Key points

Introduction

Although approximately 80% of bladder cancer is caused by “conventional” urothelial carcinoma (UC), the remaining 10% to 25% is the result of nonurothelial and “variants” of UC. Although the term “variant histology” can sometimes be used in a variety of different capacities, for the current discussion variant histology refers to any bladder malignancy other than pure UC. This simplification includes UC with aberrant differentiation in which the tumor arises from a common urothelial stem cell as well as “nonurothelial” carcinoma, which is the result of metaplasia. In reality, these histologic descriptions are based on morphologic features from hematoxylin and eosin–stained pathologic sections with little insight into their biologic derivative. Furthermore, mixed histologies are often present (including so-called urothelial and nonurothelial carcinomas), for which the term variant histology is generally used. Box 1 describes the histologic classification of tumors arising from the urinary tract and was adapted from the 2004 World Health Organization Classification of Tumors.

Introduction

Although approximately 80% of bladder cancer is caused by “conventional” urothelial carcinoma (UC), the remaining 10% to 25% is the result of nonurothelial and “variants” of UC. Although the term “variant histology” can sometimes be used in a variety of different capacities, for the current discussion variant histology refers to any bladder malignancy other than pure UC. This simplification includes UC with aberrant differentiation in which the tumor arises from a common urothelial stem cell as well as “nonurothelial” carcinoma, which is the result of metaplasia. In reality, these histologic descriptions are based on morphologic features from hematoxylin and eosin–stained pathologic sections with little insight into their biologic derivative. Furthermore, mixed histologies are often present (including so-called urothelial and nonurothelial carcinomas), for which the term variant histology is generally used. Box 1 describes the histologic classification of tumors arising from the urinary tract and was adapted from the 2004 World Health Organization Classification of Tumors.

Challenges in the study of variant histology

Studying the clinical significance of variant histology can be difficult as a result of subjectivity and challenges with diagnosis and cell type identification. As a result of sampling error and tumor heterogeneity, transurethral resection (TUR) has been reported to detect only 39% of variant cancers that present within the bladder. Adding to the inherent difficulty with detection, it has been estimated that up to 44% of cases of histologic variants are not recognized or documented by community pathologists, which further leads to underreporting and the potential for mismanagement of patients. Initial reports also suggested that variant tumors were uniformly present at a high stage with invasion into muscularis propria. However, since that time, multiple studies have been published showing variant histology present within non–muscle-invasive (NMI) tumors. This is likely a reflection of increased awareness and recognition within the scientific community. In fact, in a large registry with more than 28,000 bladder cancer patients from The Netherlands, 23% of all variant tumors identified within the registry presented with NMI disease.

Another issue complicating the study of variant histology is whether the extent of variant histology (ie, focal vs extensive) effects patient outcomes. Most studies do not control for this fact in their design. Some have explored relevant cutoffs, as was the case of 1 early study proposing that 20% variant histology was associated with worse survival outcomes. However, no consistency has been shown among subsequent studies and it has become apparent that each bladder cancer variant behaves differently and needs to be addressed individually to assess its impact on the overall biology of the disease. The significance of the extent of each specific variant remains an area under investigation and study.

These diagnostic challenges underscore the difficulty in studying variant histology. Without consistent and reliable diagnosis, it will be difficult to reliably determine the biologic origins of variant bladder tumors, or how they might respond to therapy compared with conventional UC. In an effort to combat these challenges, many groups are attempting to outline standards and guidelines formally to aid in the identification and reporting of variant histology. With the incorporation of collaborative efforts and centralized pathologic review, it is hoped that a better understanding of variant bladder cancer will result.

