Is Ultrasound Useful for Further Evaluation of Homogeneously Hyperattenuating Renal Lesions Detected on CT?

Transcription

1 Genitourinary Imaging Original Research Genitourinary Imaging Original Research Mahadevaswamy Siddaiah 1 Satheesh Krishna Matthew D. F. McInnes Jeffrey S. Quon Wael M. Shabana Demetri Papadatos Nicola Schieda Siddaiah M, Krishna S, McInnes MDF, et al. Keywords: CT, cyst, hyperattenuating, hyperdense, ultrasound DOI: /AJR Received December 12, 2016; accepted after revision February 9, All authors: The Ottawa Hospital, The University of Ottawa, 1053 Carling Ave, Ottawa, ON K1Y 4E9, Canada. Address correspondence to N. Schieda (nschieda@toh.on.ca). Supplemental Data Available online at This article is available for credit. AJR 2017; 209: X/17/ American Roentgen Ray Society Is Ultrasound Useful for Further Evaluation of Homogeneously Hyperattenuating Renal Lesions Detected on CT? OBJECTIVE. The purpose of this study was to evaluate the ability of ultrasound (US) to characterize hyperattenuating cysts detected as indeterminate hyperattenuating renal lesions on unenhanced and single phase enhanced CT. MATERIALS AND METHODS. A total of 107 consecutive homogeneously hyperattenuating renal lesions underwent gray-scale and Doppler US at our institution between 2010 and Two radiologists who were unaware of the final diagnosis retrospectively evaluated US images for visibility and diagnosis (simple cyst, intermediate complexity cyst, cystic or solid mass showing internal flow on Doppler US, or indeterminate). A third radiologist assessed lesion size, location, and distance to skin on CT and US. US visibility was compared using chi-square and independent t tests. Consensus US interpretation was compared with reference standard diagnoses, and accuracy for diagnosis of hyperattenuating cysts was tabulated. RESULTS. Mean lesion size ± SD was 20 ± 16 mm (range, 6 96 mm) and mean distance to skin on CT was 62 ± 25 mm (range, mm). In all, 89.7% (96/107) of the lesions were visible on US, including all lesions that were 15 mm or larger. Nonvisible lesions were smaller than visible ones (10.0 ± 3.6 mm vs 20.7 ± 16.3 mm, p = 0.03) regardless of location (p > 0.05). CT overestimated lesion distance to skin compared with US (46.6 ± 18.6 mm, p < 0.001). Final diagnoses for US visible lesions (n = 96) were hyperattenuating cyst (n = 66), Bosniak IIF cyst (n = 13), and cystic or solid neoplasm (n = 15); two patients were lost to follow-up. Of the 66 hyperattenuating cysts, 54 (81.8%) appeared as simple cysts on US with sensitivity and specificity for diagnosis of hyperattenuating cyst of 81.8% (95% CI, %) and 92.9% (95% CI, %), respectively. The other 12 (18.2%) hyperattenuating cysts appeared complex. Two of the 13 Bosniak IIF lesions were incorrectly classified as simple cysts with US. Including the 11 (10%) nonvisible lesions reduced sensitivity and specificity for diagnosis of hyperattenuating cyst to 73.0% (95% CI, %) and 89.7% (95% CI, %), respectively. CONCLUSION. US can further characterize hyperattenuating cysts presenting as indeterminate hyperattenuating renal lesions on CT in the majority of cases. I ncidental renal lesions are frequently detected with CT [1, 2]. In the absence of previous imaging examinations, diagnosis is established through a combination of attenuation on unenhanced images, presence or absence of enhancement, and subjective imaging findings (e.g., margin, homogeneity, presence of calcification and fat) [2, 3]. Homogeneous lesions with attenuation of HU on unenhanced CT and > 20 HU on single phase contrast-enhanced CT without ancillary subjective features indicating malignancy are generally considered indeterminate [4 7]. These hyperattenuating lesions are most frequently hyperattenuating cysts with the increased attenuation resulting from internal proteinaceous or hemorrhagic contents [8]. However, renal cell carcinoma may have a similar appearance, and further imaging is generally required for characterization [2]. Ultrasound (US) can be used for characterizing renal masses, is readily available, and is relatively inexpensive [2, 9]. US can be performed to characterize hyperattenuating lesions detected at CT, but evidence supporting its use is limited. Two studies performed by Foster et al. [10] and Zirinsky et al. [11] found that most hyperattenuating cysts appear simple on US, although their results were significantly limited by extremely small sample sizes. The American College of Radi- 604 AJR:209, September 2017

