Genitourinar y Imaging • Original Research Wehrli et al. MRI of Renal Cell Carcinoma and Urothelial Carcinoma
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
Genitourinary Imaging Original Research
Utility of MRI Features in Differentiation of Central Renal Cell Carcinoma and Renal Pelvic Urothelial Carcinoma Natasha E. Wehrli1 Min Ju Kim1,2 Brent W. Matza1 Jonathan Melamed 3 Samir S. Taneja4 Andrew B. Rosenkrantz1 Wehrli NE, Kim MJ, Matza BW, Melamed J, Taneja SS, Rosenkrantz AB
Keywords: apparent diffusion coefficient, central renal cell carcinoma, diffusion-weighted imaging, MRI, normalized apparent diffusion coefficient, normalized T2 signal, renal pelvic urothelial carcinoma DOI:10.2214/AJR.13.10673 Received January 26, 2013; accepted after revision March 9, 2013. 1
Department of Radiology, NYU Langone Medical Center, 660 First Ave, New York, NY 10016. Address correspondence to N. E. Wehrli ([email protected]). 2
Present affiliation: Department of Radiology, Anam Hospital, Korea University College of Medicine, Seoul, South Korea.
3 Department of Pathology, NYU Langone Medical Center, New York, NY. 4 Department of Urology, Division of Urologic Oncology, NYU Langone Medical Center, New York, NY.
AJR 2013; 201:1260–1267 0361–803X/13/2016–1260 © American Roentgen Ray Society
1260
OBJECTIVE. The purpose of this article is to evaluate the utility of various morphologic and quantitative MRI features in differentiating central renal cell carcinoma (RCC) from renal pelvic urothelial carcinoma. MATERIALS AND METHODS. Sixty patients (39 men and 21 women; mean [± SD] age, 65 ± 14 years; 48 with central RCC and 12 with renal pelvic urothelial carcinoma) who underwent MRI, including diffusion-weighted imaging (b values, 0, 400, and 800 s/mm2) and dynamic contrast-enhanced imaging, before histopathologic confirmation were included. Tumor T2 signal intensity and apparent diffusion coefficients (ADCs) were measured and normalized to muscle and CSF (hereafter referred to as normalized T2 signal and normalized ADC, respectively) and then were compared using receiver operating characteristic analysis. Also, two blinded radiologists independently assessed all tumors for various qualitative features, which were compared with the Fisher exact test and unpaired Student t test. RESULTS. Urothelial carcinoma exhibited significantly lower normalized ADC than did RCC (p = 0.008), but no significant difference was seen in ADC or normalized T2 signal intensity (p = 0.247–0.773). Normalized ADC had the highest area under the curve (0.757); normalized ADC below an optimal threshold of 0.451 was associated with sensitivity of 83% and specificity of 71% for diagnosing urothelial carcinoma. Features that were significantly more prevalent in urothelial carcinoma included global impression of urothelial carcinoma, location centered within the collecting system, collecting system defect, extension to the ureteropelvic junction, preserved renal shape, absence of cystic or necrotic areas, absence of hemorrhage, homogeneous enhancement, and hypovascularity (all p < 0.033). Increased T1 signal intensity suggestive of hemorrhage was significantly more prevalent in RCC (p = 0.02). Interreader agreement for the subjective features ranged from 61.7% to 98.3%. CONCLUSION. In addition to various qualitative MRI parameters, normalized ADC has utility in differentiating central RCC from renal pelvic urothelial carcinoma. Such differentiation may assist decisions regarding possible biopsy and treatment planning.
U
rothelial carcinoma of the renal pelvis is estimated to comprise approximately 8% of renal tumors [1]. Intrarenal urothelial carcinoma occurs more commonly in the renal pelvis than in the infundibular calyceal system [2]. Early in the disease process, renal pelvic urothelial carcinoma may present as broad-based or frondlike papillary filling defects within the renal collecting system or as focal thickening of the urothelium and is often associated with partial obstruction of a renal calyx. As the tumor grows, centripetal infiltration into the renal parenchyma is common [3]. In these cases, distinguishing renal pelvic urothelial carcinoma from a central endophytic renal cell carcinoma (RCC) may present a diagnostic challenge.
Differentiating renal pelvic urothelial carcinoma from central RCC is important because of the differences in surgical management and prognosis. Renal pelvic urothelial carcinoma typically portends a poorer prognosis, with renal infiltration indicative of stage III disease [4–6]. Because of the variable sensitivity of urine cytology for urothelial carcinoma [7, 8], indeterminate intrarenal masses suspicious for urothelial carcinoma usually undergo ureteroscopic biopsy before surgical management [9]. To avoid complications of biopsy, it would be helpful to improve imaging characterization of these lesions before surgery. Suspected urothelial carcinoma warrants nephroureterectomy, with resection of the bladder cuff
AJR:201, December 2013
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
MRI of Renal Cell Carcinoma and Urothelial Carcinoma as the mainstay of surgical therapy. Locally advanced central RCC has a comparably favorable prognosis and is usually managed with radical nephrectomy, although nephronsparing surgery may be performed in smaller lesions, and newer less-invasive techniques, such as radiofrequency ablation and cryotherapy, are also emerging as potential treatment options [10]. A recent study by Raza et al. [11] assessed the utility of various features on contrastenhanced CT for distinguishing renal pelvic urothelial carcinoma from central RCC, including but not limited to central location within the collecting system, preservation of renal shape, and homogeneous enhancement. However, to our knowledge, there have been no recent studies comparing features of these lesions on MRI, which offers a wider spectrum of tissue contrast mechanisms for lesion characterization. Furthermore, although a number of studies have investigated apparent diffusion coefficient (ADC) values from diffusion-weighted imaging (DWI) to help distinguish benign and malignant renal lesions, to distinguish between subtypes of RCC, and for preliminary evaluation of urinary tract tumors [12–18], there is a paucity of literature investigating the use of diffusion-based metrics for distinguishing renal pelvic urothelial carcinoma and RCC. Therefore, the purpose of our study was to retrospectively evaluate the utility of various MRI features, including DWI-based evaluation, in differentiating central RCC from renal pelvic urothelial carcinoma. We hypothesize that, in addition to anatomic features, a difference in these quantitative features will be helpful in distinguishing these two tumors.
Materials and Methods Patients
tained during the corticomedullary (30 seconds), nephrographic (60 seconds), and excretory (3 minutes) phases. Sequence parameters for our institution’s protocol are detailed in Table 1 and have been described in prior studies from our institution [19–21]. Examinations were performed using 0.1 mmol/kg gadopentetate dimeglumine (Magnevist, Bayer Pharmaceuticals) injected at a rate of 2 mL/s and followed by a 20-mL saline bolus.
This retrospective study was HIPAA compliant and was approved by our institutional review board with a waiver of the requirement for written informed consent. We searched databases from our institution from 2005 to 2012 to identify patients with pathologically proven renal urothelial carcinoma and RCC who underwent MRI, including both DWI and dynamic contrast-enhanced imaging, before resection. A fellowship-trained abdominal radiologist then reviewed imaging of all cases of RCC and selected those that were central in location; this was defined as a mass extending into the renal sinus and having a component of the mass extending into or appearing inseparable from a portion of the renal collecting system. This search initially identified 69 cases. Additional cases were excluded for the following reasons: nonstandard b values for DWI (n = 5), MRI performed at 3 T (n = 2), or mass not clearly visualized (n = 2). These exclusions provided a final cohort of 60 patients (39 men and 21 women; mean [± SD] age, 65 ± 14 years; 12 renal pelvic urothelial carcinomas and 48 central RCCs).
Image Analysis A single radiologist, who was an abdominal imaging fellow at the time of image review, reviewed all cases while blinded to the final pathologic diagnosis. This observer placed a region of interest (ROI) to measure the signal intensity (SI) of each mass on the HASTE sequence, as well as the ADC value of each lesion on the ADC map. Two ROIs were placed on each mass on each sequence, encompassing a maximal cross-sectional area of the lesion on two consecutive slices on both image sets, and the mean values of these ROIs were averaged. If the mass was visualized on only a single slice, then an ROI was placed within the lesion twice on the same slice depicting the mass. ROIs were also measured within the psoas muscle on HASTE images and within the CSF surrounding the spinal nerve roots on the ADC maps. Because T2 SI values are arbitrary and because of potential concerns regarding scanner- and sequence-related factors on ADC values, the T2 SI values and ADCs were normalized to other tissues, similar to the method described in prior studies [22–25]. Normalized tumor T2 SI was calculated as the T2 SI of the tumor divided by the T2 SI of muscle, and normalized ADC was calculated as the ADC of the tumor divided by the ADC of CSF. Two radiologists, including the previous observer (a fellowship-trained radiologist
MRI Technique The examinations were performed using a 1.5-T clinical system (Magnetom Avanto, Symphony, or Sonata, all from Siemens Healthcare). All examinations included axial or coronal HASTE; axial 2D T1-weighted in- and opposed-phase gradient-echo sequences; axial or coronal single-shot echo-planar imaging DWI with tridirectional motion-probing gradients and b values of 0, 400, and 800 s/mm2, as well as in-line reconstruction of ADC maps; and dynamic contrast-enhanced images through the kidneys using a 3D fat-suppressed spoiled gradient-echo T1-weighted sequence, with images ob-
TABLE 1: MRI Pulse Sequence Parameters 2D T1-Weighted Dual-Echo GRE
Parameter
T1-Weighted 3D Fat-Suppressed GRE
HASTE Axial or coronal or both
Axial
Breath-Hold DWI
Navigator-Triggered DWI
Axial or coronal or both
Axial or coronal or both
Imaging plane
Axial
TR (ms)
160–204
Infinite
3.1–4.5
1300–2000
One respiratory cycle
TE (ms)
2.0–2.7 (4.4–5.0)a
62–98
1.2–1.9
65–85
65–85
Flip angle (°)
80
150–180
12
NA
NA
FOV (mm)
325–500b
325–500
325–500b
325–500b
325–500b
Matrix
112–208 × 256
192–320 × 256–380
125–177 × 256
144–192 × 192
144–192 × 192
Section thickness (mm)
6–8
4–6
2–3
6–8
6–8
Parallel imaging factor
Nonec
None
2
2
2
b Values (s/mm2)
NA
NA
NA
0, 400, and 800
0, 400, and 800
Note—DWI = diffusion-weighted imaging, GRE = gradient-recalled echo, NA = not applicable. aData are opposed-phase TE, with in-phase time in parentheses. bA 70–80% rectangular FOV was used. cParallel imaging factor of 2 was used with Avanto system (Siemens Healthcare).
