Three vs. Four Cycles of Neoadjuvant Chemotherapy for Localized Muscle Invasive Bladder Cancer Undergoing Radical Cystectomy: A Retrospective Multi-Institutional Analysis

Background Three or four cycles of cisplatin-based chemotherapy is the standard neoadjuvant treatment prior to cystectomy in patients with muscle-invasive bladder cancer. Although NCCN guidelines recommend 4 cycles of cisplatin-gemcitabine, three cycles are also commonly administered in clinical practice. In this multicenter retrospective study, we assessed a large and homogenous cohort of patients with urothelial bladder cancer (UBC) treated with three or four cycles of neoadjuvant cisplatin-gemcitabine followed by radical cystectomy, in order to explore whether three vs. four cycles were associated with different outcomes. Methods Patients with histologically confirmed muscle-invasive UBC included in this retrospective study had to be treated with either 3 (cohort A) or 4 (cohort B) cycles of cisplatin-gemcitabine as neoadjuvant therapy before undergoing radical cystectomy with lymphadenectomy. Outcomes including pathologic downstaging to non-muscle invasive disease, pathologic complete response (defined as absence of disease -ypT0), overall- and cancer-specific- survival as well as time to recurrence were compared between cohorts A vs. B. Results A total of 219 patients treated at 14 different high-volume Institutions were included in this retrospective study. Patients who received 3 (cohort A) vs. 4 (cohort B) cycles of neoadjuvant cisplatin-gemcitabine were 160 (73,1%) vs. 59 (26,9%).At univariate analysis, the number of neoadjuvant cycles was not associated with either pathologic complete response, pathologic downstaging, time to recurrence, cancer specific, and overall survival. Of note, patients in cohort B vs. A showed a worse non-cancer specific overall survival at univariate analysis (HR= 2.53; 95 CI= 1.05 - 6.10; p=0.046), although this finding was not confirmed at multivariate analysis. Conclusions Our findings suggest that 3 cycles of cisplatin-gemcitabine may be equally effective, with less long-term toxicity, compared to 4 cycles in the neoadjuvant setting.

Background: Three or four cycles of cisplatin-based chemotherapy is the standard neoadjuvant treatment prior to cystectomy in patients with muscle-invasive bladder cancer. Although NCCN guidelines recommend 4 cycles of cisplatin-gemcitabine, three cycles are also commonly administered in clinical practice. In this multicenter retrospective study, we assessed a large and homogenous cohort of patients with urothelial bladder cancer (UBC) treated with three or four cycles of neoadjuvant cisplatin-gemcitabine followed by radical cystectomy, in order to explore whether three vs. four cycles were associated with different outcomes.
Methods: Patients with histologically confirmed muscle-invasive UBC included in this retrospective study had to be treated with either 3 (cohort A) or 4 (cohort B) cycles of cisplatin-gemcitabine as neoadjuvant therapy before undergoing radical cystectomy with lymphadenectomy. Outcomes including pathologic downstaging to non-muscle invasive disease, pathologic complete response (defined as absence of disease -ypT0), overalland cancer-specific-survival as well as time to recurrence were compared between cohorts A vs. B.
Results: A total of 219 patients treated at 14 different high-volume Institutions were included in this retrospective study. Patients who received 3 (cohort A) vs. 4 (cohort B) cycles of neoadjuvant cisplatin-gemcitabine were 160 (73,1%) vs. 59 (26,9%).At univariate analysis, the number of neoadjuvant cycles was not associated with either pathologic complete response, pathologic downstaging, time to recurrence, cancer specific, and overall survival. Of note, patients in cohort B vs. A showed a worse non-cancer specific overall survival at univariate analysis (HR= 2.53; 95 CI= 1.05 -6.10; p=0.046), although this finding was not confirmed at multivariate analysis.

INTRODUCTION
Bladder cancer represents approximately 3% of cancer diagnoses in the world, with a 4-time higher prevalence in males vs. females and >90% of cases diagnosed in individuals older than 55 years of age (1). Age-standardized incidence rates of approximately 20 cases per 100,000 per year in males and 4.5 cases per 100,000 per year in females are reported in Europe and North America (2), with certain areas of Italy showing remarkably higher rates [e.g. in the province of Naples, Italy, an age-standardized incidence rate of 75.3 and 16.3 cases per 100,0000 per year in males and females, respectively, has been reported (3)].
While chemotherapy agents such as mitomycin and gemcitabine (4) can be used for intravesical therapy against non-invasive bladder cancer, intravesical BCG represents the standard therapy for patients with T1 disease (5), and cystectomy (6)(7)(8), with or without perioperative chemotherapy (9), is recommended for muscle-invasive, localized bladder cancer (10). Advances in the field of peri-operative systemic therapy of muscle-invasive bladder cancer have been scarce over the past 20 years. Neoadjuvant chemotherapy regimens based on methotrexate, vinblastine, doxorubicin cisplatin or cisplatin-gemcitabine were associated with an absolute increase in 5year survival of 8% in patients with muscle invasive-bladder cancer (9). Although neoadjuvant cisplatin-based chemotherapy is currently recommended by NCCN (10) and EAU (11) guidelines in patients with muscle-invasive UBC, the optimal chemotherapy schedule and number of cycles remain to be established. According to NCCN guidelines, dose-dense methotrexate, vinblastine, doxorubicin, cisplatin combination can be administered for 3 or 4 cycles, while cisplatin, methotrexate and vinblastine combination is administered for 3 cycles. Four cycles of gemcitabine and cisplatin represent a viable option in the perioperative setting, on the grounds of the results obtained in a large randomized phase III trial (12)and of retrospective case series (13). Nevertheless, the optimal number of cycles of cisplatin-gemcitabine remains therefore to be determined, given the lack of comparative studies specifically designed to assess optimal number of cycles.
In this multicenter retrospective study, we assessed a large and homogenous cohort of patients with histologically confirmed urothelial bladder cancer (UBC) treated with 3 vs. 4 neoadjuvant cisplatin-based chemotherapy followed by radical cystectomy, to explore potential differences in outcomes in terms of cancer specific-, non-cancer specific, overall-survival, time to recurrence and pathologic response and downstaging rates.

