| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Genitourinary Imaging |
1 From the Departments of Radiology (J.K.K., S.Y.P., K.S.C.) and Urology (C.S.K., H.J.A.), Asan Medical Center, University of Ulsan, 388–1 Poongnap-dong, Songpa-gu, Seoul 138–736, Korea. Received October 1, 2002; revision requested December 12; final revision received September 24, 2003; accepted October 28. Address correspondence to K.S.C. (e-mail: kscho@amc.seoul.kr).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: In 20 patients, the attenuation value of bladder cancer was measured on dynamic contrast material–enhanced multiphasic CT images obtained with scanning delays of 40, 60, 80, and 100 seconds. In 67 patients, CT data were obtained with a 60-second scanning delay that covered the bladder (section thickness, 2.5 mm; beam pitch, 1.5) and a 180-second scanning delay that covered the abdomen (section thickness, 5 mm; beam pitch, 1.5). We prospectively evaluated CT images and compared findings at CT with findings at histologic examination. We evaluated cancer detection rate, positive predictive value of cancer detection, and sensitivity and specificity in the diagnosis of perivesical invasion.
RESULTS: The attenuation value of bladder cancers was significantly higher on 60- (105 HU ± 16) and 80-second (97 HU ± 15) delayed CT images than on the other images (P < .05). The cancer detection rate and positive predictive value for cancer detection were 97% and 95%, respectively, in 67 patients and increased to 100% and 100%, respectively, in 44 patients with a time interval of 7 or more days between transurethral resection of the bladder (TURB) and CT examination. Sensitivity and specificity in the diagnosis of perivesical invasion were 89% and 95%, respectively, in 67 patients and increased to 92% and 98%, respectively, in 44 patients with a time interval of 7 or more days between TURB and CT examination.
CONCLUSION: Bladder cancer tends to show peak enhancement with the 60-second scanning delay. Multi–detector row helical CT is useful in the detection and staging of bladder cancer.
© RSNA, 2004
Index terms: Bladder neoplasms, 83.32 • Bladder neoplasms, CT, 83.12115, 83.12112 • Computed tomography (CT), contrast enhancement, 83.12112 • Computed tomography (CT), multi–detector row, 83.12115
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Computed tomography (CT) and magnetic resonance (MR) imaging are the main radiologic examinations used in the evaluation of patients with bladder cancer. There is still controversy about which imaging modality is better. The advantages of CT include shorter acquisition time, wider coverage in a single breath hold, and lower susceptibility to various patient factors. On the other hand, CT is limited in the detection of small bladder cancers, and its staging accuracy varies from 64% to 92%, according to various authors (1–8). Moreover, to our knowledge, the enhancement pattern of bladder cancer concerning peak enhancement time and degree of enhancement on contrast material–enhanced CT images has not been analyzed. It is known, however, that bladder cancers usually enhance more intensely than adjacent normal bladder wall tissue (1).
Recent improvements in CT hardware have led to the development of the multi–detector row helical CT scanner, which can provide higher resolution and more compact volume acquisition in a shorter time. Because of these advantages, multi–detector row helical CT is anticipated to improve the evaluation of patients with bladder cancer. Thus, the purpose of our study was to evaluate the enhancement pattern of bladder cancer and the accuracy of multi–detector row helical CT in the detection and staging of bladder cancer.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Analysis of the Enhancement Pattern of Bladder Cancer
The enhancement pattern of bladder cancer was evaluated in the first 20 patients (16 men and four women; mean age, 60 years; age range, 46–70 years) with bladder cancers larger than 1.5 cm in the short diameter, as measured by a radiologist (J.K.K.). In these patients, bladder lesions were initially detected with a cystoscopic examination, and CT was performed 5–7 days later. Transurethral resection of the bladder (TURB) was subsequently performed 2–5 days after CT. In seven of 20 patients, radical cystectomy was performed because histologic examination of the TURB specimen revealed muscular invasion of bladder cancer (time interval between TURB and radical cystectomy, 7–17 days). All lesions depicted at both cystoscopic and CT examination were resected during TURB or radical cystectomy, and histologic examination confirmed the presence of transitional cell carcinoma.
