| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, Boston Medical Center, Boston University, 88 E.
Newton St., Boston. MA 02118
2
Department of Radiology, University of Melbourne, Melbourne, 3052
Australia.
3
Department of Urology, University of Melbourne, Melbourne, 3052
Australia.
Received July 21, 1999;
accepted after revision September 21, 1999.
Presented in part at the annual meeting of the International Society for
Magnetic Resonance in Medicine, Sydney, Australia, March 1998, and at the
annual meeting of the American Roentgen Ray Society, San Francisco, April-May
1998.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. In a prospective manner 32 patients (age range, 26-78 years; average age, 54 years), each with at least one suspicious mass on CT, were examined with MR imaging. The following sequences were performed: conventional spin-echo with and without fat saturation, fast spin-echo, and dynamic gadopentetate dimeglumine-enhanced infusion using a 1.5-T superconducting magnet. Results were analyzed and compared with pathologic results after resection.
RESULTS. A total of 65 renal masses of average size 2.6 cm (range, 1-10 cm) were detected with dynamic MR imaging. Seventeen of the 65 masses were malignant. Of the 17 malignant masses, three did not enhance on dynamic MR imaging (because of hemorrhage). Sixteen of the 17 malignant masses were heterogeneous on T2-weighted images. Three enhancing masses contained fat and all were angiomyolipomas. Thirty-five of the 65 masses (four with hemorrhage) did not show enhancement, all of which were homogeneous on T2-weighted images and were proven to be cysts. Five masses resulted from infections and had heterogeneous T2 appearance. The remaining masses were three hematomas with hemorrhage, one column of Bertin, and one aneurysm.
CONCLUSION. Renal masses that are interpreted as suspicious on CT may lack MR enhancement because of hemorrhage effects; heterogeneity of their T2 appearance is thus critical in differentiating malignancy from benign disease. Odds-ratio calculations give an adjusted estimate of a 3.36-fold increase (95% confidence interval, 1.8-6.27) in the likelihood of malignancy when masses are heterogeneous on T2-weighted images and a 29-fold increase (95% confidence interval, 3.67-241.8) for predicting malignancy when enhancement is present.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Because most solid renal lesions appear isointense to the surrounding normal renal parenchyma on T1-weighted images and variable in signal intensity on T2-weighted images [3], these images are often difficult to interpret and small lesions may be inconspicuous on these pulse sequences [1, 4]. Therefore, an essential component in the MR imaging evaluation of solid renal lesions is the use of IV gadopentetate dimeglumine, which increases lesion conspicuity and allows accurate assessment of the venous system when renal cell carcinoma is present [4, 5]. The IV administration of gadopentetate dimeglumine has been shown to be safe [6], even in patients with renal insufficiency [7].
As more renal masses are evaluated using MR imaging after suspicious characteristics, such as enhancement on CT, are detected, the potential of bypassing CT and pursuing MR evaluation of suspicious masses detected with other techniques raises the question of what are the potential uses, benefits, and limitations of MR imaging in first-line evaluation. Nonetheless, at this time MR imaging is useful not only in characterizing suspicious masses, but also in potentially staging them in a single examination. Consequently, this study was designed to assess the usefulness of different MR sequences for predicting potential malignancy in the characterization of lesions that are suspicious on CT.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CT
Each patient was placed on the helical CT table (Somatom Plus-S, Siemens
Medical Systems, Erlangen, Germany; HiSpeed, General Electric Medical Systems,
Milwaukee WI; or PQ2000 or PQ5000, Picker, Cleveland, OH) in the supine
position with arms placed above the head and a safety strap secured to support
the arms. If IV access was not in place, a technologist placed a minimum of a
20-gauge angiocath into an antecubital vein. This was then connected to a
power injector (Mark IV; Medrad, Indianola, PA), which had been preloaded with
100 ml of contrast material (iopamidol [Isovue 300], Bracco, Princeton, NJ; or
Optiray-320, Ioversol, Mallickrodt, St. Louis, MO). Initial helical CT from
the diaphragm proceeding caudad was performed using a table feed of 5 mm/sec
for 32 sec with a mean beam width of 5 mm to cover the complete volume of the
kidneys in a single helical acquisition. This helical CT protocol was then
repeated with the administration of IV contrast material. During acquisition
of the helical CT scan using the single breath-hold technique, the patient
received a bolus injection of IV contrast material at 3 ml/sec for 32 sec (96
ml) with image acquisition commencing after a 60-sec time delay to obtain
cortical and medullary enhancement. The images of both studies were
reconstructed at 2-mm intervals and the radiologist reviewed all axial images
(55-92 per study). Any renal lesion that had visually discernible enhancement
on abdominal windows (level, 30-40 H; window, 300-400 H) or enhancement of
more than 20 H on the interior or on a septum was defined as a
"suspicious" lesion that normally would warrant excision and was
referred to MR imaging before surgical exploration. All lesions were
correlated with final pathology. The size of the lesion and the age and sex of
the patients were recorded for subsequent analysis.
