Accuracy of Computed Tomography for Identifying Locally Advanced Disease in Patients with Muscle-Invasive Bladder Cancer


Introduction: To retrospectively evaluate the utility of computed tomography (CT) scanning in identifying patients with locally advanced bladder cancer.

Methods: We performed an Institutional Review Board-approved review of 858 patients that underwent radical cystectomy (RC) from 2000 to 2008 at our institution. We selected patients with muscle-invasive bladder cancer (MIBC) who underwent up-front RC without neoadjuvant chemotherapy and who were assessed by preoperative CT scan. We limited this analysis to 48 CT scans obtained prior to transurethral resection. All CT scans were blinded and retrospectively re-read by a dedicated genitourinary radiologist (CSN) to identify tumor location, the presence of wall thickening, and evidence of extravesical disease (stranding or nodularity) or lymph node metastases. These radiologic findings were compared with pathologic findings.

Results: Pretransurethral resection CT scans were able to accurately identify tumor location in 66.7% of patients (sensitivity = 88.9%, specificity = 33.3%) while lymph node assessment was accurate in 58.3% (sensitivity = 75%, specificity = 62.5%). However, only 16.7% of patients with pathologic T3b disease were actually identified on CT as having radiologic evidence of extravesical disease. Specific radiologic signs suggestive of local disease extension, such as wall thickening, stranding, and nodularity correlated poorly with true pathologic T3b disease.

Conclusion: CT scanning has limits in its ability to accurately identify extravesical disease and lymph node spread in patients with MIBC. Investigations into additional or alternative means of clinical staging for bladder cancer patients are incredibly crucial.

Rian J. Dickstein, Chaan S. Ng, Colin P. Dinney, Ashish M. Kamat,

Departments of Urology and Radiology, the University of Texas, M. D. Anderson Cancer Center, Houston, Texas, United States

Submitted May 29, 2013 - Accepted for Publication August 14, 2013

KEYWORDS: Urinary bladder neoplasms, X-ray computed tomography, neoplasm staging

CORRESPONDENCE: Ashish M. Kamat, Department of Urology, Unit 1373, the University of Texas, M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas 77030, United States ()

CITATION: UroToday Int J. 2013 October;6(5):art 55.



The survival of patients with bladder cancer varies greatly based on the stage of their disease [1]. Unfortunately, physicians are notoriously poor at staging patients with invasive disease [2,3,4,5,6,7,8]. It remains critically important to be able to identify those patients with locally advanced disease (such as extravesical extension of tumors or clinically node-positive disease) in order to appropriately prognosticate survival and to identify patients who would gain the most benefit from up-front chemotherapy.

Current staging of bladder tumors is based on visual assessment of the tumor on cystoscopy, pathologic assessment via transurethral resection (TUR) specimens, exam under anesthesia (EUA) to assess for evidence of a 3-dimensional mass, and computed tomography (CT) imaging. Both older and newer studies have reported that CT imaging is limited by its poor accuracy in detecting local tumor extension and lymph node metastases [9,10]. Here we sought to further evaluate the ability of CT imaging to identify locally advanced disease in a rather homogeneous population of patients with cT2 bladder cancer by comparing a blinded review of preoperative imaging to pathologic features.


Using an Institutional Review Board-approved protocol, we queried our prospectively collected database and identified 858 consecutive patients who underwent radical cystectomy (RC) for urothelial carcinoma of the bladder from January 2000 to December 2008. We isolated patients designated having cT2N0M0 disease following their TUR and EUA; those patients with a palpable 3-dimensional mass on EUA and those who received chemotherapy prior to surgery were excluded in order to study a homogeneous cohort. Patients who underwent imaging immediately following TUR (< 30 days) were also excluded in order to avoid bias from confounding extravesical inflammation or extravasation commonly seen after resection.

