not be characterized (group 3) (Fig. 3A, 3B).
As shown in Table 2, the
number of lesions that could be diagnosed as cysts with the experimental
protocol only (group 2) was five of 56 for masses less than 5 mm and 24 of 44
for masses greater than or equal to 5 mm and smaller than 10 mm, and 0/32 for
masses 10 mm or larger. These fractions were statistically significantly
greater than zero (p = 0.01 for masses < 5 mm, p <
0.005 for masses 5 mm and < 10 mm), with each result based on a
two-sided z-test of proportion.
Among 146 lesions detected with the standard protocol, 101 lesions could
not be characterized as cysts and were considered indeterminate (51 masses
< 5 mm in group 3; seven masses 5 mm and < 10 mm in group 3; five
masses < 5 mm in group 2; 24 masses 5 mm and < 10 mm in group 2;
and 14 lesions detected with the standard protocol alone). Among 161 lesions
detected with the experimental protocol, 86 lesions could not be characterized
and were considered indeterminate (51 masses < 5 mm in group 3, seven
masses 5 mm and < 10 mm in group 3, and 28 masses detected with the
experimental protocol alone).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
3 cm) masses
[3]. Reducing section thickness
reduces partial volume averaging. Using the central tenet that renal masses
should be evaluated with a maximum section thickness of half the lesion's
diameter to obtain reliable attenuation measurements, a confident diagnosis of
a 10-mm renal cyst would require a reconstructed thickness of 5 mm or thinner.
A confident diagnosis of cysts smaller than 10 mm requires a thinner
reconstruction. We therefore imaged patients using 2.5-mm collimation and
evaluated whether viewing images reconstructed with a 3-mm section thickness
and a 50% overlap (the experimental protocol) would improve the detection and
characterization of small renal masses compared with standard 5-mm
reconstructions with no overlap. Our data showed that using MDCT, thin overlapping reconstructions could improve the detection of renal masses. More masses were detected by the experimental protocol alone, but only one was 5 mm or larger and was characterized as a cyst. All other masses detected only with the experimental protocol were smaller than 5 mm and most could not be characterized as cysts. Another benefit of the experimental protocol was that some lesions identified with the standard protocol were not true masses but rather were long, linear structures that were likely normal portions of unenhanced renal medulla.
Thin overlapping reconstructions obtained with MDCT improved significantly the characterization of small renal masses. The extent of improvement depended on mass size. A significant benefit occurred among masses between 5 and 10 mm. Of these, 84% were characterized as cysts using the experimental protocol, and only 29% were characterized as cysts using the standard protocol. Characterization of lesions 10 mm and larger was not improved, and only a slight benefit occurred among masses smaller than 5 mm. Of masses smaller than 5 mm, 9% were characterized as cysts using the experimental protocol, and none could be characterized as cysts using the standard protocol.
Seven masses between 5 and 10 mm were indeterminate on the basis of their contrast-enhanced attenuation value alone, and thus could represent solid masses, hyperdense cysts, or simple cysts with attenuation that was artifactually elevated by partial volume averaging or pseudoenhancement [18, 19].
Pseudoenhancement has been cited as a cause of artifactual enhancement and is considered to be a result of an inadequate correction of beam hardening in the CT reconstruction algorithm. For this reason, some authors have recommended that the enhancement threshold value be raised to account for pseudoenhancement when imaging small renal masses [18]. However, using MDCT and thin overlapping reconstructions, 84% (38/45) of masses between 5 and 10 mm were characterized as cysts using a contrast-enhanced attenuation value of less than 20 H, a currently accepted criterion. Because we did not evaluate enhancement in masses smaller than 10 mm, the number of indeterminate lesions might have been reduced further if thin (3 mm) overlapping (1.5 mm) reconstructions had been used for the unenhanced sequence and enhancement had been evaluated in all masses.
Although our study showed that the experimental protocol enabled characterization of most cysts 5 mm and larger, most lesions smaller than 5 mm remained indeterminate. Despite finding additional lesions that were smaller than 5 mm and unable to be characterized, the overall number of indeterminate masses was actually reduced using the experimental protocol (86 lesions), compared with the standard protocol (101 lesions). In other words, the benefit of characterizing more masses between 5 and 10 mm as cysts was not offset by the detection of additional indeterminate lesions.
Using MDCT and thin overlapping reconstructions, cysts as small as 5 mm can be diagnosed with confidence. Characterizing such small renal lesions is particularly important in patients in whom a partial nephrectomy is considered. When nephron-sparing surgery is indicated in the treatment of renal cell carcinoma, the detection of additional foci of renal cell carcinoma, reported in 6–25% of cases, is important [20–26]. Using conventional nonhelical CT, additional lesions are found in the same kidney in 15–35% of cases [24–26]. However, many of these lesions are too small to be characterized as benign or malignant. If additional foci of renal cell carcinoma were preoperatively identified with confidence, they could be removed during the operation. Likewise, if additional lesions were diagnosed as cysts, a solitary renal cell carcinoma could be removed with the confidence that the remaining kidney does not contain additional foci of tumor. With single-detector CT, thin overlapping reconstructions of vascular and nephrographic phase data have been used to render 3D images to help plan surgery for partial nephrectomy [27]. Using MDCT and thin overlapping reconstructions, a comprehensive preoperative evaluation of these patients could be provided, both by characterizing lesions as small as 5 mm and by surgical planning.