AJR:201, December 2013 1261
Wehrli et al.
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
TABLE 2: Summary of Patient Demographic Features and Lesion Sizes Characteristic
Renal Pelvic Urothelial Carcinoma
Central Renal Cell Carcinoma
p
74.1 ± 15.3
61.9 ± 13.2
0.008
33
6
Patient age (y), mean ± SD Sex (no. of patients)
0.318
Male Female Lesion diameter (mm), mean ± SD
15
6
29.2 ± 15.4
71.9 ± 33.4
with 10 years of experience in abdominal radiology), independently evaluated all masses while blinded to the final pathologic diagnosis. All sequences were reviewed during a single session. During each session, the readers provided a global impression as to the diagnosis of the lesion (renal pelvic urothelial carcinoma vs central RCC) and also evaluated each lesion in terms of the presence or absence of a spectrum of features. These included the six morphologic features observed to be of greatest diagnostic utility in a previous
0.001
study—central location within the collecting system, a focal defect within the collecting system, preserved renal shape, absence of cystic or necrotic areas, homogeneous enhancement, and extension toward the ureteropelvic junction [11]—as well as an additional set of features related to the lesion’s SI on various MRI sequences, including homogeneous T2 SI on HASTE images, decreased T2 SI relative to cortex on HASTE images, loss of T1 SI on opposed-phase T1-weighted images relative to in-phase T1-weighted images suggestive of
microscopic lipid, areas of increased T1 SI on unenhanced volumetric interpolated breath-hold examination images suggestive of intratumoral hemorrhage, hypovascularity of the lesion relative to cortex on all contrast-enhanced sequences, areas of increased SI on DWI obtained using a b value of 800 s/m2, and visual areas of decreased ADC. Central location within the collecting system was judged as the center point of the tumor being located within, rather than eccentric to, the collecting system. Subsequent to their independent assessments, these two radiologists at a later date jointly reviewed the imaging for the lesions for which there was a discrepancy between radiologists and established a consensus interpretation.
Statistical Analysis The Fisher exact test was used to compare patient sex and the presence of the various binary imaging features between renal pelvic urothelial carcinoma and central RCC. An unpaired Student t test was used to compare mean age, tumor diameter,
TABLE 3: Summary Data Regarding Performance of Quantitative MRI Metrics for Differentiating Renal Pelvic Urothelial Carcinoma From Central Renal Cell Carcinoma (RCC), Including Results of Receiver Operating Characteristic (ROC) Analysis Parameter
Central RCC (Mean ± SD)
Renal Pelvic Urothelial Carcinoma (Mean ± SD)
p (RCC vs. Urothelial Carcinoma)
AUC
Optimum Threshold for Diagnosing Urothelial Carcinoma
Sensitivity (%)
Specificity (%)
Normalized T2 signal intensity
2.84 ± 1.30
3.60 ± 2.15
0.247
0.620
≤ 2.255
58
77
ADC
1.70 ± 1.20
Normalized ADC
0.49 ± 0.13
1.67 ± 1.09
0.773
0.591
≤ 1.62
83
44
0.38 ± 0.10
0.008
0.757
≤ 0.451
83
71
Note—ADC = apparent diffusion coefficient, AUC = area under the ROC curve.
TABLE 4: Performance Values for MRI Features Used to Help Distinguish Renal Pelvis Urothelial Carcinoma From Central Renal Cell Carcinoma (RCC) Frequency in Central RCC
Frequency in Renal Pelvic Urothelial Carcinoma
pa (RCC vs Urothelial Carcinoma)
Global impression of urothelial carcinoma
4 (2/48)
100 (12/12)
< 0.001a
Centered on renal collecting system
6 (3/48)
100 (12/12)
< 0.001a
Defect within the central collecting system
27 (13/48)
100 (12/12)
< 0.001a
Preserved renal shape
10 (5/48)
100 (12/12)
< 0.001a
Absence of cystic areas or necrosis
21 (10/48)
92 (11/12)
< 0.001a
Homogeneous enhancement
8 (4/48)
75 (9/12)
< 0.001a
Extending to the ureteropelvic junction
8 (4/48)
83 (10/12)
< 0.001a
Feature
Dark on HASTE sequences
21 (10/48)
42 (5/12)
0.153
Homogeneous signal on HASTE sequences
8 (4/48)
83 (10/12)
< 0.001a
Loss of signal on opposed-phase T1-weighted images suggesting intracytoplastic lipid
19 (9/48)
0 (0/12)
0.182
T1 bright areas suggestive of hemorrhage
48 (23/48)
8 (1/12)
0.019a
Hypovascular on all dynamic sequences
25 (12/48)
92 (11/12)
< 0.001a
Bright on high-b-value diffusion-weighted imaging
88 (42/48)
58 (7/12)
0.033a
Note—Except for p values, data are percentage (no./total). aStatistically significant (p < 0.05).
1262
AJR:201, December 2013
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
MRI of Renal Cell Carcinoma and Urothelial Carcinoma
A
B
C
D
Fig. 1—83-year-old woman with macroscopic hematuria and central right renal mass. Endoscopic biopsy of mass was consistent with urothelial carcinoma. A, Coronal HASTE image shows homogeneous hypointense mass (arrow) centered in right renal pelvis. B, Axial contrast-enhanced T1-weighted early nephrographic phase image shows that lesion (arrow) is homogeneously hypovascular. C, Diffusion-weighted image shows uniform increased signal intensity (arrow) in lesion. D, Apparent diffusion coefficient (ADC) map shows uniform low signal intensity (arrow) in lesion (normalized ADC = 0.202).
normalized T2 SI, ADC, and normalized ADC between the two groups. Receiver operating characteristic (ROC) curve analysis was used to assess and compare the overall diagnostic performance of each quantitative metric for differentiation of the two entities; the ROC analysis was used to derive
optimal diagnostic thresholds for each metric, defined as the value that maximizes the average of sensitivity and specificity. All p values are two-sided and are considered statistically significant at p < 0.05. Statistical analysis was performed using software (MedCalc version 10.4, Frank Schoonjans).
Interreader agreement was calculated as the percentage of cases where both readers provided the same initial assessment. A Cohen kappa coefficient was also calculated for each variable, where kappa values were interpreted as follows: 0–0.20, slight agreement; 0.21–0.40, fair agreement; 0.41–0.60,
A
B
C
D
Fig. 2—67-year-old man with mass initially identified on outside CT. Specimen obtained at radical nephrectomy was consistent with clear cell renal cell carcinoma. A, Coronal HASTE image shows hypointense endophytic lesion (arrow) in left upper renal pole. B, Axial contrast-enhanced T1-weighted early nephrographic phase image shows heterogeneous enhancement (arrow) of lesion. C, Diffusion-weighted image shows focus of increased signal intensity (arrow) in lesion. D, Apparent diffusion coefficient (ADC) map shows heterogeneous low signal intensity (arrow) in lesion (normalized ADC = 0.469).
AJR:201, December 2013 1263
Wehrli et al.
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
moderate agreement; 0.61–0.80, substantial agreement; and 0.81–1.00, almost perfect agreement [26].