Significance of variant histology

Several retrospective studies have suggested that variant histology portends worse clinical outcomes. Collectively, this is thought to be the result of a higher propensity of locally aggressive disease, higher rates of distant metastasis, and a different response to chemotherapy or radiotherapy compared with conventional UC. In a study of 448 consecutive TUR bladder tumor cases with 295 subsequent cystectomies, mixed histology was present in 25%, and the presence of variant architecture almost uniformly predicted the presence of locally advanced disease at cystectomy. A separate group observed that, among 600 cystectomy patients, variant histology (defined as squamous, glandular, or other) predicted upstaging at the time of cystectomy with an odds ratio of 2.77. The presence of variant histology has also been found to be associated with increased rates of pathologic lymph node metastasis in several multivariate analyses leading to worse survival outcomes. Finally, a multi-institutional study with approximately 1000 patients found that patients with adenocarcinoma, small cell carcinoma, or other histologic subtypes had worse disease specific survival compared with conventional UC, even after accounting for stage, adjunct treatment, and lymphovascular invasion on multivariate analyses.

However, other studies have presented conflicting evidence that these trends may not be true for all variants types of bladder cancer. A secondary analysis of the Southwest Oncology Group randomized trial S8710 was performed after the original study that showed increased survival for neoadjuvant methotrexate, vincristine, adriamycin, and cisplatin (MVAC) plus cystectomy over cystectomy alone in patients with locally advanced (cT2–T4a) bladder cancer. This secondary analysis showed that patients with mixed histology (squamous and glandular differentiation) had improved survival rates after neoadjuvant MVAC chemotherapy as well as a higher rate of pT0 downstaging (34%) versus cystectomy alone (4%). This correlated with improved survival rates for patients with mixed histology after neoadjuvant chemotherapy, although statistical significance was not attained. This secondary analysis that was based on high-quality clinical trial data opposes the idea that all variant histology leads to a worse overall prognosis and outcome.

Non–muscle-invasive variant bladder cancer

There is some controversy regarding NMI bladder cancer with variant histology, both in terms of diagnosis and the role of intravesical therapy. A critical issue in considering therapy for NMI variant bladder cancer is that of correct staging. Although this is the case for conventional UC as well, some argue that because variant tumors notoriously are associated with advanced disease, there may be a potentially greater risk of understaging variant tumors. In such situations, intravesical therapy would be less effective. Retrospective studies looking at cT1 NMI variant bladder cancer have reported local understaging rates of 27% to 57%. Rates of occult metastatic disease have also been reported as high as 27% to 44% among NMI variant tumors, and in several multivariate analyses, divergent histology has been associated with the presence of lymph node metastasis and decreased survival. This evidence of relatively high rates of understaged, locally advanced disease or occult metastatic disease raises the concern that even in the setting of NMI disease, a more aggressive treatment strategy is warranted. However, others have reported progression rates of approximately 40% (similar to conventional UC with high-risk features) and similar survival and response rates with induction and maintenance Bacillus Calmette-Guerin (BCG) treatment compared with control groups of conventional UC tumors. These studies included tumors with squamous or glandular differentiation, nested variant (NV), and micropapillary disease. Thus, a thoughtful approach to management of NMI variant bladder cancer should be employed because of the morbidity associated with radical cystectomy. In general, the role of intravesical treatment for variant NMIBC should be considered based on the unique subtype and should be weighed against the risk of understaging.

Management of variant bladder cancer

Based on existing studies, variant histology has been proposed as a clinical feature that identifies “high-risk” bladder cancer, and subsequent treatment algorithms have been presented. Conceptually, if variant histology signals high risk for advanced bladder cancer, one could argue that the mere presence of variant architecture justifies an aggressive treatment algorithm that might depart from the standard treatment for conventional UC and warrant a multimodality approach to treatment. However, if variant architecture confers resistance to chemotherapy or radiation therapy, then delaying surgery for ineffective neoadjuvant therapy may also lead to adverse events.

The reality for variant bladder cancer is that each histologic subtype (see Box 1 ) should be addressed individually because each may differ in their predisposition for metastasis and sensitivity to systemic treatment. For that reason, the common variant subtypes are addressed individually as they appear in Box 1 with an emphasis on both muscle-invasive disease and NMI disease. Treatment algorithms are summarized in Figs. 1 and 2 .

