Volume 30, Issue 1 , Pages 49-54, January 2012
Subepithelial growth patterns in urothelial carcinoma—frequency and prognostic significance
Article Outline
Abstract
Purpose
Most urothelial carcinomas are exophytic, but some tumors exhibit subepithelial components, either in the form of endophytic growth pattern (EGP) or as von Brunn's nests involvement (VBNI). The purpose of this study was to investigate the frequency, inter-relations and clinical significance of these forms of subepithelial neoplasia in urothelial carcinoma.
Patients and methods
Between June 1995 and December 2007, 760 patients (mean age of 67.5 years) underwent transurethral resection of bladder tumors in our institution, including 478, 157, and 112 patients with stage Ta, T1, and ≥T2 disease, respectively. Isolated or concomitant Tis were present in 137 (18%) patients. Median postoperative follow-up period was 53 months.
Results
EGP was found in 86 cases (11.3%) and VBNI in 30 (3.9%) patients. Both forms of subepithelial growth were significantly more common in higher stage and grade tumors and were associated with each other. Multivariate analysis showed that EGP is an independent prognostic factor of stage progression (HR 4.6, P < 0.0001) and disease specific mortality (HR 2.6, P = 0.001) but not of tumor recurrence (HR 1.2, P = 0.51). VBNI was found an independent prognostic factor of tumor progression (HR 5.1, P < 0.0001), but neither of tumor recurrence nor disease specific mortality.
Conclusions
Subepithelial growth is not an uncommon in bladder cancer. It is more frequent in high-grade and high-stage tumors. The findings of this study suggest that subepithelial growth carries a higher risk for stage progression (EGP and VBNI) and mortality (EGP), but not tumor recurrence.
Keywords: Bladder cancer , Endophytic growth pattern , von Brunn's nests involvement
1. Introduction
Most urothelial carcinomas are exophytic tumours, however some lesions exhibit an endophytic, subepithelial growth pattern (EGP). The reason for this is unknown. Possibly, microenvironmental pattern of neovascularization and growth of the lining basement membrane determine whether the tumor will grow outward or inward. Alternatively, involvement and expansion of von Brunn's nests by urothelial cancer may create the inverted pattern. Endophytic tumours pose 2 pathologic diagnostic problems: distinction from the benign tumor-inverted papilloma when anastomosing invaginations are present, and distinction from invasive cancer in cases of pushing broad front base [1].
Data regarding the frequency and prognostic significance of endophytic growth is spare. In a study by Amin et al., 18 cases of endophytic urothelial carcinoma have been reported [1]. After a mean follow-up of 15.5 months, 11/14 evaluable patients were alive with no evidence of disease, 2 patients died from other causes, and 1 patient was alive with disease. Small cohort size, short follow-up, and the absence of a control group make it impossible to draw any conclusions from this study regarding the frequency and prognostic significance of the EGP.
The frequency and prognostic significance of VBNI in bladder cancer were also addressed in only a few articles. Dinney et al. from M. D. Anderson retrospectively surveyed their database and reported on a VBNI incidence of 19.7% (73/371) among patients with non-muscle-invasive bladder cancer [2].
In this study, we investigated the phenomenon of subepithelial growth patterns in urothelial carcinoma. The frequency of the different types of subepithelial growth, their inter-relations, and their prognostic significance were systematically investigated.
2. Materials and methods
2.1. Patient population
An institutional database was retrospectively obtained from the hospital registry, which contained information on 760 consecutive patients who underwent transurethral resection of bladder cancer between June 1995 and December 2007 in our institution. All pathologic material was reviewed by a single dedicated uropathologist (GP). Pathologic staging was performed according to the TNM system and grading according to the 1973 WHO classification [3]. Endophytic growth was defined as the presence of either broad bulbous tongues of neoplastic urothelium extending deep into the lamina propria, or endophytic expansile growth with or without peripheral palisading of basal cells [4]. Tumor cells are limited by epithelial basement membrane in all cases. The study was provided with a waived approval by the IRB (number 207-31.10.08).
