Abstract
Background
Bladder cancer has notable heterogeneity. The urine-based fluorescence in situ hybridization (FISH) test can detect bladder cancer noninvasively. In this study, we aimed to construct a nomogram based on FISH results and clinical features (referred to as the FISH–clinical model) to predict the overall survival (OS) of bladder cancer patients following radical cystectomy (RC).
Methods
A total of 261 eligible patients were enrolled for this study. The SYSMH cohort was divided into training (n = 138) and internal validation (n = 70) sets; the SYSUTH cohort was used for external validation (n = 53). Multivariate Cox proportional hazards regression was applied for FISH–clinical model construction, and model performance was evaluated according to analyses of calibration, discrimination ability, and clinical usefulness.
Results
FISH-identified chromosome 7 and 17 aneuploidies correlated significantly with increased pT stage; the former was associated with lymph node metastasis. Six variables, age, tumor size, pT stage, lymphovascular invasion, chromosome 7 aneuploidy, and p16 locus loss, were found to be independent predictors of OS and were incorporated into our FISH–clinical model. The model demonstrated good calibration and discrimination, with C-indexes (95% CIs) of 0.772 (0.693–0.851), 0.712 (0.605–0.819) and 0.705 (0.587–0.822), in the training, internal validation and external validation sets respectively. Decision curve analysis demonstrated the model’s clinical utility. Furthermore, all enrolled patients were successfully categorized into high-, medium- or low-risk groups, and stratified analyses were performed.
Conclusions
Preoperative FISH has predictive value for OS, and we developed a FISH–clinical model for OS prediction in bladder cancer patients who have not received neoadjuvant chemotherapy or immunotherapy. This model showed favorable predictive efficacy with internal and external validation.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12885-025-14677-w.
Keywords: Bladder cancer, Radical cystectomy, Overall survival, Fluorescence in situ hybridization, Nomogram
Introduction
Bladder cancer is one of the most common neoplasms affecting the urinary system and is classified as non-muscle-invasive bladder cancer (NMIBC) or muscle-invasive bladder cancer (MIBC) based on detrusor muscle involvement. In 2021, the age-standardized incidence rate of bladder cancer reached 6.35 per 100,000 people [1, 2]. Radical cystectomy (RC) with pelvic lymphadenectomy is the standard surgical procedure for patients with local MIBC and is also used for some high-risk NMIBC patients [3, 4]. Bladder cancer is a highly heterogeneous disease, resulting in different prognoses even when the same stage of the tumor receives the same treatment [5, 6]. Currently, TNM staging and tumor grading are used for risk stratification in patients with bladder cancer but are unable to predict patient outcomes. Several models for overall survival (OS) prediction in individual bladder cancer patients after RC, including clinical, genomics, and radiomics models, have been reported [7–9]. However, these models are limited by their high cost, or limited in accuracy. Thus, it is imperative to discover new, accessible, and reliable predictors.
The urine-based fluorescence in situ hybridization (FISH) test can be used to diagnose bladder cancer noninvasively by detecting aneuploidy of chromosomes 3, 7, and 17 and deletion of the 9p21 band location of the p16 gene [10, 11]. As the FISH assay is more sensitive than the conventional urine cytology test in detecting bladder cancer and hematuria and urinary tract infections do not affect its performance [12], it is extensively used for detecting bladder cancer in the clinic. In addition to diagnosis, FISH tests can also assist in tumor staging, prognosis assessment, and treatment response prediction [13–15]. Previously, we reported that the presence of chromosome-specific centromeric probe 7 (CSP7) in bladder cancer patients is correlated with muscular invasion. Moreover, we revealed that the FISH test result can be used for recurrence prediction in NMIBC patients after transurethral resection of bladder tumor (TURBT) [14, 15]. However, the prognostic value of preoperative FISH results in bladder cancer patients after RC has not been determined.
Thus, in this study, we aimed to clarify the correlation of preoperative FISH results with pT stage, pN stage, lymphovascular invasion (LVI), and OS in patients with bladder cancer following RC. Moreover, we sought to construct a nomogram model based on FISH results and clinical features (referred to as the FISH–clinical model) for OS prediction in individual patients with bladder cancer after RC.
Materials and methods
Patients
This study was approved by the Ethics Review Boards at Sun Yat-Sen Memorial Hospital and the Third Affiliated Hospital of Sun Yat-Sen University, and the need to obtain informed consent from the patients was waived. In total, we enrolled 208 patients with bladder cancer who underwent RC between September 2016 and September 2021 at Sun Yat-Sen Memorial Hospital (SYSMH) and 53 patients with bladder cancer who received RC between January 2014 and July 2022 at the Third Affiliated Hospital of Sun Yat-Sen University (SYSUTH). The inclusion criteria were as follows: (1) bladder urothelial carcinoma confirmed by histology; (2) underwent RC and pelvic lymph node dissection; (3) urine-based FISH testing performed before surgery; and (4) complete follow-up and clinicopathological information. The exclusion criteria were as follows: (1) the patient had other malignant tumors and (2) the patient did not undergo surgery or only underwent transurethral resection of bladder tumor (TUBRT). The patients from SYSMH were randomly divided into a training set (n = 138) and an internal validation set (n = 70) at a ratio of 2:1. Patients from SYSUTH were used as the external validation set (n = 53). The study flowchart is presented in Fig. 1.
