- Email: [email protected]
Assessment of cystic renal masses based on Bosniak classification: Comparison of CT and contrast-enhanced US
European Journal of Radiology 61 (2007) 310–314
Assessment of cystic renal masses based on Bosniak classification: Comparison of CT and contrast-enhanced US Byung Kwan Park a,∗ , Bohyun Kim b , Seung Hyup Kim c , Kyungran Ko a , Hyun Moo Lee d , Han Yong Choi d a The Department of Radiology and Center for Imaging Science, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Kangnam-ku, Seoul, Republic of Korea b The Department of Radiology, Mayo Clinic, Minnesota, USA c The Department of Radiology, Seoul National University College of Medicine, Institute of Radiation Medicine, SNUMRC, Clinical Research Institute, Seoul National University Hospital, Seoul, Republic of Korea d The Department of Urology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
Received 30 May 2006; received in revised form 27 September 2006; accepted 3 October 2006
Abstract Objectives: To compare retrospectively the imaging features of computed tomography (CT) and contrast-enhanced US (CEUS) imaging for the assessment of cystic renal masses using the Bosniak classification system. Materials and methods: The CT and CEUS images of 31 pathologically confirmed cystic renal masses in 31 patients were retrospectively analyzed for septa numbers, wall and/or septa thickness, enhancement degree, and for the presence of a solid component by consensus between two radiologists using the Bosniak classification. Diagnostic accuracies of CT and CEUS for malignant cystic tumor were calculated and compared using McNemar test. Results: Diagnostic accuracies of CT and CEUS for malignant renal tumor were 74% and 90%, respectively, but there were not statistically different (P > 0.05). CEUS and CT images showed same Bosniak classification in 23 (74%) lesions and there were differences in 8 (26%) lesions, all of which were upgraded by CEUS; one lesion from I to IV, two lesions from II to IV, two lesions from IIF to III, and three lesions from III to IV. CEUS images depicted more septa in 10 (32%) lesions, more thickened wall and/or septa in 4 (13%) lesions, and stronger enhancement in 19 (61%) lesions. Moreover, for six lesions, solid component was detected by CEUS but not by CT. Conclusion: CEUS might better visualize septa number, septa and/or wall thickness, solid component and the enhancement of some renal cystic masses than CT, resulting in upgrade of Bosniak classification and affecting their treatment plan. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Kidney; Cysts; Neoplasms; CT; US; Staging
1. Introduction The Bosniak classification of renal cysts was first introduced in 1986 and has been commonly used to evaluate cystic renal masses. This classification is considered as an accurate and efficient method for treatment planning [1–4]. Many urologist and radiologists accept the Bosniak classification provide a useful means of assessing and managing these lesions and general interobserver agreement has been reported in most cases [4–7]. Abbreviations: CT, computed tomography; CEUS, contrast-enhanced ultrasonography; CHA, coded harmonic angio; ADI, agent detection imaging ∗ Corresponding author. Tel.: +82 2 3410 6457; fax: +82 2 3410 0084. E-mail address: [email protected] (B.K. Park). 0720-048X/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2006.10.004
Although the Bosniak classification was developed on the basis of computed tomography (CT) criteria, it can be similarly applied to magnetic resonance (MR) imaging. In some of the reports, MR imaging depicted additional septa, thickening of the wall and septa, and enhancement compared to CT [8,9]. Contrast-enhanced ultrasonography (CEUS) can be useful for evaluating the internal architectures of hepatic masses [10–12] and might be used to evaluate cystic renal masses using the Bosniak classification system. However, to our knowledge, no comparative study of CT with CEUS has been conducted in terms of evaluating pathologically confirmed cystic renal lesions. Thus, the purpose of this study was to compare the imaging features of CT and CEUS imaging for the assessment of cystic renal masses using the Bosniak classification system.
B.K. Park et al. / European Journal of Radiology 61 (2007) 310–314
311
2. Materials and methods
2.3. Image interpretation
2.1. Patients
All images were retrospectively evaluated by consensus between two radiologists with 15 and 20 years of experience at interpreting cross-sectional images, respectively. The radiologists were unaware of clinical information and histological findings and images were assessed based on the imaging features of renal cystic lesions. CT images for each patient were first interpreted and then CEUS images in the same patient were reviewed. Each lesion was categorized on CT and CEUS by using Bosniak classification system [1,9] to compare their Bosniak categorizations of these imaging modalities. For each lesion, septa number, thickness of the septa and/or wall, the presence or absence of a solid component/calcification, and enhancing degree were determined by CT and CEUS. The method used to determine septa number, and septa and/or wall thickness was as described by Israel et al. [9]. Cystic lesions were categorized based on number of septa into four groups: group 1, no septa; group 2, between one and four septa; group 3, between five and nine septa; group 4, more than nine septa. Thickness of wall and/or septa was subjectively determined as follows: category I (wall only) or II, hairline thin; category IIF, minimally thickened; category III or IV, grossly thickened and irregular. The enhancement degrees of solid component, wall and septa within lesions were subjectively determined as follows: poor (no or minimally enhancement); fair (between poor and good); good (enhancement similar to that of renal cortex). If a single septum or wall of a lesion was determined to be grossly thickened and irregular, this feature was considered as grossly thickened and irregular wall or septa for the entire lesion. The presence or absence of calcification was determined based on CT findings. Diagnostic accuracies were calculated and compared between CT and CEUS imaging using McNemar test. True positive malignant cystic renal tumors were defined as lesions were diagnosed as Bosniak category III or IV either by CT or CEUS imaging. A P value of less than 0.05 was considered statistically significant.