Evaluation of Diagnostic Accuracy of CT in Detection of Bladder Cancer
In 67 other patients (51 men and 16 women; mean age, 63 years; age range, 35–75 years), we evaluated the diagnostic accuracy of multi–detector row helical CT in the detection and staging of bladder cancer. All of these patients had bladder lesions that were found initially at cystoscopy, and each patient underwent TURB and CT thereafter. All patients underwent radical cystectomy, either because histologic examination of the TURB specimens confirmed muscular invasion of bladder cancer (n = 60) or because bladder cancers were too large to be removed completely with TURB, regardless of the depth of invasion (n = 7). In this patient group, TURB preceded CT by 1–31 days (mean, 11 days). The time interval between TURB and CT was less than 7 days in 23 patients and 7–31 days in the remaining 44 patients. Radical cystectomy was performed within 2 weeks (mean, 9 days) of CT.
CT Examination
All CT data were obtained by using a four-channel multi–detector row helical CT scanner (LightSpeed QX/i; GE Medical Systems, Milwaukee, Wis). Patients were instructed not to void for at least 2 hours before the examination. All patients received 600–900 mL of oral 2% barium sulfate suspension (E-Z CAT; E-Z-Em, Westbury, NY) 1 hour prior to CT scanning. A dose that ranged from a minimum of 2 mL/kg to a maximum of 160 mL/kg of nonionic intravenous iopromide (Ultravist; Schering, Berlin, Germany) was administered at a rate of 4 mL/sec with a power injector.
In the first 20 patients who underwent CT scanning, we obtained dynamic contrast-enhanced multiphasic CT images of bladder cancers. Scanning parameters included a multi–detector array of 2.5 x 4.0 mm, a beam pitch of 1.5 (equivalent to a section pitch of 6, high-speed mode), a gantry rotation speed of 0.8 second, an image reconstruction interval of 2.5 mm, and a voltage of 120 kVp. To reduce the radiation dose in this patient group, we used a tube current of 150 mA. First, an unenhanced CT scan was acquired over the entire bladder. When compared with the cystoscopic findings, bladder cancers were larger than 1.5 cm in the short diameter on these CT images, which allowed a sufficient area for a region of interest (ROI) to be established. Thereafter, contrast-enhanced CT images were obtained with scanning delays of 40, 60, 80, and 100 seconds. With scans obtained in every phase, only three contiguous images were obtained in the center of bladder cancers that were found on unenhanced CT images; consequently, a total of 12 contrast-enhanced CT images were obtained in each patient.
As a result of findings at dynamic contrast-enhanced multiphasic CT, in the other 67 patients, we performed two-phase CT scanning, which consisted of a 60-second delayed scan that covered the entire bladder and a 180-second delayed scan that covered the area from the diaphragm to the level from which the prior 60-second delayed scan was initiated. The parameters used for 60-second delayed scanning were the same as those used for multiphasic scanning in the prior 20 patients, except the tube current was increased to 200 mA. The parameters for the 180-second delayed scan included a multi–detector array of 5 x 4 mm, a beam pitch of 1.5, a gantry rotation speed of 0.8 second, an image reconstruction interval of 5 mm, a voltage of 120 kVp, and a tube current of 200 mA.
Image Analysis
CT images were displayed on a picture archiving and communication system (Radpia; Hyundai Information & Technology, Seoul, Korea) that made it possible to measure the size of lesions and the attenuation value (measured in Hounsfield units) in a particular ROI.
