Running Head: Significance of the NLR with CUETO model-Chung et al. Prognostic Significance of the Neutrophil-to-Lymphocyte Ratio in Patients with Non-Muscle Invasive Bladder Cancer treated with Intravesical Bacillus Calmette–Guérin and the Relationship with the CUETO Scoring Model

Purpose: In this study, we evaluated the predictability of a modified Club Urol ó gico Espa ñ ol de Tratamiento Oncol ó gico (CUETO) scoring model and preoperative neutrophil-to-lymphocyte ratio (NLR) in patients with non-muscle invasive bladder cancer (NMIBC). Materials and Methods: From August 2005 to May 2016, a total of 281 patients received intravesical bacillus Calmette–Guérin therapy after transurethral resection of a bladder tumor. The pathologic stage of all patients was Ta or T1. Of 281 patients, 84 (29.9%) experienced recurrence and 14 (5.0%) developed progression. The mean follow-up period was 46 months. The cut-off value for NLRs was 2.29. Results: One hundred-eight patients (38.4%) displayed a high NLR (> 2.29). In Kaplan–Meier curve analysis, a high NLR was associated with lower recurrence-free survival (RFS) ( P < .001) and progression-free survival (PFS) ( P = .002). CUETO scores were associated with RFS ( P < .001), but not with PFS ( P = .423). A combination of NLRs and the CUETO risk model correlated with RFS ( P < .001) and PFS ( P = .002). In multivariate analysis, female gender, concomitant carcinoma in situ (CIS), tumor number >3, recurrent tumors, and a high NLR were independent factors predicting recurrence (all P < .05). Concomitant CIS, recurrent tumors, and a high NLR were independent factors for predicting progression (all P < .05). Conclusion: In patients with NMIBC, an NLR >2.29 was identified as a significant factor for predicting tumor recurrence and progression. Inclusion of preoperative NLR enhanced the accuracy of the CUETO model to predict disease progression.


INTRODUCTION
The most common malignant tumor of the urinary tract is bladder cancer, and the fourth-most common cancer among males in developed countries. (1) Three-quarters of bladder cancer patients are diagnosed with non-muscle invasive bladder cancer (NMIBC), which includes Tis, Ta and T1 pathologic stages. (2) Transurethral resection of bladder tumor (TUBRT) is a primary surgical treatment used to treat patients with NMIBC. (3) After initial TURBT, immunotherapy with intravesical instillation of bacillus Calmette-Guérin (BCG) is the most effective adjuvant therapy for intermediate-and high-risk NMIBC. (4) Despite the effectiveness and safety of BCG, recurrence rates is 32.6% to 42.1% and progression rates is 9.5% to 13.4%. (5) The major treatment challenge with NMIBC is preventing progression to muscle invasive bladder cancer (MIBC), which rapidly worsens prognoses. (6) Thus, it is significant to predict risk factors for disease recurrence and progression in NMIBC patients according to individual characteristics, including pathology and choose optimal treatment modalities to enhance oncologic outcomes.
To predict recurrence of NMIBC and progression to MIBC, numerous clinical and pathological factors are commonly used to assign patients to different risk groups. Of these risk models, a scoring model developed by the Club Urológico Español de Tratamiento Oncológico (CUETO, or Spanish Urological Club for Oncological Treatment) is considered the most reliable. The CUETO model was developed as a risk-scoring tool that predicts the probability of disease recurrence and progression in BCG-treated patients at 1, 2, and 5 years. (7) It is now recognized that increased systemic inflammatory responses induced by tumor microenvironments trigger alteration of acute-phase reactive proteins and hematologic parameters. (8) Among these serum markers are neutrophil and lymphocyte counts, which can indicate relative neutrophilia and lymphocytopenia. In various tumor patients, a higher percentage of neutrophils than lymphocyte is associated with reduced cancer-free and overall survival. (9)(10)(11) We evaluated the efficiency of a modified CUETO scoring model combined with preoperative NLRs to predict recurrence and progression of disease in NMIBC patients.

