- Email: [email protected]
Plasmacytoid variant urothelial bladder cancer: Is it time to update the treatment paradigm?
Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838
Original article
Plasmacytoid variant urothelial bladder cancer: Is it time to update the treatment paradigm? Hristos Z. Kaimakliotis, M.D.1,a,*, M. Francesca Monn, M.D., M.P.H.1,a, K. Clint Cary, M.D., M.P.H.a, Jose A. Pedrosa, M.D.a, Kevin Rice, M.D.a, Timothy A. Masterson, M.D.a, Thomas A. Gardner, M.D.a, Noah M. Hahn, M.D.b, Richard S. Foster, M.D.a, Richard Bihrle, M.D.a, Liang Cheng, M.D.c, Michael O. Koch, M.D.a b
a Department of Urology, Indiana University School of Medicine, Indianapolis, IN Department of Genitourinary Medical Oncology, Indiana University School of Medicine, Indianapolis, IN c Department of Pathology, Indiana University School of Medicine, Indianapolis, IN
Received 31 December 2013; received in revised form 12 February 2014; accepted 8 March 2014
Abstract Objectives: Plasmacytoid variant (PCV) urothelial cancer (UC) of the bladder is rare, with poor clinical outcomes. We sought to identify factors that may better inform expectations of tumor behavior and improve management options in patients with PCV UC. Materials and methods: A retrospective analysis of the Indiana University Bladder Cancer Database between January 2008 and June 2013 was performed comparing 30 patients with PCV UC at cystectomy to 278 patients with nonvariant (NV) UC at cystectomy who underwent surgery for muscle-invasive disease. Multivariable logistic regression was used to assess precystectomy variables associated with non–organ-confined disease at cystectomy and Cox regression analysis to assess variables associated with mortality. Results: Patients with PCV UC who were diagnosed with a higher stage at cystectomy (73% pT3-4 vs. 40%, P ¼ 0.001) were more likely to have lymph node involvement (70% vs. 25%, P o 0.001), and positive surgical margins were found in 40% of patients with PCV UC vs. 10% of patients with NV UC (P o 0.001). Median overall survival and disease-specific survival were 19 and 22 months for PCV, respectively. Median overall survival and disease-specific survival had not been reached for NV at 68 months (P o 0.001). Presence of PCV UC on transurethral resection of bladder tumor was associated with non–organ-confined disease (odds ratio ¼ 4.02; 95% CI: 1.06–15.22; P ¼ 0.040), and PCV at cystectomy was associated with increased adjusted risk of mortality (hazard ratio ¼ 2.1; 95% CI: 1.2–3.8; P ¼ 0.016). Conclusions: PCV is an aggressive UC variant, predicting non–organ-confined disease and poor survival. Differentiating between non– muscle- and muscle-invasive disease in patients with PCV UC seems less important than the aggressive nature of this disease. Instead, any evidence of PCV on transurethral resection of bladder tumor may warrant aggressive therapy. r 2014 Elsevier Inc. All rights reserved.
Keywords: Bladder cancer; Plasmacytoid variant; Variant histology; Cystectomy
1. Introduction Plasmacytoid variant (PCV) urothelial cancer (UC) of the bladder is an increasingly defined, aggressive variant of UC first described in 1991 as “similar to the plasma cells seen in myeloma” [1]. Since then, knowledge regarding PCV has 1
Contributed equally to manuscript.
Corresponding author. Tel.:þ1-203-530-0445; fax: þ1-317-278-0499. E-mail address: [email protected] (H.Z. Kaimakliotis).
*
http://dx.doi.org/10.1016/j.urolonc.2014.03.008 1078-1439/r 2014 Elsevier Inc. All rights reserved.
increased, including recognition of its advanced stage at cystectomy, rate of metastasis, and clinical implications [2–7]. Optimal management strategies remain to be determined with much of the literature suggesting that PCV should be treated similarly to other aggressive variant histology, using early surgical intervention [8], while other literature point toward a greater role for novel chemotherapuetic regimens [9,10]. The largest case series range from 15 to 32 patients and begin to construct an understanding of PCV; however the need for larger, more comprehensive
834
H.Z. Kaimakliotis et al. / Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838
series remains ever present [2,5,6,11]. As we learn more about PCV, it becomes increasingly important to continue reporting outcomes so that more conclusive management guidelines can be established. We present our experience of patients diagnosed with any degree of PCV component at the time of cystectomy, and by exploring their clinical course, we hope to gain insight for improved treatment plans.
2. Materials and methods Institutional review board approval was obtained from Indiana University for this study. Using the Indiana University Bladder Cancer Database, a retrospective analysis of all patients who underwent cystectomy with curative intent between 2008 and June 2013 was performed. Patients with metastatic disease or lymphadenopathy greater than 1.5 cm on cross-sectional imaging at time of diagnosis were excluded as were patients undergoing bladder preservation protocols. Patients with pure squamous, small cell, rhabdomyosarcoma, or adenocarcinoma at transurethral resection of bladder tumor (TURBT) or cystectomy were also excluded from analysis. PCV was defined based on histology, with individual tumor cells exhibiting a prominent eccentrically placed nucleus and abundant eosinophilic cytoplasm, reminiscent of plasma cells. Most neoplastic cells exhibited nuclei of low to intermediate nuclear grade with occasional nuclear pleomorphism, with CD138 expression in some cases. Interobserver variability is always an issue for variant calling, but diagnosis of variant still relies on morphological rather than immunohistochemical characteristics [12]. All TURBT specimens underwent central review at Indiana University before cystectomy to ensure accuracy of diagnosis. Pathologic assessment of all cases was performed by dedicated genitourinary pathologists. Patients with PCV UC were defined as those whose cystectomy pathology included any component of PCV, regardless of overall percentage or mixed-variant status. As the primary aim of the project was to compare patients with PCV UC with patients with nonvariant (NV) UC, patients who had variant histology at cystectomy without the presence of PCV were eliminated (n ¼ 137). Additionally, as variant histology is often considered understaged on TURBT, we chose to use only patients with NV UC with muscle-invasive disease on TURBT as the comparison population (n ¼ 278). Although we chose to eliminate patients with PCV UC on TURBT who did not develop PCV on final cystectomy pathology (n ¼ 5), we performed a descriptive analysis of all patients with PCV who underwent TURBT to examine clinicopathologic outcomes. Data were obtained for each patient regarding sex, age, race, TURBT stage and histology, neoadjuvant and adjuvant chemotherapy, date of cystectomy, stage, histology, lymph node (LN) status, and non–organ-confined disease. Non–organ-confined disease was defined as pathologic cystectomy stage T3, T4, or LN involvement.
