Invasive low-grade papillary urothelial carcinoma: an immunohistochemical study of 26 cases

Invasive low-grade papillary urothelial carcinoma: an immunohistochemical study of 26 cases

Human Pathology (2015) 46, 1836–1841 Original contribution Invasive low-grade papillary urothelial carcinoma: an im...

860KB Sizes 12 Downloads 110 Views

Human Pathology (2015) 46, 1836–1841

Original contribution

Invasive low-grade papillary urothelial carcinoma: an immunohistochemical study of 26 cases☆,☆☆ Wei Tian MD, Jonathan I. Epstein MD ⁎ Johns Hopkins Hospital, Baltimore, MD 21231 Received 28 May 2015; revised 27 July 2015; accepted 29 July 2015

Keywords: Urothelial carcinoma; PTEN; Ki-67; E-cadherin; p53

Summary Invasive low-grade papillary urothelial carcinoma (LGPUC) is rare. We studied the immunohistochemical (IHC) expressions of CK20, p53, E-cadherin, phosphatase and tension homolog (PTEN), and Ki-67 in both noninvasive and invasive components in 26 cases. In the noninvasive component of LGPUC, 81% showed CK20 expression, and 50% showed p53 labeling. There was a wide range of Ki-67 labeling from less than 5% to 70%. All cases had intact PTEN except 1 that showed focal clonal PTEN loss in both noninvasive and subjacent invasive components. All cases had preserved strong and diffuse E-cadherin expression in both noninvasive and invasive components. There was no significant change between the noninvasive and invasive components in the IHC labeling of these markers, although 7 (33%) of 21 cases showed decreased CK20 expression to a certain extent in the invasive component. Only 2 cases showed significant increase of p53 expression in the invasive component compared with the noninvasive component. Two cases showed increase of Ki-67 labeling from less than 5% in the noninvasive to 20% and 40%, respectively, in the invasive component. High Ki-67 labeling was present in a significant portion of invasive LGPUC cases in both noninvasive and invasive components, much higher than in previous studies of noninvasive LGPUC. Whether higher Ki67 in these cases is associated with more aggressive disease warrants further study. In general, it is not very helpful to use immunostains in diagnosis and predicting prognosis. This is the largest series to study the IHC characteristics of this entity. © 2015 Elsevier Inc. All rights reserved.

1. Introduction Bladder cancer is the second most common genitourinary tract malignancy, and urothelial carcinomas (UCs) represent 90% of all primary bladder cancer. UCs carry significant

☆ Competing interests: The authors disclose no significant relationship with, or financial interest in, any commercial companies pertaining to this article. ☆☆ Funding/Support: There are no funding disclosures. ⁎ Corresponding author at: The Johns Hopkins Hospital, The Weinberg Building, Rm 2242, 401 N. Broadway Street, Baltimore, MD 21231. E-mail address: [email protected] (J. I. Epstein). 0046-8177/© 2015 Elsevier Inc. All rights reserved.

mortality and morbidity. The clinicopathological parameters such as grade, stage, and lymph node status are the most important prognostic factors to predict recurrence, progression, and survival. Low-grade papillary UC (LGPUC) has a recurrence rate of 34% to 72%, which is slightly lower or similar to that of high-grade UC (HGPUC) (43%-74%). However, LGPUC has lower disease progression rate (4%-18% compared with 28% in HGPUC) and lower disease mortality (0%-4%), whereas HGPUCs account for most of the cancer-related deaths among papillary carcinomas [1-9]. As most of the LGPUCs lack invasion (pTa), there are scant data on potential differences in the noninvasive and

Invasive low-grade urothelial carcinoma invasive components. Given the rarity of finding invasion associated with LGPUC, we have noted in our consult material that pathologists often have difficulty in establishing the diagnosis of invasion in this setting. Pathologists, when they do recognize invasion, tend to diagnose it as high grade despite overlying low-grade noninvasive papillary cancer and similarly low-grade morphology in the invasive component. The primary aim of this study was to investigate the immunohistochemical (IHC) expression of various biomarkers that, based on either their expression in flat lesions of the bladder or what is known about the markers in general, had the potential to be useful diagnostically if they were different in the invasive component. In addition, if differentially expressed, would these biomarkers have any prognostic value in helping to further understand the clinical behavior of this rare entity?