MR Imaging Technique
MR imaging was performed with a 1.5-T superconducting magnet (Signa,
General Electric Medical Systems; or ACS, Philips Medical Systems, Best, the
Netherlands) with a body coil for both transmission and reception (phased
array coils were not available by all vendors at the time of this study, and
uniformity in technique was desired). T1-weighted (with a short TR and short
TE) coronal images during breath-holding were routinely obtained as a
localizer using fast spoiled gradient-echo with fat and water in-phase, a
75° flip angle, a TR of 8 msec, a TE of 2 msec, a 48-cm field of view, 128
phase encodes (in the left-right direction), and one excitation. This sequence
was followed by axial imaging using T2-weighted fast spin-echo or turbo
spin-echo technique with prescribed superior and inferior saturation bands,
26-cm field of view, 5-mm slice thickness with a 0-mm gap, a TR range of
3000-4000 msec, a TE range of 90-104 msec, an echo train length of eight, 128
phase encodes (in left-right direction), and four excitations (total
acquisition time, approximately 9 min 4 sec). No respiratory compensation was
used.
Once all renal masses were believed to have been identified on the T2-weighted images, coronal or axial dynamic breath-hold images were obtained centered over the volume of interest using 8-11 sections for each 5-mm thickness with a 0-mm gap; images were acquired during breath-holding with a two-dimensional fast spoiled gradient-echo sequence (TR range/minimum TE, 68-150/4-4.6; flip angle, 60°-80°, number of excitations, one to two) during the administration of a bolus (0.1 mmol/kg of body weight; maximum dose, 15 ml) of gadopentetate dimeglumine (Magnevist, Schering, Berlin, Germany; or Magnevist, Berlex, Wayne, NJ) followed by a 50-ml normal-saline flush. The total 65 ml of fluid was administered over 60 sec by hand with scan acquisition for the first of the scans initiated at the beginning of bolus administration. Time of acquisition was 13-32 sec for the complete set with a predefined break of 15 sec between each acquisition; this was repeated six times sequentially followed by additional scans at 5 and 10 min. The TR and the number of excitations were tailored to give best signal-to-noise ratio based on the patient's breath-hold capability during the localizer sequence. In-plane resolution was 2 x 1 mm (matrix size, 128 x 256; field of view, 260 x 260 mm). No saturation pulses were applied.
Axial T1-weighted spin-echo MR images were obtained with the following parameters: prescribed superior and inferior saturation bands, 26-cm field of view, 5-mm slice thickness with a 1-mm gap, a TR range of 300-500 msec, a TE range of 10-20 msec, 256 phase encodes, and two excitations. Sequences were performed with flow and respiratory compensation and with and without frequency-selective fat-saturation pulses after careful shimming and tuning the pulse to 220 Hz below the water peak. Images were also obtained before and after administration of gadopentetate dimeglumine. Total acquisition time was approximately 5 min.
