Outcomes of poorly differentiated and plasmacytoid variant bladder urothelial carcinoma

Abstract Objectives The aim of this study is to assess the course and management of poorly differentiated bladder urothelial carcinoma (UC), including plasmacytoid UC (PUC), in our local area. Although bladder cancer is relatively common, PUC is a rare and aggressive subtype with a poor prognosis that is still poorly understood. Materials and Methods A retrospective assessment of all poorly differentiated high‐grade UC over the last 15 years (2005–2020) in the Hunter New England area was completed. In total, 37 patients were included, and PUC variant was compared with the remaining poorly differentiated UC. Results Of the included cases, eight were PUC, nine squamous variant, two neuroendocrine, and one sarcomatoid. Overall, 23 cases proceeded to cystectomy, 15 had chemotherapy (six neoadjuvant), and 11 had radiation therapy. In the PUC subgroup, three had metastatic disease at diagnosis (37.5%). Of the three PUC patients who underwent cystectomy, all were upstaged. Two PUC cases had adjuvant chemotherapy, and one case had radiation. Within the follow‐up period, the PUC group had a cause‐specific mortality of 50% with a mean survival in these patients of 202 days, compared with 37.9% cause‐specific mortality with survival of 671.55 days (p = 0.23) in all other undifferentiated UC cases; 5‐year cause‐specific mortality with Kaplan–Meier analysis was estimated at 26% compared with 59%, respectively (p = 0.058). Conclusion Poorly differentiated UC is demonstrated to have a poor prognosis with a high mortality rate, particularly when PUC is present. Given the rarity of these variants, further studies are necessary to explore the impact of current treatment options.

accounting for 2.1% of all cancer deaths. [3][4][5] Urothelial carcinoma (UC) accounts for 90% of all bladder cancers; of these, approximately 20-30% are invasive at diagnosis. 6 Invasive UC has a much higher rate of divergent differentiation and therefore often demonstrates histologic variants (in approximately one third of invasive UC). 6,7 Plasmacytoid UC (PUC) is one of these rare variants, accounting for 1-3% of invasive UC. 6 It was first described in 1991 and was included in the World Health Organization (WHO) classification of tumors in 2004. 8,9 Of all UC variants, it is the most likely to be found in its pure form. 8,9 PUC is characterized by infiltration of discohesive monotonous small round cells with amphophilic cytoplasm and eccentrically placed round nuclei (Figure 1). This variant of UC resembles plasma cells in appearance and lobular carcinoma of breast and signet ring carcinoma of stomach in both cell appearance and infiltrative architecture, and if no in situ urothelial component is evident, one should also consider metastasis. This differential diagnosis can be difficult to distinguish in this setting because PUC can occasionally express intestinal marker CDX2 and breast markers progesterone receptor and GCDFP-15. 10 PUC has a similar immunohistochemical profile to conventional UC with variable expression of GATA3 (80% of PUC cases), P63, CK7 (92%), S100P and CK20 (72%), and uroplakin II (33%), although the diagnosis of plasmacytoid variant is usually made morphologically. 6,8,11 PUC expresses CD138, which is also a plasma cell marker, although it is not entirely specific and stains other epithelial tumors. 12 The vast majority of PUCs are distinguished by an additional somatic mutation/promoter hypermethylation affecting the CDH1 gene, which is demonstrated by the loss of E-cadherin (a protein that maintains intercellular adhesion) and can usually aid in diagnosis of plasmacytoid variant if there is doubt ( Figure 1). 6,13 Occasional Ecadherin-positive cases have been reported, which seem to have a better prognosis. 7,14 Most PUCs also lack expression of the RB gene, which encodes for a tumor-suppressor protein that regulates cell cycle senescence and apoptosis. RB mutations are more likely to be found in muscle invasive than superficial UCs. 6,7 HER2 overexpression has also been observed in PUC. 15 Clinically, PUC has a poor prognosis, characterized by aggressive, invasive cancer and early metastasis with a predilection for peritoneal spread. 9,14,16,17 Locoregional spread is typically along pelvic fascial planes to involve perivesical, perirectal, and periureteric soft tissues; this can be seen on computed tomography (CT) and magnetic resonance imaging (MRI) as well as during intraoperative inspection. 16 PUC has a higher frequency of stage ≥pT3 tumors at diagnosis, a higher risk of lymph node metastases, positive margins, and perivesical extension at radical cystectomy than conventional UC. 18 The recent meta-analysis published this year by Kim et al. found no difference in cancer-specific mortality but worse overall survival outcomes. 18 The mainstay of treatment is radical cystectomy where patient and tumor factors allow, and (neoadjuvant) chemotherapy. Although there is still limited data available regarding PUC response to chemotherapy, it is less responsive than conventional UC in large series. Diamantopoulos et al. reported a ypT0N0 rate after neoadjuvant chemotherapy of 10% in PUC compared with 33% in conventional UC. 14,18 Nonetheless, some cases have reported responsiveness to typical chemotherapeutic agents. 19 Usually, platinum-based chemotherapy is given with regimes such as methotrexate, vinblastine, doxorubicin, and cisplatin (MVAC) or gemcitabine and cisplatin (GC). 14,20 The aim of this study is to assess the prevalence, treatment, and outcomes of plasmacytoid variant and undifferentiated high-grade UC of the bladder in our local health district.