This retrospective study was approved by our institutional review board. Thirty-one patients (23 men and eight women; range, 29–68 years; mean age, 50 years) had 31 renal cystic masses, all of which were incidentally detected by conventional US. Between January 2001 and October 2005, all of these patients underwent CT and CEUS prior to surgery. Twenty-six cystic lesions were diagnosed as Bosniak category III or IV on CT images and then CEUS was performed. For the five remaining lesions, the imaging features of CT and conventional US were discrepant. Thus, CEUS was also performed. Lesion sizes in transverse greatest diameter, measured 0.8–9.7 cm (mean ± S.D., 4.14 ± 2.05 cm) on pathologic exam. Intervals between CT and CEUS ranged between one and 18 days (mean, 9 days). All 31 lesions were pathologically proven by radical (n = 27) or partial (n = 3) nephrectomy and percutaneous biopsy (n = 1). The pathologic diagnosis included renal cell carcinomas (n = 26), metastasis from lung cancer (n = 2), hemorrhagic cyst (n = 1), xanthogranulomatous pyelonephritis (n = 1), and multilocular cystic nephroma (n = 1). 2.2. Imaging techniques The CT examinations were performed using one of four CT scanners (HiSpeed advantage, LightSpeed QX/I, LightSpeed ultra8, LightSpeed ultra16 [GE Medical Systems; Milwaukee, WI, USA]). Unenhanced CT images were obtained first and then single phase contrast-enhanced images were performed at 40–60 s following contrast administration. The section thickness ranged 2.5–5 mm and the total amount of iodine contrast material intravenously infused was 130 ml at the rate of 2–3 ml/s. The CEUS examinations were performed using one of two high-resolution US units (Logic Q [GE Medical Systems; Milwaukee, WI, USA] or Sequoia [Siemens Medical Solutions, Mountainview, CA]) equipped with the coded harmonic angio (CHA, in 12 patients) and Agent Detection Imaging (ADI, in 19 patients), respectively. All of US scans were conducted by one of two radiologists who reviewed CT images in advance for lesion location and size. SH U (4 g) 508A (Levovist; Schering AG, Berlin, Germany) in a concentration of 300 mg/ml was intravenously administered by manual bolus injection via an 18-gauge cannula inserted into an antecubital vein at a rate of 2–3 ml/s. Real-time scanning commenced immediately after the intravenous injection of the contrast agent and the elapsed time was recorded from the point when the contrast agent first passed the intravenous cannula. Since the first appearance of contrast agent in the kidney, rapid scans of the renal lesions were performed. Rapid transducer sweeping over the lesion was repeated every 10–15 s. All of the scanned images were automatically stored in the picture archiving communication systems. The mean scan time used for CEUS was about 3 min.
3. Results 3.1. Bosniak category The 31 cystic renal masses were categorized by CT as follows: category I (n = 1), category II (n = 2), category IIF (n = 2), category III (n = 9), and category IV (n = 17). These lesions were categorized by CEUS as follows: category I–IIF (n = 0), category III (n = 9), and category IV (n = 22). For 23 (74%) lesions, the CT and CEUS images had same Bosniak category, but in eight (26%) lesions, they differed. All of these discordant lesions were upgraded by CEUS: one lesion from I to IV, two lesions from II to IV, two lesions from IIF to III, and three lesions from III to IV (Fig. 1). 3.2. Number of septa In 21 (68%) lesions, septa number was equivalent by CT and CEUS imaging as follows: seven lesions in group 1, seven
312
B.K. Park et al. / European Journal of Radiology 61 (2007) 310–314
lesions in group 2, two lesions in group 3, and five lesions in group 4. In 10 (32%) lesions, CEUS images depicted more septa than did CT images, which resulted in an increased number of septa: one lesion from group 1 to group 2, one lesion from group 1 to group 4, seven lesions from group 2 to group 3, and one lesion from group 3 to group 4 (Fig. 2). 3.3. Septa and/or wall thickness For 27 (87%) lesions, CT and CEUS images were equivalent at depicting wall and septa thickness, including hairline thin (n = 2), minimally thickened (n = 0), or grossly thickened and irregular septa and/or wall (n = 25). In four (13%) lesions, CEUS depicted thicker septa and/or wall than CT: two lesions from hairline thin to grossly thickened, and two lesions from minimally thickened to grossly thickened (Fig. 2). 3.4. Enhancement, solid component, and calcification For 12 (39%) lesions, CT and CEUS images were equivalent at depicting the enhancement degree of wall, septa, and solid component including poor (n = 0), fair (n = 9), or good enhancement (n = 3). For 19 (61%) lesions, enhancement was stronger on CEUS than on CT as follows: five lesions from poor to good enhancement and 14 lesions from fair to good enhancement (Fig. 2). Enhancement of solid component within six lesions was not seen on CT images but on CEUS images, which resulted in upgrade of Bosniak classification (Fig. 1). CT images demonstrated calcification in five lesions. In four lesions of these lesions, CEUS depicted calcification. 3.5. Diagnostic accuracy
Fig. 1. A 66-year-old man with left renal cell carcinoma (patient 1). (a) Contrastenhanced CT scan shows a thin-walled cyst mass (arrows) with no septa or solid component, suggestive of Bosniak category I. (b) Longitudinal image of contrast-enhanced US demonstrates a well-enhancing mural nodule (arrows) within the lesion, suggestive of Bosniak category IV. (c) Microscopic exam obtained from partial nephrectomy shows a polypoid renal cell carcinoma (arrows), which is better correlated with that of contrast-enhanced US than CT.