Survival data were obtained through the Indiana University Cancer Registry, which gathers vital statistic information using medical records, primary care physicians, the social security death index, and electronic record searches. This registry is updated annually with greater than 90% success rate of follow-up for patients treated at Indiana University. When survival data were not available within the past month, we evaluated patient survival status using the Social Security Death Index and used disease status at time of last follow-up for disease-specific survival (DSS). We used the Pearson chi-square and Fisher exact tests for categorical variables and the Student t test and MannWhitney U tests for continuous variables to perform descriptive analysis comparing the PCV and NV groups. Multivariable logistic regression was performed to examine precystectomy factors associated with developing non– organ-confined disease on radical cystectomy. The model was built using forward and back stepwise methodology with an accepted P o 0.1. Variables included in the stepwise regression were sex, age, TURBT histology (PCV vs. NV), lymphovascular invasion, neoadjuvant chemotherapy (NACT), and number of nodes removed at cystectomy. The final model included TURBT histology, lymphovascular invasion, and NACT. We chose to include age and sex in the final model as well. Kaplan-Meier overall survival (OS) and DSS curves were generated and log-rank tests performed to evaluate differences between the PCV and NV populations. Cox proportional hazards regression was used to evaluate characteristics associated with allcause mortality. Factors used in the stepwise regression were age, sex, PCV histology, any systemic chemotherapy, and non–organ-confined disease on cystectomy. Variables accepted for use in the model were PCV histology, systemic chemotherapy, and non–organ-confined disease. Age and sex were also included in the final model as well. The proportional hazards assumption was tested using Schoenfeld residual plots and all variables included in the models did not violate the assumption. Time zero was defined as the date of radical cystectomy. A priori, P o 0.05 was considered statistically significant. Microsoft Excel (Microsoft Inc) and Stata version 12.1 (Stata Corp LP, College Station, TX) were used for all statistical analyses.
3. Results Of the 624 patients with UC who underwent radical cystectomy with curative intent, 30 (4.8%) were diagnosed with any degree of PCV on cystectomy specimen, 25 with primary PCV (450% of total variant histology), and 5 with secondary or tertiary components. A total of 137 (22.0%) patients were noted to have variant histology other than PCV and as such were excluded from the analysis. In total, 457 (73.2%) patients were diagnosed with pure urothelial histology, or NV histology. Of the 457 patients with NV
H.Z. Kaimakliotis et al. / Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838
UC, 278 were muscle-invasive on TURBT and thus were the group used for comparison in all analyses. There were no demographic differences between patients with PCV UC and patients with NV UC (Table 1). Patients with PCV UC had higher stage at time of cystectomy, with 97% demonstrating muscle-invasive disease compared with 69% of patients with NV UC (P ¼ 0.001). Non–organ-confined and LN-positive disease was more common among patients with PCV UC (P ¼ 0.001 and o0.001, respectively). Median (mean) LN density for patients with PCV UC was 23.1 (34.5) compared with 11.8 (21.1) in patients with NV UC (P ¼ 0.032) (Table 1). NACT appeared to have little effect on patients with PCV UC. Of the 5 patients with PCV UC who received NACT, 80% (4 of 5 patients) had LN involvement and 80% had ZpT3 disease, compared with 21% (13 of 63 patients) with NV UC who had LN involvement and 24% (15 of 63 patients) with ZpT3 disease following NACT (P ¼ 0.003 and 0.007, respectively). Multivariable logistic regression was performed to determine precystectomy factors associated with non– organ-confined disease at cystectomy (Table 2). Patients
Table 1 Characteristics of patients with plasmacytoid variant UC and patients with nonvariant UC at time of cystectomy P-valuea
Plasmacytoid variant, n (%)c
Nonvariant, n (%)c
Patients, n Mean age (SD), y Sex (female) Race (white) TURBT LVI TURBT CIS
30 66.2 (9.4) 6 (20.0) 27 (90.0) 1 (3.3) 8/28 (28.6)
278 66.6 (11.2) 58 (20.9) 265 (95.3) 27 (9.7) 59/255 (23.1)
Cystectomy stage rpT1 pT2 pT3 pT4
1 (3.3) 7 (23.3) 10 (33.3) 12 (40.0)
86 80 82 30
24 (80.0)
132 (47.5)
12 (40.0) 10 (33.3) 15 (50.0) 21 (70.0) 21.5 (16–36)
27 (9.7) o0.001 7 (2.5) o0.001 150 (54.0) 0.680 69 (24.8) o0.001 21.5 (14–31) 0.416
23.1 (14–42)
11.8 (6–27)
5 (16.7) 11 (36.7)
63 (22.7) 34 (12.2)
Non–organ-confined disease Positive surgical margin Positive ureteral margin Cystectomy-associated CIS Positive lymph nodes Median lymph nodes examined (IQR) Median lymph node density (IQR)b Neoadjuvant chemotherapy Adjuvant chemotherapy
(30.9) (28.8) (29.5) (10.8)
0.842 0.912 0.212 0.248 0.521 o0.001
0.001
0.038 0.452 o0.001
SD ¼ standard deviation; LVI ¼ lymphovascular invasion; CIS ¼ carcinoma in situ; IQR ¼ interquartile range. a P-value represents the chi-square test, Student t test, or the MannWhitney U test when median (IQR) is presented. b Lymph node (LN) density is calculated as the number of positive LN/ total number of LN in patients with positive LN. c Numbers in parentheses represent percentage except where otherwise noted.
835
Table 2 Multivariable logistic regression of clinical characteristics associated with non–organ-confineda disease at the time of cystectomy Characteristicsb
Odds ratio (95% CI)
P value
Age, y Sex (female) TURBT plasmacytoidc TURBT lymphovascular invasion Neoadjuvant chemotherapy
1.01 0.70 4.02 2.30 0.53
0.402 0.232 0.040 0.053 0.027
(0.99–1.03) (0.40–1.25) (1.06–15.22) (0.99–5.35) (0.30–0.93)
Non–organ-confined disease is defined as LNþ or ZpT-stage 3. Characteristics controlled for in the multivariable logistic regression model are those included in the chart. c TURBT plasmacytoid is PCV vs. NV. a
b
with PCV UC identified on TURBT had 4 times the odds of developing non–organ-confined disease compared with patients NV UC (P ¼ 0.040), and NACT was associated with decreased odds of having non–organ-confined disease (odds ratio ¼ 0.53, 95% CI: 0.30–0.93). Median follow-up time for all patients was 30 (interquartile range: 16–48) months. Both OS and DSS for patients with PCV UC were significantly worse than patients with NV UC (P o 0.001 and 0.002) (Fig. A and B). The median OS for PCV was 19 months (interquartile range: 11–28), whereas median OS for patients with NV UC was not yet reached at 68 months. Median DSS for patients with PCV UC was 22 months, whereas it was not yet reached for the patients with NV UC. After adjusting for age, sex, any systemic chemotherapy, and non–organ-confined disease, PCV histology was associated with a 2-fold increased adjusted risk of all-cause mortality (95% CI: 1.2–3.8, P ¼ 0.029) (Table 3). Significantly, when examining the full cohort of 624 patients in a separate subanalysis, we found 15 patients with PCV UC diagnosed on TURBT. The results of cystectomy pathology were negative for PCV UC for 5 of these patients; 4 had NV UC and 1 went on to develop nested variant. Among the 15 patients diagnosed with PCV UC on TURBT, 4 patients underwent NACT, 1 of whom had a complete response with pT0N0, another had pT2N0, and 2 had ZpT3Nþ. Overall, 73% of all patients with PCV UC who underwent TURBT developed non–organ-confined disease compared with 43% of patients with NV UC (P ¼ 0.032). Also, 4 of the 5 patients diagnosed with cT1 PCV UC on TURBT were upstaged to ZpT3 at the time of cystectomy, and 60% (3 of 5 patients) were found to have LN involvement on final pathology. This is in comparison with 22% (32 of 143 patients) of patients with cT1 NV UC on TURBT who were upstaged to ZpT3 (P ¼ 0.013), and 17% (25 of 143 patients) having LN involvement at time of cystectomy (P ¼ 0.047). 4. Discussion PCV is a recently described subset of UC [1], and we are only beginning to appreciate its clinicopathologic implications.