2 ology (ACR) appropriateness criteria for imaging of indeterminate renal masses suggest that US may be useful for assessing hyperattenuating lesions, but the evidence cited by the ACR criteria is also limited [2, 12]. Accordingly, the utility of US for characterization of hyperattenuating lesions has been questioned by some authors because an unknown proportion of hyperattenuating cysts show low-level internal echoes preventing the diagnosis of a simple cyst with US [8, 13 16]. In more recent studies, lower rates of diagnosis of hyperattenuating cyst have been reported using US ranging from 33% to 42% [14 16]. When one considers that a proportion of lesions detected on CT may not be visible at all with US (e.g., 63% of cysts were not seen with US in the study by Zirinsky et al.), the use of US even as an intermediate step to multiphase CT or MRI is debatable. The purpose of this study was to determine the utility of US for further characterization of hyperattenuating renal lesions detected on CT by evaluating what percentage of hyperattenuating lesions detected at CT are visible with US, determining what factors may influence US visibility, and assessing what proportion of hyperattenuating cysts will appear simple on US, enabling benign diagnosis. Materials and Methods Our institutional review board approved this retrospective study and waived the need for informed consent in all patients. Between January 2010 and December 2013, a fellowship-trained radiologist with 5 years of experience in cross-sectional imaging manually reviewed 4250 consecutive abdominal US reports retrieved from our PACS to identify studies in which the indication for US was to further characterize a renal lesion detected on CT. The study period was preselected to provide a sufficient follow-up interval (2 years) to attempt to establish a final diagnosis in the majority of lesions. A total of 146 cases were identified in which US was performed to further characterize a renal lesion detected on CT. Figure 1 provides a flow diagram of patient selection with inclusion and exclusion criteria. CT Analysis All studies were analyzed using Horizon Medical Imaging software (version 11.0, McKesson). In all 146 patients, the CT examination was available in our PACS. One reviewer extracted the following quantitative measurements from the CT examination: size (the mean of the anterior-posterior, transverse, and craniocaudal dimensions), attenuation (measured three times with the mean recorded using a circular ROI tool encompassing two-thirds 4250 Consecutive abdominal US examinations reviewed between Jan 2010 and Dec Consecutive abdominal US studies performed to specifically characterize a renal lesion detected on preceding CT examination 107 Consecutive abdominal US studies evaluating 107 homogeneously hyperattenuating renal lesions (1 lesion in each patient) 104 Lesions with reference standard diagnoses available: US (n = 59) Multiphase CT or MRI (n = 39) Histopathology (n = 6) 94 Lesions included in the diagnostic accuracy portion of the study: 66 Hyperattenuating cysts 13 Bosniak IIF cysts 15 Cystic or solid renal masses - 5 RCC - 1 AML wovf - 9 Lesions without histopathology of the lesion on the center axial slice for homogeneous lesions or the most hyperattenuating portion for heterogeneous lesions), and distance from the lesion to the skin as a surrogate marker for body mass index (BMI), measured by drawing a straight line from the outer margin of the lesion to the skin on transverse CT images [17, 18] (Fig. 2). The same reviewer extracted features related to lesion location within the kidney. The laterality was recorded. Each lesion was characterized 4104 US studies excluded because indication of study was not to evaluate renal lesion detected on CT 21 Studies excluded because of characteristic imaging features establishing benign diagnosis on CT: 15 Simple cysts 4 Hyperattenuating cysts 2 AML 18 Studies excluded because of characteristic CT imaging features establishing cystic or solid renal mass All 107 lesions included in the US visibility portion of the study 3 Patients lost to follow-up including two patients for whom lesions were not visible on US 9 Additional patients excluded from diagnostic accuracy portion of study because lesions not visible on US: 8 Hyperattenuating cysts 