Comparison of intravesical instillation of Bacille Calmette-Guérin and epirubicin: a randomized controlled trial

The optimal regimen of instillation for patients with non-muscle invasive bladder cancer (NMIBC) and the recurrence-related factors are still needed to be clear. The patients underwent transurethral resection of bladder tumor who diagnosed as intermediate or high risk NMIBC were randomized to receive bacillus Calmette–Guérin (BCG) 19, 15 times or epirubicin (EPI) 18 times (2:2:1). The primary end point was disease-free survival (DFS), the secondary end point was progression free survival (PFS). Associations between clinicopathological factors and prognosis were estimated, with accuracy of the prognostic models Spanish Urological Club for Oncological Treatment (CUETO) and European Organization for Research and Treatment of Cancer (EORTC) evaluated. a. Eastern Cooperative Oncology Group performance status score >1; b. patients with active tuberculosis or receiving treatment for anti-tuberculous; c. immune deciency or undergoing immunosuppressive therapy; d. severe complication (e.g., serious cardiovascular and cerebrovascular disease), or presence of other types of cancer; e. previous diagnosis with muscle-invasive bladder cancer; f. patients underwent treatment (e.g., chemotherapy, radiotherapy or immunotherapy) during the previous 4 weeks which may inuence the research results; f. serious intraoperative and postoperative complications (e.g., bladder perforation, serious postoperative hematuresis, bladder irritation, etc.); and g. patients not suitable to receive treatment or not able to participate in our trial due to pregnancy, severe disability, serious psychological problems, etc.


Conclusion
BCG instillation signi cantly reduced the rate of recurrence compared with epirubicin in this population.
The EORTC and CUETO models exhibited high accuracies for the prediction of BCG treatment failure.

Trial registration:
This study was approved by Center for Drug Evaluation of National Medical Products Administration (CDENMPA, China) and Ethics Committee of our institute, with the whole process supervised. The registration number in CDE-NMPA was CTR20150840 (website of CDE: http://www.cde.org.cn/).

Page 4/16
Approximately 75% of patients with bladder cancer have non-muscle invasive bladder cancer (NMIBC; Ta or T1 disease) at the time of primary diagnosis (1). These patients can be divided into three risk groups (i.e., low, intermediate, and high risk). For intermediate-and high-risk patients, transurethral resection of bladder tumor (TUR-BT) and subsequent intravesical drug instillation are the standard treatments. However, there is currently no conclusion reached regarding treatment with instillation (2).
It has been con rmed that treatment with Bacillus Calmette-Guérin (BCG) following TUR-BT is superior to TUR-BT alone or combined with chemotherapy in Caucasian patients (3,4). In addition, maintenance treatment with BCG has been recommended for intermediate-and high-risk patients (2). However, the use of BCG in Chinese patients has been limited by the high rate of intolerance and complications previously reported in clinical practice, and the design of trials of BCG instillation has been suboptimal. Moreover, despite the administration of maintenance therapy with BCG, 32.6% and 13.4% of patients continue to experience recurrence and tumor progression, respectively (5). In this study, we performed a randomized controlled trial to evaluate the e cacy and safety of adjuvant intravesical BCG or epirubicin (EPI) therapy for different regimes; we also estimated the predictive value for prognostic models to evaluate the prognostic factors for NMIBC. f. patients underwent treatment (e.g., chemotherapy, radiotherapy or immunotherapy) during the previous 4 weeks which may in uence the research results; f. serious intraoperative and postoperative complications (e.g., bladder perforation, serious postoperative hematuresis, bladder irritation, etc.); and g. patients not suitable to receive treatment or not able to participate in our trial due to pregnancy, severe disability, serious psychological problems, etc.

Methods
Post-operative immediate irrigation with EPI 50 mg was performed within 24 h after TUR-BT. After con rming the presence of intermediate-or high-risk NMIBC according to surgical ndings and the nal pathological report, patients who participated in the study were randomized into three bladder irrigation groups: a. BCG 19 group (BCG 120 mg weekly for six times, followed by biweekly irrigation for three times, and then monthly for 10 times, for a total of 19 irrigations with BCG); b. BCG 15 group (BCG 120 mg weekly for six times, followed by irrigations for the rst 3 weeks and at 3, 6, and 12 months after surgery, for a total of 15 irrigations with BCG); and c. EPI 18 group (EPI 50 mg weekly for eight times, followed by monthly irrigation for 10 times, for a total of 18 irrigations). The randomization was performed within 14 days after surgery, at a 2:2:1 ratio (Das for IWRS Version 5.0, BioVoice & BioGuider, Beijing, China). Subsequently, irrigation was initiated on day 14 post operation and the duration of treatment was 1 year.

