Comparison of Contrast-Enhanced Sonography with Unenhanced Sonography and Contrast-Enhanced CT in the Diagnosis of Malignancy in Complex Cystic Renal Masses
Abstract
OBJECTIVE. The objective of our study was to evaluate contrast-enhanced sonography in the diagnosis of malignancy in complex cystic renal masses.
MATERIALS AND METHODS. We analyzed a series of 40 cystic renal masses (diameter, 2–8 cm) with a complex pattern at contrast-enhanced CT in 40 consecutive subjects (18 men, 22 women; mean age ± SD, 62 ± 11 years). Each renal mass was scanned using sonography without contrast material and after IV injection of sulfur hexafluoride–filled microbubbles during the arterial (15–40 seconds) and venous (40–120 seconds from injection) phases. Two radiologists in consensus assessed onsite the enhancement patterns in the peripheral wall and intracystic septa and the evidence of solid endocystic components. Three blinded readers with 2, 6, and 10 years of experience in renal imaging performed a retrospective off-site interpretation of unenhanced sonography, contrast-enhanced sonography, and CT images and made a benign or malignant diagnosis according to refer ence diagnostic criteria for contrast-enhanced sonography and to the Bosniak classification for CT.
RESULTS. Final diagnoses comprised two multilocular cystic nephromas, two inflammatory and seven hemorrhagic cysts, and eight uncomplicated benign cysts and 21 cystic renal cell carcinomas. The overall diagnostic accuracy of contrast-enhanced sonography was better than unenhanced sonography and CT (contrast-enhanced sonography vs unenhanced sonography vs CT: reader 1, 83% vs 30% vs 75%; reader 2, 83% vs 30% vs 63%; reader 3, 80% vs 30% vs 70%; p < 0.05, McNemar test).
CONCLUSION. Contrast-enhanced sonography was found to be better than unenhanced sonography and CT in the diagnosis of malignancy in complex cystic renal masses.
Introduction
Renal cysts are common incidental findings during the clinical diag nostic workup of patients [1]. Renal cysts frequently present an obvious simple cystic pattern at sonography or CT and such cysts do not deserve any further imaging assessment or surgical procedure. In some cases, however, renal cysts show a complex pattern and vary in their malignant potential according to the number and thickness of intracystic septations and the presence of mural nodules and peripheral calcifications.
The Bosniak classification [2, 3] cat ego rizes renal cysts according to their CT features. Category I cysts are simple benign fluid-containing cysts with the attenuation of water and thin walls without septa or calcification. Category II cysts are minimally complicated benign cysts with hairline-thin septa and fine calcifications in the walls or septa, or segmental slightly thickened calcifications with or without minimal septal or mural enhancement. High-attenuation, sharply marginated, completely intrarenal, nonenhancing cystic masses ≤ 3 cm are also included in this category. Category IIF, with the “F” referring to follow-up, is composed of more complex cystic lesions that cannot be classified as category II or that are complex enough to be characterized as category III because they may contain an increased number of septa and an increased amount of calcifications in the walls or septa, which may be thicker and nodular; high-attenuation, sharply marginated, completely intrarenal, nonenhancing cystic masses > 3 cm are also included in this category. Category III are indeterminate cystic masses whose benign or malignant nature cannot be determined with imaging studies with thickened (> 2 mm) irregular walls or septa that may appear hyperdense on unenhanced CT and may contain either small or large amounts of calcification with septal or mural contrast enhancement. Category IV cysts show either small or large amounts of calcifications within a thickened, enhancing irregular wall or septum and present enhancing soft-tissue nodular components adjacent to or extending from the wall or septum.
The Bosniak classification is considered to be an accurate and efficient method for treatment planning. Although the Bosniak classification scheme is very useful for the clinical management of cystic renal masses, inter reader variation in distinguishing between category II, IIF, and III lesions does exist [4] and may present problems in recommending surgical versus conservative management in some cases. Moreover, CT may not reveal thin intra cystic septations due to volume averaging, which limits identification of the renal cysts that deserve further assessment by follow-up. MRI and CT reveal similar findings in most cystic renal masses, even though MRI may depict additional septa, thickening of the wall or septa, or enhancement that may lead to an upgraded Bosniak category and can affect case management [5].