1 Bosniak IIF cyst Fig. 1 Flow diagram of patient selection for this study. Benign diagnoses established at CT were simple cyst (well circumscribed, round or oval, smooth imperceptible wall, homogeneous, and attenuation between 10 and 20 HU), hyperattenuating cyst (well circumscribed, round or oval, smooth imperceptible wall, homogeneous, and attenuation > 70 HU on unenhanced CT), and angiomyolipoma (AML) (internal macroscopic fat with attenuation < 10 to 20 HU). CT features of cystic or solid renal mass included spiculated or irregular or invasive margin or contour and internal heterogeneity including visible septa or nodularity and irregular calcifications. Reference standard diagnoses were made by classic imaging features on ultrasound (US) or histopathology. RCC = renal cell carcinoma, wovf = without visible fat. as exophytic (more than two-thirds of the lesion projected beyond the renal cortical margin), endophytic (more than two-thirds of the lesion projected within the renal hilum), or mixed. Each lesion was localized to a pole of the kidney by drawing three horizontal lines on coronal CT images that divided the kidney into upper, lower, and interpolar regions. Lastly, lesions were localized to the anterior or posterior plane by drawing a line on axial CT images that divided the kidney in half. AJR:209, September

3 Twenty-one patients were excluded from the study because of a combination of attenuation measurements and subjective features on CT that enabled benign diagnoses of simple cyst, hyperattenuating cyst, and angiomyolipoma (AML) (Fig. 1). In these instances, US was recommended to confirm a provisional diagnosis provided on CT. An additional 18 patients were excluded because of CT features that indicated cystic or solid neoplasms including areas of irregular calcification (n = 2), invasive margins (n = 5), and internal heterogeneity (e.g., visible internal septa) (n = 11). The final study cohort consisted of 107 patients with 107 lesions. Fifty CT examinations were unenhanced and 57 were single phase contrast-enhanced (47 portal venous phase studies, seven arterial phase studies, and three split-bolus urographic phase studies) examinations. Ultrasound Analysis All US examinations were performed at a single tertiary care academic institution by US technologists who were directly supervised in person by fellowshiptrained radiologists. The mean time interval ± SD between CT and US was 46.6 ± 18.6 days; no patients had relevant interventions (e.g., biopsy, surgery, or chemotherapy) during that time. Patients were imaged in supine or decubitus positions with static grayscale and color or power Doppler ultrasound. Prone imaging was not performed. Movie clips were performed in 30.8% (33/107). US examinations were performed using Philips Healthcare ATL HDI 5000 (n = 29) or IU-22 (n = 33), Siemens Healthcare Sonoline Anteres (n = 19), or GE Healthcare Logiq E9 (n = 26). Curvilinear array transducers were used with a range of C4/5 1/2 MHz. Tissue harmonic imaging was performed in all patients to reduce unwanted background noise and low-level echoes that may be present in simple renal cysts [9]. Two fellowship-trained radiologists with 11 and 5 years of experience who were unaware of A the reference standard diagnoses reviewed the entire US examination and determined visibility and diagnosis. A lesion was considered visible when it was depicted in both the transverse and longitudinal axes and corresponded in location to the lesion detected on CT. CT images depicting the location of the lesion were provided during US interpretation to allow accurate coregistration of lesions. For US diagnoses, radiologists assigned the following categories. First, a simple cyst was a homogeneously anechoic round or ovoid structure with a sharply defined smooth imperceptible wall and increased sonographic through-transmission with up to two nonmeasurable septa or tiny calcifications [4, 9, 19]. Second, a cyst of intermediate complexity (roughly corresponding to Bosniak type IIF) was a cystic lesion with more than two septa and possibly minimal thickening of wall or septum and calcification but without internal flow on Doppler US [3]. The third category was a solid or part solid mass requiring internal flow on color or power Doppler US. Lastly, a lesion was categorized as indeterminate if it did not meet the above criteria. Categories were determined a priori in