Results Patient and Lesion Characteristics Patient demographic features and tumor diameters of all cases are summarized in Table 2. There was no significant difference between renal pelvic urothelial carcinoma and central RCC in terms of patient sex (p = 0.318). However, patients with renal pelvic urothelial carcinoma were significantly older than those with central RCC (74.1 ± 15.3 years vs 61.9 ± 13.2 years; p = 0.008). In addition, renal pelvic urothelial carcinoma exhibited a significantly smaller diameter than renal pelvic RCC (29.2 ± 15.4 mm vs 71.9 ± 33.4 mm; p = 0.001). Quantitative MRI Metrics Table 3 provides summary data regarding the performance of the quantitative metrics normalized T2 SI, ADC, and normalized ADC, as well as the results of the ROC analysis. There was significantly lower normalized ADC in renal pelvic urothelial carcinoma compared with central RCC (p = 0.008), but there was no significant difference between these two entities for ADC (p = 0.773) or normalized T2 SI (p = 0.247). At ROC analysis, normalized ADC had the highest area under the ROC curve (AUC) of
these three metrics (AUC = 0.757), which was associated with an optimal threshold of 0.451. A normalized ADC value below this threshold was associated with an 83% sensitivity and 71% specificity in distinguishing renal pelvic urothelial carcinoma and central RCC. The higher AUC of normalized ADC than ADC was nearly significant (p = 0.052); remaining pairwise comparisons between AUCs of these metrics were not significant (p = 0.140–0.758). Subjective Imaging Features Global impression was significantly different between RCC and urothelial carcinoma (p < 0.001), accurately identifying the correct diagnosis in 58 of 60 cases. Subjective features that were significantly more common in urothelial carcinoma included a focal collecting system defect (p < 0.001), central location on the collecting system (p < 0.001), homogeneous signal on HASTE (p < 0.001), absence of cystic or necrotic components (p < 0.001), homogeneous enhancement (p < 0.001), hypovascularity on all contrast-enhanced sequences (p < 0.001), preserved renal shape (p < 0.001), and extension to the ureteropelvic junction (p < 0.001) (Table 4). Subjective features that were significantly more common in RCC included the presence of T1 bright areas suggestive of hemorrhage (p = 0.019) and areas of increased
SI on high-b-value DWI (p = 0.033). Features that showed no significant difference between groups included decreased SI on HASTE (p = 0.153) and the presence of intratumoral lipid (p = 0.182). Characteristic features are presented in Figures 1–3. Interreader Agreement Kappa coefficient analysis showed almost perfect agreement for global impression, defect within the collecting system, preserved renal shape, and hypovascularity; substantial agreement for central location within the collecting system and homogeneous signal on HASTE; and moderate agreement for extension toward the ureteropelvic junction, homogeneous enhancement, and loss of signal on opposed-phase T1. However, the readers agreed in most cases for all qualitative features. Specifically, interreader agreement was 98.3% for global impression and ranged from 61.7% to 93.3% for the remaining subjective features (Table 5). Discussion Because central renal urothelial carcinoma comprises up to 8% of intrarenal neoplasms, noninvasive characterization of renal masses has become a topic of increasing interest among radiologists and urologists to facilitate surgical planning and patient selection for renal mass biopsy. Although urine cytolo-
A
B
C
D
Fig. 3—79-year-old woman with left intrarenal mass for which partial nephrectomy was initially considered. However, nephroureterectomy performed after surgical biopsy revealed high-grade urothelial carcinoma. A, Axial HASTE image shows hypointense endophytic mass (arrow) in left anterior renal midpole. B, Axial contrast-enhanced T1-weighted nephrographic phase shows homogeneous hypoenhancement (arrow) of lesion. C, Diffusion-weighted image shows diffuse increased signal intensity (arrow) in lesion. D, Apparent diffusion coefficient (ADC) map shows homogeneous low signal intensity (arrow) in lesion (normalized ADC = 0.38).
1264
AJR:201, December 2013
MRI of Renal Cell Carcinoma and Urothelial Carcinoma TABLE 5: Interreader Agreement for Subjective Features Interreader Agreementa
κ
Global impression of urothelial carcinoma
98.3 (59/60)
0.952
Centered on renal collecting system
90.0 (54/60)
0.707
Absence of cystic areas or necrosis
66.7 (40/60)
0.378
Defect within the central collecting system
91.7 (55/60)
0.815
Extending to the ureteropelvic junction
93.3 (56/60)
0.514
Preserved renal shape
91.7 (55/60)
0.806
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
Feature
Increased signal on high-b-value diffusion-weighted imaging
61.7 (37/60)
0.092
Decreased signal on HASTE sequences
68.3 (41/60)
0.284
T1 bright areas suggestive of hemorrhage
68.3 (41/60)
0.320
Homogeneous enhancement
80.0 (48/60)
0.514
Homogeneous signal on HASTE sequences
86.7 (52/60)
0.617
Hypovascular on all dynamic sequences
91.7 (55/60)
0.823
Loss of signal on opposed-phase T1-weighted images suggesting intracytoplastic lipid
90.0 (54/60)
0.520
aData are percentage (no./total).
gy is an inexpensive noninvasive screening test with an accuracy of up to 98% for high-grade urothelial carcinoma, sensitivities and specificities as low as 8.5% and 50%, respectively, have been reported for low-grade papillary urothelial carcinoma. In addition, distinguishing between centrally infiltrative urothelial carcinoma of the kidney and central RCC can be a diagnostic challenge on imaging. Nonetheless, differentiation is important because of differences in stage, prognosis, and surgical management of these tumors. There has been an increase in the use of MRI for the detection and characterization of solid renal tumors. One MRI technique that has received particular attention recently for lesion characterization is DWI. Investigators have shown that subjective findings on DWI and quantitative DWI-based metrics (namely, ADCs) may help to reliably distinguish solid renal neoplasms from normal parenchyma, hemorrhagic cysts, and necrotic regions of solid tumors [12–15]. For instance, Paudyal et al. [16] determined that ADC measurements can be helpful for distinguishing between clear cell and non– clear cell subtypes of RCC, and Goyal et al. [27] showed that ADC values correlate with grade of RCC. To our knowledge, ours is the first study to investigate the utility of ADC for distinguishing renal pelvic urothelial carcinoma from central RCC. In our study, we investigated the utility of standard ADC values and ADC values normalized to CSF, a technique motivated to help offset variability between examinations relating to differences in scanners and
sequence parameters, as well as possible biologic effects contributing to variability [24, 25]. Although ADC was not significantly different between the two groups, normalized ADC was significantly lower in urothelial carcinoma compared with RCC. Furthermore, normalized T2 SI measurements showed no significant difference between urothelial carcinoma and RCC. It is possible that the greater utility of the ADC-based metric relates to the stronger association between ADC and tumor cellularity, whereas T2-weighted SI measurements may be more strongly confounded by a host of additional properties, including tumor edema, inflammation, and presence of collagenous or fibrous tissue. Thus, according to our data, normalized ADC seems to be the most useful of the investigated quantitative metrics for distinguishing the two diagnoses. The role of ADC normalization is of particular note given conflicting results of the value of ADC normalization in past studies in other tissues [24, 25]. In addition to quantitative metrics, we assessed various additional features related to tumor morphologic features, signal characteristics, and enhancement patterns. Similar to the study of Raza et al. [11], which explored the CT features of centrally infiltrative urothelial carcinoma versus central RCC, global impression was the most reliable subjective feature in our series. The six additional most helpful features in the study by Raza et al., including tumor center within the collecting system, filling defect within the renal pelvis, renal shape preservation, absence of cystic or necrotic components, ho-
mogeneous enhancement, and extension toward the ureteropelvic junction, showed a similarly strong association with urothelial carcinoma in our series. The study by Raza et al. [11] did not assess the degree of lesion enhancement on CT. It is recognized that urothelial carcinoma is low in attenuation on contrast-enhanced CT [28, 29], whereas clear cell RCC, representing the most common RCC subtype, is hypervascular. (Although papillary RCC is generally hypovascular, this subtype only accounts for about 10–20% of RCCs [13, 30, 31].) In our study, we compared subjective enhancement between diagnoses and observed that urothelial carcinoma was hypovascular relative to renal parenchyma on all dynamic contrastenhanced phases on MRI at a significantly greater frequency than RCC (p < 0.001). An additional feature that showed a significantly higher frequency in urothelial carcinoma in our series was the absence of cystic or necrotic components. The study by Raza et al. [11] described this feature as difficult to assess on CT. Greater utility of this feature on MRI may relate to the greater contrast resolution and multitude of contrast mechanisms available for tissue characterization with MRI. The absence of cystic or necrotic components in urothelial carcinoma likely contributes to its typical homogeneous SI on HASTE images, which we observed with significantly higher frequency in urothelial carcinoma. An MRI feature that was much more frequent in RCC than urothelial carcinoma was the presence of high-SI areas on unenhanced T1-weighted images, a finding suggestive of