The content of renal cysts typically appears more complex at sonography than CT owing to the greater accuracy of sonography in depicting thin intracystic septations and the corpus cular content due to hemorrhage or tumor debris [1]. Contrast-enhanced sonography with microbubbles [6] has been shown to improve the visualization of renal vessels [7], to depict tumor vascularity in solid renal masses [7–11], and to improve the sonographic detection of vascularity within the septations and peri pheral wall of complex cystic renal masses [7, 11]. Recently, a classification and diagnostic workup scheme for cystic renal lesions that uses contrast-enhanced sonography as the reference technique was proposed [12, 13]. Contrast-enhanced sono graphy is appropriate in the Bosniak classification of renal cysts and was found to be superior to CT in detecting additional septa, thickening of the wall or septa, and solid components [14–16]. However, no previous study, to our know ledge, has described the diagnostic impact of contrast-enhanced sonography on the char acterization and diagnostic workup of com plex cystic renal masses. The aim of this study was to evaluate contrast-enhanced sonography in the diagnosis of malignancy in complex cystic renal masses.
Materials and Methods
Patients
This study was approved by the ethics committee, which is equivalent to an institutional review board, at our institution and patient informed consent was obtained after the procedure had been fully explained. From a computerized search of our hospital database of radiologic records between May 2001 and May 2007 performed by one reference radiologist and then cross-referenced to the histopathologic records, we identified a series of 55 subjects with 55 complex cystic renal masses ranging from 2 to 8 cm in diameter (mean ± SD, 4.3 ± 2.2 cm) that were incidentally detected during abdominal sonography or CT (n = 35 patients) or during a clinical diagnostic workup including sonography and CT of the kidneys and urinary tract for micro- or macroscopic hematuria with or without flank pain (n = 20 patients). All cystic renal masses revealed a complex pattern at the time of the initial CT interpretation by the onsite radiologists, consisting of evidence of thickened or hairlinethin intracystic septa or peripheral wall thickening > 2 mm or evidence of peripheral mural or septal nodules or calcifications.
From 3 to 25 days after detection, each renal mass was scanned using contrast-enhanced sonography after injection of a suspension of phospholipid-stabilized microbubbles filled with sulfur hexafluoride (SonoVue, Bracco). Patients with simple cysts that had a hairline-thin wall and anechoic content without septa or calcifications or those with complex renal cysts not scanned by contrast-enhanced sonography were not included in the computerized search.
We excluded 15 cystic renal masses in 15 pa tients because of suboptimal quality of the CT study (lack of unenhanced or arterial corticomedullary phase images in five patients) or of the contrast-enhanced sonography study (failure in data storage or incomplete lesion visibility due to deep position or bowel gas interposition in 10 patients). Therefore, 40 renal masses in 40 patients (18 men, 22 women; mean age ± SD, 62 ± 11 years; median age, 64 years; age range, 29–84 years) were finally included in the study.
CT Examinations
At the time of detection, all cystic renal masses were scanned by multiphase helical CT performed using either a single-detector scanner (n = 21; TomoScan Ave1, Philips Healthcare) or an MDCT scanner (n = 19; Aquilion, Toshiba Medical Systems). A beam collimation of 5 mm and a pitch factor of 1.5 were routinely used with the single-detector scanner, and dedicated thinner sections of 3 mm with 1-mm slice overlap were adopted for small masses (< 3 cm in diameter). MDCT data sets were acquired with the following technical parameters: rotation time, 400 milliseconds; beam collimation, 64 × 0.5 mm; normalized pitch, 1; reconstruction interval, 3 mm; tube voltage, 120 kVp; and tube current (effective mAs), 180–250 mAs, depending on patient size. Unenhanced images were acquired first. Afterward, each patient received 120–140 mL (300–350 mg I/mL) of nonionic iodinated contrast agent (iomeprol [Iomeron, Bracco]) at a rate of 2–3 mL/s followed by 50 mL of saline flush. Scanning delays were 30 seconds for the corticomedullary phase and 90–110 seconds for the nephrographic phase.
Contrast-Enhanced Sonography Examinations
Two radiologists who have experience in sonography (2 and 8 years, respectively) and were blinded to the findings of the reference procedures were involved in the scanning. Patients were scanned with state-of-the-art sonography digital equipment (Sequoia, Acuson–Siemens [n = 32 patients or EsaTune, Esaote [n = 8 patients]) using a 2–5-MHz wideband convex-array transducer. Each lesion was scanned first using unenhanced gray-scale sonography with noise- and speckle-reducing modes, such as tissue harmonic and compound imaging, and then using color or power Doppler sonography with spectral analysis of peripheral and intralesional tumoral vessels. Doppler parameters were optimized to detect slow flow velocities: a pulse repetition frequency of 700 Hz, medium wall filter, and low-velocity-flow optimization. Color gain was raised until color noise became evident and was then lowered slightly to clean the image.
