Prognostic value of preoperative neutrophil-to-lymphocyte ratio in histological variants of non-muscle-invasive bladder cancer.

Bladder cancer (BC) is the 10th most common form of cancer worldwide and the second most common urologic malignancy [ 1 ]. Of these, non-muscle-invasive BC (NMIBC) usually accounts for 70% to 75% of patients with BC at the time of the initial diagnosis [ 2 3 ]. Histologically, BC with variant histology accounted for approximately 15% to 25% of all patients who underwent transurethral resection of bladder tumor (TURBT) [ 3 4 ]. Of all histopathologic subtypes, squamous variant (SV) and glandular variant (GV) are the two most common histological variants. The remaining variants with significantly low incidence are collectively designated as rare histological variant (RV) [ 5 ]. Previous studies reported that most of the VH-NMIBC were more aggressive and associated with a poor response to treatment compared with pure transitional cell carcinoma (PTC) [ 6 7 8 9 ]. Consequently, urologists are striving to establish key predictors for risk stratification and therapeutic decision-making [ 10 ]. Despite identifying many prognostic factors, such as T stage, World Health Organization (WHO) grade, tumor number, BCG response, and so on, the effectiveness of some makers was under discussing in patients with VH-NMIBC [ 11 12 ]. Preoperative neutrophil-to-lymphocyte ratio (PNLR) has been discussed widely since Gondo et al. [ 13 ] reported the prognostic value of PNLR in BC. Subsequently, many studies focused on the correlation between PNLR and BC. Kaynar et al. [ 14 ] first identified that PNLR could predict the prognosis of patients with NMIBC. Although many studies reported this finding, different outcomes appeared were observed [ 13 14 15 16 17 18 ]. Meanwhile, there is insufficient evidence to suggest that PNLR can be an independent prognostic factor in patients with VH-NMIBC. Therefore, the present study aimed to investigate the prognostic value of PNLR in patients with VH-NMIBC by collecting information from our center.

Based on the results of the receiver operating characteristic curve and Youden index, the best cut-off values of PNLR were 2.2 for all groups. The median follow-up time was 50.7 months (interquartile range [IQR], 30.0–68.0 mo), 50.4 months (IQR, 30.0–67.5 mo), 54.2 months (IQR, 35.3–68.8 mo), and 43.0 months (IQR, 23.3–56.8 mo) in the total, GV, SV, and RV groups, respectively. The median follow-up time of high NLR (HNLR) and low NLR (LNLR) were 46.5 months (IQR, 28.0–58.8 mo) and 49.0 months (IQR, 31.0–78.3 mo), respectively. According to the cut-off value, patients were divided into LNLR and HNLR groups. Among 38 patients in the RV group, 13 patients were diagnosed with the sarcomatoid variant, 12 with the micropapillary variant, 5 with the plasmacytoid variant, 4 with the nested variant and 4 with the neuroendocrine variant. The detailed baseline clinicopathologic data of the patients are listed in Table 1 . Overall, 111/130 (85.4%) patients with LNLR compared with 109/113 (96.5%) patients with HNLR had cancer recurrence during the study. Kaplan-Meier analysis indicated that the HNLR group was associated with a higher recurrence rate (p<0.001) ( Fig. 2A ) compared to the LNLR group. Multivariable survival analyses revealed that PNLR (hazard ratio [HR], 2.11; 95% confidence interval [CI], 1.57–1.83; p<0.001) was an independent predictor of RFS ( Table 2 ). Further, 49/51 (96.1%) of the HNLR patients and 47/52 (90.4%) of the LNLR patients experienced cancer recurrence in the GV group, and the former was correlated with a higher recurrence rate than the latter (p=0.007) ( Fig. 2B ). In the SV group, 41/43 (95.3%) of the HNLR patients and 37/41 (90.2%) of the LNLR patients experienced cancer recurrence and the former was correlated with a lower RFS than the latter (p=0.003) ( Fig. 2C ). In the RV group, 19/19 (100%) of the HNLR patients and 17/19 (89.5%) of the LNLR patients experienced cancer recurrence. Similarly, patients with HNLR were correlated with higher recurrence than the LNLR (p=0.430) ( Fig. 2D ). Overall, disease progression was identified in 101/113 (89.4%) of the HNLR patients and 84/130 (64.6%) of the LNLR patients. Patients with HNLR had an obviously higher rate of progression (p<0.001) ( Fig. 2E ) compared to patients with LNLR. HNLR was associated with an inferior PFS compared with LNLR (HR, 2.34; 95% CI, 1.70–3.21; p<0.001) ( Table 2 ). In the subgroup analyses, progression occurred in 48/51 (94.1%) of the HNLR patients and 34/52 (65.4%) of the LNLR patients in the GV group, confirming that the former was associated with an inferior PFS (p<0.001) ( Fig. 2F ) than the latter. For the SV group, progression occurred in 36/43 (83.7%) of the HNLR patients and 37/59 (62.7%) of the LNLR patients, identifying the former had an inferior PFS than the latter (p=0.001) ( Fig. 2G ). In the RV group, progression occurred in 17/19 (89.5%) of the HNLR patients and 13/19 (68.4%) of the LNLR patients, where the former was associated with an inferior PFS than the latter (p=0.012) ( Fig. 2H ). During the follow-up period, 86/113 (76.1%) patients with HNLR compared with 53/130 (40.8%) patients with LNLR had died. Additionally, patients with HNLR had a shorter survival than patients with LNLR, both in CSS (p<0.001) and OS (p<0.001) ( Fig. 3A, Fig. 3E ; respectively). Based the outcome of multivariable survival analyses, HNLR was an independent risk predictor of CSS (HR, 2.87; 95% CI, 1.96–4.18]; p<0.001, Table 3 ) and OS (HR, 2.83; 95% CI, 1.96–4.07; p<0.001, Table 3 ). In the GV group, 35/51 (68.6%) of the HNLR patients and 18/52 (34.6%) of the LNLR patients had died. Benefits in CCS (p=0.002) and OS (p=0.004) were observed in patients with LNLR ( Fig. 3B, Fig. 3F ; respectively). In the SV group, 36/43 (83.7%) of the HNLR patients and 26/59 (44.1%) of LNLR patients had died, while patients with HNLR had a lower survival rate than those with LNLR, both in CSS (p<0.001) and OS (p<0.001) ( Fig. 3C, Fig. 3G ; respectively). In the RV group, 15/19 (78.9%) of the HNLR patients and 9/19 (47.4%) of the LNLR patients had died. Patients with LNLR had a higher survival rate than those with HNLR, both in CSS (p=0.001) and OS (p=0.001) ( Fig. 3D, Fig. 3H ; respectively). Further prognostic information was shown in Supplementary Fig. 1 and Supplementary Table 1 .