In the first 20 patients, two radiologists (J.K.K., S.Y.P.) evaluated the attenuation value of bladder cancers larger than 1.5 cm in the short diameter on the 40-, 60-, 80-, and 100-second delayed CT images. For measuring the attenuation value, a round or elliptical ROI was placed over the bladder cancer by the first radiologist (J.K.K.) and was consistent in location on all CT images; the location of the ROI was determined with consensus between the two radiologists. We attempted to cover the bladder cancer as much as possible within the ROI. Care was taken to exclude the surrounding urine or fat from the ROI. The size of the ROI was 1.8–5.4 cm2.
The two radiologists knew that the other 67 patients had undergone TURB but were unaware of the cystoscopic findings. Two reviewers prospectively evaluated CT images in consensus, with regard to tumor detection and staging. The 60-second delayed CT images were evaluated for cancer detection, staging, and pelvic lymphadenopathy, and 180-second delayed CT images were evaluated for the presence or absence of upper urinary tract abnormalities and metastases to abdominal organs or retroperitoneum.
On the 60-second delayed CT images in 67 patients, bladder cancers appeared as masses that protruded into the bladder lumen or wall thickening with greater enhancement than the adjacent bladder wall. Wall thickening without enhancement was not considered to be bladder cancer but residual inflammation after TURB. We documented the location and number of bladder cancers identified on CT images.
For all presumed bladder cancers on CT images, local stage was classified according to the TNM system (1). Because differentiation in patients with T1–T3a cancers is difficult on CT images, we only attempted to determine the presence or absence of perivesical invasion (ie, T3b or higher vs T3a or lower). Perivesical invasion was considered to be present when the interface between bladder cancer and perivesical fat was irregular or when bladder cancers showed overt growth beyond the outer margin of the bladder wall. Lymph nodes were considered to be abnormal if the short diameter was equal to or larger than 1 cm. We compared the results of CT with those of histologic examination in the specimens obtained at radical cystectomy.
Statistical Analysis
In the first 20 patients, repeated measures analysis of variance with pairwise multiple comparisons with the Tukey test were used for the comparison of the attenuation value of bladder cancers in the 40-, 60-, 80-, and 100-second delayed CT images.
In the other 67 patients, the first radiologist (K.J.K.) matched lesions on CT images to those on gross photographs of radical cystectomy specimens. Then, we evaluated the cancer detection rate and positive predictive value for cancer detection in each lesion. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for the diagnosis of perivesical invasion were calculated by using the histologic findings as the reference standard.
To evaluate the effect of TURB on cancer detection and staging, we divided the 67 patients into two groups: (a) patients with a time interval of less than 7 days between TURB and CT examination (n = 23) and (b) patients with a time interval of 7 days or more (n = 44). Thereafter, we compared the accuracy of CT in cancer detection and staging between the two patient groups by using the Fisher exact test after adjusting for the effect of clustering according to Gonen et al (9). In every analysis, statistical significance was considered to be present when the P value was less than .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
|
|
|
|
|
Table 2 demonstrates the data obtained with dedicated CT for cancer detection in 67 patients who underwent radical cystectomy. On the 60-second delayed CT images, two reviewers identified 79 lesions in 59 patients; 75 lesions were proved to be true-positive, and four were proved to be false-positive. In the remaining eight patients, reviewers could not find any lesion because the entire bladder wall showed even thickness without focus of strong enhancement. On CT images, two bladder cancers were missed in two patients; both lesions were less than 1 cm in the maximal diameter (0.3 cm and 0.5 cm). Thus, the detection rate was 97% (75 of 77) for all bladder cancers and 85% (11 of 13) for bladder cancers shorter than 1 cm in the maximal diameter (Fig 2). All four false-positive lesions were diagnosed in patients who did not have residual bladder cancer (Fig 3). In these patients, we diagnosed strongly enhancing linear lesions as bladder cancers, but they were diagnosed as local scar tissue and inflammation at histologic examination.
