Read the Full Video Transcript

Ashish Kamat: Hello, and welcome to UroToday's Bladder Cancer Center of Excellence. I'm Ashish Kamat, professor of urologic oncology and cancer research at MD Anderson. And it's a pleasure to welcome a dear friend and a colleague from MD Anderson, Professor Charles Guo, who's currently chief of the GU section. Dr. Guo has many, many years of experience when it comes to GU pathology, but specifically has been a close collaborator on a lot of groundbreaking research that we have done here at MD Anderson when it comes to bladder cancer. I could think of no better person than him to invite to share with us, our UroToday audience, his views and updates on variant histology in bladder urothelial carcinoma. Charlie, thanks for joining us, and the stage is yours.

Charles Guo: Thank you, Ashish, for the kind introduction. Today, I am delighted to talk about variant histology in urothelial carcinoma, a very common phenomenon in bladder cancer. Urothelial carcinoma is by far the most common histology in bladder cancer, and it accounts for approximately 90% of all bladder cancers. The current WHO classification of bladder cancer divides urothelial carcinoma into two major groups, noninvasive and invasive urothelial carcinoma.

The noninvasive urothelial carcinoma includes papillary urothelial tumors and urothelial carcinoma in situ. The invasive urothelial carcinoma includes the conventional type as well as a number of variant histologies. This pathological classification is a little different from the clinical classification, which divides bladder cancer into muscle-invasive and non-muscle invasive disease. The clinical classification is not as precise as the pathological classification in terms of tumor biology, but the clinical classification serves well for the purpose of patient treatment.

Conventional invasive urothelial carcinoma is typically characterized by these small irregular tumor nests, coats of single cells infiltrating the bladder wall, often associated with stronger reactions such as inflammation and fibrosis. A significant number of invasive urothelial carcinoma develops variant histology with some microscopic features that are dramatically different from the conventional type. Variant histology can be described as divergent differentiation or histological subtype. Divergent differentiation indicates that development of different epithelial lineages in the urothelial carcinoma such as squamous and glandular differentiation. Histological subtypes, they are intrinsically urothelial carcinoma and develop along the urothelial lineage. In the current WHO classification, subtype is the preferred term for this phenomenon and the variant should be avoided because it is commonly used for genomic alterations.

Divergent differentiation, it's quite common in invasive urothelial carcinoma. Up to 40% of invasive urothelial carcinoma can show focal divergent differentiation. The most common divergent differentiation is the squamous differentiation, but glandular differentiation is also quite common. And sometimes you can even see trophoblastic differentiation. Squamous differentiation is characterized by these keratin pearls or intracytoplasmic keratin production. It is not associated with HPV infection in general. Glandular differentiation, shows these kinds of intestinal-types of glands that are lined by typical columnar cells that look like the colonic adenocarcinoma.

The clinical significance of a squamous and glandular differentiation in urothelial carcinoma is largely uncertain, although some studies indicate a poor outcome compared to the pure conventional urothelial carcinoma. In pathology practice, the diagnosis of squamous cell carcinoma or adenocarcinoma is only reserved for those that show pure squamous or glandular differentiation with no urothelial carcinoma, not even urothelial carcinoma in situ. Trophoblastic differentiation, it looks like choriocarcinoma, it shows extensive hemorrhage and has these syncytiotrophoblast and the intermediate trophoblast. Patients often have a higher level of beta-hCG. Interestingly, beta-hCG level is frequently elevated in patients with invasive urothelial carcinoma.

Some studies reported in 35% of the patients have elevated beta-hCG, even in the absence of trophoblast differentiation. Trophoblast differentiation is generally associated with poor prognosis. Histological subtypes, in the current WHO classification there are about a dozen different subtypes. Some subtypes such as micropapillary, plasmacytoid, small cell carcinoma and sarcomatoid carcinoma, they show more aggressive clinical behaviors than the conventional urothelial carcinoma. Some other types such as small nested, large nested, tubular and microcystic subtypes, they show a similar clinical behavior like the conventional urothelial carcinoma but they mimic some benign or non-neoplastic lesions in the bladder. They can be challenging to diagnose on a small biopsy specimen.

So let's briefly go over the morphology of these subtypes. First micropapillary subtype, it is characterized by small tumor nests surrounded by empty space. In some cases, you can see a number of back-to-back small nests located in a single empty space. Although the micropapillary morphology looks like lymphovascular invasion, but it's not true lymphovascular invasion. On immunostains, you can see these empty spaces are not lined by endothelial cells that are usually highlighted by CD-34 endothelial markers. The empty spaces are largely caused by a retraction artifact. Micro papillary subtype is an aggressive subtype frequently metastasized to lymph nodes or distant organs. Interestingly, the metastasis often maintains this micro papillary morphology. Micro papillary subtypes, these are aggressive subtypes that show a poorer outcome than conventional urothelial carcinoma. And a study from our institute led by Ashish demonstrates that patients with T1 micro papillary subtype tumor can benefit upfront. The cystectomy patient treated with cystectomy had a significantly better outcome than those treated with intravesical BCG.