Diagnostic accuracies of CT and CEUS for malignant renal tumor were 74% and 90%, respectively. However, there was no statistical difference between CT and CEUS (P > 0.05). Hemorrhagic cyst, xanthogranulomatous pyelonephritis and multilocular cystic nephroma were falsely diagnosed as malignant cystic masses by both CT and CEUS. Table 1 shows the comparison of CT and CEUS imaging findings of the malignant renal cystic masses (n = 8) and their discordant Bosniak classifications, which all resulted from misdiagnosis by CT. Five lesions of these lesions were correctly diagnosed as Bosniak category III or IV by CEUS but were misdiagnosed as Bosniak category IIF or less by CT, which led to a change in treatment plan. Two metastatic lesions from lung cancer in patients 4 and 5 were manifested as cystic lesions with minimally thickened wall and were categorized as IIF by CT but their walls appeared grossly thickened and irregular on CEUS, suggestive of category III. Unlike CT, CEUS detection of a solid component within three lesions supported diagnostic suspicion of malignant cystic tumor before surgery. 4. Discussion CEUS imaging is a microbubble-specific harmonic US, which is designed for the optimal detection of signals from SH U 508A [10–12] and can produce signals from stationary
B.K. Park et al. / European Journal of Radiology 61 (2007) 310–314
313
Table 1 Comparison of the CT and CEUS features in the malignant cystic tumors with discrepant Bosniak classification Patients
1 2 3 4 5 6 7 8
Category
Numbera
Thicknessb
Solid componentc
CT
CEUS
CT
CEUS
CT
CEUS
CT
CEUS
I II II IIF IIF III III III
IV IV IV III III IV IV IV
1 2 2 1 1 1 2 2
1 3 2 1 1 2 3 3
M H H M M G G G
M G G G G G G G
No No No No No No No No
Yes Yes Yes No No Yes Yes Yes
a Represents the number of septa: group 1, no septa; group 2, one to four; group 3, five to nine septa; group 4, more than nine septa. b Represents the thickness of wall or septa: hairline thin (H), minimally thickened (M) and grossly thickened and irregular (G). c Represents the enhanced area within a cystic tumor.
Fig. 2. A 46-year-old man with left renal cell carcinoma (patient 7). (a) Contrastenhanced CT scan shows a complex cystic mass containing grossly thickened and irregular septa (black arrows), ranging from one to four in number (group 2). (b) Longitudinal image of contrast-enhanced US demonstrates more septa (white arrows) within the lesion than CT. These septa range between five and nine in number (group 3) and appear more thickened and more apparent on CEUS than on CT. (c) The cut surface of surgically removed specimen shows multiple septa (arrows) of five to nine in number, which is better correlated with that of contrast-enhanced US than CT.