Fig. 1.
The flowchart of the study
Clinical and pathological baseline data, including age, sex, tumor size and number, pathologic T and N stages, LVI status, and carcinoma in situ (CIS) status, were obtained. All enrolled patients underwent FISH testing, and the FISH detection result, the status of the four FISH sites, and the number of positive sites were recorded. All patients were followed up according to clinical guidelines [3]. OS refers to the period from surgery to death or the last follow-up. Follow-up personnels were blinded to the predictor variables during the follow-up.
FISH assay
At both institutions, we carried out the FISH assay by using commercial kits (GP Medical Technologies, China). The FISH site is reportedly positive if the frequency of aneuploidies on chromosomes 3, 7, 17 or p16 locus loss exceeds the diagnostic threshold values in the FISH assay. When two or more types of FISH sites were positive or when only the gene locus-specific probe p16 (GLP-p16) was positive, the FISH test result was considered positive. FISH assays were performed with consistent methodology throughout the study period. The details of the detection and diagnostic procedures are described in the Supplementary Material.
Correlations of the FISH results with pT stage, pN stage, and LVI status
Pie charts and bar charts were generated to illustrate the distributions of the FISH test results. The potential association between FISH results and pT stage was explored via the Mann‒Whitney U test. The potential correlations between FISH results and pN stage and LVI status were assessed using chi-square tests.
Development of the FISH–clinical model
With respect to the training set, we conducted Cox proportional hazards regressions to identify prognostic predictors among the FISH results and other clinicopathological factors. We used backward stepwise selection to pinpoint independent predictors to develop a FISH–clinical model for OS prediction [16]. A nomogram was subsequently constructed according to the multivariate Cox regression results, and for each patient, a risk score was calculated using the following regression formula:
Risk score = α1V1 + α2V2 +… αiVi, where Vi is the predictive variable and αi is the corresponding regression coefficient.
For the training set, we evaluated the performance of the FISH–clinical model on the basis of its calibration, discrimination ability, and clinical usefulness. To assess the calibration of the FISH–clinical model, we compared the actual survival probability and the predicted survival probability. A well-calibrated model is indicated when the model’s predicted values closely align with the actual values. We applied a calibration curve to visually determine the calibration of the model. We calculated Harrell’s C-index of the FISH–clinical model, which serves as a measure of the discriminative ability of prognostic models [17]. The C-index was calculated from 1000 bootstrap resamples to yield a stable estimate. In addition, we used decision curve analysis (DCA) to assess the–clinical usefulness of the FISH–clinical model [18].
Validation of the FISH–clinical model
Model validation was performed in the internal and external validation sets. For the patients in the validation sets, we calculated their risk scores using the proposed regression algorithm. Then, calibration curves were plotted, C-indexes were calculated, and DCAs were performed to validate the performance of our proposed model.
Patient risk stratification
In the training set, we identified optimal thresholds of risk scores to stratify patients into high-, medium-, and low-risk groups through X-tile plots [19]. The differences in OS among the three groups were evaluated by using Kaplan‒Meier survival curves and log-rank tests. Additionally, we performed stratified analyses on all enrolled patients on the basis of age and sex.
Statistical analyses
X-tile software 3.6.1 (Yale University School of Medicine, USA) was used to generate X-tile plots [19]. Other statistical analyses were conducted using R statistical software 4.2.2. All the statistical tests were 2-tailed, and P values < 0.05 denoted statistically significant differences. Supplementary Table S1 lists the R packages we used.
Results
Patient clinical characteristics
In total, 261 patients with bladder cancer (224 males and 37 females) with a median (interquartile range, IQR) age of 64 years (57, 72) were enrolled in our study. The clinicopathological characteristics of all enrolled patients are provided in Table 1. In the whole study cohort, the median follow-up time was 32.6 (IQR, 19.4–49.4) months. Eighty-six patients (33% of all patients) died during follow-up. The FISH test had a diagnostic sensitivity of 79.7% (208/261) for bladder cancer in the whole cohort.
Table 1.