2.2. Treatment protocol
Patients with T1G3 tumors underwent a second look resection. Patients with high-grade tumors, stage T1, or bladder Tis were given an induction course of 6 weekly intravesical instillations of 81 mg Connaught BCG in 50 cc of normal saline, initiated 10 to 20 days following surgery. Patients with recurrent low-grade disease occurring sooner than 24 months from the first resection were also given an induction course of BCG. Maintenance therapy (2–3 instillations every 3 months for 1 year and then every 6 months for an additional 2 years) was given to patients with high-grade cancer. Radical cystectomy was offered to patients with muscle-invasive disease who had no evidence of metastases on computerized tomography, and to patients with BCG-resistant or recurrent T1G3 or Tis.
Follow-up protocol included cystoscopy and urinary cytology every 3 months for 2 years and then every 6 months for additional 3 years. Bladder biopsies were used for monitoring response to BCG in patients with high-grade disease.
Upper tract surveillance (intravenous pyelography and, more recently, CT urography) was performed upon initial diagnosis, and then annually in cases of high-grade tumours only. The finding of an endophytic growth pattern did not impact on treatment or follow-up protocols.
2.3. Statistical analysis
The influence of the following variables on recurrence-free survival, progression-free survival, and disease-specific mortality were analyzed: patient's age, gender, tumor's stage, grade, and the presence of Tis, EGP, or VBNI on the initial pathologic specimen. Continuous variables were compared using the t-test and categorical variables using Fisher's exact test and χ2 tests when appropriate. All statistical tests were two tailed and a P value < 0.05 was considered significant. Survival curves were built using the Kaplan-Meier method. The log rank test was used for univariate analysis, and backward stepwise (likelihood ratio) logistic regression was used for multivariate analysis. SPSS software (SPSS Inc., Chicago, IL) was used for data processing.
3. Results
Study cohort included 760 patients (mean age 67.5 years, SD 12.7 years), including 622 men and 138 women. Tumor staging according to specimens obtained at initial transurethral resection were: stage “Ta” in 478 patients (62.9%), stage T1 in 157 patients (20.6%), and muscle-invasive disease (stage ≥T2) in 112 patients (14.7%). Grading according to the WHO 1973 system showed: grade 1 tumor in 166 patients (21.8%), grade 2 in 276 (36.3%), and grade 3 in 318 (41.8%). Bladder carcinoma in situ (either pure or concomitant) was found in 137 cases (18%).
EGP and VBNI were found in a total of 86 patients (11.3%) and 30 patients (3.9%), respectively (Fig. 1). The frequency of EGP and VBNI according to various patients' characteristics is presented in Table 1. EGP and VBNI were significantly more common in higher tumor stages and grades. VBNI was found almost exclusively in cases of Tis.
Fig. 1.
(A) A low grade papillary tumor with endophytic component. Part of the tumor is growing below a normal looking urothelium (arrow) H and E ×50. (B) A high grade tumor with endophytic component. Tis is seen on surface epithelium, H and E ×100. (C) Tis of the surface epithelium and involving a von Brunn's nest, H and E ×100. (Color version of figure is available online.)