Patients
All patients with UBC treated with neoadjuvant chemotherapy at the participating Institutions from January 2000 until January 2015 had to be assessed for inclusion in this retrospective study. Patients were included in this retrospective study if they had histologically confirmed muscle-invasive bladder cancer with predominant urothelial component and were treated with 3 or 4 cycles of cisplatin-gemcitabine (gemcitabine 1000-1250 mg/m2 on days 1 and 8, and cisplatin 70 mg/m2 on day 1, every 3 weeks) as neoadjuvant therapy before undergoing radical cystectomy with lymphadenectomy. Only patients with cT1-4N0M0 on whole body CT scan with and without contrast prior to chemotherapy start and with a follow-up after surgery longer than 36 months were included in this retrospective study. Follow-up was conducted with a whole CT scan with and without contrast every 3-6 months and additional tests (MRI, bone scan) if clinically indicated. Data about sex, age, ECOG performance status, Charlson Comorbidity index, previous number of cycles of cisplatin-based neoadjuvant therapy, hemoglobin, creatinine, absolute neutrophil and lymphocyte counts, total and HDL cholesterol, CRP, presence of histologically proven positive lymph-nodes, and cancer-specific survival (months) were required to have been measured within 14 days and be fully available for the patient to be included in this retrospective study. Date and cause of death were collected using death certificates and ISTAT (Italian National Institute of Statistics) cause of death records. These variables were considered as potentially prognostic and assessed for their association with cancer specific survival. Information regarding recurrence was extracted from medical charts. Retrospective observation started on the day of cystectomy until death or last follow-up.

Statistical Analysis
Descriptive statistics were used to describe the overall cohort with respect to the main demographical and clinical characteristics. Frequencies (percentages) were used for categorical variables, while medians (Q1; Q3), were used for quantitative variables. Associations between the number of cycles along with other potential predictive factors and the outcome variables downstaging and complete response were evaluated using univariate logistic regression models. Additionally, variables that presented a significant association in the univariate analysis were added to a multivariate logistic regression model. Time-to-event outcome variables were analyzed estimating survival curves with the Kaplan-Meier method and the difference between the curves was computed using log-rank test (Figures 1-4). The association of the number

Patients' Baseline Characteristics
A total of 245 patients receiving neoadjuvant cisplatingemcitabine at 14 different high-volume Institutions were initially evaluated for inclusion in this retrospective study. After excluding 7 patients who received 6 cycles, 3 patients who received 1 cycle, 13 patients who received 2 cycles and 3 patients with missing data, a total of 219 patients receiving 3 or 4 cycles were included in this retrospective study. Patients who received 3 (cohort A) vs. 4 (cohort B) cycles of neoadjuvant cisplatin-gemcitabine were 160 (73,1%) vs. 59 (26,9%). In cohort A, 83.1% were males and median age was 66 (IQR = 59,000 to 72,00).

Outcomes
After a median follow-up of 76 months, median overall survival was not reached in cohort A, with 43 reported deaths (34 because of bladder cancer), while it was 110 months (95%CI: 81-120) in cohort B, with 24 reported deaths (15 because of bladder cancer). Median cancer-specific survival and time to recurrence were not reached in either cohort. Pathologic downstaging to non-muscle invasive disease was reported in 129 and 43 patients in cohort A and B, respectively. Complete pathologic response was reported in 36 (22.5%) and 13 (22.0%) patients in cohort A and B, respectively, while recurrence was reported in 46 (28.7%) vs. 17