2. Materials and methods This study was approved by the Johns Hopkins Medicine Institutional Review Boards. A search using the Hopkins Pathology Database System from 2000 to 2014 identified 66 cases of invasive LGPUC (Inv LGPUC), of which 26 cases had tissue available for IHC studies, including 21 cases from the genitourinary pathology consult service and 5 from the Johns Hopkins Hospital. Criteria for diagnosing Inv LGPUC included low-grade cytology (low nuclear/cytoplasmic ratio and lacking pleomorphism/hyperchromasia/mitotic figures) with the usual morphologic features of invasion including single cells, small nests, retraction artifact, stromal reaction, and paradoxical differentiation. Variants of UC with lowgrade morphology, such as nested variant and large nested variant of UC, were not included in the study. Cases with only a minute focus of invasion not amenable to IHC analysis were excluded. Many of the cases were included in our previous study for morphologic and clinical characteristics of this entity by Toll and Epstein [2]. The materials were recollected for the current study. The pertinent hematoxylin and eosin–stained slides for all cases were reviewed by both authors, and diagnosis of Inv LGPUC was confirmed according criteria described in our prior study on this entity [9]. The following IHC analysis was performed at our institution using 4-μm-thick sections obtained from formalin-fixed, paraffin-embedded tissue (monoclonal antibodies ready to use unless specified): CK20 (KS20.8, prediluted; Dako, Carpinteria, CA), p53 (BP53-11, prediluted; Ventana, Tucson, AZ), E-cadherin (EP700Y, prediluted; Ventana), phosphatase and tension homolog (PTEN) (6H2.1, 1:100; Biocare, Concord, CA), MIB-1 (30-9, 1:1000; Ventana). Ki-67 labeling index was measured using the MIB-1 proliferation marker. IHC stains were performed on a Ventana BenchMark XT automated stainer (Ventana), with appropriate positive and negative controls. Immunostains were scored using a semiquantitative system considering extent and intensity of staining. Extent was graded as follows: 0, 0%; focal (F), 1% to 10% positive

1837 cells; patchy (P), 10% to 50% positive cells; diffuse (D), greater than 50% positive cells. For intensity, points were assigned as follows: 0, negative; 1+, weakly positive; 2+, moderately positive; 3+, strongly positive. p53 and CK20 were scored as follows: 0, negative; 1, weak (1+) and patchy (10% to b50%); 2, moderate (2+) and patchy (b50%) or weak (1+) and diffuse (N50%); and 3, moderate (2+) or strong (3+) and diffuse (N50%). For p53, we required score 2 and higher to be considered positive. Only moderate to strong Ki-67 nuclear stains were counted as positive.

3. Results There were 23 males and 3 females with a male predominance (88%). The median age at diagnosis was of 65 years (range, 45-92 years). The specimens consisted of 23 transurethral resection of the bladder tumors, 2 nephrectomies involving the renal pelvis, and 1 cystoprostatectomy. Two cases had definite muscularis propria invasion, 22 cases had lamina propria invasion, and 2 had muscle invasion indeterminate between muscularis mucosae and muscularis propria invasion. Diagnostic criteria for Inv LGPUC were described in our previous study (Fig. 1A) [2]. In the noninvasive component of LGPUC, CK20 expression was present in 21 cases (81%), of which 16 (61%) scored 2 and 3 (n = 7 and 9, respectively). In the invasive component, 11 (42%) scored 2 and 3. Comparing CK20 expression in the invasive versus noninvasive component, the results were as follows: unchanged (n = 18), decreased (n = 7) (Fig. 1B), and slightly increased (n = 1). For p53 labeling, 13 cases in the noninvasive component were scored 2 to 3 and considered positive (50%, score 2, n = 9; score 3, n = 4). Fifteen in the invasive component were positive (58%, score 2, n = 9; score 3, n = 6). Comparing p53 expression in noninvasive versus invasive component, the results were as follows: unchanged (n = 24) and significantly increased (n = 2) (Fig. 1C) (Table). Only a few cases (8%) showed significant gain of p53 in the invasive component. There was a wide range of Ki-67 labeling from less than 5% to 70%: Ki-67 less than or equal to 5% (n = 11, 42%), 6% to 30% (n = 8, 31%), 31% to 70% (n = 7, 27%). Two cases showed increased Ki-67 labeling (from b5% in the noninvasive to 20% and 40%, respectively, in the invasive component) (Fig. 1D); 2 showed a decrease (20%-10%, 50%-30% respectively) (Table), both with small foci of invasion. All cases had intact PTEN expression (Fig. 1E) except 1 that showed focal clonal PTEN loss in both the noninvasive and subjacent invasive components (Fig. 2A and B). All cases had preserved strong and diffuse E-cadherin expression in both the noninvasive and invasive components. The strong E-cadherin signals did spotlight the invasive nests or single cells in some difficult cases (Fig. 1F).