microbubbles as easily as from those that are moving in the vessels, whereas conventional US relies on a Doppler shift signals. CHA uses digital encoding and decoding of transmitted pulses, and can separate the contrast from the tissue and reduce unwanted echoes of the fundamental frequency and tissue harmonic signals. ADI uses two pulses of the same polarity and subtracts the signals from the two pulses and thus makes it possible to obtain only the fundamental and harmonic signals from the contrast agent while those from the tissue are removed. CEUS can improve the detection and characterization of focal liver lesions and can help to evaluate the therapeutic response to malignant hepatic tumors treated by interventional procedure because this US imaging technique can provide a better assessment of vascular morphology and enhancing patterns. However, to the best of our knowledge, the assessment of cystic renal masses with CEUS and CT has not been previously compared. The Bosniak classification system was designed to analyze the morphology of cystic renal masses solely based on CT findings, i.e., number of septa, thickness of wall and/or septa, lesion enhancement, and calcification. Israel et al suggested that MR imaging can be used to assess renal cystic lesions using Bosniak classification [9]. They found that MR imaging caused the upgrade of Bosniak classification because it demonstrated more septa, thicker septa and/or wall, and better enhancement than CT imaging. In our study, CEUS imaging was also found to be useful for evaluating cystic renal masses using Bosniak classification, and to be similar to CT imaging in the most cases. However, there was significant discrepancy of the imaging findings between CEUS and CT imaging in some cystic renal masses. Pathologic examinations revealed that CEUS imaging was superior to CT in terms of detecting additional septa, thickness of wall and/or septa, and solid component. Because US contrast agent circulating in the microvessels of septa, wall, and solid component produces strong signal intensity when it is destroyed by ultrasound, CEUS imaging could allow the sophisticated internal structures of cystic renal masses to be evaluated and result in higher resolution than CT. The demonstration of solid component within the cystic renal masses is important, as it is a key finding differentiating cate-
314
B.K. Park et al. / European Journal of Radiology 61 (2007) 310–314
gory III from IV. Lesion enhancement was more conspicuous on CEUS than on CT images in most of cases, and enhanced diagnostic confidence and eventually affected treatment plans. In the present study, several cystic lesions contained a solid component that was demonstrated by CEUS but not by CT. This finding resulted in surgical treatment although the number of patients with these lesions was small. The presence of calcification does not play a major role in the evaluation of cystic renal masses [13]. In our study, CEUS could not detect as many calcifications as CT. Calcification prevented precise evaluation of the thickness of wall and/or septa on CEUS as well as MR images [9]. Therefore, meticulous evaluation is necessary for lesions with diffusely calcified septa or wall on CEUS images. CEUS also has some disadvantages over CT. First, CEUS can be influenced by lesion location. A poor sonic window due to bowel gas or ribs prevents good quality image. Second, CEUS is an operator-dependent imaging modality and requires sufficient experience. Third, the US contrast agent used in our study has a high mechanical index and is unsuitable for continuous scan. A higher mechanical index leads to increase in the likelihood of microbubble because this index determines the level of interaction between bubbles and ultrasound. Therefore, second harmonic contrast agent requiring low mechanical index can be suitable for performing continuous scan and evaluating cystic renal masses. Our study has several limitations. First, interobserver variability was not evaluated because the cases were analyzed by consensus. In addition, US images were interpreted after CT images had been evaluated, which may have resulted in observer bias. Second, case-selection bias could have affected results because the cases were retrospectively collected and analyzed. Third, a relatively small number of pathologically proved cases were included. Thus, further investigation in surgical cases is needed to support our results. 5. Conclusion No difference in diagnostic accuracy was found between CT and CEUS for the assessment of cystic renal masses based
on Bosniak classification system. However, CEUS might better visualize septa number, septa and/or wall thickness, solid component and the enhancement of some renal cystic masses than CT, resulting in upgrade of Bosniak classification and affecting their treatment plan. References [1] Bosniak MA. The current radiological approach to renal cysts. Radiology 1986;158:1–10. [2] Bosniak MA. Diagnosis and management of patients with complicated cystic lesions of the kidney. AJR Am J Roentgenol 1997;169:819– 21. [3] Bosniak MA. The use of the Bosniak classification system for renal cysts and cystic tumors. J Urol 1997;157:1852–3. [4] Curry NS, Cochran ST, Bissada NK. Cystic renal masses: accurate Bosniak classification requires adequate renal CT. AJR Am J Roentgenol 2000;175:339–42. [5] Koga S, Nishikido M, Inuzuka S, et al. An evaluation of Bosniak’s radiological classification of cystic renal masses. BJU Int 2000;86:607– 9. [6] 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 Am J Roentgenol 1997;169:813–8. [7] Levy P, Helenon O, Merran S, et al. Cystic tumors of the kidney in adults: radio-histopathologic correlations. J Radiol 1999;80:121– 33. [8] Balci NC, Semelka RC, Patt RH, et al. Complex renal cysts: findings on MR imaging. AJR Am J Roentgenol 1999;172:1495– 500. [9] Israel GM, Hindman N, Bosniak MA. Evaluation of cystic renal masses: comparison of CT and MR imaging by using the Bosniak classification system. Radiology 2004;231:365–71. [10] Lee JY, Choi BI, Han JK, Kim AY, Shin SH, Moon SG. Improved sonographic imaging of hepatic hemangioma with contrast-enhanced coded harmonic angiography: comparison with MR imaging. Ultrasound Med Biol 2002;28:287–95. [11] Youk JH, Kim CS, Lee JM. Contrast-enhanced agent detection imaging: value in the characterization of focal hepatic lesions. J Ultrasound Med 2003;22:897–910. [12] Youk JH, Lee JM, Kim CS. Therapeutic response evaluation of malignant hepatic masses treated by interventional procedures with contrastenhanced agent detection imaging. J Ultrasound Med 2003;22:911– 20. [13] Israel GM, Bosniak MA. Calcification in cystic renal masses: is it important in diagnosis? Radiology 2003;226:47–52.
