Abstract
Aim: This study retrospectively investigated the impact of squamous differentiation on the prognosis of patients with upper tract urothelial carcinoma (UTUC) undergoing radical nephroureterectomy (RNU). Patients and Methods: Among the 244 consecutive patients who underwent RNU at our Institution from May 2005 to October 2019, 224 were analysed. Metastasis-free (MFS) and overall (OS) survival rates were evaluated using Kaplan–Meier analysis and Cox regression analysis. Results: With a median follow-up time of 58 months, the groups with pure UTUC (n=197) and UTUC with squamous differentiation (n=27) had 5-year MFS rates of 65.2% and 40.9% (p=0.005) and 5-year OS rates of 74.4% and 49.0% (p=0.002), respectively. Multivariate analyses revealed that the presence of squamous differentiation was significantly associated with poor MFS (hazard ratio=1.88; p=0.027) and OS (hazard ratio=1.70; p=0.048). Conclusion: Squamous differentiation in UTUC appears to be an independent predictor of poor prognosis after RNU for UTUC.
- Squamous differentiation
- urothelial carcinoma
- upper urinary tract cancer
- radical nephroureterectomy
- prognosis
Studies have recently focused on variant histology as a prognostic factor in patients with bladder cancer (1-3). In particular, we previously demonstrated that squamous differentiation (SD) predicted poor overall survival (OS) after radical cystectomy (4, 5). Upper tract urothelial carcinoma (UTUC) involving the renal pelvis/ureter is histologically similar to bladder cancer in morphological diversity (6). Pure UC is also a predominant histology of UTUC; however, variant histology in UC has been frequently observed (10-30%) (7). Although SD is the most common variant in UTUC (8), the clinical significance of UCSD has yet to be comprehensively elucidated given the rarity of the disease compared to bladder cancer. UTUC is an uncommon neoplasm accounting for only 5-10% of all urothelial malignancies (9). Additionally, most previous studies included SD and other variants, such as glandular, micropapillary, sarcomatoid differentiation among others, in the single-variant histology group (7). Considering this heterogeneity, analytical results regarding the predictive value of variant histology remain controversial.
The gold standard treatment for localised UTUC is radical nephroureterectomy (RNU) with ipsilateral bladder cuff excision (10). Nonetheless, despite definitive therapy, disease recurrence and cancer-related death occur in up to 30% of patients (11, 12). Rink et al. was the first to show that variant histology, including SD, adversely influenced survival outcomes in patients diagnosed with localised UTUC, although such an association was not statistically significant based on multivariate analysis (13). Moreover, given that only a few studies have shown survival outcomes focusing on patients with UCSD, the prognostic value of the presence of SD still remains unclear and controversial (14-16).
Therefore, the current study aimed to assess the impact of SD on oncological outcomes of patients with UTUC undergoing RNU.
Patients and Methods
Patient population. This study retrospectively reviewed 244 consecutive patients with UTUC who underwent RNU at the University of Occupational and Environmental Health (UOEH; Kitakyushu, Japan) between May 2005 and October 2019. Only patients who had histologically confirmed pure UC or UCSD were included. Patients with other histological variants, those with distant metastatic disease prior to RNU and those who received neoadjuvant chemotherapy were excluded from the study, resulting in a total of 224 patients for analysis. All intended procedures in the present study were approved by the University of Occupational and Environmental Health Institutional Review Board (approval no. UOEHCRB20-134).
Patient management. Patients underwent a routine RNU assessment, including physical examination, laboratory tests, confirmation of UTUC via urine cytology or ureteroscopic biopsy, chest– abdominal–pelvic computed tomography and bone scintigraphy. Tumour stage and nodal status were assigned according to the tumour, lymph node and metastasis staging system (17). RNU was performed laparoscopically in 192 patients and via open surgery in 32 patients. Lymphadenectomy was performed at the discretion of each individual surgeon when preoperative lymph node swelling was observed on computed tomography or when intraoperative lymph node swelling was observed.
The histological features of RNU specimens were reviewed again by a dedicated genitourinary pathologist who was blinded to the clinical outcomes. UCSD was defined as the mixed presence of UC and SD in the same tumour based on the World Health Organization Classification of Tumours (18). Additionally, SD required the presence of intercellular bridges or keratinisation for diagnosis (19). Based on morphological histology, all patients were divided into pure UC or UCSD groups. The histological grade was determined based on the World Health Organization grading system (20). Lymphovascular invasion (LVI) was defined as lymphatic/vascular invasion.
Patients with pathological T3–T4 or positive lymph node status were eligible for adjuvant chemotherapy. Cisplatin-based adjuvant chemotherapy was administered for two or three cycles considering factors including performance status, the patient’s wishes and comorbidities. Regarding the chemotherapy regimen, a combination of methotrexate, vinblastine, doxorubicin and cisplatin (MVAC) was administered until February 2009, after which gemcitabine and cisplatin (GC) was administered.