Several years ago we analyzed the genomic expression profile of the micro papillary subtype in comparison to conventional urothelial carcinoma. We found that the micropapillary subtype has a distinct gene expression profile, a large number of genes over-expressed, and similarly, many genes also under-expressed in the micropapillary subtype. Using ingenuity pathway analysis, we found that these differentially expressed genes are involved in several oncogenic pathways such as mechanisms of cancer, CIS checkpoint, cyclins, and cell cycle P53. We also analyzed the molecular subtype for micropapillary urothelial carcinoma, and we found that the micropapillary subtype is largely luminal subtype and some of the tumors show increased expression of the genes with P53 activation. Plasmacytoid carcinoma is also an aggressive subtype that shows a diffusely infiltrative growth pattern, but induces a very minimal stromal reaction. The tumor cells look like plasma cells with eccentrically located nuclei and abundant eosinophilic or glassy cytoplasm and signet ring-like cells are quite common in this subtype.

This is a very aggressive subtype. The tumor often infiltrates through the bladder wall and spreads along the peritoneum. So at radical cystectomy, the resection margin of paravesical soft tissue is often positive for patient with a plasmacytoid urothelial carcinoma. Several years ago, a research group at Memorial Sloan Kettering found that most plasmacytoid urothelial carcinoma carried somatic mutations at the CDH-1 gene leading to the dysfunction of the E-cadherin protein, which may underline it's aggressive behavior. Small cell carcinoma subtype. It's like the lung small cell carcinoma. The tumor cells are poorly differentiated with scant cytoplasm, high N:C ratio, nuclear moldings, numerous mitoses and coagulative necrosis. Small cell carcinoma frequently is positive for neuroendocrine markers such as synaptophys and chromogranin. It often coexists with  conventional urothelial carcinoma.

Patients with small cell carcinoma can benefit from neoadjuvant chemotherapy. Neoadjuvant chemotherapy can significantly improve the patient's outcome in terms of overall survival and disease-specific survival. A couple of years ago we studied the genomic profile of bladder small cell carcinoma and we found that bladder small cell carcinoma is likely to progress by the basal pathway with frequent gene mutations in TP-53, RB1, BRCA1/2, and it is characterized by the loss of urothelial differentiation and gain of neural differentiation. It shows activation of EMT and it also over-expressed adenosine A2A, an immune checkpoint or receptor that may contribute to its immune-desert phenotype.

Lipoid subtype. In this subtype, the tumor cells have large intracytoplasmic vacuoles that indent the nuclei. So the tumor cells look like lipo blasts. Electron microscopy studies show that these vacuoles contain lipid-like material. Osteoclast-like subtype. This subtype looks like the giant cell tumor of the bone and tumor cells have two distinct cell populations, the mononuclear, malignant epithelial cells, and the multinucleated osteoclast-like giant cells. It is associated with very aggressive behavior. Giant cell subtype. In this subtype, there are numerous, large polymorphic bizarre cells. Typically, these bizarre cells account for more than 20% of the cell population. It's also a very aggressive variant. Small nested subtype. In this subtype, the tumor cells form these small nests, but the tumor cells show very bland cytology and with very minimal cytological atypia. They look like Von Brunn's nests. So it is often difficult to diagnose this subtype because of the bland cytology. But this subtype often deeply invades urethra and involves the muscularis propria.

A large nested subtype. Like the small nested subtype, this subtype, the nest is larger. Again, the tumor cells often show bland cytology with minimal cytological atypia. The tumor is often deeply invasive, often involving the muscularis propria. Tubular and microcystic subtype. This tumor forms these small cysts or tubules. Again, the tumor cells are very plain looking with minimal cytological atypia. It mimics Cystitis cystica glandularis. So it can also be challenging to diagnose this subtype on small biopsy specimens. So like the small nested, and large nested, tubular and microcystic subtypes, they show similar clinical behavior or outcome like the conventional urothelial carcinoma when compared stage by stage. Clear cell subtype. This subtype of tumor cells has abundant glycogen. So after fixation and processing, the tumor cells have clear cytoplasma, so it looks like clear cell renal cell carcinoma.

Lymphoepithelioma-like subtype, this subtype is featured by a prominent lymphocytic infiltrate. The lymphocytic infiltrate is mostly T cells, but can be B cells, plasma cells, or eosinophils. If the tumor shows pure lymphoepithelioma-like morphology or predominant lymphoepithelioma-like morphology, the tumor often responds favorably to chemotherapy. Sarcomatoid subtype. This subtype is a biphasic tumor. Again, the epithelioid component is urothelial carcinoma and the mesenchymal component is high grade spindle cells. The high grade spindle cells typically show marked atypia with hyperchromasia, nuclear enlargement and polymorphism. And sometimes the sarcomatoid carcinoma may develop these heterologous differentiations with chondro sarcoma or osteosarcoma like components. The development of heterologous differentiation in sarcomatoid carcinoma is associated with an even worse prognosis than those with high grade spindle cells.