the study protocol on the basis of available literature. We applied the Bosniak criteria, which were initially validated for use with CT but subsequently applied to US, to classify intermediate complexity lesions not showing internal flow on color or power Doppler US. We prefer the term indeterminate complexity (rather than Bosniak IIF) cyst to acknowledge important differences when applying Bosniak criteria on CT and US [4, 9, 19]. For discordant cases, discrepancies were resolved by consensus. One reviewer measured the shortest distance from the lesion to the skin on US by drawing a straight line from the nearest edge of the lesion to the transducer. B Fig year-old man with hyperattenuating cyst in left kidney. A, Axial contrast-enhanced CT image in corticomedullary phase of enhancement depicts round, smooth margined, homogeneously hyperattenuating renal lesion in interpolar region of left kidney (arrow). Lesion attenuation was 55 HU. Line shows measurement of lesion distance to skin. B, Sagittal gray-scale sonographic image of left kidney shows round homogeneously anechoic structure with smooth imperceptible wall (solid arrow) and increased posterior through-transmission in far field (open arrow) in keeping with simple cyst. Line shows measurement of lesion distance to skin. Patients and Reference Standard Diagnoses Our study population included 60 men and 47 women; mean age was 67.7 ± 15.4 years. Final diagnoses were confirmed by histopathology (n = 6) or classic imaging features on US (requiring at least 2 years of interval follow-up stability [15]) (n = 59) or multiphase CT or MRI (n = 39). Imaging diagnoses were established by consensus interpretation of two readers performed 6 months after the initial blinded US interpretation session. Final diagnoses were hyperattenuating cyst (n = 75; 59 homogeneous hyperattenuating renal lesions with smooth wall showing features of a simple cyst on US and 16 showing no enhancement on multiphase CT or MRI [3]), Bosniak type IIF cyst (n = 14; multiple hairline-thin septa, perceived but nonmeasurable enhancement with or without minimal thickening of wall or septum, calcification [3]), cystic or solid enhancing mass on CT or MRI (n = 15; five renal cell carcinomas diagnosed by partial nephrectomy, one AML without visible fat diagnosed by partial nephrectomy, and nine enhancing lesions with no histologic confirmation). Three patients were lost to follow-up. Multiphase CT or MRI Comparator Group To assess the presence of an institutional bias toward imaging hyperattenuating lesions with US only in larger lesions or thinner patients (compared with characterization with multiphase or renal CT or MRI), a post hoc analysis was performed. One reviewer retrieved all renal protocol CT or MRI examinations performed at the same institution from January 2013 to December 2013; 215 renal CT and 116 renal MRI examinations were identified. Figure S1 (which can be viewed in the AJR electronic supplement to this article, available at provides a description of the inclusion and exclusion of studies for the post hoc analysis. Fifty-four hyperattenuating lesions (30 detected on previous unenhanced CT and 24 on single phase contrast-enhanced CT) that underwent multiphase CT or MRI for further evaluation were compared with the US cohort. One reviewer measured lesion size and distance to skin as described. 606 AJR:209, September 2017

4 Statistical Analysis Results were tabulated and categoric and nominal variables were compared with US visibility using chi-square or independent t tests, respectively. Distance of lesion to skin on CT and US were compared using a paired t test. A paired t test was also used to compare lesion size and distance to skin in the post hoc analysis. Interobserver agreement for US diagnosis was assessed using the Cohen kappa statistic. Diagnostic accuracy of US was calculated by categorizing lesions into two groups: hyperattenuating cyst or potentially malignant cystic or solid renal lesions (including Bosniak IIF cysts and cystic or solid masses). Statistical analysis was performed using STATA software (version 13.0, Statcorp). TABLE 1: Ultrasound (US) Visibility of Homogeneously Hyperattenuating Renal Lesions Detected on CT Characteristic Total No. (n = 107) Results Both readers considered 89.7% (96/107) of lesions visible with US (κ = 1.0). Mean lesion size was 20 ± 16 mm (range, 6 96 mm) with no statistically