Follow-up and outcomes
For intermediate-risk patients, cystoscopy was performed 3 and 6 months after surgery, then biannually until 5 years. For high-risk patients, cystoscopy was performed trimonthly until 2 years after surgery, then biannually until 5 years. Urinalysis was performed prior to each irrigation. Patients were followed up until recurrence, progression, or discontinuation of treatment for any reason. Adverse effects were recorded at each follow-up according to the Common Terminology Criteria for Adverse Events Version 4.0(6).
The initiation time was the day TUR-BT was performed. The primary end point was recurrence during follow-up, while the secondary end point was progression to muscle-invasive bladder cancer or therapyrelated discontinuation of treatment. Recurrence-free survival (RFS) and progression-free survival (PFS) times were recorded. BCG failure and intolerance were recorded according to the de nitions of the EAU (2).

Data collection
Recorded clinical variants included gender, age and smoking history. Intraoperative recordings included tumor size, site, and number. Tumor histopathological type, T stage (8th American Joint Committee on Cancer TNM classi cation system)(7), tumor grade (2004 World Health Organization grading system), and carcinoma in situ status. All pathological reports were provided by the Pathology Department of West China Hospital.

Statistical methods
Differences in survival between two or more subgroups were evaluated using log-rank tests. Univariate and multivariate Cox regression analyses were performed to determine the clinicopathological parameters associated with the recurrence of patients with NMIBC. Patients were strati ed using the European Organization for Research and Treatment of Cancer (EORTC) risk tables (8) and the Spanish Urological Club for Oncological Treatment (CUETO) scoring model (5). The predictive ability of these outcome prediction models was evaluated using the concordance index (c-index)(9).
Statistical analyses were performed using the SPSS Statistics version 25 (IBM, Armonk, NY, USA) and The R Programming Language 3.5.0 software (The R Foundation, Vienna, Austria). A P-value <0.05 denoted statistical signi cance.

Patient baseline data
Totally 93 patients were enrolled for analysis (BCG19, BCG15 and EPI 18 groups for 35, 37 and 21 cases respectively, Table 1). The mean age was 62.96 years (standard deviation: 8.16 years), and the median follow-up time was 33.46 months (interquartile range: 18.30-44.80 months). There was no signi cant difference noted in baseline clinicopathological factors between the BCG and EPI treatment subgroups, except for T stage. Eight patients underwent radical cystectomy after NMIBC.

E cacy and safety for BCG and EPI treatment
After follow-up for more than one year, recurrence occurred in 12 BCG treated patients (16.67%), with 9 of them suffered recurrence in the course of BCG irrigation. Progression happened in 6 BCG treated patients (8.33%). Meanwhile, number of recurrence and progression for EPI were 10 (47.61%) and 2 (9.52%) respectively. 6 cases suffered recurrence during EPI therapy (28.57%). RFS was signi cantly longer in the BCG group (Figure 1, P=0.002), but no signi cant differences were seen between BCG19 and BCG15 regime subgroups (Figure 1, P=0.975). PFS differences between BCG and EPI therapy could not be evaluated since limited ending events.
Side effects were also collected ( Table 2). Higher general complication rate was seen in BCG group compared with EPI groups (83.33% and 61.90%, respectively; P=0.022); however, most of complications for BCG group were treatment-free and there were no severe adverse events (grade 3-5) happened for these two treatment groups. Urocystitis was the most frequent complication for intravesical therapy, for which incidences were signi cant higher in BCG subgroup (75.00% vs 52.38%, P=0.047); fever was a unique complication after BCG therapy (16.67%), but all of them were treatment-free. Hematuresis occurred both after BCG and EPI treatment, with no signi cant difference demonstrated (43.1% vs 42.9%, P=0.987). There were more patients suffered rare complications such as respiratory infection, hypertension and hyperglycemia after BCG therapy, but difference wasn't signi cant (29.17% vs 9.52%, P=0.122). Six patients didn't complete full cycle of BCG therapy due to intolerance of complications, and all of them happened during the maintenance period (each had 3 patients for BCG 19 and 15 subgroups; ve were due to urocystitis and one for hematuriesis); only one patient among them suffered disease recurrence two years after treatment had ended.
In uence of clinicopathological factors and treatment methods on recurrence Among the 93 post-TUR-BT patients, the univariate analysis revealed that better survival was observed in those without smoking and previous bladder cancer history, with younger age, low tumor stage and less tumor number (Table 3). However, only prior recurrent history was signi cantly correlated with poor RFS (P=0.001). Meanwhile, when we considered these in BCG treatment subgroups (n=72), we also found better PFS for male patients and those with well nuclear grade, but weren't signi cant; and PFS for those patients with prior recurrence rate less than one per year signi cant poorer than those with recurrence rate for more than one per year (P=0.033).
The EORTC risk tables and CUETO scoring model were used to predict recurrence in patients with postoperative NMIBC. For all 93 cases, the c-index of EORTC and CUETO was 0.766 and 0.741, respectively. These numbers were obviously higher than those of prediction according to tumor stage or grade alone or by EAU risk group strati cation (Table 4). Meanwhile, higher predictive accuracy was recorded in the BCG subgroup (0.812 and 0.817, respectively) ( Table 4).