Each renal mass was scanned after the injection of an IV bolus of microbubbles (2.4 mL) followed by a 10-mL normal saline flush using a 20- or 22-gauge peripheral cannula. The technical parameters were as follows: Cadence Contrast Pulse Sequencing (Acuson–Siemens Medical Solutions) (n = 32 patients) or Contrast Tuned Imaging (Esaote) (n = 8) as contrast-specific modes with sensitivity similar to microbubble harmonic signal, low transmit power insonation (mechanical index, 0.09–0.14), dynamic range of 65 dB, temporal resolution between frames of 75–100 milliseconds (10–13 frames per second), echo-signal gain below noise visibility to allow the best visualization of the renal mass and adjacent renal parenchyma, lowest possible frame rate (7–9 Hz), signal persistence turned off, and one focus set immediately below the renal mass. The arterial phase was timed from 10 to 40 seconds after microbubble injection, after which the venous phase encompassed the time interval from 45 seconds after microbubble injection until the microbubbles disappeared.
Distinct digital cine clips for the unenhanced sonography and for the arterial and venous phases of contrast-enhanced sonography were stored on a PC (Pentium 4, Intel) connected to the sonography equipment by a high-performance hardware-based real-time Moving Picture Experts Group-2 (MPEG-2) encoder (MVR1000 card, Mediacruise software, Canopus) and frame-grabber software (Mediacruise, Canopus). Cine clips were subsequently stored on DVDs after the end of scanning.
Onsite Visual Analysis by Consensus
The digital cine clips of unenhanced sonography and contrast-enhanced sonography were preliminarily reviewed onsite by the two radiologists involved in scanning immediately after the end of the examination. The analysis was performed on screen (Intel, Pentium 4 with 19-inch [48-cm] thin-film transistor [TFT] display). The following features of each cystic renal mass were recorded: size of the mass; the number of locules in the cyst; diffuse wall or septal thickness ≤ 2 mm or > 2 mm; presence of septa (few or multiple septa); presence of focal wall or septal thickening; presence and size of solid endocystic components, including mural and septal nodules; presence and size of calcifications; echogenicity of the cyst; and presence of contrast enhancement of the peripheral wall or intracystic septa or evidence of solid endocystic components after microbubble injection. A nodule, septum, or wall was considered enhancing on visual analysis if there was evidence of microbubbles traveling inside it. Discrepant interpretations were resolved by consensus after involvement of an additional reader with similar experience in contrast-enhanced sonography.
Retrospective Off-Site Interpretation
The off-site independent review of unenhanced sonography, contrast-enhanced sonography, and CT scans was performed by three independent readers who were radiologists with 2, 6, and 10 years' experience in renal imaging and affiliated with the center where the study was performed. The readers had not been involved in the scanning and were blinded to patient identity, clinical histor ies, histology results, and other imaging findings.
All readings of unenhanced sonography and contrast-enhanced sonography digital cine clips were performed separately by each reader with a 1-day interval on the same computer (Intel, Pentium 4 with 19-inch [48-cm] TFT display and resolution of 2,560 × 1,600 pixels) at a central location using PowerDVD software (CyberLink). During each session, a sequence of the unenhanced sonography and contrast-enhanced sonography cine clips of the different cystic renal masses was randomly assigned to each reader and any patient-identifying information was masked. Readers were asked to express a diagnosis of benignancy or malignancy for each complex cystic renal mass after reviewing the unenhanced sonography and contrast-enhanced sonography cine clips.
The malignancy criteria for unenhanced sonography included evidence of thickened (> 2 mm) irregular wall or septa with or without gross calcifications or mural or septal nodules with vascular signals on color or power Doppler analysis. According to these criteria, the three readers rated their diagnostic confidence on a 5-grade scale: 1, definitely benign, evidence of hairline-thin intracystic septa; 2, probably benign, evidence of hairline-thin intracystic septa with addition of fine calcifications in the walls or septa; 3, indeterminate, increased number of septa or increased amount of calcifications or corpusc ular echoic content; 4, probably malignant, thick, irregular walls or septa; and 5, definitely malignant, peripheral mural nodules with vascular signals at color Doppler analysis.
The malignancy criteria for contrast-enhanced sonography were developed from previous studies [9, 12, 13, 15]. Figure 1A, 1B, 1C, 1D, 1E, 1F shows the different vascularity profiles that may be observed in a complex cystic renal mass after microbubble contrast agent injection. According to these criter ia, the three readers rated their diagnostic confidence on a 5-grade scale: 1, definitely benign, evidence of contrast enhancement in the hairline-thin septa or continuous or discontinuous enhancement of the peripheral wall; 2, probably benign, en hancement involving both the peripheral wall and the hair line-thin intracystic septa; 3, indeter minate, limited or absent visibility of intra cystic septa or mural nodules due to peripheral wall calcifications; 4, probably malignant, peri pheral enhancing wall combined with thickened enhancing septa; and 5, definitely malignant, evidence of peripheral enhancing mural nodules or peripheral thickened enhancing wall with thickened enhancing septa and enhancing mural and septal nodules.