AJR:201, December 2013 1265
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
Wehrli et al. intratumoral hemorrhage. It is recognized that RCCs may outgrow their blood supply as they increase in size, developing hemorrhage into areas of necrotic tumor [32]. Papillary RCC is particularly known to contain blood products of varying chronicity on pathologic assessment [33, 34]. On the other hand, intratumoral hemorrhage was observed in only one of 12 cases of urothelial carcinoma in our series. An interesting observation in our study was that, although increased SI on high-bvalue DWI was subjectively present more often in RCC than urothelial carcinoma, there was significantly lower measured normalized ADC in urothelial carcinoma than in RCC. This may relate to alternate causes for increased SI on high-b-value DWI, aside from diffusion restriction related to tumor cellularity. For instance, the increased SI on DWI in RCC may relate to T2 shine-through effect [35] from tumor necrosis, a common feature of clear cell carcinoma, rather than true diffusion restriction from densely cellular tumor. However, the lower measured normalized ADC in urothelial carcinoma likely represents true diffusion restriction in this group. In addition, we note that increased signal on high-b-value DWI had the lowest interreader agreement of any of the included subjective imaging features of only 61.7% (κ = 0.092), further raising concerns about the utility of this subjective assessment of DWI. There are several limitations to our study. First, the study was retrospective. Second, the sample size of renal pelvic urothelial carcinoma was small. However, this represents the natural low incidence of this entity. Also, we did not assess other entities that may present as a central renal mass, such as lymphoma or a metastasis to the kidney. Finally, DWI was performed using a maximal b value of 800 s/mm2 in all cases, consistent with our institutional protocol for MRI of the urinary tract. It is possible that the diagnostic performance of DWI and ADC metrics would be different in the setting of an alternate b-value acquisition scheme. In conclusion, we have presented the largest series to our knowledge to explore differences in MRI features between renal pelvic urothelial carcinoma and central RCC. Normalized ADC values, but not ADC or normalized T2 SI, were significantly different between these entities. In addition to morphologic features previously characterized on CT, urothelial carcinoma was significantly more likely to exhibit homogeneous MRI
1266
SI, absence of cystic or necrotic areas, and hypovascularity on contrast-enhanced MRI sequences, whereas RCC was significantly more likely to exhibit MRI SI characteristics suggestive of hemorrhage. Prospective differentiation of these two entities, based on cross-sectional imaging characteristics, may assist in decisions regarding biopsy and surgical management. References 1. Wagle DG, Moore RH, Murphy GP. Primary carcinoma of the renal pelvis. Cancer 1974; 33:1642– 1648 2. Nocks BN, Heney NM, Daly JJ, Perrone TA, Griffin PP, Prout GR Jr. Transitional cell carcinoma of renal pelvis. Urology 1982; 19:472–477 3. Wong-You-Cheong JJ, Wagner BJ, Davis CJ Jr. Transitional cell carcinoma of the urinary tract: radiologic-pathologic correlation. RadioGraphics 1998; 18:123–142; quiz, 148 4. Guinan P, Volgelzang NJ, Randazzo R, et al. Renal pelvic transitional cell carcinoma: the role of the kidney in tumor-node-metastasis staging. Cancer 1992; 69:1773–1775 5. Baron RL, McClennan BL, Lee JK, Lawson TL. Computed tomography of transitional-cell carcinoma of the renal pelvis and ureter. Radiology 1982; 144:125–130 6. Guinan P, Vogelzang NJ, Randazzo R, et al. Renal pelvic cancer: a review of 611 patients treated in Illinois 1975–1985. Cancer Incidence and End Results Committee. Urology 1992; 40:393–399 7. Bastacky S, Ibrahim S, Wilczynski SP, Murphy WM. The accuracy of urinary cytology in daily practice. Cancer 1999; 87:118–128 8. Raab SS, Grzybicki DM, Vrbin CM, Geisinger KR. Urine cytology discrepancies: frequency, causes, and outcomes. Am J Clin Pathol 2007; 127:946–953 9. Rouprêt M, Zigeuner R, Palou J, et al. European guidelines for the diagnosis and management of upper urinary tract urothelial cell carcinomas: 2011 update. European Association of Urology Guideline Group for urothelial cell carcinoma of the upper urinary tract (in Spanish). Actas Urol Esp 2012; 36:2–14 10. Schrader AJ, Steffens S. Renal cell carcinoma update: news from the AUA, EAU, and ASCO annual meetings 2011. ISRN Urol 2012; 2012:748235 11. Raza SA, Sohaib SA, Sahdev A, et al. Centrally infiltrating renal masses on CT: differentiating intrarenal transitional cell carcinoma from centrally located renal cell carcinoma. AJR 2012; 198:846–853 12. Kim S, Jain M, Harris AB, et al. T1 hyperintense renal lesions: characterization with diffusionweighted MR imaging versus contrast-enhanced MR imaging. Radiology 2009; 251:796–807
13. Taouli B, Thakur RK, Mannelli L, et al. Renal lesions: characterization with diffusion-weighted imaging versus contrast-enhanced MR imaging. Radiology 2009; 251:398–407 14. Squillaci E, Manenti G, Di Stefano F, Miano R, Strigari L, Simonetti G. Diffusion-weighted MR imaging in the evaluation of renal tumours. J Exp Clin Cancer Res 2004; 23:39–45 15. Zhang J, Tehrani YM, Wang L, Ishill NM, Schwartz LH, Hricak H. Renal masses: characterization with diffusion-weighted MR imaging—a preliminary experience. Radiology 2008; 247:458–464 16. Paudyal B, Paudyal P, Tsushima Y, et al. The role of the ADC value in the characterisation of renal carcinoma by diffusion-weighted MRI. Br J Radiol 2010; 83:336–343 17. Kim S, Naik M, Sigmund E, Taouli B. Diffusionweighted MR imaging of the kidneys and the urinary tract. Magn Reson Imaging Clin N Am 2008; 16:585–596; vii–viii 18. Yoshida S, Masuda H, Ishii C, Saito K, Kawakami S, Kihara K. Initial experience of functional imaging of upper urinary tract neoplasm by diffusion-weighted magnetic resonance imaging. Int J Urol 2008; 15:140–143 19. Dodelzon K, Mussi TC, Babb JS, Taneja SS, Rosenkrantz AB. Prediction of growth rate of solid renal masses: utility of MR imaging features—preliminary experience. Radiology 2012; 262:884–893 20. Rosenkrantz AB, Hindman N, Fitzgerald EF, Niver BE, Melamed J, Babb JS. MRI features of renal oncocytoma and chromophobe renal cell carcinoma. AJR 2010; 195:[web]W421–W427 21. Rosenkrantz AB, Niver BE, Fitzgerald EF, Babb JS, Chandarana H, Melamed J. Utility of the apparent diffusion coefficient for distinguishing clear cell renal cell carcinoma of low and high nuclear grade. AJR 2010; 195:[web]W344–W351 22. Choi HJ, Kim JK, Ahn H, Kim CS, Kim MH, Cho KS. Value of T2-weighted MR imaging in differentiating low-fat renal angiomyolipomas from other renal tumors. Acta Radiol 2011; 52:349–353 23. Hindman N, Ngo L, Genega EM, et al. Angiomyolipoma with minimal fat: can it be differentiated from clear cell renal cell carcinoma by using standard MR techniques? Radiology 2012; 265:468–477 24. Sasaki M, Yamada K, Watanabe Y, et al. Variability in absolute apparent diffusion coefficient values across different platforms may be substantial: a multivendor, multi-institutional comparison study. Radiology 2008; 249:624–630 25. Rosenkrantz AB, Kopec M, Kong X, et al. Prostate cancer vs. post-biopsy hemorrhage: diagnosis with T2- and diffusion-weighted imaging. J Magn Reson Imaging 2010; 31:1387–1394 26. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics
AJR:201, December 2013
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
MRI of Renal Cell Carcinoma and Urothelial Carcinoma 1977; 33:159–174 27. Goyal A, Sharma R, Bhalla AS, et al. Diffusionweighted MRI in renal cell carcinoma: a surrogate marker for predicting nuclear grade and histological subtype. Acta Radiol 2012; 53:349– 358 28. Bree RL, Schultz SR, Hayes R. Large infiltrating renal transitional cell carcinomas: CT and ultrasound features. J Comput Assist Tomogr 1990; 14:381–385 29. Browne RF, Meehan CP, Colville J, Power R, Torreggiani WC. Transitional cell carcinoma of the upper urinary tract: spectrum of imaging find-
ings. RadioGraphics 2005; 25:1609–1627 30. Herts BR, Coll DM, Novick AC, et al. Enhancement characteristics of papillary renal neoplasms revealed on triphasic helical CT of the kidneys. AJR 2002; 178:367–372 31. Rosen MA, Schnall MD. Dynamic contrast-enhanced magnetic resonance imaging for assessing tumor vascularity and vascular effects of targeted therapies in renal cell carcinoma. Clin Cancer Res 2007; 13:770s–776s 32. Sengupta S, Lohse CM, Leibovich BC, et al. Histologic coagulative tumor necrosis as a prognostic indicator of renal cell carcinoma aggressiveness.
Cancer 2005; 104:511–520 33. Prasad SR, Humphrey PA, Catena JR, et al. Common and uncommon histologic subtypes of renal cell carcinoma: imaging spectrum with pathologic correlation. RadioGraphics 2006; 26:1795– 1806; discussion, 1806–1710 34. Pedrosa I, Sun MR, Spencer M, et al. MR imaging of renal masses: correlation with findings at surgery and pathologic analysis. RadioGraphics 2008; 28:985–1003 35. Koh DM, Collins DJ. Diffusion-weighted MRI in the body: applications and challenges in oncology. AJR 2007; 188:1622–1635
F O R YO U R I N F O R M AT I O N
Unique customized medical search engine service from ARRS! ARRS GoldMiner ® is a keyword- and concept-driven search engine that provides instant access to radiologic images published in peer-reviewed journals. For more information, visit goldminer.arrs.org.