W. Tian, J. I. Epstein

Fig. 1 A, Inv LGPUC with inverted growth on top and invasion into lamina propria at the bottom. B, Weak CK20 expression. C, Increased p53 expression in invasive component. D, Increased Ki-67 labeling in invasive component. E, Intact PTEN expression. F, Intact E-cadherin expression in both noninvasive and invasive components with highlighting of invasive component. All magnifications were at ×100.

4. Discussion Urologic pathologists in the United States grade nearly all invasive UC as high grade. Most morphologically low-grade invasive carcinomas are “variants,” such as nested, tubular, microcystic, and plasmacytoid carcinomas. Once these are excluded, very few invasive tumors represent low-grade carcinomas. We have graded the invasive component strictly based on the morphology, despite that the clinical manage-


CK20 p53 Ki-67

ment is the same compared to invasive high-grade UC. Because the behavior and management of these tumors are determined by their stage, we recognize that some experts still diagnose these lesions as high grade despite their low-grade morphology. A small percentage of noninvasive low-grade papillary UCs progress to invasive carcinoma or to high-grade noninvasive papillary carcinoma [10]. Alsheikh et al [3] studied 49 LGPUCs; 7% showed both stage and grade

Comparison of IHC expression of CK20, p53, and Ki-67 in invasive versus noninvasive components Non Inv overall positive

Non Inv IHC score

Inv compared with Non Inv




2 + 3 (total)




21/26 13/26

81% 50%


7 + 9 (16) 9 + 4 (13)

1 2 2

18 24 22

7 0 2

Abbreviations: Non Inv, noninvasive; Inv, invasive; IHC, immunohistochemical stain; No., numbers.

Invasive low-grade urothelial carcinoma


Fig. 2 A, Hematoxylin and eosin section of another case with Inv LGPUC. B, Clonal PTEN loss in both noninvasive and subjacent invasive components. All magnifications were at ×100.

progression. It is hard to predict which tumor will progress based on the morphologic features. Diverse IHC markers have been studied in the diagnosis and prognostic risk stratification for UCs, especially in pTa and pT1 tumors. CK20, TP53, and Ki-67 are among the best studied markers and showed inconsistent but most promising prognostic potential for bladder cancers [11,12,13]. In normal urothelium, CK20 expression is seen in terminally differentiated superficial umbrella cells, and basal and intermediate cells are always negative. Studies have shown that CK20 expression increases with tumor grade and stage. Diffuse CK20 expression (N10%) was found in 24% to 40% of LGPUC and 47% to 69% of HGPUC, but overall positivity was found in 43% to 83% and 69% to 92% of LGPUC and HGPUC, respectively [6,11]. One study showed that CK20-positive LGPUCs were more likely to recur than CK20-negative ones [3]. In our study, abnormal CK20 expression was present in 81% of the Inv LGPUC, and 61% were scored 2 and 3, which was overall consistent with the prior studies. As CK20 expression appears to increase with stage, we expected that there would be an increase in expression in the invasive component compared with the surface noninvasive component. Comparing CK20 expression in invasive versus noninvasive component, quite the opposite occurred where the majority (69%) showed no change and 27% showed decreased expression, with only 4% slightly increased. The study of van Oers et al [13] suggested that dysregulation of CK20 expression was an early event in the carcinogenesis but occurred after the FGFR3 gene mutations, which is the molecular pathway of carcinogenesis in LGPUC. They also suggested that a combination of FGFR3 mutation and normal CK20 pattern is an excellent prognostic marker for pTa bladder tumors. However, FGFR3 mutation studies are performed by molecular testing, and many nonmutated tumors showed FGFR3 overexpression by IHC [14]. Hence, FGFR3 immunoreactivity cannot be considered as a reference method for detecting the gene mutations, and at present, FGFR3 IHC is not routinely used in the clinical setting for diagnostic and/or prognostic purposes. Although FGFR3 mutations represent the molecular oncogenic pathway for LGPUC, alteration of the tumor suppressor gene TP53 is considered a major pathway for HGPUC. A meta-analysis showed that the degree of the gene