Ethics statement
This study was approved by the institutional review board of Kyungpook National University, Hospital, Daegu, Republic of Korea (IRB Number KNUH 2020-03-042). The study was carried out in agreement with the applicable laws and regulations, good clinical practices, and ethical principles as described in the Declaration of Helsinki. The institutional review board of Kyungpook National University Chilgok Hospital waived because of the retrospective nature of the study.

Study design
The study was a single institution, retrospective observational trial. From August 2005 to May 2016, a total of 281 patients who underwent BCG induction therapy after complete TURBT were included in this study. They had not previously received intravesical BCG and showed no side effects of BCG or signs of recurrence during BCG therapy. All patients were diagnosed as histologically Ta or T1. A second TUR was performed 2 to 4 weeks after initial resection if a bladder tumor specimen did not include detrusor muscle or when a high-grade tumor was detected. Patients with only carcinoma in situ (CIS) were excluded, as were those found to have advanced bladder or ureteral tumors or non-urothelial carcinoma at the first TUR. Patients with hematologic malignance and acute or chronic infection were also excluded.
Preoperative NLR was calculated by a complete blood count with differential. Preoperative NLR was measured once at least 2 weeks before surgery. The best cut-off value of NLR was computed to be 2.29 in accordance with the receiver operating characteristic (ROC) curve. And the area under the ROC curve was 0.651 (95% CI 0.578-0.724; P < .001) (sensitivity: 59.5%, specificity: 69.5%) (Figure 1).
The follow-up period of patients was calculated from the first TURBT to the last cystoscopy examination. We performed urine cytology, cystoscopy, chest X-ray, and abdominopelvic computed tomography (CT) scans for follow-up study. During the first year after TURBT, follow-up study was conducted at 3, 6, and 12 months. Cystoscopy and urine cytology were performed every 6 months until 2 years after TURBT, and yearly thereafter. Imaging analyses, including chest X-ray and CT scans were examined every 6 months from 1 to 5 years, and annually thereafter.
Recurrence of disease was defined as a newly pathological confirmed bladder cancer regardless of stage after completion of BCG induction therapy. Progression of disease was defined as from Ta or T1 to stage T2 or higher disease (MIBC).

Intravesical bacillus Calmette-Guérin instillation
In all patients, BCG tice strain 12.5 mg (Oncotice) was used. A BCG suspension with 50 mL of 0.9% normal saline was instilled to the bladder via a 10 Fr urethral catheter. Patients were advised not to urination for two hours. Induction BCG therapy was initiated 2 weeks after TUR and repeated once a week for 6 weeks. We did not perform BCG maintenance therapy. There were no patients who received immediate postoperative instillation of chemotherapy.

Club Urológico Español de Tratamiento Oncológico scoring model
Scoring tumor recurrence and progression were calculated according to the CUETO scoring model, which includes age, gender, previous recurrence status, tumor stage (2002 TNM classification) and grade, multiplicity, and concomitant CIS. In accordance with the 2004 World Health Organization (WHO) grading system, we classified tumor grades as low or high.