Discussion
Recurrence of NMIBC is affected by numerous factors, including tumor size, number of tumors, prior recurrence rate, T stage, grade, and presence of carcinoma in situ (8). Patients at high risk of recurrence and progression can be identi ed through a comprehensive consideration of these factors. In the risk group strati cation of EAU, patients with NMIBC are strati ed into three risk groups (low, intermediate, and high risk) and different treatment strategies and follow-up measures are recommended (2). However, these strategies are not effective, with 47.8% of patients suffering at least one recurrence (8).
Since Morales rst reported the irrigation treatment with BCG in bladder cancer (10), the e cacy of this method has been widely demonstrated and became the standard therapy for intermediate-and high-risk NMIBC (11). Lower recurrence rates were observed in BCG-treated patients compared with those who underwent chemotherapy (3,12); however, higher complication rates also occurred (13). There is a limited number of studies on the use of BCG in Chinese patients due to the intolerance and contraindications of therapy with BCG. Nonetheless, this was gradually accepted by patients due to the improvement of supporting method. We rstly reported a randomized controlled trial of Chinese patients which compared the e cacy of BCG and EPI on NMIBC. The results suggested that, although six patients showed intolerance to therapy with BCG, this approach has advantages in preventing recurrence, without a signi cantly higher rate of side effects versus treatment with EPI.

Six weeks inductive BCG instillations only are not enough for intermediate and high risk NMIBC treatment,
and current evidences support 1-3 years maintenance schedule (14). A recent meta-analysis which included 1951 patients demonstrated that longer-maintenance therapy didn't increase side effect compared with induction therapy only, and longer maintenance (more than 1 years) wasn't superior to 1 years maintenance (15). However, the optimal number and frequency of maintenance intravesical therapy is still unknown (2). Since that, we compared two regime of 1 years BCG maintenance therapy, and found that there was no signi cant difference between BCG 15 and 19 regimes among recurrence and side effects.
Even treated with BCG, recurrence continues to occur in many patients treated with BCG; 32.6% as reported by Jesus et al. (5). Even after a full course of maintenance therapy with BCG, 15.7% and 26.3% of patients continue to experience early and late recurrence, respectively (16). Hence, considerable efforts were focused on the prognostic prediction of NMIBC to more rationally stratify patients, and identify those more suitable for individual surveillance and treatment programs. The EORTC scoring system, based on patients who underwent several instillation protocols, is used to predict recurrence and prognosis in patients with NMIBC (c-index of internal validation of 0.66 and 0.75, respectively) (8). Subsequently, the CUETO scoring model was designed to predict prognosis in patients receiving maintenance therapy with BCG (c-index of interval validation of 0.636 and 0.687 for early recurrence and progression, respectively)(5). Compared with risk group strati cation, the tumor recurrence status (an important prognostic factor) was included in the CUETO and EORTC. Furthermore, in our study, these two models exhibited signi cantly better differentiations than risk group strati cation, especially in BCGtreated patients, suggesting that CUETO and EORTC may be more suitable for the prediction and strati cation of patients with NMIBC.
There are limitations for our study. For one thing, the sample size is relatively small for further subgroup analysis; for another, the follow-up time is short so that the secondary end point of progression could not be evaluated since limited positive events occurred. However, since all patients had completed more than 1 year's follow-up with complications recorded in detail, present results can well re ect the feasibility of BCG instillations in Asian NMIBC populations.

Conclusion
Maintenance treatment with BCG demonstrated better control of recurrence in Chinese patients with NMIBC after TUR-BT versus intravesical therapy with EPI. Prior recurrence status is an important prognostic factor for recurrence, while the EORTC and CUETO prognostic models exhibited higher prediction accuracy for recurrence than tumor stage, grade, or risk group strati cation.

Declarations
Ethics approval and consent to participate and publications This study was approved by Center for Drug Evaluation of National Medical Products Administration (CDENMPA, China) and Ethics Committee of our institute, with the whole process supervised. The registration number in CDE-NMPA was CTR20150840 (Registered 21 December 2015, website of CDE: http://www.cde.org.cn/). All the research registration and ethical approval documents were added below. Patients and their authorized family members had been fully informed before follow-up work was performed, with informed consent signed. All procedures adhere to CONSORT guideline.

Availability of data and materials
The data that support the ndings of this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions.

Competing Interests
The authors declare that they have no competing interests.