836
H.Z. Kaimakliotis et al. / Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838
Fig. (A) Overall survival and (B) disease-specific survival. There is a significant difference in overall and disease-specific survival for the population with nonvariant histology and population with plasmacytoid variant UC (log-rank test P o 0.001 each). Median overall survival in the population with plasmacytoid variant UC was 19 months (IQR: 11–28). Median disease-specific survival in the population with plasmacytoid variant UC was 22 months. IQR ¼ interquartile range. (Color version of figure is available online.)
Dayyani et al. [2] propose PCV to be a chemosensitive cancer with a poor prognosis, noting pathologic downstaging in 80% of patients (n ¼ 4 of 5) treated with NACT and a complete response in 3. It remains unclear that patients were rendered pT0 at time of TUR, and the authors acknowledge short-lived response rates among these patients. Further, the poor response to NACT in our group compared with those of the MD Anderson study is unclear; this could be related to variation in TURBT resection, differing NACT regimens, or merely small sample sizes. The Southwest Oncology Group 8710 study found a clinically meaningful improvement in survival among patients with locally advanced UC who received NACT over
cystectomy alone [13], although fewer than 20% of patients with locally advanced node-positive bladder cancer survive for 5 years [14]. The higher incidence of extravesical disease and LN density in our PCV cohort would likely render these patients to a lesser response rate than the effect noted in the Southwest Oncology Group trial. Herein, patients with PCV UC undergoing cystectomy, regardless of systemic therapy, appear to do much worse than patients with NV UC, with few patients with PCV UC alive beyond 24 months. On multivariable analysis, presence of PCV compared with NV at cystectomy confers a 2-fold increased risk of death. The clinical significance of this may be contingent on
H.Z. Kaimakliotis et al. / Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838 Table 3 Multivariable Cox proportional hazards regression model of characteristics associated with overall mortality Characteristica
Hazard ratio (95% CI)
P value
Age, y Sex (female) Plasmacytoid variantb Systemic chemotherapy Non–organ-confined diseasec
1.01 0.99 2.07 0.57 4.87
0.534 0.978 0.016 0.029 o0.001
(0.99–1.03) (0.60–1.65) (1.15–3.75) (0.35–0.95) (2.91–8.16)
a Characteristics controlled for in the Cox regression model are those included in the chart. b Plasmacytoid vs. nonvariant histology. c Non–organ-confined disease is defined as LNþ or ZpT-stage 3. ADDIN EN.REFLIST
PCV recognition at TURBT, as there is poor concordance between TURBT and cystectomy variant histology [15]. Despite formal pathologic diagnostic criteria and increased awareness of variant histology, most UCs are diagnosed in low-volume community practice facilities, where awareness of variant UC may be less comprehensive [16]. The less frequent identification at low-volume centers results in higher potential for misclassification, which may lead to delay in therapy or inappropriate treatment algorithms [16,17]. Our results also show that 80% of patients with PCV UC and non–muscle-invasive disease on TURBT were upstaged to ZpT3 at cystectomy and that 60% had LN involvement. This is nearly 4 times the incidence encountered in cT1 NV cases. This finding is even more impressive when considering that cystectomy for all 5 patients with T1 PCV noted on TURBT was expedited without NACT or intravesical therapy, whereas most patients with NV UC with non– muscle-invasive disease were treated with cystectomy after failing intravesical therapy. The argument against delaying definitive surgery in patients with presumed organ-confined PCV is further supported by our finding that 80% (24 of 30 patients) of cystectomy-confirmed patients with PCV UC had extravesical disease compared with 47% (132 of 278 patients) in the NV population. Furthermore, after controlling for lymphovascular invasion and NACT, patients with PCV UC on TURBT still had a 4-fold increased odds of extravesical disease at cystectomy than patients with NV UC with muscle-invasive disease. The role of early cystectomy in high-grade non–muscle-invasive T1 bladder cancer has been studied, with 10-year cancer-specific survival rates of 78% for early cystectomy vs. 51% to 65% for deferred cystectomy [18,19]. Our demonstration of the high rate of upstaging from non–muscle-invasive to muscle-invasive disease and LN involvement in the population with PCV UC may argue against differentiating patients with PCV UC between invasive or not. Instead, any evidence of PCV histology during TURBT may warrant aggressive surgical therapy. Certain variants of UC, such as micropapillary and sarcomatoid, have been hypothesized to be chemoresistant
837
with worse overall clinical outcomes than patients with NV UC [8,20,21]. Among patients undergoing NACT, patients with PCV UC on cystectomy were more than twice as likely to have extravesical disease as patients with NV UC (80% vs. 35%). Our results herein liken PCV to such aggressive variants, as evidenced by higher LN involvement and LN density, higher pathologic staging, and poor DSS and OS. A number of patients with PCV UC on cystectomy were noted to have other variant histology concurrently. Additionally, a small number of patients with NV UC or nonPCV variant UC on TURBT went on to develop PCV on final cystectomy pathology. Presence of variant histology at cystectomy would likely imply either its unidentified presence on TURBT or divergent differentiation at some point thereafter. Divergent differentiation is well described with molecular genetic evidence supporting a close relationship between urothelial and divergent elements that may give rise to pure or mixed-variant UC [22]. Patients with any degree of PCV on cystectomy were analyzed in this cohort, and the variant with most composition was designated as the primary. Pure PCV without any other variant on TURBT or cystectomy pathology was identified in 18 patients. An analysis with this smaller number of patients produced similar results, although less meaningful comparisons could be made given the cohort size. The same conclusions are drawn from both plasmacytoid cohorts, whether identified as the sole variant or as mixed. The presence of any component PCV appears to confer a dismal prognosis. As this is a retrospective, single institution study, there are inherent limitations. It is difficult to identify an optimal therapeutic protocol from such