Visual analysis of CT images was retrospectively performed by the same readers 4 weeks after they had reviewed the digital cine clips recorded during the sonography examination. All readings were performed separately by each reader with a 1-day interval on a PACS-integrated workstation (19-inch [48-cm] TFT display; resolu tion, 2,560 × 1,600 pixels; EbitAET, Esaote) at a central location using a pro prietary soft ware pack age (EbitAET, Esaote). The diagnostic criteria included evaluation of both unenhanced and contrast-enhanced CT scans and corresponded to the Bosniak classification system. Enhancement of the solid endocystic components was assessed by measuring the region of interest on CT images, and enhancement was considered present when attenuation increased by more than 15 HU. According to these criteria, both readers rated their diagnostic confidence on a 5-grade scale: definitely (grade 1) or probably (grade 2) benign for cystic renal masses exhibiting a Bosniak I, II, or IIF pattern; indeterminate (grade 3) for those with a Bosniak III pattern; probably (grade 4) or definitely (grade 5) malignant for those exhibiting a Bosniak IV pattern.
Reference Standards for Diagnosis
The diagnosis and clinical management of the complex renal cysts were determined by the findings at the time of the initial interpretation of the CT scan. The cystic renal masses that were classified as Bosniak category III or IV (n = 28) were surgically resected, whereas those classified as Bosniak category IIF (n = 5) or II (n = 7) underwent imaging follow-up consisting of CT alone or conventional sonography and CT for a minimum of 12 months and a maximum of 24 months (mean follow-up time, 18 months). All follow-up examinations were interpreted by on-site radiologists with at least 5 years' experience. The criteria used to assess the stability of a mass were the absence of an increase in size, as measured on transverse CT images, and the absence of morphologic and contrast-enhancement changes. Cystic renal masses that showed morphologic changes at follow-up, including the appearance of septal thickening or of mural and septal nodules (n = 4), were surgically resected.
Statistical Analysis
A biostatistician participated in the statistical analysis performed by a computer software package (Analyze-It, version 1.63, Analyze-It Software). Sample size was determined with the assumption of an estimated diagnostic accuracy of 85% and a 95% CI of 10% [17]. Retrospective benign or malignant diagnosis was considered true-positive (lesion correctly assessed as malignant with confidence level of 4 or 5), false-negative (malignant lesion incorrectly assessed as benign with confidence level of 1 or 2 or assessed as indeterminate with confidence level of 3), true-negative (lesion correctly assessed as benign with confidence level of 1 or 2), or false-positive (benign lesion incorrectly assessed as malignant with confidence level of 4 or 5 or assessed as indeterminate with confidence level of 3).
The weighted kappa statistic was calculated to assess interreader agreement in diagnostic confidence for unenhanced sonography, contrast-enhanced sonography, and CT. Agreement was graded as poor (κ < 0.20), fair (≥ 0.20 and < 0.40), moderate (≥ 0.40 and < 0.60), good (≥ 0.60 and < 0.80), or very good (≥ 0.8–1). The McNemar test was used to compare the sensitivity and specificity of unenhanced sonography, contrast-enhanced sonography, and CT, whereas improvement in diagnostic confidence was assessed using receiver operating characteristic (ROC) curve analysis by plotting the sensitivity (true-positive fraction) against 1 – specificity (false-positive fraction). The area under each ROC curve was calculated using a nonparametric method [18], and the method proposed by Hanley and McNeil [19] was used to compare areas under each ROC curve. A p value < 0.05 was considered to indicate a statistically significant difference.
Results
Onsite Visual Analysis by Consensus
Table 1 shows the final diagnoses obtained by reference standard procedures. After microbubble injection, benign lesions revealed peripheral wall enhancement involving (n = 8) or not involving (n = 11) the intracystic septa (Fig. 2A, 2B, 2C). Inflammatory and hemorrhagic cysts revealed peripheral wall enhancement without any intracystic septa (Fig. 3A, 3B, 3C). In multilocular cystic nephromas, hairline-thin and thickened enhancing septa were observed together with a thickened peripheral enhancing wall (Fig. 4A, 4B, 4C, 4D). After microbubble injection, cystic renal cell carcinomas revealed peripheral enhancing mural nodules with or without thickened enhancing septa (n = 3; Fig. 5A, 5B, 5C), thickened peripheral enhancing wall combined with thickened enhancing intratumoral septa (n = 9), or peripheral thickened enhancing wall with enhancing mural and septal nodules (n = 9; Fig. 6A, 6B, 6C).