Image Analysis
All MR imaging sequences were evaluated by two radiologists who were
unaware of the presumed clinical diagnosis and individual opinion was formed
and recorded. The pathologic diagnosis formed the true diagnosis for
evaluation of the MR images. Two months later, the same reviewers were
independently presented with MR images alone and asked to make a diagnostic
evaluation of the renal masses, which was subsequently recorded. The kappa
statistic was applied to test the observer agreement for the MR techniques
[8]. The number, size, and
enhancement pattern of the lesions, whether the lesions contained fat, and
their heterogeneity on T2-weighted images were recorded. The severity of
hemorrhage was not evaluated on MR images.
Visual Rating
All T2-weighted and dynamic studies were rated using the following 3-point
quality scale: 1, not satisfactory for diagnosis; 2, adequate image quality
for diagnosis; and 3, excellent image quality. The difference between a rating
of 2 or 3 points was that a score of 2 points was given if it was believed
that it was possible to improve the quality of the images. The validity of
this scoring scheme has been validated elsewhere
[9,
10] with kappa statistic
measured [8].
Quantitative Analysis
The diameter of all intrarenal lesions was obtained from the hard copy of
the T2-weighted images by a single reviewer using handheld calipers. The
average size was 2.6 cm (range, 1-10 cm). The largest possible oval or
circular region of interest located within the confines of the lesion was
selected. The signal intensity of each lesion on conventional spin-echo images
before and after contrast material administration, with and without fat
saturation, and on the dynamic sequences was measured and recorded. Dynamic
sequences showed enhancement that was rapid in a mass of the right kidney
(Fig.
1A,1B,1C,1D)
yet was less than that in the cortex. The dynamic enhancement curves were
calculated and plotted against time. Note that lack of prescanning makes
objective comparison of signal-intensity measurements between sequences
impossible.
|
|
|
|
Visual Rating
Dynamic studies.—All dynamic studies were viewed as
sequential images on film and enhancement pattern was classified compared with
cortical and medullary enhancement. Classification types
(Fig. 2) included M, enhancing
comparable with the medulla; C, enhancing comparable with the cortex; 1,
enhancing comparable with the medulla but less than the cortex (i.e.,
discernible from cortex and medulla with some enhancement); 2, nonenhancing;
and 3, enhancing more rapidly than cortex. Inter- and intraobserver error were
measured and kappa statistic was calculated. Visual and quantitative
enhancement evaluations were compared with the kappa statistic, measured by
defining the classification of the quantitative curves as M, enhancing between
20% and 100% of medullary signal; C, enhancing between 90% and 110% of
cortical signal; 1, enhancing between cortex and medulla; 2, enhancing less
than 20% of medullary signal; and 3, peak before renal parenchyma.
|
T2 character.—All lesions on conventional T2-weighted image sequences were scored as heterogeneous or homogeneous and the classification was recorded. A homogeneous lesion was rated as such if it was of uniform intensity on each image throughout the lesion volume, without the presence of a sharply demarcated signal change. Thus, slow gradations caused by signal drop-off or respiratory phase encode artifact were not counted, but nodules, layered intensity levels, or sharp chemical-shift effects would all be rated as heterogeneous. Inter- and intraobserver error were measured with the kappa statistic.
Statistical analysis.—All data collected were organized on a spreadsheet (Excel; Microsoft, Redmond, WA), and logistic regression analysis was performed using software (version 5.0 STATA; STATA, College Park, TX). Univariate analysis was used to determine the significance of each factor in predicting malignancy by odds-ratio evaluation. The most significant variables, those at the cutoff p value of less than 0.05, were collected together, and colinearity and confounding and effect modification were assessed by forward-selection, backward-selection, and selective elimination [11]. Odds-ratios were then calculated for the resulting best-fit model with receiver operating characteristic and the Hosmer-Leme show goodness-of-fit test was calculated [11].
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Subjective Evaluation
Study distribution indicated that 0% of images were nondiagnostic, 90% were
diagnostic, and 10% were of excellent quality, with an inter- and
intraobserver kappa statistic for study quality that ranged from 0.77 to 1.0.
Rating of T2 heterogeneity had inter- and intraobserver kappa of 0.83.