AJR:201, December 2013 1267
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
Genitourinary Imaging Original Research
Utility of MRI Features in Differentiation of Central Renal Cell Carcinoma and Renal Pelvic Urothelial Carcinoma Natasha E. Wehrli1 Min Ju Kim1,2 Brent W. Matza1 Jonathan Melamed 3 Samir S. Taneja4 Andrew B. Rosenkrantz1 Wehrli NE, Kim MJ, Matza BW, Melamed J, Taneja SS, Rosenkrantz AB
Keywords: apparent diffusion coefficient, central renal cell carcinoma, diffusion-weighted imaging, MRI, normalized apparent diffusion coefficient, normalized T2 signal, renal pelvic urothelial carcinoma DOI:10.2214/AJR.13.10673 Received January 26, 2013; accepted after revision March 9, 2013. 1
Department of Radiology, NYU Langone Medical Center, 660 First Ave, New York, NY 10016. Address correspondence to N. E. Wehrli ([email protected]). 2
Present affiliation: Department of Radiology, Anam Hospital, Korea University College of Medicine, Seoul, South Korea.
3 Department of Pathology, NYU Langone Medical Center, New York, NY. 4 Department of Urology, Division of Urologic Oncology, NYU Langone Medical Center, New York, NY.
AJR 2013; 201:1260–1267 0361–803X/13/2016–1260 © American Roentgen Ray Society
1260
OBJECTIVE. The purpose of this article is to evaluate the utility of various morphologic and quantitative MRI features in differentiating central renal cell carcinoma (RCC) from renal pelvic urothelial carcinoma. MATERIALS AND METHODS. Sixty patients (39 men and 21 women; mean [± SD] age, 65 ± 14 years; 48 with central RCC and 12 with renal pelvic urothelial carcinoma) who underwent MRI, including diffusion-weighted imaging (b values, 0, 400, and 800 s/mm2) and dynamic contrast-enhanced imaging, before histopathologic confirmation were included. Tumor T2 signal intensity and apparent diffusion coefficients (ADCs) were measured and normalized to muscle and CSF (hereafter referred to as normalized T2 signal and normalized ADC, respectively) and then were compared using receiver operating characteristic analysis. Also, two blinded radiologists independently assessed all tumors for various qualitative features, which were compared with the Fisher exact test and unpaired Student t test. RESULTS. Urothelial carcinoma exhibited significantly lower normalized ADC than did RCC (p = 0.008), but no significant difference was seen in ADC or normalized T2 signal intensity (p = 0.247–0.773). Normalized ADC had the highest area under the curve (0.757); normalized ADC below an optimal threshold of 0.451 was associated with sensitivity of 83% and specificity of 71% for diagnosing urothelial carcinoma. Features that were significantly more prevalent in urothelial carcinoma included global impression of urothelial carcinoma, location centered within the collecting system, collecting system defect, extension to the ureteropelvic junction, preserved renal shape, absence of cystic or necrotic areas, absence of hemorrhage, homogeneous enhancement, and hypovascularity (all p < 0.033). Increased T1 signal intensity suggestive of hemorrhage was significantly more prevalent in RCC (p = 0.02). Interreader agreement for the subjective features ranged from 61.7% to 98.3%. CONCLUSION. In addition to various qualitative MRI parameters, normalized ADC has utility in differentiating central RCC from renal pelvic urothelial carcinoma. Such differentiation may assist decisions regarding possible biopsy and treatment planning.
U
rothelial carcinoma of the renal pelvis is estimated to comprise approximately 8% of renal tumors [1]. Intrarenal urothelial carcinoma occurs more commonly in the renal pelvis than in the infundibular calyceal system [2]. Early in the disease process, renal pelvic urothelial carcinoma may present as broad-based or frondlike papillary filling defects within the renal collecting system or as focal thickening of the urothelium and is often associated with partial obstruction of a renal calyx. As the tumor grows, centripetal infiltration into the renal parenchyma is common [3]. In these cases, distinguishing renal pelvic urothelial carcinoma from a central endophytic renal cell carcinoma (RCC) may present a diagnostic challenge.
Differentiating renal pelvic urothelial carcinoma from central RCC is important because of the differences in surgical management and prognosis. Renal pelvic urothelial carcinoma typically portends a poorer prognosis, with renal infiltration indicative of stage III disease [4–6]. Because of the variable sensitivity of urine cytology for urothelial carcinoma [7, 8], indeterminate intrarenal masses suspicious for urothelial carcinoma usually undergo ureteroscopic biopsy before surgical management [9]. To avoid complications of biopsy, it would be helpful to improve imaging characterization of these lesions before surgery. Suspected urothelial carcinoma warrants nephroureterectomy, with resection of the bladder cuff
AJR:201, December 2013
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
MRI of Renal Cell Carcinoma and Urothelial Carcinoma as the mainstay of surgical therapy. Locally advanced central RCC has a comparably favorable prognosis and is usually managed with radical nephrectomy, although nephronsparing surgery may be performed in smaller lesions, and newer less-invasive techniques, such as radiofrequency ablation and cryotherapy, are also emerging as potential treatment options [10]. A recent study by Raza et al. [11] assessed the utility of various features on contrastenhanced CT for distinguishing renal pelvic urothelial carcinoma from central RCC, including but not limited to central location within the collecting system, preservation of renal shape, and homogeneous enhancement. However, to our knowledge, there have been no recent studies comparing features of these lesions on MRI, which offers a wider spectrum of tissue contrast mechanisms for lesion characterization. Furthermore, although a number of studies have investigated apparent diffusion coefficient (ADC) values from diffusion-weighted imaging (DWI) to help distinguish benign and malignant renal lesions, to distinguish between subtypes of RCC, and for preliminary evaluation of urinary tract tumors [12–18], there is a paucity of literature investigating the use of diffusion-based metrics for distinguishing renal pelvic urothelial carcinoma and RCC. Therefore, the purpose of our study was to retrospectively evaluate the utility of various MRI features, including DWI-based evaluation, in differentiating central RCC from renal pelvic urothelial carcinoma. We hypothesize that, in addition to anatomic features, a difference in these quantitative features will be helpful in distinguishing these two tumors.
Materials and Methods Patients
tained during the corticomedullary (30 seconds), nephrographic (60 seconds), and excretory (3 minutes) phases. Sequence parameters for our institution’s protocol are detailed in Table 1 and have been described in prior studies from our institution [19–21]. Examinations were performed using 0.1 mmol/kg gadopentetate dimeglumine (Magnevist, Bayer Pharmaceuticals) injected at a rate of 2 mL/s and followed by a 20-mL saline bolus.
This retrospective study was HIPAA compliant and was approved by our institutional review board with a waiver of the requirement for written informed consent. We searched databases from our institution from 2005 to 2012 to identify patients with pathologically proven renal urothelial carcinoma and RCC who underwent MRI, including both DWI and dynamic contrast-enhanced imaging, before resection. A fellowship-trained abdominal radiologist then reviewed imaging of all cases of RCC and selected those that were central in location; this was defined as a mass extending into the renal sinus and having a component of the mass extending into or appearing inseparable from a portion of the renal collecting system. This search initially identified 69 cases. Additional cases were excluded for the following reasons: nonstandard b values for DWI (n = 5), MRI performed at 3 T (n = 2), or mass not clearly visualized (n = 2). These exclusions provided a final cohort of 60 patients (39 men and 21 women; mean [± SD] age, 65 ± 14 years; 12 renal pelvic urothelial carcinomas and 48 central RCCs).
Image Analysis A single radiologist, who was an abdominal imaging fellow at the time of image review, reviewed all cases while blinded to the final pathologic diagnosis. This observer placed a region of interest (ROI) to measure the signal intensity (SI) of each mass on the HASTE sequence, as well as the ADC value of each lesion on the ADC map. Two ROIs were placed on each mass on each sequence, encompassing a maximal cross-sectional area of the lesion on two consecutive slices on both image sets, and the mean values of these ROIs were averaged. If the mass was visualized on only a single slice, then an ROI was placed within the lesion twice on the same slice depicting the mass. ROIs were also measured within the psoas muscle on HASTE images and within the CSF surrounding the spinal nerve roots on the ADC maps. Because T2 SI values are arbitrary and because of potential concerns regarding scanner- and sequence-related factors on ADC values, the T2 SI values and ADCs were normalized to other tissues, similar to the method described in prior studies [22–25]. Normalized tumor T2 SI was calculated as the T2 SI of the tumor divided by the T2 SI of muscle, and normalized ADC was calculated as the ADC of the tumor divided by the ADC of CSF. Two radiologists, including the previous observer (a fellowship-trained radiologist
MRI Technique The examinations were performed using a 1.5-T clinical system (Magnetom Avanto, Symphony, or Sonata, all from Siemens Healthcare). All examinations included axial or coronal HASTE; axial 2D T1-weighted in- and opposed-phase gradient-echo sequences; axial or coronal single-shot echo-planar imaging DWI with tridirectional motion-probing gradients and b values of 0, 400, and 800 s/mm2, as well as in-line reconstruction of ADC maps; and dynamic contrast-enhanced images through the kidneys using a 3D fat-suppressed spoiled gradient-echo T1-weighted sequence, with images ob-
TABLE 1: MRI Pulse Sequence Parameters 2D T1-Weighted Dual-Echo GRE
Parameter
T1-Weighted 3D Fat-Suppressed GRE
HASTE Axial or coronal or both
Axial
Breath-Hold DWI
Navigator-Triggered DWI
Axial or coronal or both
Axial or coronal or both
Imaging plane
Axial
TR (ms)
160–204
Infinite
3.1–4.5
1300–2000
One respiratory cycle
TE (ms)
2.0–2.7 (4.4–5.0)a
62–98
1.2–1.9
65–85
65–85
Flip angle (°)
80
150–180
12
NA
NA
FOV (mm)
325–500b
325–500
325–500b
325–500b
325–500b
Matrix
112–208 × 256
192–320 × 256–380
125–177 × 256
144–192 × 192
144–192 × 192
Section thickness (mm)
6–8
4–6
2–3
6–8
6–8
Parallel imaging factor
Nonec
None
2
2
2
b Values (s/mm2)
NA
NA
NA
0, 400, and 800
0, 400, and 800
Note—DWI = diffusion-weighted imaging, GRE = gradient-recalled echo, NA = not applicable. aData are opposed-phase TE, with in-phase time in parentheses. bA 70–80% rectangular FOV was used. cParallel imaging factor of 2 was used with Avanto system (Siemens Healthcare).