|
|
|
Table 3 demonstrates the data from dedicated CT for the diagnosis of perivesical invasion in 67 patients with bladder cancer at histologic analysis. In the 75 bladder cancers depicted on the 60-second delayed CT images, histologic examination demonstrated that 18 cancers invaded perivesical fat (Fig 4), while the others were confined to the bladder wall. Findings at CT and histologic examination agreed on 70 bladder cancers, including 16 bladder cancers with perivesical invasion and 54 without. In the five bladder cancers in which CT and histologic findings disagreed, two cancers with perivesical invasion were understaged as being confined to the bladder wall, and three cancers without perivesical invasion were overstaged as having invaded the perivesical fat (Fig 5). Thus, the sensitivity, specificity, and overall accuracy for the diagnosis of perivesical invasion were 89%, 95%, and 93%, respectively.
|
|
|
Histologic examination revealed six pelvic lymph nodes with metastasis in five patients; four of these nodes were demonstrated on the CT images. The other two metastatic pelvic lymph nodes were not considered to be abnormal on CT images because they were smaller than 1 cm in the short diameter. There were no false-positive findings at CT of lymph nodes. Because of the small number of patients with lymph node metastasis, statistical analysis was not performed.
On abdominal CT images obtained with the 180-second scanning delay, there was no evidence of metastasis in the upper urinary tract, abdominal organs, or retroperitoneum.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Among previous reports in which the diagnostic accuracy of CT for detection and staging of bladder cancer was evaluated, we believe the report of Kim et al (8) showed the best results: The cancer detection rate was 93%, and the accuracy for diagnosis of perivesical invasion was 83%. In comparison with their report, our results showed slightly higher accuracy, as the cancer detection rate was 97% and the accuracy for diagnosing perivesical invasion was 93%. Furthermore, the cancer detection rate and the accuracy for diagnosis of perivesical invasion could rise to 100% and 96%, respectively, in patients with a time interval of 7 or more days between TURB and CT. The detection rate for bladder cancers smaller than 1 cm in diameter was also higher in our study (85%) than in the study of Tanimoto et al (53%) (7). Thus, dedicated multi–detector row helical CT seems to improve the performance of CT in the evaluation of patients with bladder cancer.
There is still confusion and controversy about whether CT or MR imaging is better for evaluation of bladder cancer. Although some authors have demonstrated that MR imaging is more accurate than CT in the detection and staging of bladder cancer (10–12), Husband et al (5) argued against the superiority of MR imaging and demonstrated that MR imaging was slightly inferior to CT. In the most recent comparative study between CT and MR imaging that we found, Kim et al (8) showed that dynamic gadolinium-enhanced MR imaging was slightly more accurate than CT and other MR techniques, although the difference was not statistically significant. According to their results with dynamic gadolinium-enhanced MR imaging, the cancer detection rate was 100%, and the sensitivity, specificity, and overall accuracy for diagnosing perivesical invasion were 100%, 73%, and 93%, respectively. Thus, our study shows slightly lower sensitivity (89%) and higher specificity (95%) for the diagnosis of perivesical invasion, whereas the cancer detection rate and overall accuracy for perivesical invasion are similar. We suggest that these differences may result from the different imaging properties of CT and MR imaging. In general, MR imaging is superior to CT in soft-tissue resolution and can be used to better visualize perivesical fat invasion. On the other hand, even subtle inflammation around bladder cancers may easily mimic perivesical invasion at MR imaging, thereby resulting in overstaging.
The interference of TURB on the evaluation of bladder cancer at CT or MR imaging has been commented on by many researchers (8,13–15). In fact, it is hardly possible to distinguish bladder cancer from TURB-associated inflammation. In our study, TURB also negatively affected cancer detection and diagnosis of perivesical invasion; however, our data shows that perivesical changes or bladder wall thickening caused by TURB may reduce enough so as not to confuse cancer detection when CT images are obtained a considerable time after TURB.