significant difference between measurements on unenhanced and contrastenhanced CT (23 ± 18 mm [range, 4 96 mm] vs 17 ± 13 mm [range, mm], p = 0.07). Mean lesion distance to skin measured on CT was 61 ± 23 mm (range, mm) with no statistically significant difference between measurements on unenhanced and contrastenhanced CT (66 ± 21 mm vs 58 ± 25 mm, p = 0.14). No statistically significant difference was seen in lesion location between unenhanced and contrast-enhanced CT (p > 0.05). Because of the lack of statistically significant differences for these characteristics on unenhanced and contrast-enhanced CT, lesions were combined for statistical analyses; Table 1 summarizes those results. Lesions that were not visible on US were smaller than US-visible lesions (10 ± 4 mm [range, 4 14 mm] vs 21 ± 16 mm [range, mm], p = 0.03). All lesions 15 mm or larger were visible with US. No statistically significant difference was seen in US visibility by distance to skin measured on CT (p = 0.38). Lesion distance to skin was larger when measured on CT compared with US (61 ± 23 mm vs 47 ± 19 mm, p < 0.001) (Fig. 3). No statistically significant difference in US visibility was seen with respect to location of lesions (p > 0.05). One reviewer identified 57 simple cysts, three intermediate complexity cysts, and 34 cystic or solid renal masses and considered lesions in two patients to be indeterminate on US. The other reviewer identified 49 simple cysts, five intermediate complexity cysts, and 35 cystic or solid renal masses and considered lesions in seven patients to be indeterminate on US. Interobserver agreement was strong (κ = 0.81). Both readers identified the same 11 cases in which renal lesions were not visible with US. The final diagnoses based on reference standards for the 11 US nonvisible lesions were eight hyperattenuating cysts, one Bosniak IIF cyst, and two lesions lost to follow-up. No. Not Visible on US (n = 11) No. Visible on US (n = 96) p a Laterality 0.29 Right 45 (42.1) 3 (27.3) 42 (43.8) Left 62 (57.9) 8 (72.7) 54 (56.3) Anterior or posterior location 0.07 Anterior 50 (46.7) 8 (72.7) 42 (43.8) Posterior 57 (53.3) 3 (27.3) 54 (56.3) Craniocaudal location 0.50 Upper pole 37 (34.6) 5 (45.5) 32 (33.3) Interpolar region 34 (31.8) 4 (36.4) 30 (31.3) Lower pole 36 (33.6) 2 (18.2) 34 (35.4) Endophytic or exophytic b 0.18 Exophytic 53 (49.5) 5 (45.5) 48 (50.0) Mixed 17 (15.9) 0 17 (17.7) Endophytic 37 (34.6) 6 (55.5) 31 (32.3) Mean size ± SD (mm) 20 ± ± 4 21 ± Mean distance to skin ± SD c (mm) 62 ± ± ± Note Values in parentheses are percentages. Some percentages do not add up to 100 due to rounding. a Comparisons performed between groups using chi-square or independent t tests for categoric and nominal data, respectively. b Endophytic lesions were defined as projecting two-thirds beyond the renal cortical margin; endophytic lesions were defined as projecting two-thirds within the renal hilum. c Lesion distance to skin was measured as a straight line from the outer margin of the lesion to the overlying skin on transverse CT images (Fig. 2). Fig. 3 Scatter diagram of lesion distance to skin as measured on transverse CT and ultrasound images showing weak positive correlation (ß = 0.3) and overestimation of distance on CT compared with ultrasound (US) Lesion Distance to Skin on CT (mm) Lesion Distance to Skin on US (mm) Of the 96 lesions that were visible with US, two were lost to follow-up, and the remaining 94 were used to calculate the accuracy of US (Fig. 1). The sensitivity and specificity (95% CIs) of US for diagnosis of hyperattenuating cyst were 81.8% ( %) and 92.9% ( %). Therefore, 81.8% (54/66) hyperattenuating cysts were correctly classified and appeared simple on US (Fig. 2). Twelve AJR:209, September