Agreement between subjective and objective classification of lesion
enhancement was 100%. These findings are shown in
Table 2 in which the mean
signal intensity for the type-1 and type-2 lesions is presented. The lack of
discordance may be because of the lack of any lesions that could be classified
as type-C or type-M. All results are hence based on the subjective
classification.
|
Logistic Regression
Analysis of univariate significance of the odds-ratio revealed that
significant variables at the 0.05 cutoff were heterogeneity on T2-weighted
images, with an odds-ratio value of 39 (95% confidence interval [CI],
6.97-218); presence of enhancement, with an odds-ratio value of 78 (95% CI,
11.6-523). Nonsignificant variables were the sex and age of the patient and
the size of the lesion. Analysis of colinearity showed a strong correlation
between size and enhancement, with large lesions being more likely to enhance
(p = 0.001) and to be heterogeneous on T2-weighted images (p
= 0.002), and older patients being more likely to have enhancing lesions
(p = 0.003) that are heterogeneous on T2-weighted images (p
= 0.009). No confounding or effect modification was seen with adjusted odds
ratios of 3.36 (95% CI, 1.8-6.27) for heterogeneity on T2-weighted images, and
29 (95% CI, 3.67-241.8) for presence of enhancement for predicting malignancy.
Receiver-operating-characteristic curve analysis showed that use of these
factors allows accurate differentiation of benign from malignant lesions with
an Az (area under the curve) of 0.96
(Fig. 3).
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
When performing MR imaging of renal lesions, contrast resolution can be improved by using breath-hold and fat-suppressed techniques to reduce artifacts and by the administration of IV gadopentetate dimeglumine. In fact, gadolinium-enhanced sequences with fat saturation have been shown to be more sensitive than contrast-enhanced CT (single phase; collimation, 10 mm) for the detection of tumors of less than 2 cm in diameter [4]. This finding carries great significance when nephron-sparing surgery is being considered for a renal cell carcinoma because synchronous lesions measuring 1-15 mm have been reported to occur with a frequency as high as 19.7% [13]. It is in this setting that this study was performed, because improving the accuracy of lesion classification before nephron-sparing surgery minimizes unnecessary procedures and should improve outcome in patients in whom lesions are detected earlier.
Treatment of renal cell carcinoma is based on tumor stage, which includes evaluation of tumor extent and venous involvement. The multiplanar imaging capabilities of MR imaging are well suited to the evaluation of tumor extent and gradient-recalled echo sequences are 80-100% accurate for detecting the presence and extent of tumor into venous structures [2, 14]. It is imperative to observe that lack of enhancement of a renal lesion on administration of gadopentetate dimeglumine, particularly for small lesions (<1 cm), not be considered a sufficient criterion for excluding malignancy, as the hemorrhagic lesions in our study show. Figure 4A,4B,4C,4D,4E illustrates a case of an 8-mm hemorrhagic lesion that did not show enhancement on gadopentetate dimeglumine administration. It is in this setting that the appearance on T2-weighted images must be considered.
|
|
|
|
|
The use of logistic regression analysis with adjusted odds-ratio calculations allows the determination of what factors are significant predictors for malignancy while adjusting for bias in referral patterns. In particular, if a patient population has a baseline probability of malignancy, determined using a test that indicates that a positive result has an odds ratio of two, then the ratio of the post-test probability of malignancy to benign (the odds of having malignancy) is doubled. For example, using the adjusted odds ratios of this study, if we have a population with a priori probability of renal malignancy of 0.1, then enhancement on MR imaging increases the probability to 0.76 and heterogenecity on T2-weighted images increases the probability to 0.27, whereas a heterogeneous mass on T2-weighted images with enhancement has a probability of 0.92 of being malignant. In essence, if we know the odds ratio (OR) we can calculate the probability of malignancy in enhancing masses (P2) if we know the probability of malignancy if there is no enhancement (P1), calculated by the following formula: P2 = P1 x OR / (OR x P1 + 1 - P1).