AJR:201, December 2013 1261
Wehrli et al.
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
TABLE 2: Summary of Patient Demographic Features and Lesion Sizes Characteristic
Renal Pelvic Urothelial Carcinoma
Central Renal Cell Carcinoma
p
74.1 ± 15.3
61.9 ± 13.2
0.008
33
6
Patient age (y), mean ± SD Sex (no. of patients)
0.318
Male Female Lesion diameter (mm), mean ± SD
15
6
29.2 ± 15.4
71.9 ± 33.4
with 10 years of experience in abdominal radiology), independently evaluated all masses while blinded to the final pathologic diagnosis. All sequences were reviewed during a single session. During each session, the readers provided a global impression as to the diagnosis of the lesion (renal pelvic urothelial carcinoma vs central RCC) and also evaluated each lesion in terms of the presence or absence of a spectrum of features. These included the six morphologic features observed to be of greatest diagnostic utility in a previous
0.001
study—central location within the collecting system, a focal defect within the collecting system, preserved renal shape, absence of cystic or necrotic areas, homogeneous enhancement, and extension toward the ureteropelvic junction [11]—as well as an additional set of features related to the lesion’s SI on various MRI sequences, including homogeneous T2 SI on HASTE images, decreased T2 SI relative to cortex on HASTE images, loss of T1 SI on opposed-phase T1-weighted images relative to in-phase T1-weighted images suggestive of
microscopic lipid, areas of increased T1 SI on unenhanced volumetric interpolated breath-hold examination images suggestive of intratumoral hemorrhage, hypovascularity of the lesion relative to cortex on all contrast-enhanced sequences, areas of increased SI on DWI obtained using a b value of 800 s/m2, and visual areas of decreased ADC. Central location within the collecting system was judged as the center point of the tumor being located within, rather than eccentric to, the collecting system. Subsequent to their independent assessments, these two radiologists at a later date jointly reviewed the imaging for the lesions for which there was a discrepancy between radiologists and established a consensus interpretation.
Statistical Analysis The Fisher exact test was used to compare patient sex and the presence of the various binary imaging features between renal pelvic urothelial carcinoma and central RCC. An unpaired Student t test was used to compare mean age, tumor diameter,
TABLE 3: Summary Data Regarding Performance of Quantitative MRI Metrics for Differentiating Renal Pelvic Urothelial Carcinoma From Central Renal Cell Carcinoma (RCC), Including Results of Receiver Operating Characteristic (ROC) Analysis Parameter
Central RCC (Mean ± SD)
Renal Pelvic Urothelial Carcinoma (Mean ± SD)
p (RCC vs. Urothelial Carcinoma)
AUC
Optimum Threshold for Diagnosing Urothelial Carcinoma
Sensitivity (%)
Specificity (%)
Normalized T2 signal intensity
2.84 ± 1.30
3.60 ± 2.15
0.247
0.620
≤ 2.255
58
77
ADC
1.70 ± 1.20
Normalized ADC
0.49 ± 0.13
1.67 ± 1.09
0.773
0.591
≤ 1.62
83
44
0.38 ± 0.10
0.008
0.757
≤ 0.451
83
71
Note—ADC = apparent diffusion coefficient, AUC = area under the ROC curve.
TABLE 4: Performance Values for MRI Features Used to Help Distinguish Renal Pelvis Urothelial Carcinoma From Central Renal Cell Carcinoma (RCC) Frequency in Central RCC
Frequency in Renal Pelvic Urothelial Carcinoma
pa (RCC vs Urothelial Carcinoma)
Global impression of urothelial carcinoma
4 (2/48)
100 (12/12)
< 0.001a
Centered on renal collecting system
6 (3/48)
100 (12/12)
< 0.001a
Defect within the central collecting system
27 (13/48)
100 (12/12)
< 0.001a
Preserved renal shape
10 (5/48)
100 (12/12)
< 0.001a
Absence of cystic areas or necrosis
21 (10/48)
92 (11/12)
< 0.001a
Homogeneous enhancement
8 (4/48)
75 (9/12)
< 0.001a
Extending to the ureteropelvic junction
8 (4/48)
83 (10/12)
< 0.001a
Feature
Dark on HASTE sequences
21 (10/48)
42 (5/12)
0.153
Homogeneous signal on HASTE sequences
8 (4/48)
83 (10/12)
< 0.001a
Loss of signal on opposed-phase T1-weighted images suggesting intracytoplastic lipid
19 (9/48)
0 (0/12)
0.182
T1 bright areas suggestive of hemorrhage
48 (23/48)
8 (1/12)
0.019a
Hypovascular on all dynamic sequences
25 (12/48)
92 (11/12)
< 0.001a
Bright on high-b-value diffusion-weighted imaging
88 (42/48)
58 (7/12)
0.033a
Note—Except for p values, data are percentage (no./total). aStatistically significant (p < 0.05).
1262
AJR:201, December 2013
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
MRI of Renal Cell Carcinoma and Urothelial Carcinoma
A
B
C
D
Fig. 1—83-year-old woman with macroscopic hematuria and central right renal mass. Endoscopic biopsy of mass was consistent with urothelial carcinoma. A, Coronal HASTE image shows homogeneous hypointense mass (arrow) centered in right renal pelvis. B, Axial contrast-enhanced T1-weighted early nephrographic phase image shows that lesion (arrow) is homogeneously hypovascular. C, Diffusion-weighted image shows uniform increased signal intensity (arrow) in lesion. D, Apparent diffusion coefficient (ADC) map shows uniform low signal intensity (arrow) in lesion (normalized ADC = 0.202).
normalized T2 SI, ADC, and normalized ADC between the two groups. Receiver operating characteristic (ROC) curve analysis was used to assess and compare the overall diagnostic performance of each quantitative metric for differentiation of the two entities; the ROC analysis was used to derive
optimal diagnostic thresholds for each metric, defined as the value that maximizes the average of sensitivity and specificity. All p values are two-sided and are considered statistically significant at p < 0.05. Statistical analysis was performed using software (MedCalc version 10.4, Frank Schoonjans).
Interreader agreement was calculated as the percentage of cases where both readers provided the same initial assessment. A Cohen kappa coefficient was also calculated for each variable, where kappa values were interpreted as follows: 0–0.20, slight agreement; 0.21–0.40, fair agreement; 0.41–0.60,
A
B
C
D
Fig. 2—67-year-old man with mass initially identified on outside CT. Specimen obtained at radical nephrectomy was consistent with clear cell renal cell carcinoma. A, Coronal HASTE image shows hypointense endophytic lesion (arrow) in left upper renal pole. B, Axial contrast-enhanced T1-weighted early nephrographic phase image shows heterogeneous enhancement (arrow) of lesion. C, Diffusion-weighted image shows focus of increased signal intensity (arrow) in lesion. D, Apparent diffusion coefficient (ADC) map shows heterogeneous low signal intensity (arrow) in lesion (normalized ADC = 0.469).
AJR:201, December 2013 1263
Wehrli et al.
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
moderate agreement; 0.61–0.80, substantial agreement; and 0.81–1.00, almost perfect agreement [26].