We also studied the genomic expression profile of bladder sarcomatoid carcinoma. Again, we found that sarcomatoid carcinoma is likely to progress through the basal pathway with frequented gene mutations in TP53, PIK3CA, and RB1. It shows an immune infiltrate phenotype with over expression of PD-L1, and it shows the activation of the EMT network through multiple molecular mechanisms. The sarcomatoid carcinoma can be divided into two distinct subsets. The mesenchymal subset has a higher level of EMT activation and also has a worse prognosis than the basal subset. So in summary bladder urothelial carcinoma frequently develops variant histology such as divergent differentiation and histological subtypes. Although they are biologically different, they are all lumped together in the literature. Urothelial carcinoma subtypes demonstrate different histological and clinical features from conventional urothelial carcinoma and urothelial carcinoma subtypes. Show distinct genomic alterations and molecular signatures which may outline the aggressive behavior. Thank you.

Ashish Kamat: Thank you so much, Charles, for covering such a very dense and important topic in such a succinct manner. You highlighted a lot of the important points and I just wanted to ask you a few points of clarification for our audience and listeners. When we talk about histologic subtypes and variants, right, that terminology has changed over the years. Could you just enlighten our audience that, especially for the young investigators that are writing grants or papers currently, the term variant and subtypes, can that be used interchangeably or do you recommend that one is used more for genotyping and one is left for immunohistochemistry?

Charles Guo: In the past, the subtype used to be called variants, but in the current WHO classification, now in its fifth edition, covering all human tumors. They uniformly use the term subtype to replace the variant terminology. At least in the pathology community, the variant is replaced by subtype when you describe a histological variant. But in literature, the term variant is still prevalent. So also I think the other point I want to mention is in many of the studies, particularly some of the clinical studies, they often mix divergent differentiation, and histological subtype. So such when we are talking about divergent differentiation, we are talking about different epithelial lineages from urothelial carcinoma such as squamous or glandular. And when we talk about histological subtype, they are intrinsically urothelial carcinoma. They develop along the urothelial lineage. They are two different biological phenomena, but in the literature they are often lumped together. But that can be problematic because they are two different biological processes.

Ashish Kamat: And that's a very important distinction and I'm glad you emphasized that. The other question that comes up sometimes is what percentage of a histologic subtype is relevant? I know here you report with actual percentages, but out in the community or where patients are having path reports that don't actually give a percentage, is that because the pathologist is looking for a certain threshold beyond which they report it? Or should they be reporting any incidence and just actually reporting the percentages?

Charles Guo: So for squamous or glandular differentiation, if it's focal, probably the effect on the clinical outcome is minimal. We often see focal squamous differentiation or glandular differentiation in urothelial carcinoma. Their clinical significance probably is minimal but for these aggressive subtypes like micro papillary, plasmacytoid, small cell, or sarcomatoid, even a very tiny amount can affect the outcome of the patient. We often see patients have... you don't see aggressive variants in the primary tumor, but you see the aggressive morphology in the metastasis. When you go back to look at the primary tumors, you often find the subtype morphology may have been missed in the primary tumor. So even a small amount of subtype is significant. So it's all depending. The divergent differentiation probably is okay, the amount, probably needs to be significant to see a significant effect on the clinical outcome.

Ashish Kamat: And last question, Charles, and this is going to be a little provocative, but since you've done so much work in this area, I think you're more than qualified to answer this question, which comes up in a lot of people's minds. If you have an excellent GU pathologist that can look at the tumor under the microscope, and of course with some ancillary staining that's needed, give a diagnosis of histological variants, grade, stage, et cetera. But here we're focusing on histologic subtypes. Is there anything extra to be gained outside of a research area from sending that patient's tumor off for genomic examination?

Charles Guo: Genomic examination. So at MD Anderson, we often do genomic analysis, particularly next generation sequencing on those advanced bladder cancers because the clinical oncologists, they use the gene expression profile to select some of the targeted therapy such as particularly these immune checkpoint therapies, using MSI, tumor mutation burden, and a number of other factors. So it depends. So if the clinical oncologist thinks this analysis can assist the decision on the treatment. So, we do all the molecular tests on the tumor.

Ashish Kamat: Separate from that though, Charles, from the clinical perspective, absolutely. But separate from that, is there any role for genomic testing or genomic profiling to make the diagnosis?

Charles Guo: Diagnosis usually is not a problem. So based on the morphology, we can make a relatively accurate diagnosis. So although some of the molecular alterations may occur prior to the appearance of typical morphology, but in most cases we don't use these genomic molecular analyses to assist the diagnosis. They are more often used to predict the patient's response to targeted therapy, but in diagnosis, so generally, it's not necessary, we don't need the test for diagnosis.

Ashish Kamat: Right. Thank you for clarifying that question, because it does come up quite often. Once again, Dr. Guo, thank you for taking the time and spending it with us today. This lecture and presentation of yours is going to be a great resource, I'm sure, which a lot of people will access over and over again. So thank you.

Charles Guo: Well, thank you. It's my pleasure.