a retrospective study and the multivariate analyses should be cautiously interpreted given the small size. It remains unclear whether this aggressive variant responds to chemotherapy, and whether any such response contributes to survival. PCV is rare, with a limited number of cases to present here, which is a limitation to our study. Owing to the limited number of patients with PCV UC and because we believe that any component of PCV histology would drive oncologic outcomes, we chose not to use a cutoff for component of PCV in the specimen, which means that some of our patients had 5% PCV, whereas others had 100%. Despite these limitations, we feel strongly that this study represents an important addition to the literature as it is one of the largest series of plasmacytoid cystectomy cases yet presented, with insight into clinocopathologic staging and survival outcomes. We chose to use patients with NV UC with muscle-invasive disease only on TURBT as the comparison cohort to accurately demonstrate the disparity in outcomes between patients with PCV UC and patients with NV UC, which we believe our study succeeds in doing. The low incidence of patients with PCV UC precludes the conduct of prospective clinical studies. In fact most scholarly publications regarding variant histology UC are generated from small retrospective series. The need
838
H.Z. Kaimakliotis et al. / Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838
for collaborative studies and the development of tissue repositories to assist in significant prospective trial and biomolecular investigational research is imperative. Our findings join the growing body of literature that defines PCV UC as an extremely aggressive variant, and we suggest that timely and aggressive surgical treatment be the mainstay in management for PCV until novel chemotherapeutic regimens and treatment algorithms are developed. 5. Conclusion PCV is an aggressive subset of UC of the bladder, predicting both extravesical disease and poor survival. The role of NACT is yet to be determined, and early definitive surgery remains the mainstay of therapy. Upstaging of non– muscle-invasive to muscle-invasive and extravesical disease may be evidence against differentiating between non– muscle-invasive and muscle-invasive disease. Instead, any evidence of PCV on TURBT may warrant aggressive therapy, and clinical understaging may be the rule. Appendix A. Supplementary materials Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j. urolonc.2014.03.008.
References [1] Sahin AA, Myhre M, Ro JY, et al. Plasmacytoid transitional cell carcinoma. Report of a case with initial presentation mimicking multiple myeloma. Acta Cytol 1991;35:277–80. [2] Dayyani F, Czerniak BA, Sircar K, et al. Plasmacytoid urothelial carcinoma, a chemosensitive cancer with poor prognosis, and peritoneal carcinomatosis. J Urol 2013;189:1656–61. [3] Fritsche HM, Burger M, Denzinger S, et al. Plasmacytoid urothelial carcinoma of the bladder: histological and clinical features of 5 cases. J Urol 2008;180:1923–7. [4] Lopez-Beltran A, Requena MJ, Montironi R, et al. Plasmacytoid urothelial carcinoma of the bladder. Hum Pathol 2009;40:1023–8. [5] Nigwekar P, Tamboli P, Amin MB, et al. Plasmacytoid urothelial carcinoma: detailed analysis of morphology with clinicopathologic correlation in 17 cases. Am J Surg Pathol 2009;33:417–24.
[6] Ricardo-Gonzalez RR, Nguyen M, Gokden N, et al. Plasmacytoid carcinoma of the bladder: a urothelial carcinoma variant with a predilection for intraperitoneal spread. J Urol 2012;187:852–5. [7] Ro JY, Shen SS, Lee HI, et al. Plasmacytoid transitional cell carcinoma of urinary bladder: a clinicopathologic study of 9 cases. Am J Surg Pathol 2008;32:752–7. [8] Willis DL, Porten SP, Kamat AM. Should histologic variants alter definitive treatment of bladder cancer? Curr Opin Urol 2013;23:435–43. [9] Hayashi T, Tanigawa G, Fujita K, et al. Two cases of plasmacytoid variant of urothelial carcinoma of urinary bladder: systemic chemotherapy might be of benefit. Int J Clin Oncol 2011;16:759–62. [10] Keck B, Wach S, Stoehr R, et al. Plasmacytoid variant of bladder cancer defines patients with poor prognosis if treated with cystectomy and adjuvant cisplatin-based chemotherapy. BMC Cancer 2013;13:71. [11] Keck B, Stoehr R, Wach S, et al. The plasmacytoid carcinoma of the bladder—rare variant of aggressive urothelial carcinoma. Int J Cancer 2011;129:346–54. [12] Cheng L, Lopez-Beltran A, Bostwick DG. Histologic variants of urothelial carcinoma. Cheng L, Lopez-Beltran A, Bostwick DG, editors. Bladder Pathology. Hoboken, NJ: Wiley-Blackwell; 2012;257–60. [13] Grossman HB, Natale RB, Tangen CM, et al. Neoadjuvant chemotherapy plus cystectomy compared with cystectomy alone for locally advanced bladder cancer. N Engl J Med 2003;349:859–66. [14] Vieweg J, Gschwend JE, Herr HW, et al. Pelvic lymph node dissection can be curative in patients with node positive bladder cancer. J Urol 1999;161:449–54. [15] Abd El-Latif A, Watts KE, Elson P, et al. The sensitivity of initial transurethral resection or biopsy of bladder tumor(s) for detecting bladder cancer variants on radical cystectomy. J Urol 2013;189:1263–7. [16] Shah RB, Montgomery JS, Montie JE, et al. Variant (divergent) histologic differentiation in urothelial carcinoma is under-recognized in community practice: impact of mandatory central pathology review at a large referral hospital. Urol Oncol 2013;31:1650–5. [17] Sangoi AR, Beck AH, Amin MB, et al. Interobserver reproducibility in the diagnosis of invasive micropapillary carcinoma of the urinary tract among urologic pathologists. Am J Surg Pathol 2010;34:1367–76. [18] Denzinger S, Fritsche HM, Otto W, et al. Early versus deferred cystectomy for initial high-risk pT1G3 urothelial carcinoma of the bladder: do risk factors define feasibility of bladder-sparing approach? Eur Urol 2008;53:146–52. [19] Hautmann RE, Volkmer BG, Gust K. Quantification of the survival benefit of early versus deferred cystectomy in high-risk non-muscle invasive bladder cancer (T1 G3). World J Urol 2009;27:347–51. [20] Kamat AM, Gee JR, Dinney CP, et al. The case for early cystectomy in the treatment of nonmuscle invasive micropapillary bladder carcinoma. J Urol 2006;175:881–5. [21] Kamat AM, Dinney CP, Gee JR, et al. Micropapillary bladder cancer: a review of the University of Texas M. D. Anderson Cancer Center experience with 100 consecutive patients. Cancer 2007;110:62–7. [22] Shanks JH, Iczkowski KA. Divergent differentiation in urothelial carcinoma and other bladder cancer subtypes with selected mimics. Histopathology 2009;54:885–900.