product p53 expression was significantly associated with pathologic tumor stage, grade, and disease progression [15,16]. The prevalence of p53 by IHC is 30% in pTa tumors and 50% in pT1 tumors [15]. The study of Cina et al [17] found that p53 is positive in 32% of LGPUC and 50% of HGPUC. As most of the pT1 tumors are HGPUC, we were interested in investigating the p53 prevalence in Inv LGPUCs [15]. Evaluation and interpretation of p53 IHC results lack standardization and vary dramatically in cutoff values between studies. Normal urothelium frequently demonstrates weak nuclear p53 immunoreactivity, especially in the basal layers. In our study, the prevalence of p53 was 50% and 58% in the noninvasive and invasive components, respectively, which is consistent with the prevalence in pT1 tumors in general. Rare cases showed significant increase in p53 expression from 0 to 3 in the invasive component, which may potentially indicate newly acquired TP53 gene alterations in the process of developing invasion. However, this was not a sufficiently common finding to be of general diagnostic utility. Proliferation index (Ki-67) is another well-studied prognostic marker in UC by nuclear immunoreactivity with the MIB-1 monoclonal antibody, which detects Ki-67 antigen on formalin-fixed, paraffin-embedded tissue [18]. Many studies have showed that high Ki-67 is associated with tumor grade, stage, recurrence, and progression, especially for pT1 tumors, but most cases were high-grade UC [17,19-21]. Otto et al [19] studied a large series of pT1 UCs, and only 5 (1.6%) of 306 were LGPUC. In the noninvasive UC, the Ki-67 was reported to be 0.5% to 38.5% in LGPUC and 1% to 65% in HGPUC in the study of Cina et al of 151 papillary UCs [17]. Because of the rarity of the Inv LGPUC, our study is the first to report the Ki-67 data in a large cohort. It showed a wide range of Ki-67 labeling up to 70%, with 7 cases (27%) showing a high Ki-67 of 31% to 70%, which is much higher than the reported data on noninvasive LGPUC. Although some cases with high Ki-67 showed slight increase in cellularity in the hot spot, there was no significant increase in mitotic figures, altered cellular polarity, or any other morphologic finding that would indicate the elevated Ki-67 proliferation rate. A few cases