5 of 66 (18.2%) hyperattenuating cysts were incorrectly classified as complex lesions with US. The false-positive rate of diagnosis of hyperattenuating cyst with US was 2.1% (2/94); in both instances the reference standard diagnosis was Bosniak IIF cyst (Fig. 4). Twenty-six of 28 (92.9%) lesions that were not hyperattenuating cysts appeared complex on US. Including the 11 nonvisible lesions (10%) in the diagnostic accuracy assessment reduced the sensitivity and specificity of US for diagnosis of hyperattenuating cyst to 73.0% (95% CI, %) and 89.7% (95% CI, %), respectively. Our post hoc analysis showed mean lesion sizes and distance to skin for patients imaged with multiphase CT or MRI of 29 ± 18 mm (range, 6 74 mm) and 58 ± 20 mm (range, mm). Lesions imaged by US were significantly smaller (p = 0.001) with no significant difference in lesion distance to skin seen between groups (p = 0.15). Discussion This study evaluated the role of US for evaluation of hyperattenuating renal lesions detected on CT, specifically evaluating the accuracy of US for diagnosis of hyperattenuating cysts. A majority of lesions were visible on US, and size was the only factor associated with reduced visibility. Over 80% of hyperattenuating cysts appeared simple on US. The number of false-positive diagnoses of simple cysts on US was low (2 of 94 cases). Including the 10% of lesions that were not visible with US in the accuracy portion of our analysis reduced the sensitivity and specificity of US for diagnosing hyperattenuating cysts, but overall accuracy was still very good. Some debate continues regarding the utility of US (compared with multiphase CT and MRI) for characterizing hyperattenuating lesions detected on CT, with limited evidence supporting the use of US in this setting. Although contrastenhanced US is considered valuable for characterizing indeterminate renal lesions [14, 20, 21], routine sonographic assessment of hyperattenuating lesions detected on CT has been questioned because of the reported phenomenon of hyperattenuating cysts showing variable internal complexity on US and a tendency for US to erroneously upgrade simple cysts [8, 13 16, 22]. Our results suggest US could be the next-step imaging test for characterizing indeterminate homogeneously hyperattenuating renal lesions encountered on CT because it can accurately depict and characterize hyperattenuating cysts in most cases. The rate of visibility of lesions with US in our study compares favorably to what has been reported, namely that a majority of renal lesions that are not isoechoic are detectable [9, 19]. In this study, radiologists could only retrospectively review images that were mainly static, so with real-time assessment or movie clips the rate of visualization could be higher than what we report. The rate of visualization in our study is higher than the 37% reported by Zirinsky et al. [11], but differences likely relate to advances in US technology since their publication in Previously described features associated with limited US visibility of renal lesions include habitus, multiple lesions, polar location, smaller lesion size, and isoechogenicity to renal cortex [2, 9, 19]. In our study, the only factor associated with reduced US visibility was lesion size. The difference could be because the lesions in our study population consisted mainly of cysts and cystic lesions, which may be more readily seen on US compared with solid lesions, and because none of our patients had multiple lesions. Moreover, in our study, US was performed to further characterize lesions already identified on CT. This finding suggests that a priori knowledge of a lesion may improve its detectability with US. Lesion distance to skin (measured with CT), a surrogate marker for BMI, was not associated with subsequent lesion visibility on US, which may be explained by our observation that measurements on CT were larger when compared with US. These differences are likely the result of the ability to compress tissues with transducer pressure and alter patient position during US. Among lesions visible on US, 81.8% of hyperattenuating cysts appeared simple on US, thus allowing a diagnosis of benignity. Reported accuracies of US for diagnosis of hyperattenuating cyst have varied widely [9 11, 14 16]. However, our results support findings by Foster et al. [10] and Zirinsky et al. [11], who showed that five out of seven and four out of five US-visible hyperattenuating cysts appeared simple, respectively. Our study yielded a very low false-positive rate of diagnosis of hyperattenuating cysts. In both false-positive interpretations, lesions were diagnosed as Bosniak IIF cysts. Given the low risk of malignancy among Bos- A B C Fig year-old woman with Bosniak IIF cystic lesion in upper pole of left kidney and simple cyst in upper pole of right kidney. A, Axial contrast-enhanced CT image shows homogeneously hyperattenuating renal lesion in upper pole of left kidney (black arrow). Lesion measures 52 HU in density. White arrow depicts simple cyst in upper pole of right kidney. B, Corresponding sagittal gray-scale image shows lesion as simple cyst without internal complexity (arrow). C, Coronal image from follow-up multiphase CT study performed 3 months after ultrasound (US) for unexplained anemia shows multiple subjectively (but nonmeasurable) enhancing septa within lesion (arrowheads) that were not visible on US. Lesion (black arrow) was classified as Bosniak IIF lesion. White arrow depicts simple cyst in upper pole of right kidney. 608 AJR:209, September 2017