Baseline clinical characteristics of the enrolled patients
| All patients (n = 261) |
Training set (n = 138) |
Internal validation set (n = 70) |
External validation set (n = 53) |
P | |
|---|---|---|---|---|---|
| Age, years | 0.127 | ||||
| Median (interquartile range) | 64 (56, 72) | 64 (55, 70) | 64 (57, 71) | 67 (59, 74) | |
| Sex | 0.199 | ||||
| Male | 224 (85.8%) | 123 (89.1%) | 56 (80.0%) | 45 (84.9%) | |
| Female | 37 (14.2%) | 15 (10.9%) | 14 (20.0%) | 8 (15.1%) | |
| Tumor size, cm | 0.821 | ||||
| ≤ 3 | 120 (46.0%) | 61 (44.2%) | 34 (48.6%) | 25 (47.2%) | |
| >3 | 141 (54.0%) | 77 (55.8%) | 36 (51.4%) | 28 (52.8%) | |
| Tumor number | 0.726 | ||||
| Single | 192 (73.6%) | 100 (72.5%) | 54 (77.1%) | 38 (71.7%) | |
| Multiple | 69 (26.4%) | 38 (27.5%) | 16 (22.9%) | 15 (28.3%) | |
| pT stage | 0.449 | ||||
| 1 | 65 (24.9%) | 36 (26.1%) | 15 (21.4%) | 14 (26.4%) | |
| 2 | 94 (36.0%) | 47 (34.1%) | 26 (37.2%) | 21 (39.6%) | |
| 3 | 71 (27.2%) | 34 (24.6%) | 21 (30.0%) | 16 (30.2%) | |
| 4 | 31 (11.9%) | 21 (15.2%) | 8 (11.4%) | 2 (3.8%) | |
| pN stage | 0.268 | ||||
| 0 | 208 (79.7%) | 113 (81.9%) | 57 (81.4%) | 38 (71.7%) | |
| 1–3 | 53 (20.3%) | 25 (18.1%) | 13 (18.6%) | 15 (28.3%) | |
| Associated CIS | 0.333 | ||||
| No | 230 (88.1%) | 120 (87.0%) | 65 (92.9%) | 45 (84.9%) | |
| Yes | 31 (11.9%) | 18 (13.0%) | 5 (7.1%) | 8 (15.1%) | |
| LVI status | 0.586 | ||||
| Negative | 180 (69.0%) | 95 (68.8%) | 51 (72.9%) | 34 (64.2%) | |
| Positive | 81 (31.0%) | 43 (31.2%) | 19 (27.1%) | 19 (35.8%) | |
| CSP3 | 0.023* | ||||
| Negative | 76 (29.1%) | 32 (23.2%) | 29 (41.4%) | 15 (28.3%) | |
| Positive | 185 (70.9%) | 106 (76.8%) | 41 (58.6%) | 38 (71.7%) | |
| CSP7 | 0.049* | ||||
| Negative | 74 (28.4%) | 31 (22.5%) | 27 (38.6%) | 16 (30.2%) | |
| Positive | 187 (71.6%) | 107 (77.5%) | 43 (61.4%) | 37 (69.8%) | |
| CSP17 | 0.299 | ||||
| Negative | 77 (29.5%) | 35 (25.4%) | 24 (34.3%) | 18 (34.0%) | |
| Positive | 184 (70.5%) | 103 (74.6%) | 46 (65.7%) | 35 (66.0%) | |
| GLP-p16 | 0.672 | ||||
| Negative | 196 (75.1%) | 103 (74.6%) | 55 (78.6%) | 38 (71.7%) | |
| Positive | 65 (24.9%) | 35 (25.4%) | 15 (21.4%) | 15 (28.3%) | |
| No. of positive FISH site | 0.376 | ||||
| 0 | 40 (15.3%) | 17 (12.3%) | 16 (22.8%) | 7 (13.2%) | |
| 1 | 31 (11.9%) | 14 (10.1%) | 10 (14.3%) | 7 (13.2%) | |
| 2 | 15 (5.7%) | 7 (5.1%) | 3 (4.3%) | 5 (9.4%) | |
| 3 | 138 (52.9%) | 77 (55.8%) | 35 (50.0%) | 26 (49.1%) | |
| 4 | 37 (14.2%) | 23 (16.7%) | 6 (8.6%) | 8 (15.1%) | |
| FISH test | 0.021* | ||||
| Negative | 53 (20.3%) | 24 (17.4%) | 22 (31.4%) | 7 (13.2%) | |
| Positive | 208 (79.7%) | 114 (82.6%) | 48 (68.6%) | 46 (86.8%) | |
| Follow time, months | 0.065 | ||||
| Median (interquartile range) |
32.6 (19.4, 49.4) |
31.6 (19.3, 48.1) |
32.2 (18.8, 46.1) |
39.2 (21.2, 69.3) |
|
| Survival status | 0.707 | ||||
| Alive | 175 (67.0%) | 94 (68.1%) | 48 (68.6%) | 33 (62.3%) | |
| Dead | 86 (33.0%) | 44 (31.9%) | 22 (31.4%) | 20 (37.7%) |
* P < 0.05
The distribution of the status of the four FISH sites across the whole cohort is depicted in Fig. 2A. A total of 70.9%, 71.6%, and 70.5% of the patients presented aneuploidy on chromosomes 3, 7, and 17, respectively, whereas only 25% presented loss of the p16 locus in the 9p21 region. Figure 2B shows the distribution of positive combinations of the four FISH sites. The most common FISH profile was CSP3(+), CSP7(+), CSP17(+), and p16(-), which was observed in 123 patients (47.1% of the whole cohort).
Fig. 2.
Distribution of FISH results across the whole cohort.(A) Pie plots illustrating the status of the four FISH sites. (B) UpSet plot showing the distribution of different combinations of FISH-positive sites. The horizontal axis represents different combinations of FISH-positive sites. The vertical axis shows the patient number for each combination. FISH: fluorescence in situ hybridization analysis; CSP: chromosome-specific centromeric probe; GLP: gene locus-specific probe
Correlations of the FISH results with pT stage, pN stage, and LVI status
As shown in Fig. 3, only CSP7 status and CSP17 status were significantly associated with pT stage (Mann‒Whitney U tests, P = 0.033 and 0.046, respectively). Patients with aneuploidy on chromosome 7 or 17 were more likely to have a high pT stage. Despite a trend whereby patients who had aneuploidy according to chromosome 3 or positive FISH results had a higher pT stage, the differences were not statistically significant. Furthermore, we explored the correlation of the FISH test results with pN stage and LVI status (Fig. 4). Only the status of CSP7 was significantly related to lymph node metastasis (P = 0.006; Fig. 4A). Notably, none of the four FISH sites was significantly correlated with LVI status (Fig. 4B). Although the patients with positive FISH results tended to have lymph node metastasis or LVI, the differences were not statistically significant (P = 0.075 and 0.070, respectively).