Statistical analysis
Non-continuous variables of patient characteristics, including gender, T stage, tumor grade, concomitant CIS, size (≦ 3 cm vs > 3 cm), number (≦ 3 vs > 3), recurrence status and CUETO risk model (categorical) were analyzed using the chi-square test. Student's t-test was used to analyze continuous variables such as age, body mass index (BMI), CUETO risk model (noncategorical) and follow up periods. In addition, univariate and multivariate Cox regression model was used for analysis of tumor recurrence and progression, and Kaplan-Meier curves via a log-rank test were used for analysis of recurrence-free survival (RFS) and progressionfree survival (PFS). Statistical analysis was performed using SPSS 16.0 for Windows (SPSS Inc., Chicago, IL, USA), and a P value < .05 was considered statistically significant. Table 1 lists the characteristics of patients with NMIBC. A total of 173 patients had an NRL ≤ 2.29 (61.6%) and 108 (38.4%) had an NLR > 2.29. Eighty-four patients (29.9%) experienced recurrence and 14 (5.0%) showed progression after TUR followed by BCG therapy. Female patients accounted for 12.8% (36/281) of the study groups. No significant differences were evident in gender, age, BMI, and follow-up periods between the two NLR groups. There were no significant differences in tumor T stage, grade, concomitant CIS, size, number, and prior recurrence status between the two NLR groups. No significant differences were found in categorical or non-categorical CUETO scores between the two NLR groups. Table 2 shows subgroup analysis of high risk NMIBC patients (n = 251). High risk group was defined if any high grade tumor or CIS is present. (12) Subgroup analysis of high risk group showed the similar results to the overall group. Table 3 shows absolute neutrophil, lymphocyte count and NLR according to recurrence and progression. NRL was significantly higher in patients who showed recurrence or progression.  Kaplan-Meier curve analyses with a log-rank test are shown in Figure 2, 3, and 4. A high NLR (> 2.29) were associated with significantly low RFS and PFS (P < .001, P = .002) ( Figure   2). A high CUETO was associated with a significantly low RFS (P < .001), but there was no significant association between CUETO scores and PFS (P = .423) (Figure 3). After combining NLRs (cutoff value = 2.29) and CUETO scoring (cut-off value = 7), the modified risk model showed that high NLR and CUETO scores were significantly associated with low RFS and PFS (P < .001, P = .002) (Figure 4).