Other patients with small incidentally discovered renal lesions might also benefit from using MDCT and thin overlapping reconstruction. However, the management of lesions smaller than 10 mm is controversial [2]. On the one hand, most of these lesions are cysts, and the likelihood of missing a small cancer is small. Furthermore, many small renal cancers are indolent [28, 29]. On the other hand, the growth rate of small renal masses is variable and, although uncommon, some grow rapidly [30]. As a result, when radiologists encounter small lesions that cannot be fully characterized, either some form of follow-up (sonography or CT) is suggested, or the radiology report leaves the referring physician and patient with some degree of uncertainty as to the diagnosis of the lesions. Our study shows that MDCT can reconstruct the data with a variety of section thicknesses; the thin sections can be used to further decrease the number of indeterminate lesions.
Because of our findings, we now use MDCT and thin overlapping
reconstructions routinely when evaluating patients with suspicious renal
masses. Using soft-copy viewing, we routinely view all (
60–80)
images using a one-on-one format. Alternatively, particularly if soft-copy
viewing is not available, MDCT and thin overlapping reconstructions could be
used in selected patients, such as those who need nephron-sparing surgery.
Another approach might include reconstructing and viewing images in the
standard way (5-mm-thick sections, 5-mm collimation) and obtaining thin
overlapping reconstructions retrospectively of only those lesions between 5
and 10 mm.
A data set obtained with 2.5-mm collimation and reconstructed in multiple ways, as previously described, can be achieved without an increase in radiation dose. Our study simply analyzed the effect of reconstructing a data set differently. When we used the experimental reconstruction protocol, patient dose was not affected.
Our study has some limitations. First, histologic proof was available in only one mass that proved to be renal cell carcinoma. However, the criteria we used to diagnose simple cysts are well established, have been used successfully in clinical practice for many years, and are used in other studies [5–7, 19]. The rationale for using CT criteria to serve as the gold standard is that, to our knowledge, no lesions fulfilling these criteria for simple cysts have been reported to be malignant. Furthermore, surgical removal of these lesions is not feasible or clinically indicated. Second, attenuation measurements were obtained by a single reviewer. Different reviewers may differ in their measurements of attenuation and this method may have introduced some variation in our results [31]. However, the same observer analyzed both protocols, using the same methods for obtaining attenuation measurements in both. Third, in assessing enhancement in masses 10 mm or larger, 3-mm enhanced images were compared with 5-mm unenhanced images, making assessment of enhancement potentially inaccurate. However, partial volume averaging of masses 10 mm and larger should be insignificant because the section thickness is less than half the diameter of the lesion. Also, we evaluated the ability of our protocol to allow us to characterize small masses as cysts with more confidence. However, we did not assess whether thin reconstructions would help characterize lesions with septa or nodularity or other indeterminate lesions. Thin overlapping reconstructions were not used to detect more solid masses. Whether thin reconstructions improve our ability to characterize complex cysts or solid masses awaits future study. Finally, our analysis examined the combined effect of both a thin reconstructed section and a 50% overlap, so we do not know the benefit of each factor alone. We overlapped the reconstructed images to improve lesion analysis (by reducing partial volume averaging further) and to optimize the quality of multiplanar or 3D images should they be desired. Multiplanar and 3D images should be reconstructed with at least a 50% overlap [12].
In summary, when evaluating small renal masses with MDCT and a 2.5-mm collimation, reconstruction with a 3-mm thickness and a 50% overlap improves both the detection and characterization of renal masses, compared with images reconstructed with 5-mm thickness and no overlap. Although more masses are detected, virtually all are smaller than 5 mm and most cannot be characterized. When MDCT and thin overlapping reconstructions are used, cysts as small as 5 mm can be diagnosed with more confidence than is possible with standard reconstructions and the overall number of indeterminate masses is reduced.
Acknowledgments
We thank Donna L. Vega for help with manuscript preparation.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
3.0 cm) renal parenchymal tumor:
detection, diagnosis, and controversies. Radiology1991; 179:307
–317 3 cm) renal
masses: correlation of spiral CT features and pathologic findings.
AJR 1994;163:597
–605 3 cm)
hyperechoic renal masses: comparison of helical and conventional CT for
diagnosing angiomyolipoma. AJR1996; 167:877
–881
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. G. Silverman, J. R. Leyendecker, and E. S. Amis Jr What Is the Current Role of CT Urography and MR Urography in the Evaluation of the Urinary Tract? Radiology, February 1, 2009; 250(2): 309 - 323. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Silverman, G. M. Israel, B. R. Herts, and J. P. Richie Management of the Incidental Renal Mass Radiology, October 1, 2008; 249(1): 16 - 31. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Iezzi, A. R. Cotroneo, A. Filippone, F. Di Fabio, M. Santoro, and M. L. Storto MDCT Angiography in Abdominal Aortic Aneurysm Treated with Endovascular Repair: Diagnostic Impact of Slice Thickness on Detection of Endoleaks Am. J. Roentgenol., December 1, 2007; 189(6): 1414 - 1420. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Gervais, F. J. McGovern, R. S. Arellano, W. S. McDougal, and P. R. Mueller Radiofrequency Ablation of Renal Cell Carcinoma: Part 1, Indications, Results, and Role in Patient Management over a 6-Year Period and Ablation of 100 Tumors Am. J. Roentgenol., July 1, 2005; 185(1): 64 - 71. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