Fig. 3.
Correlations of the FISH assay results with the pT stage in the whole cohort. (A-D) Bar plots showing the correlations between the four FISH sites status and pT stage. (E) Bar plots showing the correlation between the FISH test result and pT stage. (F) Bar plots showing the distribution of the number of positive FISH sites in bladder cancer patients with different pT stages. FISH: fluorescence in situ hybridization analysis; CSP: chromosome-specific centromeric probe; GLP: gene locus-specific probe
Fig. 4.
Correlations of the FISH assay results with pN stage and LVI status. (A) Bar plots showing the correlations between the FISH assay results and pN stage. (B) Bar plots showing the correlations between the FISH assay results and LVI status. FISH: fluorescence in situ hybridization analysis; CSP: chromosome-specific centromeric probe; GLP: gene locus-specific probe; LNM: lymph node metastasis; LVI: lymphovascular invasion
Development of the FISH–clinical model
Six variables, namely, age, tumor size, pT stage, LVI status, CSP7 status, and GLP-p16 status, were determined to be independent predictors of OS (Table 2). A FISH–clinical model was subsequently constructed for OS prediction in bladder cancer patients who underwent RC. In addition, we performed univariate Cox proportional hazards regression analyses of FISH results in pT stage-stratified analyses in all enrolled patients (Supplementary Table S2). Among the different subgroups, only CSP7 status was associated with OS. To facilitate user-friendly graphical interfaces, a nomogram was constructed (Fig. 5). The risk score calculation formula is as follows:
Table 2.
Univariate and multivariate Cox proportional hazards regression analyses of candidate predictors in the training set
| Univariate regression | Multivariate regression | ||||
|---|---|---|---|---|---|
| Variables | HR (95% CI) | P | HR (95% CI) | P | |
| Age, years | 1.034 (1.007, 1.061) | 0.012* | 1.048 (1.021, 1.077) | < 0.001* | |
| Sex (male vs. female) | 1.177 (0.464, 2.991) | 0.731 | - | - | |
| Tumor size, cm (≤ 3 vs. >3) | 1.846 (0.989, 3.449) | 0.054 | 2.432 (1.247, 4.743) | 0.009* | |
| Tumor number (single vs. multiple) | 0.933 (0.487, 1.788) | 0.834 | - | - | |
| pT stage | |||||
| 1 | Reference | Reference | |||
| 2 | 2.781 (1.078, 7.176) | 0.034 | 3.084 (0.879, 10.819) | 0.079 | |
| 3 | 5.219 (2.162, 12.596) | < 0.001* | 6.381 (1.793, 22.710) | 0.004* | |
| 4 | 4.086 (1.517, 11.005) | 0.005* | 3.397 (1.525, 5.389) | 0.070 | |
| pN stage (0 vs. 1–3) | 2.415 (1.279, 4.561) | 0.007* | - | - | |
| Associated CIS (no vs. yes) | 0.502 (0.155, 1.621) | 0.249 | - | - | |
| LVI status (negative vs. positive) | 3.306 (1.820, 6.006) | < 0.001* | 2.867 (1.525, 5.389) | 0.001* | |
| CSP3 status (negative vs. positive) | 2.622 (1.033, 6.656) | 0.043* | - | - | |
| CSP7 status (negative vs. positive) | 3.292 (1.177, 9.206) | 0.023* | 3.266 (1.148, 9.293) | 0.027* | |
| CSP17 status (negative vs. positive) | 1.568 (0.728, 3.374) | 0.250 | - | - | |
| GLP-p16 status (negative vs. positive) | 0.575 (0.267, 1.238) | 0.157 | 0.403 (0.171, 0.949) | 0.038* | |
| No. of positive FISH site | 1.197 (0.928, 1.545) | 0.167 | - | - | |
| FISH test (negative vs. positive) | 2.399 (0.858, 6.708) | 0.095 | - | - | |
* P < 0.05
Fig. 5.
The FISH–clinical nomogram for OS prediction. The FISH–clinical nomogram was developed to predict the 1-year, 2-year, and 3-year OS of bladder cancer patients who underwent RC. CSP: chromosome-specific centromeric probe; GLP: gene locus‐specific probe; LVI: lymphovascular invasion; OS: overall survival; RC: radical cystectomy
Risk score = 0.047 × age + 0.889 × I (tumor size > 3 cm) + 1.184 × I (CSP7 [+])– 0.908 × I (p16 [+]) + 1.053 × I (LVI [+]) + 1.126 × I (T2 stage) + 1.853 × I (T3 stage) + 1.223 × I (T4 stage). If the statement in the parentheses is true, the indicator function (I) value will be 1; otherwise, it will be 0.