| Cyst No. | Diameter (cm) | Bosniak Categorya | Reference | Final Diagnosis |
|---|---|---|---|---|
| 1 | 3.0 | IV | Histology | Clear cell type carcinoma |
| 2 | 4.0 | III | Follow-up | Hemorrhagic cyst |
| 3 | 6.0 | IV | Histology | Clear cell type carcinoma |
| 4 | 4.5 | IV | Histology | Clear cell type carcinoma |
| 5 | 2.0 | IIF | Follow-up | Uncomplicated cystb |
| 6 | 7.4 | III | Histology | Clear cell type carcinoma |
| 7 | 5.2 | IV | Histology | Clear cell type carcinoma |
| 8 | 1.5 | II | Follow-up | Hemorrhagic cyst |
| 9 | 8.0 | III | Histology | Clear cell type carcinoma |
| 10 | 2.2 | II | Follow-up | Uncomplicated cystb |
| 11 | 4.1 | IV | Histology | Clear cell type carcinoma |
| 12 | 3.2 | IV | Histology | Clear cell type carcinoma |
| 13 | 4.0 | III | Histology | Hemorrhagic cyst |
| 14 | 3.0 | IV | Histology | Clear cell type carcinoma |
| 15 | 4.0 | III | Histology | Multilocular cystic nephroma |
| 16 | 3.0 | IV | Histology | Papillary cell type carcinoma |
| 17 | 2.5 | IV | Histology | Papillary cell type carcinoma |
| 18 | 1.6 | IIF | Follow-up | Uncomplicated cystb |
| 19 | 4.5 | IV | Histology | Papillary cell type carcinoma |
| 20 | 3.0 | IV | Histology | Papillary cell type carcinoma |
| 21 | 4.6 | IV | Histology | Papillary cell type carcinoma |
| 22 | 3.0 | IIF | Follow-up | Uncomplicated cystb |
| 23 | 5.0 | IV | Histology | Papillary cell type carcinoma |
| 24 | 8.0 | IIF | Follow-up | Uncomplicated cystb |
| 25 | 6.0 | IV | Histology | Papillary cell type carcinoma |
| 26 | 3.0 | III | Histology | Papillary cell type carcinoma |
| 27 | 5.0 | III | Histology | Multilocular cystic nephroma |
| 28 | 2.3 | III | Histology | Inflammatory cyst |
| 29 | 5.0 | IV | Histology | Clear cell type carcinoma |
| 30 | 2.0 | III | Histology | Hemorrhagic cyst |
| 31 | 6.0 | IV | Histology | Papillary cell type carcinoma |
| 32 | 3.0 | III | Histology | Hemorrhagic cyst |
| 33 | 2.0 | III | Histology | Inflammatory cyst |
| 34 | 1.8 | III | Histology | Hemorrhagic cyst |
| 35 | 5.4 | IV | Histology | Papillary cell type carcinoma |
| 36 | 3.0 | III | Histology | Hemorrhagic cyst |
| 37 | 6.0 | IIF | Follow-up | Uncomplicated cystb |
| 38 | 2.8 | IV | Histology | Papillary cell type carcinoma |
| 39 | 6.0 | IIF | Follow-up | Uncomplicated cystb |
| 40 | 3.2 | II | Follow-up | Uncomplicated cystb |
a
Bosniak category at the time of the initial CT interpretation.
b
Uncomplicated benign cysts revealed intracystic septations or corpuscular pattern on unenhanced sonography and appeared hyperdense on unenhanced CT due to high protein content.
Retrospective Off-Site Interpretation
Table 2 shows the results of the retrospective image interpretation of unenhanced sonog raphy, contrast-enhanced sonography, and CT. Two (according to reader 1), 10 (reader 2), and four (reader 3) cystic renal masses, respect ively, were correctly classified as benign after review of the contrast-enhanced sonog raphy scans, whereas the same renal masses were classified as indeterminate after review of the CT images. On the other hand, one (according to reader 3) or two (readers 1 and 2) cystic renal masses were correctly characterized as malignant after review of the contrast-enhanced sonog raphy scans, which showed a solid enhancing mural nodule missed at CT. After review of the contrast-enhanced sonog raphy scans, all readers misclassified two benign cystic renal masses—corresponding to multilocular cystic nephromas—with evidence of en hancing peripheral wall and thickened intracystic septations and one (according to reader 3) or three (readers 1 and 2) malignant renal masses with diffuse mural calcifications, which limited visibility of the enhancing peripheral wall, mural nodules, and intracystic septa. The same cystic renal masses were classified as Bosniak category IIF or III after review of the CT scans.