Table 1. The frequency of endophytic growth pattern and von Brunn's nests involvement according to various parameters
| Endophytic growth pattern no/yes | P value | von Brunn's nests involvement no/yes | P value | |
|---|---|---|---|---|
| Total | 674 (88.7%)/86(11.3%) | — | 730(96.1%)/30(3.9%) | — |
| Sex | ||||
| Male | 557(89.5%)/65(10.5) | — | 596(95.8%)/26(4.2%) | — |
| Female | 117(84.8%)/21(15.2%) | 0.14⁎ | 134(97.1%)/4(2.9%) | 0.63⁎ |
| Age | ||||
| 25–55 | 144(83.8%)/22(16.2%) | — | 131(96.3%)/5(3.7%) | — |
| 56–65 | 144(88.9%)/18(11.1%) | — | 155(95.7%)/7(4.3%) | — |
| 66–75 | 221(89.1%)/27(10.9%) | — | 236(95.2%)/12(4.8%) | — |
| >75 | 195(91.1%)/86(11.3%) | 0.210⁎⁎ | 208(97.2%)/6(2.8%) | 0.719⁎⁎ |
| TNM Stage | ||||
| a | 433(90.6%)/45(9.4%) | — | 471(98.5%)/7(1.5%) | — |
| 1 | 125(79.6%)/32(20.4%) | — | 142(90.4%)/15(9.6%) | — |
| 2 | 103(92%)/9(8%) | 0.001⁎⁎ | 106(94.6%)/6(5.4%) | — |
| Tis (pure) | — | — | 11(84.6%)/2(15.4%) | <0.0001⁎⁎ |
| Tis (concomitant and pure) | ||||
| No | 553(88.8%)/70(11.2%) | 615(98.7%)/8(1.3%) | ||
| Yes | 121(88.3%)/16(11.7%) | 0.88⁎ | 115(83.9%)/22(16.1%) | <0.0001⁎ |
| Grade | ||||
| 1 | 159(95.8%)/7(4.2%) | — | 166(100%)/0(0%) | — |
| 2 | 238(86.2%)/38(13.8%) | — | 273(98.9%)/3(1.1%) | — |
| 3 | 277(87.1%)/41(12.9%) | 0.005⁎⁎ | 291(96.1%)/27(8.5%) | <0.0001⁎⁎ |
| von Brunn's nests involvement | ||||
| No | 652(96.7%)/22(3.3%) | — | — | — |
| Yes | 72(83.7%)/8(9.3%) | 0.009⁎ | — | — |
| Endophytic growth pattern | ||||
| No | — | — | 652(89.3%)/72(9.7%) | — |
| Yes | — | — | 22(73.3%)/8(26.7%) | 0.009⁎ |
⁎ Fisher's exact test. |
⁎⁎ χ2 test. |
After a median follow-up of 53 months, 202 patients (26.6%) suffered from tumor recurrence (without stage progression), 58 (7.6%) from stage progression, and 86 patients (11.3%) died from bladder cancer. Univariate analysis (Table 2) demonstrated that EGP and VBNI are both significant prognostic factors of stage progression (HR 5.9 and 10.9 respectively, P < 0.0001 for both) and disease specific mortality (HR 1.8 and 2.4, P = 0.036 and 0.017, respectively) but not of tumor recurrence. When considering patients with stage a, grades 1-2 only, stage progression was found in 11/38 patients (28.9%) with EGP compared with 8/376 patients (2.1%) without EGP (P < 0.0001). Among the patients with T1G3 disease (141 cases), 14 had VBNI and 28 EGP. VBNI was a risk factor of progression (P = 0.0004) and mortality (P = 0.017), but not of recurrence (P = 0.056). EGP was a risk factor of mortality (P = 0.02), but not of recurrence (P = 0.74) or progression (P = 0.06).