In the training set, the model achieved a C-index (95% CI) of 0.772 (0.693–0.851), suggesting good discrimination ability. Furthermore, the calibration curve suggested favorable agreement between the model predictions and actual observations (Fig. 6A). DCA for the FISH–clinical model is shown in Fig. 6D. The results indicated that using the proposed model to predict OS provided greater net benefits than the “treat all” or “treat none” strategy over a wide range of threshold probabilities, indicating the good clinical utility of the FISH–clinical model.
Fig. 6.
Calibration curve analysis and DCA of the FISH–clinical model. (A-C) Calibration curves of the FISH–clinical model in the training, internal validation, and external validation sets. The 45-degree gray line represents perfect prediction. The broken line represents the model prediction, which has a closer fit to the gray line, indicating a better prediction. (D-F) DCA results of the FISH–clinical model in the training, internal validation, and external validation sets. The x-axis shows the threshold probability. The y-axis represents the net benefit, which is calculated across a range of threshold probabilities. The black and gray lines represent the hypothesis that no patients or all patients died, respectively. The solid red line represents the FISH–clinical model
Validation of the FISH–clinical model
Our prediction model yielded C-indexes (95% CI) of 0.712 (0.605–0.819) and 0.705 (0.587–0.823) in the internal and external validation sets, respectively, and the calibration curves confirmed the favorable calibration of the FISH–clinical model (Fig. 6B‒C). Moreover, similar DCA results were obtained in the two validation sets, confirming the favorable clinical utility of our model (Fig. 6E-F).
In addition, we also developed a clinical model incorporating pT stage, pN stage, and LVI status. The C-indexes (95% CIs) of the clinical model were 0.715 (0.642–0.788), 0.704 (0.650–0.758) and 0.683 (0.576–0.790), in the training, internal validation and external validation sets respectively. Therefore, our FISH-clinical model outperformed the clinical model.
Patient risk stratification
X-tile plot analysis indicates optimal risk score thresholds of 2.74 and 3.82 (Supplementary Figure S1). On the basis of these thresholds, patients were categorized into groups with high (≥ 3.82), moderate (≥ 2.74 and < 3.82), and low (< 2.74) risk. There was a significant difference in OS among the three risk groups in the training set (P < 0.001; Fig. 7A). This finding was validated in the internal validation set (P < 0.001; Fig. 7B), the external validation set (P = 0.014; Fig. 7C), and the whole cohort (P < 0.001; Fig. 7D). Furthermore, similar results were obtained when age and sex were considered (Supplementary Figure S2). These findings confirm the reliability and robustness of the proposed model.
Fig. 7.
Kaplan‒Meier survival curves for patients in the high-, moderate-, and low-risk groups. Kaplan‒Meier survival curves indicating significant differences in OS among patients in the high-, moderate-, and low-risk groups in the training set (A), the internal validation set (B), the external validation set (C), and the whole cohort (D)
Discussion
Previous studies have indicated that bladder cancer patients have the highest tumor-related mortality in the first few years after RC and that the mortality of survivors gradually decreases over time [20, 21]. Therefore, accurately predicting the prognosis at the early stage after RC can facilitate optimization of disease management and follow-up. The TNM staging system is considered a standard prognostic tool for bladder cancer. However, recent advances have led to the development of individualized prediction models that integrate comprehensive clinicopathological factors to enhance survival prediction accuracy.
In this study, a positive preoperative FISH status for CSP7 or CSP17 was significantly correlated with increased pT stage in bladder cancer patients, which was consistent with our previous findings [14, 15] and studies conducted by other researchers [22, 23]. A potential mechanism is that the copy number of oncogenes on chromosomes 7 and 17 increases, subsequently promoting tumor growth. Indeed, EGFR on chromosome 7 and ERBB2 on chromosome 17 have been reported to promote bladder cancer development [23, 24]. In addition, CSP7 aneuploidy showed a consistent association with overall survival across all pT stage subgroups, suggesting a stage-independent prognostic value. Moreover, it was significantly associated with lymph node metastasis (P = 0.006), which may serve as a novel preoperative predictor of lymph node metastasis in these patients. However, this correlation needs to be validated in larger cohorts, and its potential mechanism needs to be further investigated.
Previously, we demonstrated that CSP7 and GLP-p16 status based on preoperative FISH testing is related to unfavorable recurrence-free survival (RFS) in NMIBC patients after TURBT [15]. However, no study has explored whether preoperative FISH results can serve as prognostic predictors in bladder cancer patients after RC. Indeed, we also found in the present study that the status of CSP7 and GLP-p16 are independent predictors of OS in bladder cancer patients following RC. In addition to the FISH results, four clinicopathological factors, namely, age, bladder tumor size, pT stage, and LVI status, were determined to be independent prognostic predictors. These clinicopathological variables have been confirmed as useful prognostic predictors for patients with bladder cancer in many previous studies [25–30], which proves the reliability of our findings. By incorporating these six independent predictors, we developed the FISH–clinical model for OS prediction in bladder cancer patients after RC.