| Performance Measure | Unenhanced Sonography | Contrast-Enhanced Sonography | CT |
|---|---|---|---|
| Sensitivity, % (no.) | |||
| Reader 1 | 48 (10/21) | 86 (18/21) | 81 (17/21) |
| Reader 2 | 48 (10/21) | 86 (18/21) | 81 (17/21) |
| Reader 3 | 43 (9/21) | 95 (20/21) | 95 (20/21) |
| Specificity, % (no.) | |||
| Reader 1 | 11 (2/19) | 79 (15/19) | 68 (13/19) |
| Reader 2 | 11 (2/19) | 79 (15/19) | 42 (8/19) |
| Reader 3 | 16 (3/19) | 63 (12/19) | 42 (8/19) |
| Positive predictive value, % (no.) | |||
| Reader 1 | 37 (10/27) | 82 (18/22) | 74 (17/23) |
| Reader 2 | 37 (10/27) | 82 (18/22) | 61 (17/28) |
| Reader 3 | 36 (9/25) | 74 (20/27) | 65 (20/31) |
| Negative predictive value, % (no.) | |||
| Reader 1 | 15 (2/13) | 83 (15/18) | 76 (13/17) |
| Reader 2 | 15 (2/13) | 83 (15/18) | 67 (8/12) |
| Reader 3 | 20 (3/15) | 92 (12/13) | 89 (8/9) |
| Accuracy, % (no.) | |||
| Reader 1 | 30 (12/40) | 83 (33/40) | 75 (30/40) |
| Reader 2 | 30 (12/40) | 83 (33/40) | 63 (25/40) |
| Reader 3 | 30 (12/40) | 80 (32/40) | 70 (28/40) |
| Diagnostic confidence, Az(95% CI) | |||
| Reader 1 | 0.793 (0.656-0.930) | 0.985 (0.958-1.000) | 0.970 (0.928-1.000) |
| Reader 2 | 0.779 (0.638-0.921) | 0.990 (0.969-1.000) | 0.945 (0.878-1.000) |
| Reader 3 | 0.759 (0.612-0.906) | 0.991 (0.969-1.000) | 0.986 (0.954-1.000) |
Note—Az = area under the receiver operating characteristic curve.
Contrast-enhanced sonography was bet ter than unenhanced sonography in both diagnostic performance and confidence (p < 0.05) for all readers. Contrast-enhanced sonography had a higher diagnostic per formance—sensitivity, specificity, positive and negative predictive values, and overall accuracy—than CT (p < 0.05) for all readers except for reader 3 who showed the same value of sensitivity for contrast-enhanced sonography and CT. Contrast-enhanced sono g raphy and CT yielded similar diagnos tic confidence (p > 0.05) for all readers with a high value of the area under the ROC curve (Fig. 7A, 7B, 7C). Table 3 shows interreader agree ment in the characterization of cystic renal tumors on unenhanced sonography, contrast-enhanc ed sonography, and CT.
| Weighted κ Statistic | |||
|---|---|---|---|
| Reader | Sonography | Contrast-Enhanced Sonography | CT |
| 1 vs 2 | 0.64 | 0.81 | 0.58 |
| 2 vs 3 | 0.86 | 0.67 | 0.59 |
| 1 vs 3 | 0.5 | 0.13 | 0.55 |
Discussion
Sonography is very effective in revealing the cystic component of a renal mass, even though it is limited in the characterization of complex renal cysts [1]. Contrast-enhanced sonography with microbubbles was recently proposed for the characterization of complex renal cysts [7–10, 12, 13]. Microbubble contrast agents are true intravascular agents and have a short lifetime (< 10 minutes) and, not being nephrotoxic, can be administered to patients with renal insufficiency. Microbubble contrast agents have been widely used in the heart and liver, whereas their use in the kidney has been less exhaustively studied because of the multiple competing imaging techniques that allow complete assessment of renal pathology. A classification and workup scheme based on contrast-enhanced sonography as the reference technique has recently been proposed for renal cystic lesions [12, 13], in which the absence of enhancement after microbubble injection implies no further workup, whereas evidence of thickened septations or mural nodules with contrast enhancement after microbubble injection is considered a reliable criterion for malignancy.
The results of our study confirm that benign and malignant complex renal cysts present typical enhancement patterns at contrast-enhanced sonography [14–16]. In cases of malignancy, the peripheral wall or intratumoral septa and the mural or septal nodules mainly revealed intense contrast enhancement after microbubble injection. Some minimal septal enhancement may also occur in benign cystic renal lesions, although a few microbubbles traveling in the septa are unlikely to represent a malignancy. In our series, evidence of microbubbles in the peripheral wall or septa led to the misdiagnosis of two benign renal cysts. Also multilocular cystic nephromas revealed hairline-thin and thick enhancing septa with peripheral wall enhancement and were misclassified as malignant after review of contrast-enhanced sonography scans. Inflammatory or hemorrhagic cysts showed exclusively a peripheral enhancing regular wall and were not misdiagnosed because they did not reveal intracystic septations. Cystic renal masses with diffuse mural calcification were misdiagnosed because of the impossibility of assessing the cystic content and intracystic components at sonography. Malignant cystic renal masses revealed peripheral wall enhancement, even though this finding was always associated with thickened enhancing intratumoral septa or with enhancing mural and septal nodules.