Assessment of cystic renal masses based on Bosniak classification: Comparison of CT and contrast-enhanced US Byung Kwan Park a,∗ , Bohyun Kim b , Seung Hyup Kim c , Kyungran Ko a , Hyun Moo Lee d , Han Yong Choi d a The Department of Radiology and Center for Imaging Science, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Kangnam-ku, Seoul, Republic of Korea b The Department of Radiology, Mayo Clinic, Minnesota, USA c The Department of Radiology, Seoul National University College of Medicine, Institute of Radiation Medicine, SNUMRC, Clinical Research Institute, Seoul National University Hospital, Seoul, Republic of Korea d The Department of Urology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
Received 30 May 2006; received in revised form 27 September 2006; accepted 3 October 2006
Abstract Objectives: To compare retrospectively the imaging features of computed tomography (CT) and contrast-enhanced US (CEUS) imaging for the assessment of cystic renal masses using the Bosniak classification system. Materials and methods: The CT and CEUS images of 31 pathologically confirmed cystic renal masses in 31 patients were retrospectively analyzed for septa numbers, wall and/or septa thickness, enhancement degree, and for the presence of a solid component by consensus between two radiologists using the Bosniak classification. Diagnostic accuracies of CT and CEUS for malignant cystic tumor were calculated and compared using McNemar test. Results: Diagnostic accuracies of CT and CEUS for malignant renal tumor were 74% and 90%, respectively, but there were not statistically different (P > 0.05). CEUS and CT images showed same Bosniak classification in 23 (74%) lesions and there were differences in 8 (26%) lesions, all of which were upgraded by CEUS; one lesion from I to IV, two lesions from II to IV, two lesions from IIF to III, and three lesions from III to IV. CEUS images depicted more septa in 10 (32%) lesions, more thickened wall and/or septa in 4 (13%) lesions, and stronger enhancement in 19 (61%) lesions. Moreover, for six lesions, solid component was detected by CEUS but not by CT. Conclusion: CEUS might better visualize septa number, septa and/or wall thickness, solid component and the enhancement of some renal cystic masses than CT, resulting in upgrade of Bosniak classification and affecting their treatment plan. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Kidney; Cysts; Neoplasms; CT; US; Staging
1. Introduction The Bosniak classification of renal cysts was first introduced in 1986 and has been commonly used to evaluate cystic renal masses. This classification is considered as an accurate and efficient method for treatment planning [1–4]. Many urologist and radiologists accept the Bosniak classification provide a useful means of assessing and managing these lesions and general interobserver agreement has been reported in most cases [4–7]. Abbreviations: CT, computed tomography; CEUS, contrast-enhanced ultrasonography; CHA, coded harmonic angio; ADI, agent detection imaging ∗ Corresponding author. Tel.: +82 2 3410 6457; fax: +82 2 3410 0084. E-mail address: [email protected] (B.K. Park). 0720-048X/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2006.10.004
Although the Bosniak classification was developed on the basis of computed tomography (CT) criteria, it can be similarly applied to magnetic resonance (MR) imaging. In some of the reports, MR imaging depicted additional septa, thickening of the wall and septa, and enhancement compared to CT [8,9]. Contrast-enhanced ultrasonography (CEUS) can be useful for evaluating the internal architectures of hepatic masses [10–12] and might be used to evaluate cystic renal masses using the Bosniak classification system. However, to our knowledge, no comparative study of CT with CEUS has been conducted in terms of evaluating pathologically confirmed cystic renal lesions. Thus, the purpose of this study was to compare the imaging features of CT and CEUS imaging for the assessment of cystic renal masses using the Bosniak classification system.
B.K. Park et al. / European Journal of Radiology 61 (2007) 310–314
311
2. Materials and methods
2.3. Image interpretation
2.1. Patients
All images were retrospectively evaluated by consensus between two radiologists with 15 and 20 years of experience at interpreting cross-sectional images, respectively. The radiologists were unaware of clinical information and histological findings and images were assessed based on the imaging features of renal cystic lesions. CT images for each patient were first interpreted and then CEUS images in the same patient were reviewed. Each lesion was categorized on CT and CEUS by using Bosniak classification system [1,9] to compare their Bosniak categorizations of these imaging modalities. For each lesion, septa number, thickness of the septa and/or wall, the presence or absence of a solid component/calcification, and enhancing degree were determined by CT and CEUS. The method used to determine septa number, and septa and/or wall thickness was as described by Israel et al. [9]. Cystic lesions were categorized based on number of septa into four groups: group 1, no septa; group 2, between one and four septa; group 3, between five and nine septa; group 4, more than nine septa. Thickness of wall and/or septa was subjectively determined as follows: category I (wall only) or II, hairline thin; category IIF, minimally thickened; category III or IV, grossly thickened and irregular. The enhancement degrees of solid component, wall and septa within lesions were subjectively determined as follows: poor (no or minimally enhancement); fair (between poor and good); good (enhancement similar to that of renal cortex). If a single septum or wall of a lesion was determined to be grossly thickened and irregular, this feature was considered as grossly thickened and irregular wall or septa for the entire lesion. The presence or absence of calcification was determined based on CT findings. Diagnostic accuracies were calculated and compared between CT and CEUS imaging using McNemar test. True positive malignant cystic renal tumors were defined as lesions were diagnosed as Bosniak category III or IV either by CT or CEUS imaging. A P value of less than 0.05 was considered statistically significant.