The present study represents the first attempt to investigate the associations of preoperative FISH results with LVI and lymph node metastasis in bladder cancer patients. Moreover, for the first time, we innovatively introduced postoperative FISH site status into a prediction model and developed the FISH–clinical model for predicting OS in bladder cancer patients after RC. In recent years, several clinical prediction models have been developed for OS prediction in patients with bladder cancer. Most of these studies incorporated clinical variables, such as age, sex, pathologic T stage, soft tissue surgical margin status, and LVI [31–33]. However, these clinical models still fall short in predictive accuracy. In addition, some genomics-based and radiomic-based models with better performance than clinical models have been reported [34, 35]. However, their clinical application is limited by high cost, limited accessibility, unclear biological mechanisms and low repeatability. The presented FISH–clinical model is specific for bladder cancer patients after RC, which well reflects the clinical characteristics of this population. In addition, the urine-based FISH assay is a currently used noninvasive routine medical test for bladder cancer patients, and all variables in the FISH–clinical model are easily assessed in routine clinical practice. Previous studies have shown that routine FISH applied for risk stratification or early recurrence detection can be cost-effective [36, 37].We also provide a user-friendly nomogram for expedient use of the model. Therefore, the model can serve as a noninvasive tool to predict the prognosis of bladder cancer after RC, which may offer both clinical and economic benefits.
Despite the strengths of our study, there are several limitations. First, potential selection biases might have occurred due to the retrospective study design and the strict inclusion and exclusion criteria of our study, which might influence the generalizability of our findings. In addition, the sample sizes of patients with LVI and lymph node metastasis were relatively small in our study, which might affect the statistical power and robustness of the stratified analyses. Further validation studies are supposed to involved larger cohort and multi-institutional collaborations to ensure more robust and statistically reliable result. Second, our model achieved moderate discrimination accuracy, with C-indexes ranging from 0.7 to 0.8, suggesting room for further improvement. Third, the present study revealed the associations of preoperative FISH site status with pT stage, pN stage, and OS in patients with bladder cancer, but we did not explore the potential underlying mechanism involved. Relevant bioinformatic and basic experimental research should be conducted to reveal the underlying mechanism, which may provide useful information for the development of novel therapies and diagnostic approaches. Fourth, treatments such as chemotherapy and radiotherapy may alter chromosomal structure and potentially affect FISH results [38, 39]. In our study, none of the enrolled patients received neoadjuvant chemotherapy. Although this ensures that the FISH results reflect the characteristics of untreated tumor, it limits our opportunities to assess dynamic changes in FISH status in response to therapy. These dynamic changes in FISH test may provide significant insights into therapy response and warrant further investigation in future research. In addition, due to the retrospective design and extended time span of our study, detailed adjuvant treatment data were incomplete, which was not included in the analysis. This may potentially affect the performance of our model.
In conclusion, we found that the preoperative FISH site status is associated with pT stage, pN stage, and OS in patients with bladder cancer after RC. In addition, we propose a FISH–clinical model for OS prediction in bladder cancer patients undergoing RC without prior neoadjuvant chemotherapy or immunotherapy; this model exhibits favorable predictive efficacy and has been internally and externally validated. The FISH–clinical model can be applied for tumor prognosis prediction at an early stage after surgery, which is beneficial for formulating or optimizing individualized treatment and follow-up strategies in a timely manner.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Acknowledgements
We would like to thank Figdraw (www.figdraw.com) for assistance in drawing Figure 1.
Abbreviations
- CSP7
Chromosome-specific centromeric probe 7
- CIS
Carcinoma in situ
- DCA
Decision curve analysis
- FISH
Fluorescence in situ hybridization
- GLP-p16
Gene locus-specific probe p16
- LVI
Lymphovascular invasion
- MIBC
Muscle-invasive bladder cancer
- NMIBC
Non-muscle-invasive bladder cancer
- OS
Overall survival
- RC
Radical cystectomy
- RFS
Recurrence-free survival
- SYSMH
Sun Yat-sen Memorial Hospital
- SYSUTH
Third Affiliated Hospital of Sun Yat-Sen University
- TURBT
Transurethral resection of bladder tumor
Author contributions
Tianxin Lin, Junjiong Zheng, Sihong Lu, and Yun Luo conceptualized and designed the study. Yi Huang, Jianqiu Kong, Xu Chen, Jie Zhang, and Yuhui Yao acquired the data. Junjiong Zheng, Sihong Lu, Qihang Zhang, and Long Zhang analyzed the data. All the authors wrote, reviewed and/or revised the manuscript.
Funding
This study was supported by the National Natural Science Foundation of China (Grant No. 82203188), the National Key R&D Program of China (Grant No. 2023YFC2507002), Young Science and Technology Talent Support Program of Guangdong Precision Medicine Application Association (YSTTGDPMAA202502), and the Guangdong Provincial Clinical Research Center for Urological Diseases (2020B1111170006).
Data availability
All the data generated or analyzed during this study are included in this published article and its supplementary information files.
Declarations
Ethics approval and consent to participate
Approval for this study was obtained from the Ethics Review Boards at Sun Yat-Sen Memorial Hospital and the Third Affiliated Hospital of Sun Yat-Sen University and in conformity to the Declaration of Helsinki. The need to obtain informed consent from the patients was waived because of the retrospective nature of the study.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Junjiong Zheng, Sihong Lu and Qihang Zhang are Co-first authors.
Contributor Information
Yun Luo, Email: luoyun8@mail.sysu.edu.cn.
Tianxin Lin, Email: lintx@mail.sysu.edu.cn.