6 niak IIF lesions (approximately 6% [3, 23]) our results suggest that even when an erroneous diagnosis of a simple cyst is made after US to evaluate a homogeneously hyperattenuating lesion, the overall risk to the patient is low. Given that our study was a retrospective analysis of mainly static images, use of real-time evaluation or movie clips may further reduce the false-positive interpretations by depicting more internal complexity than conveyed on static images; however, this hypothesis would require further analysis in prospective studies. One potential caveat to our results would occur in the setting of a clinically suspected acutely hemorrhagic cyst, which appears indeterminately hyperattenuating on CT. In this instance, US may be of limited value because internal complexity could be a result of acute hemorrhage. Nevertheless, this uncommon clinical presentation would not apply to the intended study population, which consists of incidentally discovered hyperattenuating lesions seen on CT. Our study has limitations. The single-institution study design limits the applicability of our results to other centers. Although we found no difference in US visibility related to lesion distance to skin (a surrogate marker for body mass index [17, 18]), we were not able to report BMI because it was not available in patient medical records. The recorded distance to skin on US suggests that our patient population was of average size, so applicability of our findings to obese and morbidly obese patients requires further validation. To strengthen generalizability of our results, we performed a post hoc analysis that showed no bias toward imaging thinner patients or larger lesions with US. The reference standard in the majority of cases was by characteristic imaging appearance and not histopathology. Moreover, US was the reference standard for diagnosis of approximately 80% of hyperattenuating cysts. We acknowledge that this situation is suboptimal because US was also the diagnostic test being studied. Nevertheless, we feel this limitation is to be expected because multiphase CT or MRI characterization of lesions that US has characterized as simple cysts is unrealistic (given the high accuracy of US for diagnosis of a simple cyst [9, 19]) and because 2-year follow-up stability was required for instances where US was used to form the reference standard for diagnoses [15]. Approximately half of the cystic or solid renal masses in this study did not have histopathologic correlation, which can be expected in the era of surveillance of small (< 4 cm) renal masses [24]. Lastly, we combined hyperattenuating lesions detected at unenhanced CT and single phase contrast-enhanced CT to improve sample size and because we found no statistically significant differences in lesion size or location between differing CT techniques. Some hyperattenuating cysts measuring > 20 HU on single phase contrast-enhanced CT examinations may have measured 10 and 20 HU at unenhanced CT and attenuation may have been increased by pseudoenhancement [25, 26]. In conclusion, this study shows that US visualizes a substantial majority of hyperattenuating renal lesions including all lesions measuring 15 mm or greater. Among homogeneously hyperattenuating lesions, US was accurate for diagnosis of hyperattenuating simple cysts with a very low false-positive rate and a false-negative rate of diagnosis of approximately 20%. Our study results, combined with what has already been reported, suggest that ultrasound should be the nextstep imaging test for characterization of indeterminate homogeneously hyperattenuating renal lesions encountered at CT. References 1. Berland LL, Silverman SG, Gore RM, et al. Managing incidental findings on abdominal CT: white paper of the ACR incidental findings committee. J Am Coll Radiol 2010; 7: Heilbrun ME, Casalino DD, Beland MD, et al. American College of Radiology appropriateness