All CT images were then de-identified and reviewed by a dedicated genitourinary radiologist (CSN), who was blinded to the pathologic features and clinical outcome. The CT images were acquired from various institutions, using a variety of helical and multidetector scanners, and thus there was no uniform protocol that was used. Axial CT images of 2.5 mm to 10 mm sections were reviewed. Endpoints of interest included tumor location, radiographic suspicion of locally advanced disease, and individual radiographic signs potentially predictive of extravesical disease or lymph node metastases.

The presence or absence and location of focal bladder wall thickening greater than that of the adjacent bladder wall, intraluminal bladder wall nodules/nodularity, and/or focal post-contrast enhancement of the bladder wall was documented, and the bladder location recorded and taken to represent the site of tumor(s). Local extension of disease was considered if there was stranding and/or nodularity in the adjacent perivesical fat. For this study, pelvic lymph nodes were considered to be positive for disease if ≥ 0.5 cm in the short axis.

We then compared the radiologic impression and findings to the pathologic results. Summary statistics were used to describe demographical and clinical characteristics of the study population. Basic statistical measures were used to decipher the accuracy of CT imaging in detecting locally advanced bladder cancer.


Using our stringent selection criteria for this study, we identified 48 patients with complete clinical and pathologic data who also had a pre-TUR CT scan available for review. For these patients, the median time from CT to TUR was only 3 days and from TUR to RC was 42. Despite this selection, 45.8% of our cT2 patients were found to be upstaged to pT3 disease pathologically and 16.7% were node positive (Table 1).

None of the patients were originally thought to have CT findings consistent with advanced disease, but for the purposes of this study the CT scans were re-evaluated for radiologic findings consistent with extravesical disease (i.e., ≥ rT3b) or nodal metastases (rN+). The repeat radiographic impressions indicated that 14.6% of patients had rT3b disease and 31.3% rN+ (Table 1). However, only 2 of 12 patients (16.7%) with pathologic T3b disease were correctly identified as having radiologic signs suspicious for extravesical disease. Furthermore, 5 of 7 patients (71.4%) thought to have rT3b disease and 11 of 15 (73.3%) thought to have rN+ disease were overstaged. Finally, 12 of 29 patients (41.4%) thought to have organ-confined disease (< rT3b and rN0) were understaged.

EUA is an underutilized but important part of the staging evaluation for muscle-invasive bladder cancer (MIBC). However, the accuracy of EUA can vary based on tumor location and body habitus, with poor sensitivity in obese patients. Here we found that EUA was unable to identify macroscopic tumor extension in 12 of the 48 patients (25%). However, tumor location was accurately identified by CT in 66.7% of patients (sensitivity = 88.9%, specificity = 33.3%), and lymph node assessment was correct in 58.3% (sensitivity = 75%, specificity = 62.5%) (Table 2).

We then evaluated the correlation of extravesical extension with bladder wall thickening, extravesical stranding, and extravesical nodularity as individual variables. As shown in Table 3, these criteria were not sufficient to accurately identify extravesical disease as only 28.2, 27.6, and 28.6% of patients with pT3b disease were found to have these radiographic findings, respectively. Furthermore, 28 patients had at least pT2 disease; among these, 11 of 23 (47.8%) with bladder wall thickening, 8 of 19 (42.1%) with extravesical stranding, and 2 of 5 (40%) with extravesical nodularity had extravesical disease.


Accurate clinical staging is essential for the appropriate selection of therapy for all cancers, and it is especially pertinent to bladder cancer given that current staging paradigms result in pathologic upstaging (extravesical disease extension or lymph node positivity) ranging from 36 to 73% following radical cystectomy [3,4,5,6]. It is partly due to this limitation in clinical staging that many experts recommend the use of neoadjuvant chemotherapy (NAC) to all patients with MIBC [11,12,13,14]. Most, however, will agree that the true benefit of NAC is in patients who harbor occult nodal metastases and/or have tumors that extend beyond the confines of the bladder. Thus, the hope is that refinements in technology allow CT imaging to better delineate cases of locally advanced disease and therefore increase utilization of preoperative chemotherapy. At our cancer center, we have reported on the use of clinical parameters to risk-stratify these patients—namely, the presence of lymphovascular invasion on TUR specimens, hydroureteronephrosis, variant histologic subtypes, and suspicion for extravesical disease on EUA—in an attempt to identify patients judged to be at the highest risk and who would likely benefit from NAC [15].