References
- 1.Siegel RL, Miller KD, Wagle NS, Jemal A. Cancer statistics, 2023. CA Cancer J Clin. 2023;73(1):17–48. [DOI] [PubMed] [Google Scholar]
- 2.Feng DC, Li DX, Wu RC, Wang J, Xiao YH, Yoo KH, Ye X, Wei WR, Kong DP, Tuo ZT. Global burden and cross-country inequalities in urinary tumors from 1990 to 2021 and predicted incidence changes to 2046. Mil Med Res. 2025;12(1):12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Alfred Witjes J, Max Bruins H, Carrion A, Cathomas R, Comperat E, Efstathiou JA, Fietkau R, Gakis G, Lorch A, Martini A, et al. European association of urology guidelines on Muscle-invasive and metastatic bladder cancer: summary of the 2023 guidelines. Eur Urol. 2024;85(1):17–31. [DOI] [PubMed] [Google Scholar]
- 4.Babjuk M, Burger M, Capoun O, Cohen D, Compérat EM, Dominguez Escrig JL, Gontero P, Liedberg F, Masson-Lecomte A, Mostafid AH, et al. European association of urology guidelines on Non-muscle-invasive bladder Cancer (Ta, T1, and carcinoma in Situ). Eur Urol. 2022;81(1):75–94. [DOI] [PubMed] [Google Scholar]
- 5.Ma G, Yang X, Liang Y, Wang L, Li D, Chen Y, Liang Z, Wang Y, Niu H. Precision medicine and bladder cancer heterogeneity. Bull Cancer. 2018;105(10):925–31. [DOI] [PubMed] [Google Scholar]
- 6.Aziz A, Gierth M, Fritsche HM, May M, Otto W, Denzinger S, Wieland WF, Merseburger A, Riedmiller H, Kocot A, et al. Oncological outcome of primary versus secondary muscle-invasive bladder cancer is comparable after radical cystectomy. Urol Int. 2013;91(1):97–102. [DOI] [PubMed] [Google Scholar]
- 7.Shariat SF, Karakiewicz PI, Palapattu GS, Amiel GE, Lotan Y, Rogers CG, Vazina A, Bastian PJ, Gupta A, Sagalowsky AI, et al. Nomograms provide improved accuracy for predicting survival after radical cystectomy. Clin Cancer Res. 2006;12(22):6663–76. [DOI] [PubMed] [Google Scholar]
- 8.Lin J, Yang J, Xu X, Wang Y, Yu M, Zhu Y. A robust 11-genes prognostic model can predict overall survival in bladder cancer patients based on five cohorts. Cancer Cell Int. 2020;20:402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Woźnicki P, Laqua FC, Messmer K, Kunz WG, Stief C, Nörenberg D, Schreier A, Wójcik J, Ruebenthaler J, Ingrisch M et al. Radiomics for the Prediction of Overall Survival in Patients with Bladder Cancer Prior to Radical Cystectomy. Cancers (Basel) 2022, 14(18). [DOI] [PMC free article] [PubMed]
- 10.Bubendorf L. Multiprobe fluorescence in situ hybridization (UroVysion) for the detection of urothelial carcinoma - FISHing for the right catch. Acta Cytol. 2011;55(2):113–9. [DOI] [PubMed] [Google Scholar]
- 11.Soria F, Droller MJ, Lotan Y, Gontero P, D’Andrea D, Gust KM, Rouprêt M, Babjuk M, Palou J, Shariat SF. An up-to-date catalog of available urinary biomarkers for the surveillance of non-muscle invasive bladder cancer. World J Urol. 2018;36(12):1981–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lin T, Liu Z, Liu L, Yang L, Han P, Zhang P, Wei Q. Prospective evaluation of fluorescence in situ hybridization for diagnosing urothelial carcinoma. Oncol Lett. 2017;13(5):3928–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Liem E, Oddens JR, Vernooij RWM, Li R, Kamat A, Dinney CP, Mengual L, Alcaraz A, Izquierdo L, Savic S, et al. The role of fluorescence in situ hybridization for predicting recurrence after adjuvant bacillus Calmette-Guérin in patients with intermediate and high risk nonmuscle invasive bladder cancer: A systematic review and Meta-Analysis of individual patient data. J Urol. 2020;203(2):283–91. [DOI] [PubMed] [Google Scholar]
- 14.Diao X, Cai J, Zheng J, Kong J, Wu S, Yu H, Huang H, Xie W, Chen X, Huang C et al. Association of chromosome 7 aneuploidy measured by fluorescence in situ hybridization assay with muscular invasion in bladder cancer. Cancer Commun 2020. 40(4):167-180. [DOI] [PMC free article] [PubMed]
- 15.Zheng J, Lu S, Huang Y, Chen X, Zhang J, Yao Y, Cai J, Wu J, Kong J, Lin T. Preoperative fluorescence in situ hybridization analysis as a predictor of tumor recurrence in patients with non-muscle invasive bladder cancer: a bi-institutional study. J Transl Med. 2023;21(1):685. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Akaike H. Information theory and an extension of the maximum likelihood principle. In: 2nd international symposium on information theory Abstracts of papers. edn.; 1971: 276–276.