Results Patient and Lesion Characteristics Patient demographic features and tumor diameters of all cases are summarized in Table 2. There was no significant difference between renal pelvic urothelial carcinoma and central RCC in terms of patient sex (p = 0.318). However, patients with renal pelvic urothelial carcinoma were significantly older than those with central RCC (74.1 ± 15.3 years vs 61.9 ± 13.2 years; p = 0.008). In addition, renal pelvic urothelial carcinoma exhibited a significantly smaller diameter than renal pelvic RCC (29.2 ± 15.4 mm vs 71.9 ± 33.4 mm; p = 0.001). Quantitative MRI Metrics Table 3 provides summary data regarding the performance of the quantitative metrics normalized T2 SI, ADC, and normalized ADC, as well as the results of the ROC analysis. There was significantly lower normalized ADC in renal pelvic urothelial carcinoma compared with central RCC (p = 0.008), but there was no significant difference between these two entities for ADC (p = 0.773) or normalized T2 SI (p = 0.247). At ROC analysis, normalized ADC had the highest area under the ROC curve (AUC) of
these three metrics (AUC = 0.757), which was associated with an optimal threshold of 0.451. A normalized ADC value below this threshold was associated with an 83% sensitivity and 71% specificity in distinguishing renal pelvic urothelial carcinoma and central RCC. The higher AUC of normalized ADC than ADC was nearly significant (p = 0.052); remaining pairwise comparisons between AUCs of these metrics were not significant (p = 0.140–0.758). Subjective Imaging Features Global impression was significantly different between RCC and urothelial carcinoma (p < 0.001), accurately identifying the correct diagnosis in 58 of 60 cases. Subjective features that were significantly more common in urothelial carcinoma included a focal collecting system defect (p < 0.001), central location on the collecting system (p < 0.001), homogeneous signal on HASTE (p < 0.001), absence of cystic or necrotic components (p < 0.001), homogeneous enhancement (p < 0.001), hypovascularity on all contrast-enhanced sequences (p < 0.001), preserved renal shape (p < 0.001), and extension to the ureteropelvic junction (p < 0.001) (Table 4). Subjective features that were significantly more common in RCC included the presence of T1 bright areas suggestive of hemorrhage (p = 0.019) and areas of increased
SI on high-b-value DWI (p = 0.033). Features that showed no significant difference between groups included decreased SI on HASTE (p = 0.153) and the presence of intratumoral lipid (p = 0.182). Characteristic features are presented in Figures 1–3. Interreader Agreement Kappa coefficient analysis showed almost perfect agreement for global impression, defect within the collecting system, preserved renal shape, and hypovascularity; substantial agreement for central location within the collecting system and homogeneous signal on HASTE; and moderate agreement for extension toward the ureteropelvic junction, homogeneous enhancement, and loss of signal on opposed-phase T1. However, the readers agreed in most cases for all qualitative features. Specifically, interreader agreement was 98.3% for global impression and ranged from 61.7% to 93.3% for the remaining subjective features (Table 5). Discussion Because central renal urothelial carcinoma comprises up to 8% of intrarenal neoplasms, noninvasive characterization of renal masses has become a topic of increasing interest among radiologists and urologists to facilitate surgical planning and patient selection for renal mass biopsy. Although urine cytolo-
A
B
C
D
Fig. 3—79-year-old woman with left intrarenal mass for which partial nephrectomy was initially considered. However, nephroureterectomy performed after surgical biopsy revealed high-grade urothelial carcinoma. A, Axial HASTE image shows hypointense endophytic mass (arrow) in left anterior renal midpole. B, Axial contrast-enhanced T1-weighted nephrographic phase shows homogeneous hypoenhancement (arrow) of lesion. C, Diffusion-weighted image shows diffuse increased signal intensity (arrow) in lesion. D, Apparent diffusion coefficient (ADC) map shows homogeneous low signal intensity (arrow) in lesion (normalized ADC = 0.38).
1264
AJR:201, December 2013
MRI of Renal Cell Carcinoma and Urothelial Carcinoma TABLE 5: Interreader Agreement for Subjective Features Interreader Agreementa
κ
Global impression of urothelial carcinoma
98.3 (59/60)
0.952
Centered on renal collecting system
90.0 (54/60)
0.707
Absence of cystic areas or necrosis
66.7 (40/60)
0.378
Defect within the central collecting system
91.7 (55/60)
0.815
Extending to the ureteropelvic junction
93.3 (56/60)
0.514
Preserved renal shape
91.7 (55/60)
0.806
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
Feature
Increased signal on high-b-value diffusion-weighted imaging
61.7 (37/60)
0.092
Decreased signal on HASTE sequences
68.3 (41/60)
0.284
T1 bright areas suggestive of hemorrhage
68.3 (41/60)
0.320
Homogeneous enhancement
80.0 (48/60)
0.514
Homogeneous signal on HASTE sequences
86.7 (52/60)
0.617
Hypovascular on all dynamic sequences
91.7 (55/60)
0.823
Loss of signal on opposed-phase T1-weighted images suggesting intracytoplastic lipid
90.0 (54/60)
0.520
aData are percentage (no./total).
gy is an inexpensive noninvasive screening test with an accuracy of up to 98% for high-grade urothelial carcinoma, sensitivities and specificities as low as 8.5% and 50%, respectively, have been reported for low-grade papillary urothelial carcinoma. In addition, distinguishing between centrally infiltrative urothelial carcinoma of the kidney and central RCC can be a diagnostic challenge on imaging. Nonetheless, differentiation is important because of differences in stage, prognosis, and surgical management of these tumors. There has been an increase in the use of MRI for the detection and characterization of solid renal tumors. One MRI technique that has received particular attention recently for lesion characterization is DWI. Investigators have shown that subjective findings on DWI and quantitative DWI-based metrics (namely, ADCs) may help to reliably distinguish solid renal neoplasms from normal parenchyma, hemorrhagic cysts, and necrotic regions of solid tumors [12–15]. For instance, Paudyal et al. [16] determined that ADC measurements can be helpful for distinguishing between clear cell and non– clear cell subtypes of RCC, and Goyal et al. [27] showed that ADC values correlate with grade of RCC. To our knowledge, ours is the first study to investigate the utility of ADC for distinguishing renal pelvic urothelial carcinoma from central RCC. In our study, we investigated the utility of standard ADC values and ADC values normalized to CSF, a technique motivated to help offset variability between examinations relating to differences in scanners and
sequence parameters, as well as possible biologic effects contributing to variability [24, 25]. Although ADC was not significantly different between the two groups, normalized ADC was significantly lower in urothelial carcinoma compared with RCC. Furthermore, normalized T2 SI measurements showed no significant difference between urothelial carcinoma and RCC. It is possible that the greater utility of the ADC-based metric relates to the stronger association between ADC and tumor cellularity, whereas T2-weighted SI measurements may be more strongly confounded by a host of additional properties, including tumor edema, inflammation, and presence of collagenous or fibrous tissue. Thus, according to our data, normalized ADC seems to be the most useful of the investigated quantitative metrics for distinguishing the two diagnoses. The role of ADC normalization is of particular note given conflicting results of the value of ADC normalization in past studies in other tissues [24, 25]. In addition to quantitative metrics, we assessed various additional features related to tumor morphologic features, signal characteristics, and enhancement patterns. Similar to the study of Raza et al. [11], which explored the CT features of centrally infiltrative urothelial carcinoma versus central RCC, global impression was the most reliable subjective feature in our series. The six additional most helpful features in the study by Raza et al., including tumor center within the collecting system, filling defect within the renal pelvis, renal shape preservation, absence of cystic or necrotic components, ho-
mogeneous enhancement, and extension toward the ureteropelvic junction, showed a similarly strong association with urothelial carcinoma in our series. The study by Raza et al. [11] did not assess the degree of lesion enhancement on CT. It is recognized that urothelial carcinoma is low in attenuation on contrast-enhanced CT [28, 29], whereas clear cell RCC, representing the most common RCC subtype, is hypervascular. (Although papillary RCC is generally hypovascular, this subtype only accounts for about 10–20% of RCCs [13, 30, 31].) In our study, we compared subjective enhancement between diagnoses and observed that urothelial carcinoma was hypovascular relative to renal parenchyma on all dynamic contrastenhanced phases on MRI at a significantly greater frequency than RCC (p < 0.001). An additional feature that showed a significantly higher frequency in urothelial carcinoma in our series was the absence of cystic or necrotic components. The study by Raza et al. [11] described this feature as difficult to assess on CT. Greater utility of this feature on MRI may relate to the greater contrast resolution and multitude of contrast mechanisms available for tissue characterization with MRI. The absence of cystic or necrotic components in urothelial carcinoma likely contributes to its typical homogeneous SI on HASTE images, which we observed with significantly higher frequency in urothelial carcinoma. An MRI feature that was much more frequent in RCC than urothelial carcinoma was the presence of high-SI areas on unenhanced T1-weighted images, a finding suggestive of