Table 2. Univariate analysis of factors influencing disease recurrence, progression, and disease-specific survival
| Tumor recurrence hazard ratio (95% CI) | P value | Tumor progression hazard ratio (95% CI) | P value | Disease-specific mortality hazard ratio (95% CI) | P value | |
|---|---|---|---|---|---|---|
| TNM Stage | ||||||
| a | 1 | — | 1 | — | 1 | — |
| 1 relative to a | 0.79(0.56–1.13) | 0.198 | 5.01(2.89–8.65) | <0.0001 | 10.92(3.963–30.091)⁎ | <0.0001 |
| WHO 1973 grade | ||||||
| 1 | 1 | — | 1 | — | 1 | — |
| 2 relative to 1 | 1.14(0.822–1.59) | 0.428 | 4.07(1.19–13.81) | 0.025 | — | — |
| 3 relative to 1 | 0.87(0.59–1.28) | 0.481 | 12.51(3.92–40.97) | <0.0001 | 36.42(13.335–99.467)⁎⁎ | <0.0001 |
| Tis | ||||||
| No | 1 | — | 1 | — | 1 | — |
| Yes relative to no | 1.13(0.78–1.63) | 0.511 | 3.98(2.30–6.87) | <0.0001 | 1.93(1.22–3.05) | 0.005 |
| Endophytic growth pattern | ||||||
| No | 1 | — | 1 | — | 1 | — |
| Yes relative to no | 1.14(0.75–1.75) | 0.537 | 5.94(3.41–10.35) | <0.0001 | 1.82(1.04–3.17) | 0.036 |
| von Brunn's nests involvement | ||||||
| No | 1 | — | 1 | — | 1 | — |
| Yes relative to no | 1.45(0.75–2.82) | 0.278 | 10.89(5.70–20.78) | <0.0001 | 2.42(1.17–5.02) | 0.017 |
⁎ Stage 2 relative to stage a, relative risk 85.642 (95% CI 34.344–213.562), P < 0.0001. |
⁎⁎ Grade 3 relative to grades 1 and 2 combined. |
The 5-year progression-free survival of patients with and without VBNI was 41% and 92%, respectively. Similarly, the 5-year progression-free survival of patients with and without EGP was 70% and 92%, respectively (Fig. 2). Multivariate analysis (Table 3) showed that EGP is an independent prognostic factor of stage progression (HR 4.6, P < 0.0001) and disease-specific mortality (HR 2.6, P = 0.001). VBNI was found an independent prognostic factor of progression (HR 5.1, P < 0.0001), but not of disease specific mortality.
Fig. 2.
Kaplan Meier curves for tumor recurrence, progression, and disease specific survival (A–C) for patients with endophytic growth pattern and D–F for patients with von Brunn's nests involvement. The numbers above and below the curves represent the number of “patients at risk” without and with the feature. (Color version of figure is available online.)
Table 3. Multivariate analysis (final model) of factors influencing disease recurrence, progression, and disease-specific survival
| Tumor recurrence hazard ratio (95% CI) | P value | Tumor progression hazard ratio (95% CI) | P value | Disease-specific mortality hazard ratio (95% CI) | P value | |
|---|---|---|---|---|---|---|
| TNM stage | ||||||
| a | 1 | — | — | — | 1 | — |
| 1 vs. a | 0.82(0.49–1.34) | 0.42 | — | 3.33(0.88–12.54) | 0.076 | |
| 2 vs. a | — | — | 27.59(7.65–99.57) | <0.0001 | ||
| WHO 1973 grade | — | |||||
| 1 | 1 | — | 1 | — | 1 | — |
| 2 vs. 1 | 1.11(0.78–1.56) | 0.53 | 3.14(0.92–10.76) | 0.068 | — | — |
| 3 vs. 1 | 0.84(0.50–1.42) | 0.52 | 7.20(2.13–24.32) | 0.001 | 36.419(13.335–99.467)⁎ | <0.0001 |
| Tis | ||||||
| No | 1 | — | — | — | — | — |
| Yes vs. no | 1.23(0.85–1.98) | 0.23 | — | — | — | |
| Endophytic growth pattern | ||||||
| No | 1 | — | 1 | — | 1 | — |
| Yes vs. no | 1.16(0.75–1.79) | 0.51 | 4.63(2.63–8.16) | <0.0001 | 2.6(1.47–4.61) | 0.001 |
| von Brunn's nests involvement | ||||||
| No | 1 | — | 1 | — | — | — |
| Yes vs. no | 1.53(0.75–3.13) | 0.24 | 5.15(2.56–10.37) | <0.0001 | — | — |
⁎ Grade 3 vs. grades 1 or 2. |
4. Discussion
The urothelium has a tendency of inward growth. This tendency is reflected both in the normal epithelium, in benign conditions and in malignancy. The normal urothelium is characterized by a high prevalence of von Brunn's nests (89%) and cystitis cystica (60%) [5]. The benign tumor-inverted papilloma is also well characterized; it comprises 1% of all bladder tumours, and typically arises in the trigone. After transurethral resection, it rarely recurs [6]. The endophytic growth pattern of malignant tumours was reported only in 1982 [7]. Despite accurate morphologic descriptions [1], [4], the frequency and clinical significance of this phenomenon remained relatively unknown.