The Bosniak classification of the appearance of renal cysts at CT is accurate for predicting malignancy at contrast-enhanced CT [2, 3], and the presence of either nodular or septal enhancement has shown the highest diagnostic accuracy [20]. However, there is a problem in confidently diagnosing cystic renal masses that were classified as Bosniak category III at CT [20, 21]. Several invasive methods have been proposed to address this problem, including laparoscopic evaluation with biopsy of the mass [22] and CT-guided biopsy [23]. All of these diagnostic methods are invasive, and use of a noninvasive imaging technique that could increase the diagnostic confidence of CT is advocated.
In our study, contrast-enhanced sonography had superior diagnostic accuracy in comparison with CT and improved the characterization of complex renal cysts that had been classified as indeterminate at CT. This finding was due to the discrepancy between CT and contrast-enhanced sonography in depicting septal vascularity and may be explained by the high sensitivity of contrast-enhanced sonography both in detecting microbubble signals in the peripheral wall or in the intracystic septa and in revealing the solid enhancing component in the renal cysts that was not shown by CT. The high sensitivity of contrast-enhanced sonography is determined by the contrast-specific techniques that suppress the signals from the stationary tissues and enhance the harmonic signals produced by the microbubbles. However, the same high sensitivity of the microbubble signal from contrast-enhanced sonography led to the misclassification of a minority of benign cystic renal masses with enhancing peripheral wall and thickened intracystic septa.
Differently from the previous studies [14, 16] in which investigators described the capabilities of contrast-enhanced sonography compared with CT in the classification of complex cystic renal masses with the Bosniak system [2, 3], in our study we used an independent diagnostic classification system for contrast-enhanced sonography that was based on the enhancement patterns involving or not involving the peripheral wall, the intracystic septa, and the mural or septal nodules. We consider this independent classification system necessary because of the higher sensitivity to contrast provided by contrast-enhanced sonography compared with CT, which results from the different physical principles and pharmacokinetic properties of microbubbles versus iodinated contrast agents (resonance phenomenon vs x-ray attenuation; blood-pool distribution vs leakage in the interstitial space).
Interreader agreement was progressively lower in contrast-enhanced sonography according to the increasing difference in reader experience, which indicates that a learning curve in image interpretation must be considered in contrast-enhanced sonography. On the other hand, the interreader agreement was moderate also in the analysis of the CT images in which the Bosniak classification system was used. This was determined by the 19 cases that were classified as Bosniak category IIF or III in which the interreader variation is known to be quite high [4, 24]. Moreover, on contrast-enhanced sonography the diffuse mural calcifications in some cystic renal masses prevented the thickness of the wall or septa (or both) from being evaluated because of back-shadowing. Therefore, meticulous evaluation is necessary for lesions showing diffusely calcified septa or wall on contrast-enhanced sonography images.
According to the results of this study, contrast-enhanced sonography should be used to evaluate every renal mass with a complex cystic appearance at unenhanced sonography provided that the lesion can be explored adequately. CT should be used for staging complex cystic renal masses that reveal a malignant enhancement pattern at contrast-enhanced sonography. A cystic renal mass exhibiting an enhancing peripheral wall and thickened intracystic septa or mural nodules (or both) after microbubble injection should be considered malignant. Contrast-enhanced sonography should be considered an alternative to CT [23] in the follow-up of complex renal cysts whenever these cysts can be adequately evaluated with contrast-enhanced sonography. Contrast-enhanced sonography can be used in the char acterization of complex cystic renal masses in patients with renal insufficiency and with contraindications to contrast-enhanced CT or MRI. However, contrast-enhanced sonography presents some clear limitations due to deep lesion location and bowel gas interposition that limit the visibility of the enhancement pattern after microbubble injection.
The present study has some limitations. First, its retrospective nature limits extrapolation of the results to a clinical setting. Second, the analysis of contrast en hancement was merely visual and quantitative analysis of the solid portion of each cystic renal mass was not performed. Quantitative analysis could allow a more reproducible study of peripheral or septal enhancement. Third, the number of lesions included was relatively small, but this is because complex renal masses are infrequently identified. Fourth, not all diagnoses were confirmed pathologically.
In conclusion, contrast-enhanced sonography allows reliable characterization of indeterminate cystic renal masses, and in this study it was found to perform better than unenhanced sonography and CT in the diagnosis of malignancy in complex cystic renal masses.
Footnotes
Address correspondence to E. Quaia ([email protected]).
Presented at the 2007 annual meeting of the American Roentgen Ray Society, Orlando, FL.
References
1.
Hélénon O, Correas JM, Balleyguier C, Ghoudani M, Cornud F. Ultrasound of renal tumors. Eur Radiol 2001; 11:1890 –1901
2.