Management of non–muscle-invasive and variant bladder cancer. BCG, Bacillus Calmette-Guerin; TUR, transurethral resection; UC, urothelial carcinoma.

Management of muscle-invasive bladder cancer and variant histology. Consider neoadjuvant chemotherapy for select patients. LVI, lymphatic invasion; UC, urothelial carcinoma; XRT, x-ray therapy.

Squamous/glandular differentiation

Because the biologic origins of variant tumors are still in question, squamous and glandular differentiation have been included in the present discussion of variant histology. Each is typically present in the background of conventional UC in varying degrees. Squamous and glandular differentiation are discussed together here because many studies have commonly grouped them as “mixed” histology in the current literature. In reality, squamous differentiation is more prevalent and dominates studies with “mixed” histology. In fact, squamous differentiation has been reported to be present in as many as 60% of urothelial tumors specimens, although it is less likely to be reported unless the architecture predominates the specimen as low proportions of squamous differentiation have not been shown to be clinically relevant. Glandular differentiation has not been adequately studied independently to allow any conclusions. Interestingly, in many postcystectomy bladder cancer series, squamous differentiation correlates with advanced stage and poor prognosis in multivariate models, suggesting it is a clinically relevant entity. However, most studies are retrospective in nature and many other conflicting reports exist.

As mentioned, secondary analysis of the Southwest Oncology Group neoadjuvant MVAC trial shows relatively high rates of response to neoadjuvant chemotherapy among tumors with squamous/glandular differentiation. Separate, large, retrospective, case control studies also shown that despite an increased rate of locally advance tumors and nodal metastasis, patients with squamous/glandular differentiation had similar survival rates after controlling for clinical parameters such as T stage. One study even stratified by the extent of differentiation and found no survival difference between tumors with less than 30% versus 30% or greater squamous differentiation. In general, based on the current evidence, squamous or glandular differentiation may indicate advanced disease or a risk for understaging. However, these tumors are thought to respond to current treatment modalities, including neoadjuvant chemotherapy, and so they should be treated similarly to stage matched conventional UC. It has been suggested that extensive squamous differentiation may clinically resemble pure squamous cell carcinoma (SCC) with a predisposition for local recurrence, for which other treatment algorithms could be considered (ie, radiation), although this has not been reliably demonstrated in the literature.

Although the majority of tumors with squamous and glandular differentiation present with muscle-invasive disease (70%), there is no evidence to suggest that NMI tumors with squamous or glandular differentiation do not respond to intravesical treatments. Most studies have shown outcomes similar to pure UC with comparable rates of recurrence and progression. It is therefore acceptable to pursue intravesical therapy for NMI tumors with “mixed histology” with the caveat that close surveillance and a high suspicion for understaging be maintained. Patients with low-volume, completely resected tumors with a small focus of squamous or glandular elements would be the best candidates for bladder preserving therapies. One should be cautious to delay definitive treatment in the form of radical cystectomy for those patients not responding to intravesical treatments.

Nested variant

NV bladder cancer is unique among the variants because of its deceptively benign appearance, which can easily be confused with von Brunn nests. As has been consistent with other variants of bladder cancer, most studies show a predisposition to present with advanced disease compared with conventional UC with high rates of muscle invasion at TUR (70% vs 31%), extravesical disease (83% vs 33%), and metastasis (67% vs 19%) when compared with high-grade UC, respectively. Nevertheless, stage for stage, NV does not seem to be any more aggressive than conventional UC with matched cohort analyses (n = 52) showing that, despite high rates of locally advanced (69%) and node-positive (19%) disease (10-year cancer-specific survival of 41%), NV has equivalent survival outcomes after controlling for pathologic stage. There is very little published regarding the efficacy of neoadjuvant chemotherapy for NV. Unfortunately, little work has been done on NV bladder cancer and there is little evidence to support a treatment algorithm outside that employed for high-risk conventional UC for either muscle invasive or NMI NV. However, the high rates of extravesical and metastatic disease suggest a need for multimodal therapy.