Prognostic assessment. Postoperative follow-up was performed every 3 months for 2 years, every 6 months until 5 years and then annually thereafter. Routine follow-up consisted of physical examinations, laboratory tests, urinary cytology, cystoscopy and computed tomography. When symptoms appeared, appropriate additional examinations were conducted. The intravesical recurrence-free survival duration was calculated from the date of RNU to the date of local recurrence of urinary bladder cancer, death due to any cause or last follow-up in patients with no known recurrence. The duration of metastasis-free survival (MFS) was calculated from the date of RNU to the date of extravesical recurrence, death due to any cause or last follow-up in patients with no known metastasis. The OS duration was calculated from the date of RNU to the date of death due to any cause or last follow-up in patients who survived. Moreover, follow-up information was obtained through the medical records of our institute or local hospitals. For patients whose medical records did not include follow-up data, information was obtained through telephone contact.
Statistical analysis. All statistical analyses were performed using EZR ver.1.40 (Easy R, Vienna, Austria), a graphical user interface for R (The R Foundation for Statistical Computing). Associations between categorical variables were examined using the Fisher exact test, whereas continuous variables were compared using Mann– Whitney U-test. Intravesical recurrence-free survival, MFS and OS were estimated using the Kaplan–Meier method and log-rank test. Univariable and multivariable Cox proportional hazards models were utilised to determine time to recurrence and mortality. A value of p<0.05 was considered statistically significant.
Results
Patient characteristics. Among the 224 patients, 197 (87.9%) and 27 (12.1%) had pure UC and UCSD, respectively. The clinicopathological characteristics of the patients stratified by pure UC and UCSD are summarised in Table I. No significant differences in age, sex, concomitant carcinoma in situ and tumour grade were observed between the groups. Only 44 (19.3%) patients (pure UC: 35; UCSD: nine) underwent lymphadenectomy. UCSD was significantly associated with an advanced tumour stage and a higher rate of LVI. In the UCSD group, patients with ≤pT2 and ≥pT3 had an LVI rate of 22.2% (n=9) and 100% (n=18), respectively. Adjuvant chemotherapy was provided for 35 (17.8%) and 14 (51.9%) patients with pure UC and UCSD, respectively (p<0.001).
Patient characteristics.
Association between UCSD and disease recurrence and mortality. The median follow-up time was 58 months (interquartile range=30-84 months), during which 103 (45.9%) and 66 (29.5%) patients experienced intravesical and extravesical recurrence, respectively, whereas 62 (27.7%) died. No significant differences in intravesical recurrence-free survival rates were observed between the pure UC and UCSD groups (Figure 1A). However, patients with UCSD had poorer MFS and OS rates compared to those with pure UC (Figure 1B and C). The pure UC and UCSD groups had 5-year MFS rates of 65.2% (median MFS=135 months) and 40.9% (median MFS=42 months), respectively (p=0.005), and 5-year OS rates of 74.4% (median OS=159 months) and 49.0% (median OS=61 months), respectively (p=0.002). Patients with UCSD were more likely to receive adjuvant chemotherapy, with no significant difference in intravesical recurrence-free survival, MFS and OS rates between such patients with pure UC and UCSD (Figure 2).
Kaplan–Meier curves for the studied patients stratified according to pure urothelial carcinoma (UC) and that with squamous differentiation (UCSD) after radical nephroureterectomy. A: Intravesical recurrence-free survival. B: Metastasis-free survival. C: Overall survival.
Kaplan-Meier curves for the studied patients who underwent adjuvant chemotherapy stratified according to pure urothelial carcinoma (UC) and that with squamous differentiation (UCSD) after radical nephroureterectomy. A: Intravesical recurrence-free survival. B: Metastasis-free survival. C: Overall survival.
Univariate and multivariate analyses for factors associated with clinical outcomes. Table II and Table III detail the results of univariate and multivariate Cox proportional hazards regression analyses predicting MFS and OS after adjusting for clinicopathological characteristics. Accordingly, the presence of SD was identified as a significant independent predictor of MFS (hazard ratio=1.88, 95% confidence intervaI=1.18-3.65; p=0.027) and OS (hazard ratio=1.70; 95% confidence intervaI=1.01-2.88; p=0.048). Moreover, advanced tumour stage and LVI were significantly associated with both survival outcomes.
Results of univariate and multivariate analyses for clinicopathological factors associated with metastasis-free survival in the studied patients.
Results of univariate and multivariate analyses for clinicopathologic factors associated with overall survival in the studied patients.
Discussion
To assess the influence of SD on oncological outcomes of UC, the current study pathologically reviewed RNU specimens and determined the clinical characteristics of patients treated with RNU for pure UC or UCSD. Patients with UCSD had poorer MFS and OS compared to those with pure UC. Moreover, the presence of SD was significantly associated with oncological outcomes based on multivariate analyses. The presence of SD was identified as an independent prognostic factor in patients with UTUC.