This retrospective study was approved by our institutional review board. Thirty-one patients (23 men and eight women; range, 29–68 years; mean age, 50 years) had 31 renal cystic masses, all of which were incidentally detected by conventional US. Between January 2001 and October 2005, all of these patients underwent CT and CEUS prior to surgery. Twenty-six cystic lesions were diagnosed as Bosniak category III or IV on CT images and then CEUS was performed. For the five remaining lesions, the imaging features of CT and conventional US were discrepant. Thus, CEUS was also performed. Lesion sizes in transverse greatest diameter, measured 0.8–9.7 cm (mean ± S.D., 4.14 ± 2.05 cm) on pathologic exam. Intervals between CT and CEUS ranged between one and 18 days (mean, 9 days). All 31 lesions were pathologically proven by radical (n = 27) or partial (n = 3) nephrectomy and percutaneous biopsy (n = 1). The pathologic diagnosis included renal cell carcinomas (n = 26), metastasis from lung cancer (n = 2), hemorrhagic cyst (n = 1), xanthogranulomatous pyelonephritis (n = 1), and multilocular cystic nephroma (n = 1). 2.2. Imaging techniques The CT examinations were performed using one of four CT scanners (HiSpeed advantage, LightSpeed QX/I, LightSpeed ultra8, LightSpeed ultra16 [GE Medical Systems; Milwaukee, WI, USA]). Unenhanced CT images were obtained first and then single phase contrast-enhanced images were performed at 40–60 s following contrast administration. The section thickness ranged 2.5–5 mm and the total amount of iodine contrast material intravenously infused was 130 ml at the rate of 2–3 ml/s. The CEUS examinations were performed using one of two high-resolution US units (Logic Q [GE Medical Systems; Milwaukee, WI, USA] or Sequoia [Siemens Medical Solutions, Mountainview, CA]) equipped with the coded harmonic angio (CHA, in 12 patients) and Agent Detection Imaging (ADI, in 19 patients), respectively. All of US scans were conducted by one of two radiologists who reviewed CT images in advance for lesion location and size. SH U (4 g) 508A (Levovist; Schering AG, Berlin, Germany) in a concentration of 300 mg/ml was intravenously administered by manual bolus injection via an 18-gauge cannula inserted into an antecubital vein at a rate of 2–3 ml/s. Real-time scanning commenced immediately after the intravenous injection of the contrast agent and the elapsed time was recorded from the point when the contrast agent first passed the intravenous cannula. Since the first appearance of contrast agent in the kidney, rapid scans of the renal lesions were performed. Rapid transducer sweeping over the lesion was repeated every 10–15 s. All of the scanned images were automatically stored in the picture archiving communication systems. The mean scan time used for CEUS was about 3 min.
3. Results 3.1. Bosniak category The 31 cystic renal masses were categorized by CT as follows: category I (n = 1), category II (n = 2), category IIF (n = 2), category III (n = 9), and category IV (n = 17). These lesions were categorized by CEUS as follows: category I–IIF (n = 0), category III (n = 9), and category IV (n = 22). For 23 (74%) lesions, the CT and CEUS images had same Bosniak category, but in eight (26%) lesions, they differed. All of these discordant lesions were upgraded by CEUS: one lesion from I to IV, two lesions from II to IV, two lesions from IIF to III, and three lesions from III to IV (Fig. 1). 3.2. Number of septa In 21 (68%) lesions, septa number was equivalent by CT and CEUS imaging as follows: seven lesions in group 1, seven
312
B.K. Park et al. / European Journal of Radiology 61 (2007) 310–314
lesions in group 2, two lesions in group 3, and five lesions in group 4. In 10 (32%) lesions, CEUS images depicted more septa than did CT images, which resulted in an increased number of septa: one lesion from group 1 to group 2, one lesion from group 1 to group 4, seven lesions from group 2 to group 3, and one lesion from group 3 to group 4 (Fig. 2). 3.3. Septa and/or wall thickness For 27 (87%) lesions, CT and CEUS images were equivalent at depicting wall and septa thickness, including hairline thin (n = 2), minimally thickened (n = 0), or grossly thickened and irregular septa and/or wall (n = 25). In four (13%) lesions, CEUS depicted thicker septa and/or wall than CT: two lesions from hairline thin to grossly thickened, and two lesions from minimally thickened to grossly thickened (Fig. 2). 3.4. Enhancement, solid component, and calcification For 12 (39%) lesions, CT and CEUS images were equivalent at depicting the enhancement degree of wall, septa, and solid component including poor (n = 0), fair (n = 9), or good enhancement (n = 3). For 19 (61%) lesions, enhancement was stronger on CEUS than on CT as follows: five lesions from poor to good enhancement and 14 lesions from fair to good enhancement (Fig. 2). Enhancement of solid component within six lesions was not seen on CT images but on CEUS images, which resulted in upgrade of Bosniak classification (Fig. 1). CT images demonstrated calcification in five lesions. In four lesions of these lesions, CEUS depicted calcification. 3.5. Diagnostic accuracy
Fig. 1. A 66-year-old man with left renal cell carcinoma (patient 1). (a) Contrastenhanced CT scan shows a thin-walled cyst mass (arrows) with no septa or solid component, suggestive of Bosniak category I. (b) Longitudinal image of contrast-enhanced US demonstrates a well-enhancing mural nodule (arrows) within the lesion, suggestive of Bosniak category IV. (c) Microscopic exam obtained from partial nephrectomy shows a polypoid renal cell carcinoma (arrows), which is better correlated with that of contrast-enhanced US than CT.