Original article
Plasmacytoid variant urothelial bladder cancer: Is it time to update the treatment paradigm? Hristos Z. Kaimakliotis, M.D.1,a,*, M. Francesca Monn, M.D., M.P.H.1,a, K. Clint Cary, M.D., M.P.H.a, Jose A. Pedrosa, M.D.a, Kevin Rice, M.D.a, Timothy A. Masterson, M.D.a, Thomas A. Gardner, M.D.a, Noah M. Hahn, M.D.b, Richard S. Foster, M.D.a, Richard Bihrle, M.D.a, Liang Cheng, M.D.c, Michael O. Koch, M.D.a b
a Department of Urology, Indiana University School of Medicine, Indianapolis, IN Department of Genitourinary Medical Oncology, Indiana University School of Medicine, Indianapolis, IN c Department of Pathology, Indiana University School of Medicine, Indianapolis, IN
Received 31 December 2013; received in revised form 12 February 2014; accepted 8 March 2014
Abstract Objectives: Plasmacytoid variant (PCV) urothelial cancer (UC) of the bladder is rare, with poor clinical outcomes. We sought to identify factors that may better inform expectations of tumor behavior and improve management options in patients with PCV UC. Materials and methods: A retrospective analysis of the Indiana University Bladder Cancer Database between January 2008 and June 2013 was performed comparing 30 patients with PCV UC at cystectomy to 278 patients with nonvariant (NV) UC at cystectomy who underwent surgery for muscle-invasive disease. Multivariable logistic regression was used to assess precystectomy variables associated with non–organ-confined disease at cystectomy and Cox regression analysis to assess variables associated with mortality. Results: Patients with PCV UC who were diagnosed with a higher stage at cystectomy (73% pT3-4 vs. 40%, P ¼ 0.001) were more likely to have lymph node involvement (70% vs. 25%, P o 0.001), and positive surgical margins were found in 40% of patients with PCV UC vs. 10% of patients with NV UC (P o 0.001). Median overall survival and disease-specific survival were 19 and 22 months for PCV, respectively. Median overall survival and disease-specific survival had not been reached for NV at 68 months (P o 0.001). Presence of PCV UC on transurethral resection of bladder tumor was associated with non–organ-confined disease (odds ratio ¼ 4.02; 95% CI: 1.06–15.22; P ¼ 0.040), and PCV at cystectomy was associated with increased adjusted risk of mortality (hazard ratio ¼ 2.1; 95% CI: 1.2–3.8; P ¼ 0.016). Conclusions: PCV is an aggressive UC variant, predicting non–organ-confined disease and poor survival. Differentiating between non– muscle- and muscle-invasive disease in patients with PCV UC seems less important than the aggressive nature of this disease. Instead, any evidence of PCV on transurethral resection of bladder tumor may warrant aggressive therapy. r 2014 Elsevier Inc. All rights reserved.
Keywords: Bladder cancer; Plasmacytoid variant; Variant histology; Cystectomy
1. Introduction Plasmacytoid variant (PCV) urothelial cancer (UC) of the bladder is an increasingly defined, aggressive variant of UC first described in 1991 as “similar to the plasma cells seen in myeloma” [1]. Since then, knowledge regarding PCV has 1
Contributed equally to manuscript.
Corresponding author. Tel.:þ1-203-530-0445; fax: þ1-317-278-0499. E-mail address: [email protected] (H.Z. Kaimakliotis).
*
http://dx.doi.org/10.1016/j.urolonc.2014.03.008 1078-1439/r 2014 Elsevier Inc. All rights reserved.
increased, including recognition of its advanced stage at cystectomy, rate of metastasis, and clinical implications [2–7]. Optimal management strategies remain to be determined with much of the literature suggesting that PCV should be treated similarly to other aggressive variant histology, using early surgical intervention [8], while other literature point toward a greater role for novel chemotherapuetic regimens [9,10]. The largest case series range from 15 to 32 patients and begin to construct an understanding of PCV; however the need for larger, more comprehensive
834
H.Z. Kaimakliotis et al. / Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838
series remains ever present [2,5,6,11]. As we learn more about PCV, it becomes increasingly important to continue reporting outcomes so that more conclusive management guidelines can be established. We present our experience of patients diagnosed with any degree of PCV component at the time of cystectomy, and by exploring their clinical course, we hope to gain insight for improved treatment plans.
2. Materials and methods Institutional review board approval was obtained from Indiana University for this study. Using the Indiana University Bladder Cancer Database, a retrospective analysis of all patients who underwent cystectomy with curative intent between 2008 and June 2013 was performed. Patients with metastatic disease or lymphadenopathy greater than 1.5 cm on cross-sectional imaging at time of diagnosis were excluded as were patients undergoing bladder preservation protocols. Patients with pure squamous, small cell, rhabdomyosarcoma, or adenocarcinoma at transurethral resection of bladder tumor (TURBT) or cystectomy were also excluded from analysis. PCV was defined based on histology, with individual tumor cells exhibiting a prominent eccentrically placed nucleus and abundant eosinophilic cytoplasm, reminiscent of plasma cells. Most neoplastic cells exhibited nuclei of low to intermediate nuclear grade with occasional nuclear pleomorphism, with CD138 expression in some cases. Interobserver variability is always an issue for variant calling, but diagnosis of variant still relies on morphological rather than immunohistochemical characteristics [12]. All TURBT specimens underwent central review at Indiana University before cystectomy to ensure accuracy of diagnosis. Pathologic assessment of all cases was performed by dedicated genitourinary pathologists. Patients with PCV UC were defined as those whose cystectomy pathology included any component of PCV, regardless of overall percentage or mixed-variant status. As the primary aim of the project was to compare patients with PCV UC with patients with nonvariant (NV) UC, patients who had variant histology at cystectomy without the presence of PCV were eliminated (n ¼ 137). Additionally, as variant histology is often considered understaged on TURBT, we chose to use only patients with NV UC with muscle-invasive disease on TURBT as the comparison population (n ¼ 278). Although we chose to eliminate patients with PCV UC on TURBT who did not develop PCV on final cystectomy pathology (n ¼ 5), we performed a descriptive analysis of all patients with PCV who underwent TURBT to examine clinicopathologic outcomes. Data were obtained for each patient regarding sex, age, race, TURBT stage and histology, neoadjuvant and adjuvant chemotherapy, date of cystectomy, stage, histology, lymph node (LN) status, and non–organ-confined disease. Non–organ-confined disease was defined as pathologic cystectomy stage T3, T4, or LN involvement.