Despite CT imaging demonstrating a respectable positive predictive value for identifying tumor extension into perivesical fat, it displayed a dismal specificity. Thus CT imaging adds minimal benefit for local tumor staging, which highlights the importance of incorporating EUA into the staging regimen to help identify locally advanced tumors. Nevertheless, CT imaging was able to exclude lymph node metastases in most patients (negative predictive value of 92.6%), indicating that it remains an important part of this aspect of staging.

Ideally, one would like to use preoperative cross-sectional imaging as the primary modality to accurately stage the extent of local tumor invasion and patterns of lymph node spread or distant metastases (bony or visceral). However, accuracy of CT for bladder cancer staging has been reported to be at best 50% with at least a 25% rate of understaging [9,10]. Since some of these series included patients that were treated prior to 2000, and CT scanning technology has advanced since then, we decided to re-evaluate the utility of CT scanning (from 2000 until now) in the hands of a dedicated genitourinary radiologist blinded to patient outcomes and pathologic features. Unfortunately, we have corroborated prior studies showing that cross sectional imaging with CT scans remains unreliable in identifying the presence of extravesical disease. Moreover, we found no reliable radiologic signs that consistently identified the extravesical extent of disease. Finally, identification of small-volume lymph node metastases is barely adequate with today’s CT scans.

Our study cohort consisted of our most recent 12 years of patient data. However, our findings are primarily limited by the heterogeneity in the available CT imaging quality and technique. Some of our CT images were derived from scanning techniques as far back as 2000. Furthermore, only axial CT images were reviewed since the majority of scans were obtained from routine staging protocols rather than any kind of specialized or research thin-cut section protocols with the possibility of 3-D reformations. Nevertheless, we specifically did not exclude CT studies from other institutions in order to replicate real world results.

We do recognize that the number of patients selected for this re-review can be considered small. However, we used strict criteria to avoid the biases of a post-TUR CT scan, and it is possible that we have not been able to demonstrate an improvement in staging due to the small number of CT scans reviewed; however, it is likely that inclusion of post-TUR scans would only lower the accuracy of CT scanning due to procedure-introduced artifacts.

Ultimately, CT imaging is not yet sensitive enough to distinguish between organ-confined and extravesical bladder cancer. Moreover, CT is especially difficult to interpret in the post-biopsy or transurethral setting secondary to inflammatory changes. Magnetic resonance imaging has been proposed as a better cross-sectional imaging modality to stage patients given the high tissue contrast with contrast enhancement and diffusion-weighted imaging. Unfortunately, studies have been unable to prove the significant clinical benefit of magnetic resonance over CT that justifies its use [16,17,18,19,20]. These limitations in cross-sectional imaging highlight our desperate need for better patient stratification, whether it is through molecular analyses of tumor characteristics on TUR specimens, circulating tumor cells, or molecular imaging.


CT scanning has limits in its ability to properly identify extravesical disease or lymph node spread in patients with MIBC. Investigations into additional or alternative means of clinical staging for bladder cancer patients are incredibly crucial.


This research was supported by the Genitourinary SPORE (NIH Grant # P50CA091846). There were no additional financial relationships sponsoring this research.