1840 showed variation of Ki-67 labeling in superficial and invasive components. However, because of the heterogeneous nature of the Ki-67 labeling and the relatively small component of the invasive component, this was not a useful test for diagnostic purposes and did not provide evidence that cellular proliferation rate plays a major role in the process of developing invasion. E-cadherin is a cell adhesion molecule that plays an important role in maintaining normal epithelial phenotype, organization, and tissue structure [22]. Loss or reduction of E-cadherin has been linked to tumor invasiveness, tumor progression, disease recurrence, metastasis, and decreased survival in UC [23-26]. Loss of E-cadherin expression mainly occurs in high-grade UC. Our study showed strong E-cadherin expression in 100% of the Inv LGPUC, which is consistent with the study of Watts et al [27]. Although superficial tumors (stage Ta and Tis) frequently carry gain-of-function mutations, such as FGFR3 and RAS, invasive tumors and CIS (T1 and above) frequently harbor loss-of-function mutations affecting tumor suppressor genes, including TP53, Retinoblastoma (RB), and PTEN [28]. Recently, studies of PTEN loss in UC suggested its potential prognostic value. The study of Schultz et al [29] showed that PTEN loss was seen in 42% of bladder UC, and the complete loss was more likely in invasive compared with noninvasive tumor component within a given tissue microarray (TMA) spot. In our study, the incidence of PTEN loss in LGPUC was very low (3.8%), with only 1 case showing loss; the loss was focally present in the noninvasive and subjacent invasive components (Fig. 2B). The remaining noninvasive component showed intact PTEN expression. Although this 1 case suggests that PTEN loss may have played a role in developing invasiveness in this case, the lack of this finding in other cases limits its diagnostic utility. In summary, in this large cohort of Inv LGPUC, IHC was not helpful in general in identifying invasion in LGPUC. However, high Ki-67 labeling was present in a significant portion of Inv LGPUC cases in both noninvasive and invasive components, much higher than previous studies of noninvasive LGPUC. Whether higher Ki-67 in these cases is associated with more aggressive disease warrants further study.

References [1] Pan CC, Chang YH, Chen KK, Yu HJ, Sun CH, Ho DM. Prognostic significance of the 2004 WHO/ISUP classification for prediction of recurrence, progression, and cancer-specific mortality of non–muscleinvasive urothelial tumors of the urinary bladder: a clinicopathologic study of 1,515 cases. Am J Clin Pathol 2010;133:788-95. [2] Toll AD, Epstein JI. Invasive low-grade papillary urothelial carcinoma: a clinicopathologic analysis of 41 cases. Am J Surg Pathol 2012;36:1081-6. [3] Alsheikh A, Mohamedali Z, Jones E, Masterson J, Gilks CB. Comparison of the WHO/ISUP classification and cytokeratin 20 expression in predicting the behavior of low-grade papillary urothelial tumors. World/Health Organization/International Society of Urologic Pathology. Mod Pathol 2001;14:267-72.