DISCUSSION
This study identified NLR as a significant factor for predicting tumor recurrence and progression, and inclusion of preoperative NLR enhanced the accuracy of the CUETO model to predict progression in patients with NMIBC.
In the early stages, NMIBC is not life-threatening, but it will recur in more than half of patients and progress from 10% to 20% to MIBC. (13) Although numerous efforts have been made to predict and prevent tumor recurrence and progression, the exact characteristics of NMIBC are unknown due to its heterogeneity.
The European Organization for Research and Treatment of Cancer (EORTC) has developed a simple scoring system that uses information such as tumor size and number, prior recurrence rate, stage, and concomitant CIS and WHO grade based on data of 2596 patients with NMIBC, to predict the risk of relapse and progression. (14) The CUETO scoring model was created to compensate for the EORTC with low rates of BCG treatment, using information from 1062 patients who received BCG treatment. (7) Compared to EORTC, where most of the 78% patients received intravesical chemotherapy, all patients in the CUETO study received BCG instillation, and 15% of them received mitomycin C.
Intravesical instillation of BCG is a standard treatment for CIS and an adjuvant option for T1 and higher-grade Ta bladder tumors after TUR. (15) The CUETO model is thought to be more suitable for patients treated with BCG. In this study, as with the CUETO study, we included patients who completed 6 BCG instillations. However, compared with the CUETO scoring model, only female gender, concomitant CIS, multiplicity (> 3), and prior recurrence status were significant factors for predicting tumor recurrence (all P < .05). About tumor progression, only concomitant CIS and prior recurrence status were significant factors (all P < .05). A Kaplan-Meier curve analysis demonstrated that the CUETO score was associated with RFS (P < .05), but not with PFS (P = .423). We therefore decided to add the NLR ratio to the CUETO scoring model if inclusion of an NLR would enhance the predictability of CUETO scoring.
Preoperative NLR has proven to be a useful marker and a high NLR has been linked to higher tumor stages and adverse oncologic outcomes in numerous cancers, including not only the gastrointestinal cancer but genitourinary tract cancer, such as urothelial carcinoma of the bladder. (16)(17)(18)(19) Although the pathophysiology is not understood clearly, relative neutrophilia may increase inflammatory markers that include proangiogenic factors, growth factors, proteases, and antiapoptotic markers, which facilitate tumor growth and progression. (20) In addition, lymphocytopenia may destroy cell-mediated immune responses and therefore worsen prognoses. (21) In bladder tumors, several previous studies have evaluated the predictive value of NLRs (22) ; most focused on MIBC and were conducted mainly on patients who underwent radical cystectomy. (23)(24)(25)(26) (29) evaluated whether an NLR ratio can predict the response to BCG in high-risk NMIBC patients. One hundred consecutive patients with newly diagnosed high-risk NMIBC were analyzed retrospectively. All received an induction course of intravesical immunotherapy with BCG followed by a maintenance course for at least a year. Forty-eight patients underwent radical cystectomy for high-grade recurrence or progression to muscle invasive disease (BCG non-responder group). The mean NLR was 2.61 ± 0.77 in the BCG responder group and 3.65 ± 1.16 in the BCG non-responder group (P = .01). The NLR was associated with both recurrence (P = .01) and progression (P = .01). A Kaplan-Meier analysis with a log-rank test showed statistically significant differences between the curves for an NLR < 3 and an NLR ≥ 3 (P < .05).
Based on the ability of the NLR to predict tumor recurrence and progression, we added the NLR to the CUETO scoring model. Using the CUETO scoring model alone, a significant association was observed with low RFS (P < .001), but not with PFS (P = .423) (Figure 2).
However, after combining the NLR (cut-off value 2.29) and CUETO scoring model (cut-off value 7), the resulting modified risk model showed that a high NLR and high CUETO score were significantly associated with both low RFS and PFS (P < .001 and P = .002, respectively) ( Figure 3). Other combined risk models have been shown to enhance the predictability of each risk model. (8,21) Getzler et al. provided statistical evidence that an NLR > 2.5 may improve the predictive power of an EORTC score when the two are calculated together. In 2019, Aydin et al. evaluated the correlation between NLR and EORTC recurrence and progression scores. They reported that as the NLR increased, recurrence (P < .001) and progression (P = .034) scores increased significantly. Nevertheless, this study is the first to analyze the prognostic significance of the NLR and its synergic relation with the CUETO scoring model in patients with NMIBC after intravesical BCG instillation. Furthermore, interestingly, there were no differences in the clinical and pathological findings between the two NLR groups. And this highlights that the biological properties of tumor cells may be very different from the pathological and anatomical characteristics of the tumor. As such patients within a specific pathological classification may have differing prognosis due to differing biological properties such as the degree of immune dysfunction.
There were several limitations to be considered in this study. First, it was based on a retrospective analysis of the records of patients treated at a single institution with unavoidable selection biases. Small numbers and heterogeneous patients are also weak points. It should also be noted that in many previous studies, various NLR cut-off values were evaluated and utilized. (30) Each study's results should be interpreted carefully. Because the idealized and generalized NLR have not yet been established, each study selected cut-off values with different sensitivities and specificities. Furthermore, the main limitation concerning NLRs is the volatile counts of neutrophils and lymphocytes. Although we excluded patients with hematologic malignances and acute or chronic infections, it is possible that individual chronic medications, herbs, or antibiotics affected the NLR value. As the NLR is a dynamic parameter (unlike standard pathological parameters), the dynamic changes of NLR after various treatments of bladder cancer may be important in the clinical day-to-day management of patients. Few studies have been reported on the NLR measured after TURBT or BCG instillation, therefore, studies comparing NLR before and after treatment of bladder cancer or optimal timing of NLR determination are also essential, either. A prospective study with a larger cohort is required to solidify the place of NLR in predicting disease recurrence and progression in patients with NMIBC in the future.

CONCLUSIONS
Our study showed that in patients with NMIBC, the NLR was identified as a significant factor for predicting tumor recurrence and progression. Furthermore, inclusion of a preoperative NLR enhanced the accuracy of the CUETO model to predict disease progression. NLR is promising and inexpensive hematologic biomarker which can be applied to clinical decision making and estimation of oncologic outcomes in the bladder cancer patients. We therefore recommend that patients with a high NLR receive more aggressive management.