Survival data were obtained through the Indiana University Cancer Registry, which gathers vital statistic information using medical records, primary care physicians, the social security death index, and electronic record searches. This registry is updated annually with greater than 90% success rate of follow-up for patients treated at Indiana University. When survival data were not available within the past month, we evaluated patient survival status using the Social Security Death Index and used disease status at time of last follow-up for disease-specific survival (DSS). We used the Pearson chi-square and Fisher exact tests for categorical variables and the Student t test and MannWhitney U tests for continuous variables to perform descriptive analysis comparing the PCV and NV groups. Multivariable logistic regression was performed to examine precystectomy factors associated with developing non– organ-confined disease on radical cystectomy. The model was built using forward and back stepwise methodology with an accepted P o 0.1. Variables included in the stepwise regression were sex, age, TURBT histology (PCV vs. NV), lymphovascular invasion, neoadjuvant chemotherapy (NACT), and number of nodes removed at cystectomy. The final model included TURBT histology, lymphovascular invasion, and NACT. We chose to include age and sex in the final model as well. Kaplan-Meier overall survival (OS) and DSS curves were generated and log-rank tests performed to evaluate differences between the PCV and NV populations. Cox proportional hazards regression was used to evaluate characteristics associated with allcause mortality. Factors used in the stepwise regression were age, sex, PCV histology, any systemic chemotherapy, and non–organ-confined disease on cystectomy. Variables accepted for use in the model were PCV histology, systemic chemotherapy, and non–organ-confined disease. Age and sex were also included in the final model as well. The proportional hazards assumption was tested using Schoenfeld residual plots and all variables included in the models did not violate the assumption. Time zero was defined as the date of radical cystectomy. A priori, P o 0.05 was considered statistically significant. Microsoft Excel (Microsoft Inc) and Stata version 12.1 (Stata Corp LP, College Station, TX) were used for all statistical analyses.
3. Results Of the 624 patients with UC who underwent radical cystectomy with curative intent, 30 (4.8%) were diagnosed with any degree of PCV on cystectomy specimen, 25 with primary PCV (450% of total variant histology), and 5 with secondary or tertiary components. A total of 137 (22.0%) patients were noted to have variant histology other than PCV and as such were excluded from the analysis. In total, 457 (73.2%) patients were diagnosed with pure urothelial histology, or NV histology. Of the 457 patients with NV
H.Z. Kaimakliotis et al. / Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838
UC, 278 were muscle-invasive on TURBT and thus were the group used for comparison in all analyses. There were no demographic differences between patients with PCV UC and patients with NV UC (Table 1). Patients with PCV UC had higher stage at time of cystectomy, with 97% demonstrating muscle-invasive disease compared with 69% of patients with NV UC (P ¼ 0.001). Non–organ-confined and LN-positive disease was more common among patients with PCV UC (P ¼ 0.001 and o0.001, respectively). Median (mean) LN density for patients with PCV UC was 23.1 (34.5) compared with 11.8 (21.1) in patients with NV UC (P ¼ 0.032) (Table 1). NACT appeared to have little effect on patients with PCV UC. Of the 5 patients with PCV UC who received NACT, 80% (4 of 5 patients) had LN involvement and 80% had ZpT3 disease, compared with 21% (13 of 63 patients) with NV UC who had LN involvement and 24% (15 of 63 patients) with ZpT3 disease following NACT (P ¼ 0.003 and 0.007, respectively). Multivariable logistic regression was performed to determine precystectomy factors associated with non– organ-confined disease at cystectomy (Table 2). Patients
Table 1 Characteristics of patients with plasmacytoid variant UC and patients with nonvariant UC at time of cystectomy P-valuea
Plasmacytoid variant, n (%)c
Nonvariant, n (%)c
Patients, n Mean age (SD), y Sex (female) Race (white) TURBT LVI TURBT CIS
30 66.2 (9.4) 6 (20.0) 27 (90.0) 1 (3.3) 8/28 (28.6)
278 66.6 (11.2) 58 (20.9) 265 (95.3) 27 (9.7) 59/255 (23.1)
Cystectomy stage rpT1 pT2 pT3 pT4
1 (3.3) 7 (23.3) 10 (33.3) 12 (40.0)
86 80 82 30
24 (80.0)
132 (47.5)
12 (40.0) 10 (33.3) 15 (50.0) 21 (70.0) 21.5 (16–36)
27 (9.7) o0.001 7 (2.5) o0.001 150 (54.0) 0.680 69 (24.8) o0.001 21.5 (14–31) 0.416
23.1 (14–42)
11.8 (6–27)
5 (16.7) 11 (36.7)
63 (22.7) 34 (12.2)
Non–organ-confined disease Positive surgical margin Positive ureteral margin Cystectomy-associated CIS Positive lymph nodes Median lymph nodes examined (IQR) Median lymph node density (IQR)b Neoadjuvant chemotherapy Adjuvant chemotherapy
(30.9) (28.8) (29.5) (10.8)
0.842 0.912 0.212 0.248 0.521 o0.001
0.001
0.038 0.452 o0.001
SD ¼ standard deviation; LVI ¼ lymphovascular invasion; CIS ¼ carcinoma in situ; IQR ¼ interquartile range. a P-value represents the chi-square test, Student t test, or the MannWhitney U test when median (IQR) is presented. b Lymph node (LN) density is calculated as the number of positive LN/ total number of LN in patients with positive LN. c Numbers in parentheses represent percentage except where otherwise noted.
835
Table 2 Multivariable logistic regression of clinical characteristics associated with non–organ-confineda disease at the time of cystectomy Characteristicsb
Odds ratio (95% CI)
P value
Age, y Sex (female) TURBT plasmacytoidc TURBT lymphovascular invasion Neoadjuvant chemotherapy
1.01 0.70 4.02 2.30 0.53
0.402 0.232 0.040 0.053 0.027
(0.99–1.03) (0.40–1.25) (1.06–15.22) (0.99–5.35) (0.30–0.93)
Non–organ-confined disease is defined as LNþ or ZpT-stage 3. Characteristics controlled for in the multivariable logistic regression model are those included in the chart. c TURBT plasmacytoid is PCV vs. NV. a
b
with PCV UC identified on TURBT had 4 times the odds of developing non–organ-confined disease compared with patients NV UC (P ¼ 0.040), and NACT was associated with decreased odds of having non–organ-confined disease (odds ratio ¼ 0.53, 95% CI: 0.30–0.93). Median follow-up time for all patients was 30 (interquartile range: 16–48) months. Both OS and DSS for patients with PCV UC were significantly worse than patients with NV UC (P o 0.001 and 0.002) (Fig. A and B). The median OS for PCV was 19 months (interquartile range: 11–28), whereas median OS for patients with NV UC was not yet reached at 68 months. Median DSS for patients with PCV UC was 22 months, whereas it was not yet reached for the patients with NV UC. After adjusting for age, sex, any systemic chemotherapy, and non–organ-confined disease, PCV histology was associated with a 2-fold increased adjusted risk of all-cause mortality (95% CI: 1.2–3.8, P ¼ 0.029) (Table 3). Significantly, when examining the full cohort of 624 patients in a separate subanalysis, we found 15 patients with PCV UC diagnosed on TURBT. The results of cystectomy pathology were negative for PCV UC for 5 of these patients; 4 had NV UC and 1 went on to develop nested variant. Among the 15 patients diagnosed with PCV UC on TURBT, 4 patients underwent NACT, 1 of whom had a complete response with pT0N0, another had pT2N0, and 2 had ZpT3Nþ. Overall, 73% of all patients with PCV UC who underwent TURBT developed non–organ-confined disease compared with 43% of patients with NV UC (P ¼ 0.032). Also, 4 of the 5 patients diagnosed with cT1 PCV UC on TURBT were upstaged to ZpT3 at the time of cystectomy, and 60% (3 of 5 patients) were found to have LN involvement on final pathology. This is in comparison with 22% (32 of 143 patients) of patients with cT1 NV UC on TURBT who were upstaged to ZpT3 (P ¼ 0.013), and 17% (25 of 143 patients) having LN involvement at time of cystectomy (P ¼ 0.047). 4. Discussion PCV is a recently described subset of UC [1], and we are only beginning to appreciate its clinicopathologic implications.