  1. Hollenbeck, B. K., D. C. Miller, et al. (2005). "The effects of stage divergence on survival after radical cystectomy for urothelial cancer." Urol Oncol 23(2): 77-81. PubMed | CrossRef
  2. Cheng, L., R. M. Neumann, et al. (2000). "Grading and staging of bladder carcinoma in transurethral resection specimens. Correlation with 105 matched cystectomy specimens." Am J Clin Pathol 113(2): 275-279. PubMed | CrossRef
  3. Chang, B. S., H. L. Kim, et al. (2001). "Correlation between biopsy and radical cystectomy in assessing grade and depth of invasion in bladder urothelial carcinoma." Urology 57(6): 1063-1066; discussion 1066-1067. PubMed
  4. Ficarra, V., O. Dalpiaz, et al. (2005). "Correlation between clinical and pathological staging in a series of radical cystectomies for bladder carcinoma." BJU Int 95(6): 786-790. PubMed | CrossRef
  5. McLaughlin, S., J. Shephard, et al. (2007). "Comparison of the clinical and pathologic staging in patients undergoing radical cystectomy for bladder cancer." Int Braz J Urol 33(1): 25-31; discussion 31-22. PubMed
  6. Mehrsai, A., D. Mansoori, et al. (2007) "A Comparison between Clinical and Pathologic Staging in Patients with Bladder Cancer." Urol J 1: 85-89.
  7. Shariat, S. F., G. S. Palapattu, et al. (2007). "Discrepancy between clinical and pathologic stage: impact on prognosis after radical cystectomy." Eur Urol 51(1): 137-149; discussion 149-151. PubMed | CrossRef
  8. Svatek, R. S., S. F. Shariat, et al. (2011). "Discrepancy between clinical and pathological stage: external validation of the impact on prognosis in an international radical cystectomy cohort." BJU Int 107(6): 898-904. PubMed | CrossRef
  9. Paik, M., M. Scolieri, et al. (2011) "Limitations of computerized tomography in staging invasive bladder cancer before radical cystectomy." J Urol 163: 1693-1696.
  10. Tritschler, S., C. Mosler, et al. (2012). "Staging of muscle-invasive bladder cancer: can computerized tomography help us to decide on local treatment?" World J Urol 30(6): 827-831. PubMed | CrossRef
  11. (2003). "Neoadjuvant chemotherapy in invasive bladder cancer: a systematic review and meta-analysis." Lancet 361(9373): 1927-1934. PubMed
  12. Winquist, E., T. S. Kirchner, et al. (2004). "Neoadjuvant chemotherapy for transitional cell carcinoma of the bladder: a systematic review and meta-analysis." J Urol 171(2 Pt 1): 561-569. PubMed | CrossRef
  13. Grossman, H. B., R. B. Natale, et al. (2003). "Neoadjuvant chemotherapy plus cystectomy compared with cystectomy alone for locally advanced bladder cancer." N Engl J Med 349(9): 859-866. PubMed | CrossRef
  14. Griffiths, G., R. Hall, et al. (2011). "International phase III trial assessing neoadjuvant cisplatin, methotrexate, and vinblastine chemotherapy for muscle-invasive bladder cancer: long-term results of the BA06 30894 trial." J Clin Oncol 29(16): 2171-2177. PubMed
  15. Dickstein, R., H. Grossman, et al. (2011) "Can We Reliably Identify Patients for Radical Cystectomy Without Neoadjuvant Chemotherapy?" J Clin Oncol 29: 7854.
  16. Kim, B., R. C. Semelka, et al. (1994). "Bladder tumor staging: comparison of contrast-enhanced CT, T1- and T2-weighted MR imaging, dynamic gadolinium-enhanced imaging, and late gadolinium-enhanced imaging." Radiology 193(1): 239-245. PubMed
  17. Barentsz, J. O., G. J. Jager, et al. (1996). "Primary staging of urinary bladder carcinoma: the role of MRI and a comparison with CT." Eur Radiol 6(2): 129-133. PubMed
  18. Husband, J. E., J. F. Olliff, et al. (1989). "Bladder cancer: staging with CT and MR imaging." Radiology 173(2): 435-440. PubMed
  19. Takeuchi, M., S. Sasaki, et al. (2009). "Urinary bladder cancer: diffusion-weighted MR imaging--accuracy for diagnosing T stage and estimating histologic grade." Radiology 251(1): 112-121. PubMed | CrossRef
  20. El-Assmy, A., M. E. Abou-El-Ghar, et al. (2009). "Bladder tumour staging: comparison of diffusion- and T2-weighted MR imaging." Eur Radiol 19(7): 1575-1581. PubMed | CrossRef