criteria: indeterminate renal mass. American College of Radiology website. acsearch.acr.org/ docs/69367/narrative/. Published Updated Accessed May 26, Israel GM, Bosniak MA. How I do it: evaluating renal masses. Radiology 2005; 236: Bosniak MA. The current radiological approach to renal cysts. Radiology 1986; 158: Pooler BD, Pickhardt PJ, O Connor SD, Bruce RJ, Patel SR, Nakada SY. Renal cell carcinoma: attenuation values on unenhanced CT. AJR 2012; 198: O Connor SD, Silverman SG, Ip IK, Maehara CK, Khorasani R. Simple cyst-appearing renal masses at unenhanced CT: can they be presumed to be benign? Radiology 2013; 269: Jonisch AI, Rubinowitz AN, Mutalik PG, Israel GM. Can high-attenuation renal cysts be differentiated from renal cell carcinoma at unenhanced CT? Radiology 2007; 243: Hartman DSCP, Hartman S. From the RSNA Refresher Courses: a practical approach to the cystic renal mass. RadioGraphics 2004; 24:S101 S Hélénon O, Correas JM, Balleyguier C, Ghouadni M, Cornud F. Ultrasound of renal tumors. Eur Radiol 2001; 11: Foster WL Jr, Roberts L Jr, Halvorsen RA Jr, Dunnick NR. Sonography of small renal masses with indeterminant density characteristics on computed tomography. Urol Radiol 1988; 10: Zirinsky K, Auh YH, Rubenstein WA, Williams JJ, Pasmantier MW, Kazam E. CT of the hyperdense renal cyst: sonographic correlation. AJR 1984; 143: Silverman SG, Lee BY, Seltzer SE, Bloom DA, Corless CL, Adams DF. Small ( 3 cm) renal masses: correlation of spiral CT features and pathologic findings. AJR 1994; 163: Silverman SG, Mortele KJ, Tuncali K, Jinzaki M, Cibas ES. Hyperattenuating renal masses: etiologies, pathogenesis, and imaging evaluation. RadioGraphics 2007; 27: Bertolotto M, Cicero C, Perrone R, Degrassi F, Cacciato F, Cova MA. Renal masses with equivocal enhancement at CT: characterization with contrastenhanced ultrasound. AJR 2015; 204:[web] W557 W Xue LY, Lu Q, Huang BJ, et al. Contrast-enhanced ultrasonography for evaluation of cystic renal mass: in comparison to contrast-enhanced CT and conventional ultrasound. Abdom Imaging 2014; 39: Nicolau C, Bunesch L, Pano B, et al. Prospective evaluation of CT indeterminate renal masses using US and contrast-enhanced ultrasound. Abdom Imaging 2015; 40: Kim J, Lim H, Lee SI, Kim YJ. Thickness of rectus abdominis muscle and abdominal subcutaneous fat tissue in adult women: correlation with age, pregnancy, laparotomy, and body mass index. Arch Plast Surg 2012; 39: Borruel S, Molto JF, Alpanes M, et al. Surrogate markers of visceral adiposity in young adults: waist circumference and body mass index are more accurate than waist hip ratio, model of adipose distribution and visceral adiposity index. PLoS One 2014; 9:e Kim SH, Cho JY, Kim SY, Moon KC, Kwak C, Kim HH. Ultrasound evaluation of renal masses: gray-scale, Doppler, and more. Ultrasound Clin 2013; 8: Setola SV, Catalano O, Sandomenico F, Siani A. Contrast-enhanced sonography of the kidney. Abdom Imaging 2007; 32: Cazals X, Bleuzen A, Tranquart F, et al. Value of contrast-enhanced ultrasound in the evaluation of AJR:209, September

7 indeterminate renal cysts on CT. [in French] J Radiol 2011; 92: Hindman NM. Imaging of cystic renal masses. Radiol Clin North Am 2017; 55: Whelan TF. Guidelines on the management of renal cyst disease. Can Urol Assoc J 2010; 4: Pierorazio PM, Hyams ES, Mullins JK, Allaf ME. Active surveillance for small renal masses. Rev Urol 2012; 14: Israel GM, Bosniak MA. Pitfalls in renal mass evaluation and how to avoid them. RadioGraphics 2008; 28: Maki DD, Birnbaum BA, Chakraborty DP, Jacobs JE, Carvalho BM, Herman GT. Renal cyst pseudoenhancement: beam-hardening effects on CT numbers. Radiology 1999; 213: Schieda N, Vakili M, Dilauro M, Hodgdon T, Flood TA, Shabana WM. Solid renal cell carcinoma measuring water attenuation ( 10 to 20 HU) on unenhanced CT. AJR 2015; 205: Schieda N, Kielar AZ, Al Dandan O, McInnes MD, Flood TA. Ten uncommon and unusual variants of renal angiomyolipoma (AML): radiologic-pathologic correlation. Clin Radiol 2015; 70: FOR YOUR INFORMATION A data supplement for this article can be viewed in the online version of the article at: This article is available for CME and Self-Assessment (SA-CME) credit that satisfies Part II requirements for maintenance of certification (MOC). To access the examination for this article, follow the prompts associated with the online version of the article. 610 AJR:209, September 2017