Diagnostic accuracies of CT and CEUS for malignant renal tumor were 74% and 90%, respectively. However, there was no statistical difference between CT and CEUS (P > 0.05). Hemorrhagic cyst, xanthogranulomatous pyelonephritis and multilocular cystic nephroma were falsely diagnosed as malignant cystic masses by both CT and CEUS. Table 1 shows the comparison of CT and CEUS imaging findings of the malignant renal cystic masses (n = 8) and their discordant Bosniak classifications, which all resulted from misdiagnosis by CT. Five lesions of these lesions were correctly diagnosed as Bosniak category III or IV by CEUS but were misdiagnosed as Bosniak category IIF or less by CT, which led to a change in treatment plan. Two metastatic lesions from lung cancer in patients 4 and 5 were manifested as cystic lesions with minimally thickened wall and were categorized as IIF by CT but their walls appeared grossly thickened and irregular on CEUS, suggestive of category III. Unlike CT, CEUS detection of a solid component within three lesions supported diagnostic suspicion of malignant cystic tumor before surgery. 4. Discussion CEUS imaging is a microbubble-specific harmonic US, which is designed for the optimal detection of signals from SH U 508A [10–12] and can produce signals from stationary
B.K. Park et al. / European Journal of Radiology 61 (2007) 310–314
313
Table 1 Comparison of the CT and CEUS features in the malignant cystic tumors with discrepant Bosniak classification Patients
1 2 3 4 5 6 7 8
Category
Numbera
Thicknessb
Solid componentc
CT
CEUS
CT
CEUS
CT
CEUS
CT
CEUS
I II II IIF IIF III III III
IV IV IV III III IV IV IV
1 2 2 1 1 1 2 2
1 3 2 1 1 2 3 3
M H H M M G G G
M G G G G G G G
No No No No No No No No
Yes Yes Yes No No Yes Yes Yes
a Represents the number of septa: group 1, no septa; group 2, one to four; group 3, five to nine septa; group 4, more than nine septa. b Represents the thickness of wall or septa: hairline thin (H), minimally thickened (M) and grossly thickened and irregular (G). c Represents the enhanced area within a cystic tumor.
Fig. 2. A 46-year-old man with left renal cell carcinoma (patient 7). (a) Contrastenhanced CT scan shows a complex cystic mass containing grossly thickened and irregular septa (black arrows), ranging from one to four in number (group 2). (b) Longitudinal image of contrast-enhanced US demonstrates more septa (white arrows) within the lesion than CT. These septa range between five and nine in number (group 3) and appear more thickened and more apparent on CEUS than on CT. (c) The cut surface of surgically removed specimen shows multiple septa (arrows) of five to nine in number, which is better correlated with that of contrast-enhanced US than CT.