The use of logistic regression analysis allows the separation of the most important and independent variables for predicting malignancy; any form of enhancement on gadopentetate dimeglumine-enhanced T1-weighted MR sequences, followed by heterogeneous appearance on T2-weighted sequences. These factors thereby indicate that although enhancement is sufficient for diagnosing malignancy, nonenhancement is not sufficient to exclude malignancy. Although these lesions are clearly hypervascular given their enhancement on CT, the presence of hemorrhagic products may cause localized signal nulling that presents as nonenhancement on post-gadopentetate dimeglumine MR images. It is thus necessary to include T2-weighted imaging of renal masses so that T2 heterogeneity can be evaluated in nonenhancing masses. In particular, although a large number of lesions were benign cysts, the use of stringent criteria—enhancement of at least 20 H or enhancing thick septae or nodules—did result in many of these cysts being classified as suspicious on CT. However, the multiplanar ability of MR imaging, lack of enhancement, and homogeneity on T2-weighted images were indicative of the benign nature of the masses, thus validating the usefulness of MR in further characterizing suspicious lesions. In the setting of a T2 heterogeneous nonenhancing mass, careful follow-up after an antibiotic trial may be a prudent recommendation to avoid nephrectomy of renal abscess and to avoid misdiagnosis of a hemorrhagic renal cell carcinoma. Lastly, although newer imaging techniques to improve the quality and shorten the time for MR image acquisition are continuously being proposed, the ability to show a statistical correlation between pathologic results and older imaging techniques merely establishes another standard that newer techniques must meet. Specifically, although we used standard T1-weighted spin-echo images and nongated T2-weighted images, the statistical correlation indicates that any newer renal imaging must be similarly validated before it replaces these methods.
Limitations in this work stem from the potentially small number of patients (n = 32); however, this was a very selected population. The lack of phased array coils and respiratory compensation techniques may help in improving sensitivity, but we doubt that enhancement would be shown in small hemorrhagic lesions or that the potential usefulness of T2 heterogeneity would be obviated. The need to use T2 heterogeneity in addition to enhancement character appears to be a relatively discordant result with the prevailing literature [3]; however, these findings can be reconciled in that the effects of the hemorrhagic products that can null gadolinium chelate enhancement also may contribute to heterogeneity on T2-weighted images. In prior studies, this finding may have been limited by the small number of lesions and the requisite need for pathologic correlation, which is not difficult but can be troublesome [15].
In conclusion, although hemorrhagic products may obscure enhancement on dynamic gadopentetate dimeglumine T1-weighted MR imaging, the integration of T2 appearance can be clinically useful in improving the ability to differentiate benign from malignant renal lesions regardless of the size of the lesion or the sex or age of the patient.
Acknowledgments
We thank Ian Hood for photographic assistance; General Electric, Australia,
for technical support for this research project; and Schering Pty. Ltd.,
Australia, and Berlex USA for supplies.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Kim, M. Jain, A. B. Harris, V. S. Lee, J. S. Babb, E. E. Sigmund, L. E. Rueff, and B. Taouli T1 Hyperintense Renal Lesions: Characterization with Diffusion-weighted MR Imaging versus Contrast-enhanced MR Imaging Radiology, June 1, 2009; 251(3): 796 - 807. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. R. M. Sun, L. Ngo, E. M. Genega, M. B. Atkins, M. E. Finn, N. M. Rofsky, and I. Pedrosa Renal Cell Carcinoma: Dynamic Contrast-enhanced MR Imaging for Differentiation of Tumor Subtypes--Correlation with Pathologic Findings Radiology, March 1, 2009; 250(3): 793 - 802. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J. Beer, M. Dobritz, N. Zantl, G. Weirich, J. Stollfuss, and E. J. Rummeny Comparison of 16-MDCT and MRI for characterization of kidney lesions. Am. J. Roentgenol., June 1, 2006; 186(6): 1639 - 1650. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. E. Applegate, R. Tello, and J. Ying Hypothesis Testing III: Counts and Medians Radiology, September 1, 2003; 228(3): 603 - 608. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. B. Ho, S. F. Allen, M. N. Hood, and P. L. Choyke Renal Masses: Quantitative Assessment of Enhancement with Dynamic MR Imaging Radiology, September 1, 2002; 224(3): 695 - 700. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