AJR:201, December 2013 1265
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
Wehrli et al. intratumoral hemorrhage. It is recognized that RCCs may outgrow their blood supply as they increase in size, developing hemorrhage into areas of necrotic tumor [32]. Papillary RCC is particularly known to contain blood products of varying chronicity on pathologic assessment [33, 34]. On the other hand, intratumoral hemorrhage was observed in only one of 12 cases of urothelial carcinoma in our series. An interesting observation in our study was that, although increased SI on high-bvalue DWI was subjectively present more often in RCC than urothelial carcinoma, there was significantly lower measured normalized ADC in urothelial carcinoma than in RCC. This may relate to alternate causes for increased SI on high-b-value DWI, aside from diffusion restriction related to tumor cellularity. For instance, the increased SI on DWI in RCC may relate to T2 shine-through effect [35] from tumor necrosis, a common feature of clear cell carcinoma, rather than true diffusion restriction from densely cellular tumor. However, the lower measured normalized ADC in urothelial carcinoma likely represents true diffusion restriction in this group. In addition, we note that increased signal on high-b-value DWI had the lowest interreader agreement of any of the included subjective imaging features of only 61.7% (κ = 0.092), further raising concerns about the utility of this subjective assessment of DWI. There are several limitations to our study. First, the study was retrospective. Second, the sample size of renal pelvic urothelial carcinoma was small. However, this represents the natural low incidence of this entity. Also, we did not assess other entities that may present as a central renal mass, such as lymphoma or a metastasis to the kidney. Finally, DWI was performed using a maximal b value of 800 s/mm2 in all cases, consistent with our institutional protocol for MRI of the urinary tract. It is possible that the diagnostic performance of DWI and ADC metrics would be different in the setting of an alternate b-value acquisition scheme. In conclusion, we have presented the largest series to our knowledge to explore differences in MRI features between renal pelvic urothelial carcinoma and central RCC. Normalized ADC values, but not ADC or normalized T2 SI, were significantly different between these entities. In addition to morphologic features previously characterized on CT, urothelial carcinoma was significantly more likely to exhibit homogeneous MRI
1266
SI, absence of cystic or necrotic areas, and hypovascularity on contrast-enhanced MRI sequences, whereas RCC was significantly more likely to exhibit MRI SI characteristics suggestive of hemorrhage. Prospective differentiation of these two entities, based on cross-sectional imaging characteristics, may assist in decisions regarding biopsy and surgical management. References 1. Wagle DG, Moore RH, Murphy GP. Primary carcinoma of the renal pelvis. Cancer 1974; 33:1642– 1648 2. Nocks BN, Heney NM, Daly JJ, Perrone TA, Griffin PP, Prout GR Jr. Transitional cell carcinoma of renal pelvis. Urology 1982; 19:472–477 3. Wong-You-Cheong JJ, Wagner BJ, Davis CJ Jr. Transitional cell carcinoma of the urinary tract: radiologic-pathologic correlation. RadioGraphics 1998; 18:123–142; quiz, 148 4. Guinan P, Volgelzang NJ, Randazzo R, et al. Renal pelvic transitional cell carcinoma: the role of the kidney in tumor-node-metastasis staging. Cancer 1992; 69:1773–1775 5. Baron RL, McClennan BL, Lee JK, Lawson TL. Computed tomography of transitional-cell carcinoma of the renal pelvis and ureter. Radiology 1982; 144:125–130 6. Guinan P, Vogelzang NJ, Randazzo R, et al. Renal pelvic cancer: a review of 611 patients treated in Illinois 1975–1985. Cancer Incidence and End Results Committee. Urology 1992; 40:393–399 7. Bastacky S, Ibrahim S, Wilczynski SP, Murphy WM. The accuracy of urinary cytology in daily practice. Cancer 1999; 87:118–128 8. Raab SS, Grzybicki DM, Vrbin CM, Geisinger KR. Urine cytology discrepancies: frequency, causes, and outcomes. Am J Clin Pathol 2007; 127:946–953 9. Rouprêt M, Zigeuner R, Palou J, et al. European guidelines for the diagnosis and management of upper urinary tract urothelial cell carcinomas: 2011 update. European Association of Urology Guideline Group for urothelial cell carcinoma of the upper urinary tract (in Spanish). Actas Urol Esp 2012; 36:2–14 10. Schrader AJ, Steffens S. Renal cell carcinoma update: news from the AUA, EAU, and ASCO annual meetings 2011. ISRN Urol 2012; 2012:748235 11. Raza SA, Sohaib SA, Sahdev A, et al. Centrally infiltrating renal masses on CT: differentiating intrarenal transitional cell carcinoma from centrally located renal cell carcinoma. AJR 2012; 198:846–853 12. Kim S, Jain M, Harris AB, et al. T1 hyperintense renal lesions: characterization with diffusionweighted MR imaging versus contrast-enhanced MR imaging. Radiology 2009; 251:796–807
13. Taouli B, Thakur RK, Mannelli L, et al. Renal lesions: characterization with diffusion-weighted imaging versus contrast-enhanced MR imaging. Radiology 2009; 251:398–407 14. Squillaci E, Manenti G, Di Stefano F, Miano R, Strigari L, Simonetti G. Diffusion-weighted MR imaging in the evaluation of renal tumours. J Exp Clin Cancer Res 2004; 23:39–45 15. Zhang J, Tehrani YM, Wang L, Ishill NM, Schwartz LH, Hricak H. Renal masses: characterization with diffusion-weighted MR imaging—a preliminary experience. Radiology 2008; 247:458–464 16. Paudyal B, Paudyal P, Tsushima Y, et al. The role of the ADC value in the characterisation of renal carcinoma by diffusion-weighted MRI. Br J Radiol 2010; 83:336–343 17. Kim S, Naik M, Sigmund E, Taouli B. Diffusionweighted MR imaging of the kidneys and the urinary tract. Magn Reson Imaging Clin N Am 2008; 16:585–596; vii–viii 18. Yoshida S, Masuda H, Ishii C, Saito K, Kawakami S, Kihara K. Initial experience of functional imaging of upper urinary tract neoplasm by diffusion-weighted magnetic resonance imaging. Int J Urol 2008; 15:140–143 19. Dodelzon K, Mussi TC, Babb JS, Taneja SS, Rosenkrantz AB. Prediction of growth rate of solid renal masses: utility of MR imaging features—preliminary experience. Radiology 2012; 262:884–893 20. Rosenkrantz AB, Hindman N, Fitzgerald EF, Niver BE, Melamed J, Babb JS. MRI features of renal oncocytoma and chromophobe renal cell carcinoma. AJR 2010; 195:[web]W421–W427 21. Rosenkrantz AB, Niver BE, Fitzgerald EF, Babb JS, Chandarana H, Melamed J. Utility of the apparent diffusion coefficient for distinguishing clear cell renal cell carcinoma of low and high nuclear grade. AJR 2010; 195:[web]W344–W351 22. Choi HJ, Kim JK, Ahn H, Kim CS, Kim MH, Cho KS. Value of T2-weighted MR imaging in differentiating low-fat renal angiomyolipomas from other renal tumors. Acta Radiol 2011; 52:349–353 23. Hindman N, Ngo L, Genega EM, et al. Angiomyolipoma with minimal fat: can it be differentiated from clear cell renal cell carcinoma by using standard MR techniques? Radiology 2012; 265:468–477 24. Sasaki M, Yamada K, Watanabe Y, et al. Variability in absolute apparent diffusion coefficient values across different platforms may be substantial: a multivendor, multi-institutional comparison study. Radiology 2008; 249:624–630 25. Rosenkrantz AB, Kopec M, Kong X, et al. Prostate cancer vs. post-biopsy hemorrhage: diagnosis with T2- and diffusion-weighted imaging. J Magn Reson Imaging 2010; 31:1387–1394 26. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics
AJR:201, December 2013
Downloaded from www.ajronline.org by University of Melbourne-Faculties on 10/14/14 from IP address 128.250.144.144. Copyright ARRS. For personal use only; all rights reserved
MRI of Renal Cell Carcinoma and Urothelial Carcinoma 1977; 33:159–174 27. Goyal A, Sharma R, Bhalla AS, et al. Diffusionweighted MRI in renal cell carcinoma: a surrogate marker for predicting nuclear grade and histological subtype. Acta Radiol 2012; 53:349– 358 28. Bree RL, Schultz SR, Hayes R. Large infiltrating renal transitional cell carcinomas: CT and ultrasound features. J Comput Assist Tomogr 1990; 14:381–385 29. Browne RF, Meehan CP, Colville J, Power R, Torreggiani WC. Transitional cell carcinoma of the upper urinary tract: spectrum of imaging find-
ings. RadioGraphics 2005; 25:1609–1627 30. Herts BR, Coll DM, Novick AC, et al. Enhancement characteristics of papillary renal neoplasms revealed on triphasic helical CT of the kidneys. AJR 2002; 178:367–372 31. Rosen MA, Schnall MD. Dynamic contrast-enhanced magnetic resonance imaging for assessing tumor vascularity and vascular effects of targeted therapies in renal cell carcinoma. Clin Cancer Res 2007; 13:770s–776s 32. Sengupta S, Lohse CM, Leibovich BC, et al. Histologic coagulative tumor necrosis as a prognostic indicator of renal cell carcinoma aggressiveness.
Cancer 2005; 104:511–520 33. Prasad SR, Humphrey PA, Catena JR, et al. Common and uncommon histologic subtypes of renal cell carcinoma: imaging spectrum with pathologic correlation. RadioGraphics 2006; 26:1795– 1806; discussion, 1806–1710 34. Pedrosa I, Sun MR, Spencer M, et al. MR imaging of renal masses: correlation with findings at surgery and pathologic analysis. RadioGraphics 2008; 28:985–1003 35. Koh DM, Collins DJ. Diffusion-weighted MRI in the body: applications and challenges in oncology. AJR 2007; 188:1622–1635
F O R YO U R I N F O R M AT I O N
Unique customized medical search engine service from ARRS! ARRS GoldMiner ® is a keyword- and concept-driven search engine that provides instant access to radiologic images published in peer-reviewed journals. For more information, visit goldminer.arrs.org.
AJR:201, December 2013 1267