Several studies indicated that PNLR was associated with an increased risk of disease recurrence, progression and survival in patients with NMIBC [ 17 18 22 ]. However, none of these studies focused on the effectiveness of PNLR in VH-NMIBC. Therefore, this study aimed to investigate the prognostic value of PNLR in VH-NMIBC. Finally, the present study for the first time identified that patients with HNLR have a worse prognosis, including RFS, PFS, CSS and OS, than of patients with LNLR. Many studies indicated that the appearance of VH in NMIBC usually meant advent disease and poor prognosis, prompting surgeons to perform early RC [ 22 ]. Urologists sought to determine what patient subset has aggressive properties and discover the useful biomarkers for identifying a higher risk population, especially in patients with VH-NMIBC. Hence, prognostic factors, such as gene, RNA, protein, and clinicopathological parameters, were validated. Of these, inflammatory markers were widely discussed and reported. Through cell–cell contact and/or secretion of inhibitory factors, immune cells could create a pro-tumor inflammatory state, leading an active role of systemic inflammation in tumor growth, recurrence, and progression [ 23 ]. As significant immune cells, neutrophils and lymphocytes have inhibitory and activating function, respectively, where in the former is usually associated with a poor prognosis [ 24 ]. After being identified in other tumors, the prognostic function of PNLR was increasingly appreciated by urologists. The cut-off value of PNLR in NMIBC usually ranged from 2 to 3; however, Yuk et al. [ 21 ] reported a value of 1.5. In addition, Buisan et al. [ 25 ] believed 5 was the optimal cut-off value in SV-MIBC before performing RC because they believed that the SV was linked to chronic inflammation. Similar to most studies, the optimal cut-off value in the present study was 2.2, but this difference might be explained by sample size and disease stage. According to the results of multivariable analyses, most studies demonstrated that HNLR was associated with worse prognosis [ 14 15 18 26 ]. Recently, a meta-analysis indicated that HNLR was associated with an increased risk of disease recurrence. Similarly, the prognostic effect of PNLR in RFS was proved in this study. Some discrepancy existed in the risk of progression and many studies identified that HNLR could predict an inferior PFS, but this has been challenged by some studies. Of note, all of the studies, opposing PNLR as a predictor excluded patients with VH-NMIBC in their patient selection criteria [ 16 17 21 27 ]. Conversely, studies that confirmed the statistical significance of PNLR included VH-NMIBC patients in the final analysis [ 14 15 16 26 28 ]. This difference maybe the cause of the discrepancy and supports the prognostic value of PNLR in VH-NMIBC. In this study on VH-NMIBC, data on progression validated the independent prognostic value of PNLR. Due to the long survival time, only a few studies have analyzed survival status. A multi-institutional study reported that HNLR only predicted inferior CSS, but a significant difference regarding OS was not noted [ 18 ]. The remaining two studies found an association between PNLR and the survival of NMIBC both in CSS and OS [ 21 29 ]. The present study also identified that HNLR was associated with worse CSS and OS. Therefore, the PNLR level could assist with therapeutic decision-making and identify appropriate candidates for more aggressive therapy, such as early RC for high-risk VH-NMIBC. This study has some limitations, including those inherent to the retrospective design of the study. The bias of data might be noticed because the study data was derived from a hospital information system. Additionally, due to sample size limitations for different subgroups, multivariable survival analyses could not be performed. However, of the prognostic value of the PNLR in VH-NMIBC was identified.