Bosniak MA. The current radiological approach to renal cysts. Radiology 1986; 158:1–10
3.
Bosniak MA. Difficulties in classifying cystic lesions of the kidney. Urol Radiol 1991; 13:91–93
4.
Siegel CL, McFarland EG, Brink JA, Fisher AJ, Humphrey P, Heiken JP. CT of cystic renal masses: analysis of diagnostic performance and interobserver variation. AJR 1997; 169:813–818
5.
Israel G, Hindman N, Bosniak MA. Evaluation of cystic renal masses: comparison of CT and MR imaging by using the Bosniak classification system. Radiology 2004; 231:365–371
6.
Correas JM, Claudon M, Tranquart F, Hélénon O. Contrast-enhanced ultrasonography: renal applications [in French]. J Radiol 2003; 84 (12 Pt 2):2041 –2054
7.
Correas JM, Claudon M, Tranquart F, Hélénon AO. The kidney: imaging with microbubble contrast agents. Ultrasound Q 2006; 22:53 –66
8.
Ascenti G, Gaeta M, Magno C, et al. Contrast-enhanced second-harmonic sonography in the detection of pseudocapsule in renal cell carcinoma. AJR 2004; 182:1525 –1530
9.
Quaia E, Siracusano S, Bertolotto M, Monduzzi M, Mucelli RP. Characterization of renal tumours with pulse inversion harmonic imaging by intermittent high mechanical index technique: initial results. Eur Radiol 2003; 13:1402 –1412
10.
Tamai H, Takiguchi Y, Oka M, et al. Contrast-enhanced ultrasonography in the diagnosis of solid renal tumors. J Ultrasound Med 2005; 24:1635 –1640
11.
Kim AY, Kim SH, Kim YJ, Lee IH. Contrast-enhanced power Doppler sonography for the differentiation of cystic renal lesions: preliminary study. J Ultrasound Med 1999; 18:581–588
12.
Robbin ML. Ultrasound contrast agents: a promising future. Radiol Clin North Am 2001; 39:399–414
13.
Robbin ML, Lockhart ME, Barr RG. Renal imaging with ultrasound contrast: current status. Radiol Clin North Am 2003; 41:963 –978
14.
Bertolotto M, Barozzi L, Valentino M, et al. Evaluation of cystic renal masses with contrast-enhanced US comparison with CT. In: Radiological Society of North America scientific assembly and annual meeting program. Oak Brook, IL: Radiological Society of North America, 2006:404 –405
15.
Park BK, Kim B, Kim SH, Ko K, Lee HM, Choi HY. Assessment of cystic renal masses based on Bosniak classification: comparison of CT and contrast-enhanced US. Eur J Radiol 2007; 61:310–314
16.
Ascenti G, Mazziotti S, Zimbaro G, et al. Complex cystic renal masses: characterization with contrast-enhanced US. Radiology 2007; 24:158–165
17.
Eng J. Sample size estimation: how many individuals should be studied? Radiology 2003; 227:309–313
18.
Beck JR, Shultz EK. The use of relative operating characteristic (ROC) curves in test performance evaluation. Arch Pathol Lab Med 1986; 110:13 –20
19.
Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983; 148:839–843
20.
Benjaminov O, Atri M, O'Malley M, Lobo K, Tomlinson G. Enhancing component on CT to predict malignancy in cystic renal masses and interobserver agreement of different CT features. AJR 2006; 186:665 –672
21.
Harisinghani MG, Maher MM, Gervais DA, et al. Incidence of malignancy in complex cystic renal masses (Bosniak category III): should imaging-guided biopsy precede surgery? AJR 2003; 180:755 –758
22.
Limb J, Santiago L, Kaswick J, Bellman GC. Laparoscopic evaluation of indeterminate renal cysts: long term follow-up. J Endourol 2002; 16:79 –82
23.
Lang EK, Macchia RJ, Gayle B, et al. CT-guided biopsy of indeterminate renal cystic masses (Bosniak 3 and 2F): accuracy and impact on clinical management. Eur Radiol 2002; 12:2518 –2524
24.
Israel GM, Bosniak MA. Follow-up CT of moderately complex cystic lesions of the kidney (Bosniak category IIF). AJR 2003; 181:627 –633
25.
Quaia E, ed. Contrast media in ultrasonography: basic principles and clinical applications. Heidelberg, Germany: Springer-Verlag, 2005
Information & Authors
Information
Published In
Copyright
© American Roentgen Ray Society.
History
Submitted: December 14, 2007
Accepted: May 1, 2008
First published: November 23, 2012
Keywords
Authors
Metrics & Citations
Metrics
Citations
Export Citations
To download the citation to this article, select your reference manager software.