microbubbles as easily as from those that are moving in the vessels, whereas conventional US relies on a Doppler shift signals. CHA uses digital encoding and decoding of transmitted pulses, and can separate the contrast from the tissue and reduce unwanted echoes of the fundamental frequency and tissue harmonic signals. ADI uses two pulses of the same polarity and subtracts the signals from the two pulses and thus makes it possible to obtain only the fundamental and harmonic signals from the contrast agent while those from the tissue are removed. CEUS can improve the detection and characterization of focal liver lesions and can help to evaluate the therapeutic response to malignant hepatic tumors treated by interventional procedure because this US imaging technique can provide a better assessment of vascular morphology and enhancing patterns. However, to the best of our knowledge, the assessment of cystic renal masses with CEUS and CT has not been previously compared. The Bosniak classification system was designed to analyze the morphology of cystic renal masses solely based on CT findings, i.e., number of septa, thickness of wall and/or septa, lesion enhancement, and calcification. Israel et al suggested that MR imaging can be used to assess renal cystic lesions using Bosniak classification [9]. They found that MR imaging caused the upgrade of Bosniak classification because it demonstrated more septa, thicker septa and/or wall, and better enhancement than CT imaging. In our study, CEUS imaging was also found to be useful for evaluating cystic renal masses using Bosniak classification, and to be similar to CT imaging in the most cases. However, there was significant discrepancy of the imaging findings between CEUS and CT imaging in some cystic renal masses. Pathologic examinations revealed that CEUS imaging was superior to CT in terms of detecting additional septa, thickness of wall and/or septa, and solid component. Because US contrast agent circulating in the microvessels of septa, wall, and solid component produces strong signal intensity when it is destroyed by ultrasound, CEUS imaging could allow the sophisticated internal structures of cystic renal masses to be evaluated and result in higher resolution than CT. The demonstration of solid component within the cystic renal masses is important, as it is a key finding differentiating cate-
314
B.K. Park et al. / European Journal of Radiology 61 (2007) 310–314
gory III from IV. Lesion enhancement was more conspicuous on CEUS than on CT images in most of cases, and enhanced diagnostic confidence and eventually affected treatment plans. In the present study, several cystic lesions contained a solid component that was demonstrated by CEUS but not by CT. This finding resulted in surgical treatment although the number of patients with these lesions was small. The presence of calcification does not play a major role in the evaluation of cystic renal masses [13]. In our study, CEUS could not detect as many calcifications as CT. Calcification prevented precise evaluation of the thickness of wall and/or septa on CEUS as well as MR images [9]. Therefore, meticulous evaluation is necessary for lesions with diffusely calcified septa or wall on CEUS images. CEUS also has some disadvantages over CT. First, CEUS can be influenced by lesion location. A poor sonic window due to bowel gas or ribs prevents good quality image. Second, CEUS is an operator-dependent imaging modality and requires sufficient experience. Third, the US contrast agent used in our study has a high mechanical index and is unsuitable for continuous scan. A higher mechanical index leads to increase in the likelihood of microbubble because this index determines the level of interaction between bubbles and ultrasound. Therefore, second harmonic contrast agent requiring low mechanical index can be suitable for performing continuous scan and evaluating cystic renal masses. Our study has several limitations. First, interobserver variability was not evaluated because the cases were analyzed by consensus. In addition, US images were interpreted after CT images had been evaluated, which may have resulted in observer bias. Second, case-selection bias could have affected results because the cases were retrospectively collected and analyzed. Third, a relatively small number of pathologically proved cases were included. Thus, further investigation in surgical cases is needed to support our results. 5. Conclusion No difference in diagnostic accuracy was found between CT and CEUS for the assessment of cystic renal masses based
on Bosniak classification system. However, CEUS might better visualize septa number, septa and/or wall thickness, solid component and the enhancement of some renal cystic masses than CT, resulting in upgrade of Bosniak classification and affecting their treatment plan. References [1] Bosniak MA. The current radiological approach to renal cysts. Radiology 1986;158:1–10. [2] Bosniak MA. Diagnosis and management of patients with complicated cystic lesions of the kidney. AJR Am J Roentgenol 1997;169:819– 21. [3] Bosniak MA. The use of the Bosniak classification system for renal cysts and cystic tumors. J Urol 1997;157:1852–3. [4] Curry NS, Cochran ST, Bissada NK. Cystic renal masses: accurate Bosniak classification requires adequate renal CT. AJR Am J Roentgenol 2000;175:339–42. [5] Koga S, Nishikido M, Inuzuka S, et al. An evaluation of Bosniak’s radiological classification of cystic renal masses. BJU Int 2000;86:607– 9. [6] 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 Am J Roentgenol 1997;169:813–8. [7] Levy P, Helenon O, Merran S, et al. Cystic tumors of the kidney in adults: radio-histopathologic correlations. J Radiol 1999;80:121– 33. [8] Balci NC, Semelka RC, Patt RH, et al. Complex renal cysts: findings on MR imaging. AJR Am J Roentgenol 1999;172:1495– 500. [9] Israel GM, Hindman N, Bosniak MA. Evaluation of cystic renal masses: comparison of CT and MR imaging by using the Bosniak classification system. Radiology 2004;231:365–71. [10] Lee JY, Choi BI, Han JK, Kim AY, Shin SH, Moon SG. Improved sonographic imaging of hepatic hemangioma with contrast-enhanced coded harmonic angiography: comparison with MR imaging. Ultrasound Med Biol 2002;28:287–95. [11] Youk JH, Kim CS, Lee JM. Contrast-enhanced agent detection imaging: value in the characterization of focal hepatic lesions. J Ultrasound Med 2003;22:897–910. [12] Youk JH, Lee JM, Kim CS. Therapeutic response evaluation of malignant hepatic masses treated by interventional procedures with contrastenhanced agent detection imaging. J Ultrasound Med 2003;22:911– 20. [13] Israel GM, Bosniak MA. Calcification in cystic renal masses: is it important in diagnosis? Radiology 2003;226:47–52.