In our study, 10 patients were found to be free of bladder cancer at histologic examination of the radical cystectomy specimens because the cancers were completely removed at TURB. In those patients, all four pseudolesions were demonstrated in patients with a time interval of less than 7 days. In contrast, in the six patients with a time interval of 7 or more days, CT correctly demonstrated no residual bladder cancer. These data also indicate that an adequate interval between TURB and CT examination may improve the accuracy of CT in the follow-up of patients who have undergone TURB.
A major limitation of our study is a bias in the patient population. Advanced bladder cancers had already been diagnosed in 67 patients at cystoscopy or TURB, and these patients were scheduled to undergo radical cystectomy. We did not include patients with superficial, small, or lower-grade bladder cancers that could be treated only with TURB or those who initially had no bladder cancer at cystoscopy. Thus, our results for CT in the detection of bladder cancer may be biased and cannot provide the value of CT in the screening of patients with bladder cancer. Considering this issue, a recent study with multi–detector row CT urography has shown that eight of 10 patients with focal or asymmetric bladder wall thickening at CT urography had bladder cancers (16).
To reduce the radiation dose in the first 20 patients who underwent CT for evaluation of the enhancement pattern of bladder cancer, we applied a low tube current (150 mA) and a high-speed mode (beam pitch of 1.5) and obtained only three images in each phase of contrast-enhanced scanning. A criticism concerning an increased radiation dose due to rescanning the same region cannot be avoided, because repeated scanning of the same area may increase the radiation dose two- or threefold (17).
In this study, we attempted to use the stronger enhancement of bladder cancer and its morphology to differentiate bladder cancer from adjacent normal bladder wall or surrounding connective tissue. There may be some doubts with regard to whether a stronger-enhancing area exactly corresponds to the extent of the bladder cancer, however, because we did not perform a direct histologic correlation. Furthermore, in patients with large bladder cancers that may appear heterogeneously enhanced, it becomes more difficult to distinguish cancerous tissues from surrounding structures by using the different degree of enhancement.
In summary, bladder cancers strongly enhance, particularly around the 60-second scanning delay, and dedicated multi–detector row helical CT can provide a high cancer detection rate and accuracy in the diagnosis of perivesical invasion. Although TURB is still likely to result in pseudolesions and inaccurate staging, multi–detector row helical CT can improve the performance of CT in the evaluation of patients with bladder cancer.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, J.K.K., K.S.C.; study concepts and design, J.K.K., K.S.C.; literature research, J.K.K., S.Y.P.; clinical studies, H.J.A., C.S.K.; data acquisition and analysis/interpretation, J.K.K., S.Y.P.; statistical analysis, J.K.K.; manuscript preparation, J.K.K.; manuscript definition of intellectual content and editing, J.K.K., K.S.C.; manuscript revision/review, K.S.C.; manuscript final version approval, J.K.K., K.S.C.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. B. Park, J. K. Kim, H. J. Lee, H. J. Choi, and K.-S. Cho Hematuria: Portal Venous Phase Multi Detector Row CT of the Bladder A Prospective Study Radiology, December 1, 2007; 245(3): 798 - 805. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Jinzaki, A. Tanimoto, H. Shinmoto, Y. Horiguchi, K. Sato, S. Kuribayashi, and S. G. Silverman Detection of Bladder Tumors with Dynamic Contrast-Enhanced MDCT Am. J. Roentgenol., April 1, 2007; 188(4): 913 - 918. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. J. Wong-You-Cheong, P. J. Woodward, M. A. Manning, and I. A. Sesterhenn From the Archives of the AFIP: Neoplasms of the Urinary Bladder: Radiologic-Pathologic Correlation RadioGraphics, March 1, 2006; 26(2): 553 - 580. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. J. Matta, A. J. Kenney, G. M. Barre, and R. M. Vanlangendonck Jr Best Cases from the AFIP: Intradiverticular Bladder Carcinoma RadioGraphics, September 1, 2005; 25(5): 1397 - 1403. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