DISCUSSION
Neo-adjuvant chemotherapy based on cisplatin has represented the standard of care in patients with T2-T4 UBC for the past two decades (14), although the optimal schedule and number of cycles to administer remain to be determined (9). In a landmark meta-analysis (9) that analyzed data collected in 15 randomized trials enrolling a total of 3,285 patients neoadjuvant regimens based on cisplatin alone were not associated with any survival benefit, which was only provided by cisplatin-containing regimens including cisplatin-gemcitabine or MVAC -like chemotherapy (HR, 0.82; 95% CI, 0.74-0.91; p <.001; p = .99 for heterogeneity, I2 = 0%). This meta-analysis did not identify any differences in outcomes associated with GC vs. MVAC. In a phase II trial that randomized 237 bladder cancer patients to dose dense MVAC, administered every 14 days for 4 cycles, or gemcitabine-cisplatin, administered every 21 days for 4 cycles, the pT0 rates for ddMVAC and GC were 32% and 35%, respectively (15). In a recently published randomized phase III trial (12) designed to compare the efficacy of dose dense (dd)-MVAC or GC in the neoadjuvant/adjuvant setting, 500 patients were randomized to either either six cycles of dd-MVAC every 2 weeks or four cycles of GC every 3 weeks. Of note, in the neoadjuvant group, 218 patients were treated with dd-MVAC (60% received the planned six cycles) and 219 were treated with GC (84% received the planned four cycles). A complete pathological response rate of 42% and 36% was obtained in the dd-MVAC vs. GC arms. In a retrospective observational study that included data from 212 patients muscle-invasive UBC treated with neoadjuvant chemotherapy (146 patients treated with GC and 66 patients treated with MVAC, the pCR rate was 29% in the MVAC cohort and 31% in the GC cohort, with a median of 3 cycles of chemotherapy administered and no significant difference in the pathologic response rate among the two regimens (16). In another retrospective study (13) including 42 patients receiving 4 cycles of neoadjuvant GC, the complete pathologic response rate was 26% (95% confidence interval [CI], , and no residual muscle-invasive disease proportion (<pT2) was 36% (95% CI, 21-52); pT0 was achieved in 28% (95% CI,  and <pT2 in 35% (95% CI, 23-49) of 54 MVAC-treated patients. All 15 GC patients achieving <pT2 pathologic stage remained disease-free at a median follow-up of 30 months. In another retrospective study, including 58 patients treated with neoadjuvant GC therapy and 74 treated with neoadjuvant MVAC, similar pathologic complete response rates were obtained (20.7% vs. 18.9%, P = 0.83, respectively). Of note, while neoadjuvant GC yielded improved 2-year OS rate than neoadjuvant MVAC for clinical T2 disease (95.2% vs. 70.8%, P = 0.036), in patients with T3 or more advanced disease, neoadjuvant MVAC provided more pT0 (20.0% vs. 5.6%, P = 0.07) and better 2-year OS than neoadjuvant GC (71.1% vs. 55.0%, P = 0.142), although the difference was not statistically significant (17).
Of note, in our retrospective study, we obtained a complete pathologic response rate of 22,5% in patients receiving 3 cycles of neoadjuvant cisplatin-gemcitabine and of 22% in patients receiving 4 cycles of cisplatin-gemcitabine. These results are consistent with the published data reviewed. Pathologic complete response rates obtained with immune checkpoint inhibitors may be higher. While The PURE-01 trial reported that pembrolizumab was associated with a 37% complete response (pT0) after neoadjuvant therapy, the ABACUS trial showed that atezolizumab yielded a complete response rate of 31%. Conversely, preoperative combination of ipilimumab + nivolumab (18), cisplatin plus gemcitabine plus pembrolizumab (19), durvalumab + tremelimumab (20), was associated with a complete pathologic response rate of 46%, 44, 34.8%, respectively. Combination of 4 cycles of cisplatin-gemcitabine plus 4 cycles of nivolumab in the BLASST-1 trial was able to yield a complete pathologic response rate in 20 of 41 of patients or 49% of patients (21). Of note, the results obtained in our retrospective study in terms of pathologic complete response rate are lower than expected in both cohorts, although we cannot provide an explanation for this finding. Importantly, not only did we not identify any differences in cancer-specific survival or time to recurrence among patients treated with 3 vs. 4 cycles, but we reported that non-cancer specific survival was worse in patients receiving an additional cycle(HR= 2.53; 95 CI= 1.05-6.10; p=0.046). When we analyzed all available baseline variables between the two cohorts, we did not identify any difference that may explain this finding. Considering that long-term platinuminduced cardiovascular toxicity has been shown in young patients treated for germ-cell tumors (22), we speculate that an additional cycle may increase mid-and long-term cardiovascular toxicity with an increased risk of death in a population who is generally at high or very cardiovascular high risk (males, elderly, heavy smokers). As reported by others (13), we confirmed that pathologic downstaging and complete response were associated   In conclusion, although we are well aware of the limitations of our retrospective study, which include the study design and the limited sample size, this study represents the largest specifically designed to capture potential differences in outcomes between 3 and 4 cycles of neo-adjuvant cisplatingemcitabine. We found that 3 vs. 4 cycles may be equally effective, with a signal of decreased overall mortality in patients who received less cycles. Our finding may be incorporated in novel combination prospective trials based on cisplatin+gemcitabine +immune checkpoint inhibitors. Prospective studies are warranted.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
MF and CB contributed to the conception of the work and final version approval. All authors contributed to the article and approved the submitted version.