| RESULTS
A total of 37 patients with high-grade poorly differentiated UC were included over a 15-year period from January 2005 to December 2020. Of the included cases, eight were PUC. In the comparison UC group, nine were noted to have squamous variant, two neuroendocrine, and one sarcomatoid, and the remaining were poorly differentiated without identifiable variant pathology. The average age at diagnosis was 71.13 years, and average Charlson Comorbidity Index score was 4.73. Six patients had metastatic disease at diagnosis-three of these patients were plasmacytoid variant (37.5% of this group). All of the patients were discussed in our uro-oncology multidisciplinary meeting before proceeding to definitive treatment options.
Overall, 23 cases proceeded to cystectomy, 15 had chemotherapy (six neoadjuvant), and 11 had radiation therapy. There were a total of 18 deaths (48.65%) in the study period, with two additional patients lost to follow-up but metastatic progression demonstrated at last review. Mean overall survival of those who died in the study period was 24.9 months, and mean cause-specific survival was 18.2 months.
The mean age of diagnosis in the PUC subgroup was 74, with a mean Charlson Comorbidity Index of 5.25 (compared with 4.76 in all other UCs, p = 0.54). Of the three patients with metastatic PUC at diagnosis, one had palliative radiotherapy to manage significant hematuria. The remaining two had palliative management. One with pT2 PUC was not deemed suitable (in shared decision making according to patient preference) for surgical intervention or chemoradiotherapy due to comorbidities and had further staging with PET within 3 months demonstrating nodal disease. Another with pT2 disease on diagnosis was found to have concurrent lung cancer on further staging; this took treatment priority, and the patient went on to have palliative chemoradiotherapy (chemotherapeutic agents carboplatin and etoposide). Of the three PUC patients who had a cystectomy, all were upstaged-one had adjuvant palliative chemotherapy for T4b disease and positive margins (six cycles cisplatin and gemcitabine completed 5 months postoperatively), and two went on to have imaging surveillance with Oncology without further chemoradiotherapy to date. Table 1 compares the characteristics of the PUC group and all other poorly differentiated UC cases. When considered indicated at the time by the reviewing Pathologist, immunohistochemistry profiles of the included PUC cases were performed; these are outlined in Table 2; many were able to be diagnosed predominantly on morphological features.
Within the follow-up period, the PUC group had a cause-specific mortality of 50% (four cases; three of these had metastatic disease at diagnosis, and one was upstaged to T4b disease at cystectomy). The mean survival in these patients was 6 months, compared with 37.

We thank Bethan Major and Dr Simon Palfreeman at Douglass Hanly
Moir Pathology for assisting in the data collection.

CONFLICT OF INTERESTS
All authors declare that they have no conflicts of interest or additional sources of support or funding to declare.