In this study, analyzing pathologic material from 760 patients, we found that subepithelial growth is not an uncommon component in bladder cancer. EGP was found in 11.3% of the patients. To the best of our knowledge, this is the first study systematically evaluating the incidence of EGP. EGP was found to be a strong independent predictor of tumor progression (HR 4.6) and disease-specific mortality (HR 2.6).
VBNI was a less common finding inflicting 3.9% of the patients in this study. This incidence is far less than the 19.7% reported by Dinney et al. from M.D. Anderson [2]. This may be related to differences in study populations. While patients attending our institution represent the general population of bladder cancer patients, attendees of the M. D. Anderson institution are selectively higher risk patients.
We also found that VBNI is an independent prognostic factor for tumor progression (HR 5.15), in contrast to the conclusions of Dinney et al. [2]. They concluded that VBNI is neither a risk factor for disease progression nor an absolute indication for radical cystectomy. The reason for this disparity can be a difference in treatment protocols. While Dinney et al. practiced a very aggressive approach toward patients with VBNI (driven by the opinion that early cystectomy is the optimal treatment for these patients), in the current study, presence of VBNI did not influence the treatment protocol.
The explanation for the worse prognosis of patients with subepithelial growth pattern is not known, but several possibilities may be raised. Due to their irregular morphology, tumors with subepithelial component are more difficult to stage, and therefore, a T1 tumor can be understaged as Ta. Additionally, the subepithelial component of the tumor may be more difficult to eradicate by transurethral surgery, and tumor remnants that are not exposed to the bladder lumen may be less responsive to intravesical treatment with BCG.
Another finding of the study is the association of the 2 types of subepithelial growth with each other. A patient with EGP has a higher risk to be inflicted by VBNI (Table 1). This could be another explanation for the worst prognosis of patients with EGP.
This study has several limitations, including the retrospective nature of the study and the source of data from a single institution. Therefore, verification of the findings by other investigators is necessary.
In conclusion, subepithelial growth is found in more than 10% of patients with bladder cancer. This condition is associated with an independently higher risk for stage progression (EGP and VBNI) and mortality (EGP), but not tumor recurrence. Patients with subepithelial tumor growth patterns must be carefully observed. A liberal use of “second look” biopsies and biopsies for verification of treatment success must be taken in order to avoid stage progression and disease specific mortality.
References
- . Urothelial transitional cell carcinoma with endophytic growth patterns: a discussion of patterns of invasion and problems associated with assessment of invasion in 18 cases . Am J Surg Pathol . 1997;21:1057–1068
- Management of transitional cell carcinoma involving von Brunn's nests . J Urol . 1995;153:944–949
- . International histologic classification of tumors, No. 10. Histologic typing of urinary bladder tumours . Geneva: World Health Organization; 1973;
- . Bladder Biopsy Interpretation . Philadelphia, PA: Lippincott Williams and Wilkins; 2004;
- The prevalence and significance of Brunn's nests, cystitis cystica, and squamous metaplasia in normal bladders . J Urol . 1979;122:317–321
- Inverted papilloma of the urinary bladder: A molecular genetic appraisal . Mod Pathol . 2006;19:1289–1294
- Malignant inverted papilloma and carcinoma in situ of the bladder . J Urol . 1982;128:816–818
PII: S1078-1439(09)00363-9
doi:10.1016/j.urolonc.2009.11.010
© 2012 Elsevier Inc. All rights reserved.
Volume 30, Issue 1 , Pages 49-54, January 2012