836
H.Z. Kaimakliotis et al. / Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838
Fig. (A) Overall survival and (B) disease-specific survival. There is a significant difference in overall and disease-specific survival for the population with nonvariant histology and population with plasmacytoid variant UC (log-rank test P o 0.001 each). Median overall survival in the population with plasmacytoid variant UC was 19 months (IQR: 11–28). Median disease-specific survival in the population with plasmacytoid variant UC was 22 months. IQR ¼ interquartile range. (Color version of figure is available online.)
Dayyani et al. [2] propose PCV to be a chemosensitive cancer with a poor prognosis, noting pathologic downstaging in 80% of patients (n ¼ 4 of 5) treated with NACT and a complete response in 3. It remains unclear that patients were rendered pT0 at time of TUR, and the authors acknowledge short-lived response rates among these patients. Further, the poor response to NACT in our group compared with those of the MD Anderson study is unclear; this could be related to variation in TURBT resection, differing NACT regimens, or merely small sample sizes. The Southwest Oncology Group 8710 study found a clinically meaningful improvement in survival among patients with locally advanced UC who received NACT over
cystectomy alone [13], although fewer than 20% of patients with locally advanced node-positive bladder cancer survive for 5 years [14]. The higher incidence of extravesical disease and LN density in our PCV cohort would likely render these patients to a lesser response rate than the effect noted in the Southwest Oncology Group trial. Herein, patients with PCV UC undergoing cystectomy, regardless of systemic therapy, appear to do much worse than patients with NV UC, with few patients with PCV UC alive beyond 24 months. On multivariable analysis, presence of PCV compared with NV at cystectomy confers a 2-fold increased risk of death. The clinical significance of this may be contingent on
H.Z. Kaimakliotis et al. / Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838 Table 3 Multivariable Cox proportional hazards regression model of characteristics associated with overall mortality Characteristica
Hazard ratio (95% CI)
P value
Age, y Sex (female) Plasmacytoid variantb Systemic chemotherapy Non–organ-confined diseasec
1.01 0.99 2.07 0.57 4.87
0.534 0.978 0.016 0.029 o0.001
(0.99–1.03) (0.60–1.65) (1.15–3.75) (0.35–0.95) (2.91–8.16)
a Characteristics controlled for in the Cox regression model are those included in the chart. b Plasmacytoid vs. nonvariant histology. c Non–organ-confined disease is defined as LNþ or ZpT-stage 3. ADDIN EN.REFLIST
PCV recognition at TURBT, as there is poor concordance between TURBT and cystectomy variant histology [15]. Despite formal pathologic diagnostic criteria and increased awareness of variant histology, most UCs are diagnosed in low-volume community practice facilities, where awareness of variant UC may be less comprehensive [16]. The less frequent identification at low-volume centers results in higher potential for misclassification, which may lead to delay in therapy or inappropriate treatment algorithms [16,17]. Our results also show that 80% of patients with PCV UC and non–muscle-invasive disease on TURBT were upstaged to ZpT3 at cystectomy and that 60% had LN involvement. This is nearly 4 times the incidence encountered in cT1 NV cases. This finding is even more impressive when considering that cystectomy for all 5 patients with T1 PCV noted on TURBT was expedited without NACT or intravesical therapy, whereas most patients with NV UC with non– muscle-invasive disease were treated with cystectomy after failing intravesical therapy. The argument against delaying definitive surgery in patients with presumed organ-confined PCV is further supported by our finding that 80% (24 of 30 patients) of cystectomy-confirmed patients with PCV UC had extravesical disease compared with 47% (132 of 278 patients) in the NV population. Furthermore, after controlling for lymphovascular invasion and NACT, patients with PCV UC on TURBT still had a 4-fold increased odds of extravesical disease at cystectomy than patients with NV UC with muscle-invasive disease. The role of early cystectomy in high-grade non–muscle-invasive T1 bladder cancer has been studied, with 10-year cancer-specific survival rates of 78% for early cystectomy vs. 51% to 65% for deferred cystectomy [18,19]. Our demonstration of the high rate of upstaging from non–muscle-invasive to muscle-invasive disease and LN involvement in the population with PCV UC may argue against differentiating patients with PCV UC between invasive or not. Instead, any evidence of PCV histology during TURBT may warrant aggressive surgical therapy. Certain variants of UC, such as micropapillary and sarcomatoid, have been hypothesized to be chemoresistant
837
with worse overall clinical outcomes than patients with NV UC [8,20,21]. Among patients undergoing NACT, patients with PCV UC on cystectomy were more than twice as likely to have extravesical disease as patients with NV UC (80% vs. 35%). Our results herein liken PCV to such aggressive variants, as evidenced by higher LN involvement and LN density, higher pathologic staging, and poor DSS and OS. A number of patients with PCV UC on cystectomy were noted to have other variant histology concurrently. Additionally, a small number of patients with NV UC or nonPCV variant UC on TURBT went on to develop PCV on final cystectomy pathology. Presence of variant histology at cystectomy would likely imply either its unidentified presence on TURBT or divergent differentiation at some point thereafter. Divergent differentiation is well described with molecular genetic evidence supporting a close relationship between urothelial and divergent elements that may give rise to pure or mixed-variant UC [22]. Patients with any degree of PCV on cystectomy were analyzed in this cohort, and the variant with most composition was designated as the primary. Pure PCV without any other variant on TURBT or cystectomy pathology was identified in 18 patients. An analysis with this smaller number of patients produced similar results, although less meaningful comparisons could be made given the cohort size. The same conclusions are drawn from both plasmacytoid cohorts, whether identified as the sole variant or as mixed. The presence of any component PCV appears to confer a dismal prognosis. As this is a retrospective, single institution study, there are inherent limitations. It is difficult to identify an optimal therapeutic protocol from such a