The incidence rate of UCSD in the present study was 12.1% of specimens, which is similar to that reported by Shibing et al. (11.5%) (21). Moreover, the association between the presence of SD and poor prognostic features, including advanced tumour stage, higher grade, LVI and lymph node status, was consistent with that reported in previous results (13, 16, 21). Interestingly, our findings showed that all patients with ≥pT3 UCSD had LVI, suggesting that patients with UCSD in UTUC have more aggressive tumour-specific characteristics.
Only a few studies on patients with UTUC have investigated the effect of SD on postoperative prognosis. Notably, our findings showed a significant difference in MFS and OS between patients with pure UC and UCSD. Previous studies by Yu et al. (22) and Tang et al. (16) found 5-year cancer-specific survival rates of 47% and 48.9%, respectively, following RNU for UCSD. Although most previous studies indicated that patients with UCSD had lower survival rates compared to those with pure UC based on univariate analysis, some studies failed to identify significant differences based on multivariate analysis (15, 16, 23). In contrast, our results revealed that the presence of SD was associated with increased risk of mortality, similar to the findings reported by Lee et al. (14). Inconsistent results may be attributed to several factors, such as retrospective study design, differences in patient populations, pathological recognition for SD and covariates included in the multivariate analysis. A recent meta-analysis by Mori et al. (7) of six retrospective studies (15, 16, 24-27) showed that the presence of SD and glandular differentiation was associated with worse CSS (pooled hazard ratio=1.48, 95% confidence intervaI=1.14–1.92).
The present study found no difference in intravesical recurrence-free survival between the pure UC and UCSD groups. Although detailed analysis was not performed, this discrepancy implies that salvage treatment, such as intravesical chemotherapy for bladder recurrence, may affect clinical outcomes. Similar results were found by Makise et al. (15), who reported no significant association between SD and bladder recurrence.
Our findings showed that patients with UCSD had a higher probability of receiving platinum-based chemotherapy (MVAC or GC regimen) after RNU compared to those with pure UC. Unfortunately, the latest prospective randomised controlled trial (The POUT trial) that assessed the efficacy of platinum-based perioperative chemotherapy in UTUC did not compare the oncological outcomes according to different histological types (28). Given the lack of recommended salvage regimens for UCSD, evaluating the effects of adjuvant chemotherapy after RNU in patients with UCSD may be of interest. In our study cohort, the limited number of patients with UCSD did not allow to the stratification according to the presence (n=14) or absence of adjuvant chemotherapy (n=4). However, no significant differences in MFS and OS were observed between the group with pure UC and that with UCSD in patients who underwent adjuvant chemotherapy. Although UCSD was associated with poor oncological outcomes, our results may suggest that adjuvant chemotherapy increases survival in patients with UCSD after RNU. A multi-institutional cohort study by Chung et al. demonstrated a significant difference in survival outcomes after RNU between patients with pure UC and variant histology, including SD. However, among the 84 patients who underwent adjuvant chemotherapy, recurrence-free, cancer-specific and overall survival rates did not significantly differ between the groups (29). Similar results were reported in another study conducted by Kim et al. (30). Murakami et al. showed that the adjuvant chemotherapy group had significantly better CSS compared to the non-adjuvant chemotherapy group in patients with ≥pT3 or positive lymph node status in variant histology (31). These studies recommended adjuvant chemotherapy in patients with variant histology in UTUC. Therefore, examining the difference in response to chemotherapy according to histological subtype is imperative. Interestingly, the data from the KEYNOTE-045 study showed that patients with variant histology were more likely to benefit from pembrolizumab compared to those with pure UC (32). Further investigations comparing the effects of immune checkpoint inhibitor on adjuvant chemotherapy for UCSD in UTUC are also warranted.
This study has several limitations worth noting, including its retrospective non-randomised single-institutional design and small sample size. In particular, lymphadenectomy was performed in only 20% of the cases given the lack of standard criteria, which may have resulted in biased prognostic outcomes. Moreover, the adjuvant chemotherapy regimen was not uniform, with some patients receiving MVAC and others GC. Despite these limitations, our results suggest that patients with UCSD are at high risk for metastasis and short survival. As such, evaluating differences in oncological outcomes between histological variants and pure UC in UTUC is important. We believe that our study provides a better understanding of the significance of morphological diversity in terms of the biological behaviour of UTUC. Understanding the oncological effects of SD may help individualise treatment approaches and improve intensive follow-up schedules in patients undergoing RNU. Multi-institutional studies with larger cohorts and preferably a prospective study design are warranted to further validate our results.
In conclusion, the presence of SD in UC was found to be an independent predictor of metastasis and mortality after RNU.
Footnotes
Authors’ Contributions
AM: Conceptualisation, investigation, data curation, formal analysis and writing of the original manuscript. HN: Pathologic assessment. RK, MH and YN: Investigation. YH: Data curation. IT, KH and NF: Supervision. All Authors discussed, verified and approved the final version of the article.
Conflicts of Interest
The Authors declare that they have no competing interests in relation to this study.
- Received October 16, 2021.
- Revision received November 2, 2021.
- Accepted November 3, 2021.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