PNLR level was correlated with the prognosis of patients with VH-NMIBC, which could assist with therapeutic decision-making and identify appropriate candidates for more aggressive therapy. Prospective clinical trials are warranted to provide a higher level of evidence.

With Institutional Research Ethics Committee of the West China Hospital, Sichuan University approval (approval number: 20201045), patients with VH-NMIBC were included from January 2009 to May 2019 ( Fig. 1 ). Written informed consent was obtained from all participants. To avoid bias, we only included patients whose PNLR was determined 2 weeks before surgery. The exclusion criteria were as follows: patients with a diagnosis of prostatitis or cystitis, urinary tract infection, yeast infections, endometriosis, systemic inflammatory disease; those with missing data, other sites tumor; or those with a need for preforming radical cystectomy (RC) without muscle invasion. All patients in this study were classified for total group and subgroup analyses, including those with SV categorized as the SV group; with GV categorized as the GV group, and those with RV categorized as the RV group. All specimens with VH were diagnosed by experienced pathologists according to the 2016 WHO BC classification [ 19 ]. VH-NMIBC was defined as any VH appearing in the specimen and a sample with mixed-variant would be classified as the VH with the highest proportion. Tumor stage, and carcinoma in situ were evaluated by the 2016 American Joint Committee on Cancer (AJCC). In the TNM staging system, the tumor was graded as per the 2016 WHO grading system [ 19 20 ]. Demographic, clinical and pathologic outcomes were determined in this study. Intravesical therapies (IVTs), including bacillus Calmette–Guérin (BCG) and chemotherapeutic drugs, were utilized in patients with T1 stage unless patients refused or their physical status did not permit. Chemotherapeutic drugs, including epirubicin and gemcitabine, were administered for 12 cycles after TURBT. Under the recommendation of the European Association of Urology guidelines, patients would receive a 6-week course of intravesical BCG induction followed by a standard maintenance regimen [ 11 ]. Follow-up was performed according to the European Association of Urology guidelines, including cystoscopy every 3 months for the first 2 years, every 6 months 2 to 5 years after TURBT, and then annually [ 11 ]. A phone or face-to-face interview was used for patient follow-up. Recurrence-free survival (RFS) was defined as the time from the date of surgery to local or distant recurrence. Progression-free survival (PFS) was defined as an increase in the stage to MIBC and/or metastasis. Cancer-specific survival (CSS) was defined as the time from the date of surgery to death from BC. Overall survival (OS) was defined as the time from the date of TURBT to death due to any cause. Chi-square and Student's t-tests were used for analyzing the associations between categorical and continuous variables, respectively. Fisher's exact test was applied to estimate categorical variables in the RV group due to the limited number of patients. The best cut-off value of PNLR was calculated by the receiver operating characteristic curve and Youden index. The cut-off value of body mass index was 23.4 kg/m 2 [ 21 ]. The survival of patients with VH-NMIBC was validated using the Kaplan-Meier method and comparisons between groups were performed by log-rank tests. Multivariable Cox proportional hazard regression models were employed to evaluate the association of PNLR with RFS, PFS, CSS, and OS. We did not perform multivariable survival analysis in subgroups because of the limited number of patients. A p<0.05 was considered significant for all analyses, which were performed using SPSS Version 25.0 (IBM Corp., Armonk, NY, USA).