retrospective study and the multivariate analyses should be cautiously interpreted given the small size. It remains unclear whether this aggressive variant responds to chemotherapy, and whether any such response contributes to survival. PCV is rare, with a limited number of cases to present here, which is a limitation to our study. Owing to the limited number of patients with PCV UC and because we believe that any component of PCV histology would drive oncologic outcomes, we chose not to use a cutoff for component of PCV in the specimen, which means that some of our patients had 5% PCV, whereas others had 100%. Despite these limitations, we feel strongly that this study represents an important addition to the literature as it is one of the largest series of plasmacytoid cystectomy cases yet presented, with insight into clinocopathologic staging and survival outcomes. We chose to use patients with NV UC with muscle-invasive disease only on TURBT as the comparison cohort to accurately demonstrate the disparity in outcomes between patients with PCV UC and patients with NV UC, which we believe our study succeeds in doing. The low incidence of patients with PCV UC precludes the conduct of prospective clinical studies. In fact most scholarly publications regarding variant histology UC are generated from small retrospective series. The need
838
H.Z. Kaimakliotis et al. / Urologic Oncology: Seminars and Original Investigations 32 (2014) 833–838
for collaborative studies and the development of tissue repositories to assist in significant prospective trial and biomolecular investigational research is imperative. Our findings join the growing body of literature that defines PCV UC as an extremely aggressive variant, and we suggest that timely and aggressive surgical treatment be the mainstay in management for PCV until novel chemotherapeutic regimens and treatment algorithms are developed. 5. Conclusion PCV is an aggressive subset of UC of the bladder, predicting both extravesical disease and poor survival. The role of NACT is yet to be determined, and early definitive surgery remains the mainstay of therapy. Upstaging of non– muscle-invasive to muscle-invasive and extravesical disease may be evidence against differentiating between non– muscle-invasive and muscle-invasive disease. Instead, any evidence of PCV on TURBT may warrant aggressive therapy, and clinical understaging may be the rule. Appendix A. Supplementary materials Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j. urolonc.2014.03.008.
References [1] Sahin AA, Myhre M, Ro JY, et al. Plasmacytoid transitional cell carcinoma. Report of a case with initial presentation mimicking multiple myeloma. Acta Cytol 1991;35:277–80. [2] Dayyani F, Czerniak BA, Sircar K, et al. Plasmacytoid urothelial carcinoma, a chemosensitive cancer with poor prognosis, and peritoneal carcinomatosis. J Urol 2013;189:1656–61. [3] Fritsche HM, Burger M, Denzinger S, et al. Plasmacytoid urothelial carcinoma of the bladder: histological and clinical features of 5 cases. J Urol 2008;180:1923–7. [4] Lopez-Beltran A, Requena MJ, Montironi R, et al. Plasmacytoid urothelial carcinoma of the bladder. Hum Pathol 2009;40:1023–8. [5] Nigwekar P, Tamboli P, Amin MB, et al. Plasmacytoid urothelial carcinoma: detailed analysis of morphology with clinicopathologic correlation in 17 cases. Am J Surg Pathol 2009;33:417–24.
[6] Ricardo-Gonzalez RR, Nguyen M, Gokden N, et al. Plasmacytoid carcinoma of the bladder: a urothelial carcinoma variant with a predilection for intraperitoneal spread. J Urol 2012;187:852–5. [7] Ro JY, Shen SS, Lee HI, et al. Plasmacytoid transitional cell carcinoma of urinary bladder: a clinicopathologic study of 9 cases. Am J Surg Pathol 2008;32:752–7. [8] Willis DL, Porten SP, Kamat AM. Should histologic variants alter definitive treatment of bladder cancer? Curr Opin Urol 2013;23:435–43. [9] Hayashi T, Tanigawa G, Fujita K, et al. Two cases of plasmacytoid variant of urothelial carcinoma of urinary bladder: systemic chemotherapy might be of benefit. Int J Clin Oncol 2011;16:759–62. [10] Keck B, Wach S, Stoehr R, et al. Plasmacytoid variant of bladder cancer defines patients with poor prognosis if treated with cystectomy and adjuvant cisplatin-based chemotherapy. BMC Cancer 2013;13:71. [11] Keck B, Stoehr R, Wach S, et al. The plasmacytoid carcinoma of the bladder—rare variant of aggressive urothelial carcinoma. Int J Cancer 2011;129:346–54. [12] Cheng L, Lopez-Beltran A, Bostwick DG. Histologic variants of urothelial carcinoma. Cheng L, Lopez-Beltran A, Bostwick DG, editors. Bladder Pathology. Hoboken, NJ: Wiley-Blackwell; 2012;257–60. [13] Grossman HB, Natale RB, Tangen CM, et al. Neoadjuvant chemotherapy plus cystectomy compared with cystectomy alone for locally advanced bladder cancer. N Engl J Med 2003;349:859–66. [14] Vieweg J, Gschwend JE, Herr HW, et al. Pelvic lymph node dissection can be curative in patients with node positive bladder cancer. J Urol 1999;161:449–54. [15] Abd El-Latif A, Watts KE, Elson P, et al. The sensitivity of initial transurethral resection or biopsy of bladder tumor(s) for detecting bladder cancer variants on radical cystectomy. J Urol 2013;189:1263–7. [16] Shah RB, Montgomery JS, Montie JE, et al. Variant (divergent) histologic differentiation in urothelial carcinoma is under-recognized in community practice: impact of mandatory central pathology review at a large referral hospital. Urol Oncol 2013;31:1650–5. [17] Sangoi AR, Beck AH, Amin MB, et al. Interobserver reproducibility in the diagnosis of invasive micropapillary carcinoma of the urinary tract among urologic pathologists. Am J Surg Pathol 2010;34:1367–76. [18] Denzinger S, Fritsche HM, Otto W, et al. Early versus deferred cystectomy for initial high-risk pT1G3 urothelial carcinoma of the bladder: do risk factors define feasibility of bladder-sparing approach? Eur Urol 2008;53:146–52. [19] Hautmann RE, Volkmer BG, Gust K. Quantification of the survival benefit of early versus deferred cystectomy in high-risk non-muscle invasive bladder cancer (T1 G3). World J Urol 2009;27:347–51. [20] Kamat AM, Gee JR, Dinney CP, et al. The case for early cystectomy in the treatment of nonmuscle invasive micropapillary bladder carcinoma. J Urol 2006;175:881–5. [21] Kamat AM, Dinney CP, Gee JR, et al. Micropapillary bladder cancer: a review of the University of Texas M. D. Anderson Cancer Center experience with 100 consecutive patients. Cancer 2007;110:62–7. [22] Shanks JH, Iczkowski KA. Divergent differentiation in urothelial carcinoma and other bladder cancer subtypes with selected mimics. Histopathology 2009;54:885–900.