W. Tian, J. I. Epstein [4] Yin H, Leong AS. Histologic grading of noninvasive papillary urothelial tumors: validation of the 1998 WHO/ISUP system by immunophenotyping and follow-up. Am J Clin Pathol 2004;121: 679-87. [5] Oosterhuis JW, Schapers RF, Janssen-Heijnen ML, Pauwels RP, Newling DW, ten Kate F. Histological grading of papillary urothelial carcinoma of the bladder: prognostic value of the 1998 WHO/ISUP classification system and comparison with conventional grading systems. J Clin Pathol 2002;55:900-5. [6] Desai S, Lim SD, Jimenez RE, et al. Relationship of cytokeratin 20 and CD44 protein expression with WHO/ISUP grade in pTa and pT1 papillary urothelial neoplasia. Mod Pathol 2000;13:1315-23. [7] Holmang S, Andius P, Hedelin H, Wester K, Busch C, Johansson SL. Stage progression in Ta papillary urothelial tumors: relationship to grade, immunohistochemical expression of tumor markers, mitotic frequency and DNA ploidy. J Urol 2001;165:1124-8 [discussion 1128–30]. [8] Herr HW, Donat SM, Reuter VE. Management of low grade papillary bladder tumors. J Urol 2007;178:1201-5 [discussion 1205]. [9] Miyamoto H, Brimo F, Schultz L, et al. Low-grade papillary urothelial carcinoma of the urinary bladder: a clinicopathologic analysis of a post–World Health Organization/International Society of Urological Pathology classification cohort from a single academic center. Arch Pathol Lab Med 2010;134:1160-3. [10] Epstein J, Amin M, Reuter V. Biopsy interpretation of the bladder. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 2010. [11] Mumtaz S, Hashmi AA, Hasan SH, Edhi MM, Khan M. Diagnostic utility of p53 and CK20 immunohistochemical expression grading urothelial malignancies. Int Arch Med 2014;36 [7682-7-36. eCollection 2014]. [12] Ramos D, Navarro S, Villamon R, Gil-Salom M, Llombart-Bosch A. Cytokeratin expression patterns in low-grade papillary urothelial neoplasms of the urinary bladder. Cancer 2003;97:1876-83. [13] van Oers JM, Wild PJ, Burger M, et al. FGFR3 mutations and a normal CK20 staining pattern define low-grade noninvasive urothelial bladder tumours. Eur Urol 2007;52:760-8. [14] Guancial EA, Werner L, Bellmunt J, et al. FGFR3 expression in primary and metastatic urothelial carcinoma of the bladder. Cancer Med 2014;3:835-44. [15] Goebell PJ, Groshen SG, Schmitz-Drager BJ, International StudyInitiative on Bladder Cancer (ISBC). P53 immunohistochemistry in bladder cancer—a new approach to an old question. Urol Oncol 2010; 28:377-88. [16] Shariat SF, Weizer AZ, Green A, et al. Prognostic value of P53 nuclear accumulation and histopathologic features in T1 transitional cell carcinoma of the urinary bladder. Urology 2000;56:735-40. [17] Cina SJ, Lancaster-Weiss KJ, Lecksell K, Epstein JI. Correlation of Ki-67 and p53 with the new World Health Organization/International Society of Urological Pathology Classification System for Urothelial Neoplasia. Arch Pathol Lab Med 2001;125:646-51. [18] Mazerolles C, Rishmann P, Chopin D, et al. Usefulness of MIB1 monoclonal antibody in assessing the proliferative index in human bladder carcinoma: comparison with Ki-67 antibody. Histopathology 1994;25:563-8. [19] Otto W, Denzinger S, Fritsche HM, et al. Introduction and first clinical application of a simplified immunohistochemical validation system confirms prognostic impact of KI-67 and CK20 for stage T1 urothelial bladder carcinoma: single-center analysis of eight biomarkers in a series of three hundred six patients. Clin Genitourin Cancer 2013;11: 537-44. [20] Goyal S, Singh UR, Sharma S, Kaur N. Correlation of mitotic indices, AgNor count, Ki-67 and Bcl-2 with grade and stage in papillary urothelial bladder cancer. Urol J 2014;11:1238-47. [21] Bertz S, Otto W, Denzinger S, et al. Combination of CK20 and Ki-67 immunostaining analysis predicts recurrence, progression, and cancerspecific survival in pT1 urothelial bladder cancer. Eur Urol 2014;65: 218-26.

Invasive low-grade urothelial carcinoma [22] Takeichi M. Cadherin cell adhesion receptors as a morphogenetic regulator. Science 1991;251:1451-5. [23] Sun W, Herrera GA. E-cadherin expression in invasive urothelial carcinoma. Ann Diagn Pathol 2004;8:17-22. [24] Cui D, Han BM, Jing YF, Xia SJ. Analysis of N-E cadherin switch as an independent predictive parameter of bladder cancer survival outcomes. Zhonghua Yi Xue Za Zhi 2012;92:380-3. [25] Omran OM. CD10 and E-cad expression in urinary bladder urothelial and squamous cell carcinoma. J Environ Pathol Toxicol Oncol 2012; 31:203-12. [26] Muramaki M, Miyake H, Terakawa T, Kumano M, Sakai I, Fujisawa M. Expression profile of E-cadherin and N-cadherin in non–muscle-

1841 invasive bladder cancer as a novel predictor of intravesical recurrence following transurethral resection. Urol Oncol 2012;30:161-6. [27] Watts KE, Montironi R, Mazzucchelli R, et al. Clinicopathologic characteristics of 23 cases of invasive low-grade papillary urothelial carcinoma. Urology 2012;80:361-6. [28] Cordon-Cardo C. Molecular alterations associated with bladder cancer initiation and progression. Scand J Urol Nephrol Suppl 2008:154-65. [29] Schultz L, Albadine R, Hicks J, et al. Expression status and prognostic significance of mammalian target of rapamycin pathway members in urothelial carcinoma of urinary bladder after cystectomy. Cancer 2010; 116:5517-26.