- 17.Harrell FE Jr., Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med. 1996;15(4):361–87. [DOI] [PubMed] [Google Scholar]
- 18.Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making: Int J Soc Med Decis Mak. 2006;26(6):565–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Camp RL, Dolled-Filhart M, Rimm DL. X-tile: a new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization. Clin Cancer Res. 2004;10(21):7252–9. [DOI] [PubMed] [Google Scholar]
- 20.Soukup V, Čapoun O, Cohen D, Hernández V, Burger M, Compérat E, Gontero P, Lam T, Mostafid AH, Palou J, et al. Risk stratification tools and prognostic models in Non-muscle-invasive bladder cancer: A critical assessment from the European association of urology Non-muscle-invasive bladder Cancer guidelines panel. Eur Urol Focus. 2020;6(3):479–89. [DOI] [PubMed] [Google Scholar]
- 21.Witjes JA, Bruins HM, Cathomas R, Compérat EM, Cowan NC, Gakis G, Hernández V, Linares Espinós E, Lorch A, Neuzillet Y, et al. European association of urology guidelines on Muscle-invasive and metastatic bladder cancer: summary of the 2020 guidelines. Eur Urol. 2021;79(1):82–104. [DOI] [PubMed] [Google Scholar]
- 22.Waldman FM, Carroll PR, Kerschmann R, Cohen MB, Field FG, Mayall BH. Centromeric copy number of chromosome 7 is strongly correlated with tumor grade and labeling index in human bladder cancer. Cancer Res. 1991;51(14):3807–13. [PubMed] [Google Scholar]
- 23.Ohta JI, Miyoshi Y, Uemura H, Fujinami K, Mikata K, Hosaka M, Tokita Y, Kubota Y. Fluorescence in situ hybridization evaluation of c-erbB-2 gene amplification and chromosomal anomalies in bladder cancer. Clin Cancer Res. 2001;7(8):2463–7. [PubMed] [Google Scholar]
- 24.Ying X, Liu B, Yuan Z, Huang Y, Chen C, Jiang X, Zhang H, Qi D, Yang S, Lin S, et al. METTL1-m(7) G-EGFR/EFEMP1 axis promotes the bladder cancer development. Clin Transl Med. 2021;11(12):e675. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kluth LA, Black PC, Bochner BH, Catto J, Lerner SP, Stenzl A, Sylvester R, Vickers AJ, Xylinas E, Shariat SF. Prognostic and prediction tools in bladder cancer: A comprehensive review of the literature. Eur Urol. 2015;68(2):238–53. [DOI] [PubMed] [Google Scholar]
- 26.Lotan Y, Gupta A, Shariat SF, Palapattu GS, Vazina A, Karakiewicz PI, Bastian PJ, Rogers CG, Amiel G, Perotte P, et al. Lymphovascular invasion is independently associated with overall survival, cause-specific survival, and local and distant recurrence in patients with negative lymph nodes at radical cystectomy. J Clin Oncol. 2005;23(27):6533–9. [DOI] [PubMed] [Google Scholar]
- 27.Canter D, Guzzo T, Resnick M, Magerfleisch L, Sonnad S, Bergey M, Tomazewski J, Vaughn D, Van Arsdalen K, Malkowicz B. The presence of lymphovascular invasion in radical cystectomy specimens from patients with urothelial carcinoma portends a poor clinical prognosis. BJU Int. 2008;102(8):952–7. [DOI] [PubMed] [Google Scholar]
- 28.Bolenz C, Herrmann E, Bastian PJ, Michel MS, Wülfing C, Tiemann A, Buchner A, Stief CG, Fritsche HM, Burger M, et al. Lymphovascular invasion is an independent predictor of oncological outcomes in patients with lymph node-negative urothelial bladder cancer treated by radical cystectomy: a multicentre validation trial. BJU Int. 2010;106(4):493–9. [DOI] [PubMed] [Google Scholar]
- 29.Mathieu R, Lucca I, Rouprêt M, Briganti A, Shariat SF. The prognostic role of lymphovascular invasion in urothelial carcinoma of the bladder. Nat Rev Urol. 2016;13(8):471–9. [DOI] [PubMed] [Google Scholar]
- 30.Werntz RP, Smith ZL, Packiam VT, Smith N, Steinberg GD. The impact of lymphovascular invasion on risk of upstaging and lymph node metastasis at the time of radical cystectomy. Eur Urol Focus. 2020;6(2):292–7. [DOI] [PubMed] [Google Scholar]
- 31.Sonpavde G, Khan MM, Svatek RS, Lee R, Novara G, Tilki D, Lerner SP, Amiel GE, Skinner E, Karakiewicz PI, et al. Prognostic risk stratification of pathological stage T3N0 bladder cancer after radical cystectomy. J Urol. 2011;185(4):1216–21. [DOI] [PubMed] [Google Scholar]
- 32.Xylinas E, Cha EK, Sun M, Rink M, Trinh QD, Novara G, Green DA, Pycha A, Fradet Y, Daneshmand S, et al. Risk stratification of pT1-3N0 patients after radical cystectomy for adjuvant chemotherapy counselling. Br J Cancer. 2012;107(11):1826–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Rink M, Hansen J, Cha EK, Green DA, Babjuk M, Svatek RS, Xylinas E, Tagawa ST, Faison T, Novara G, et al. Outcomes and prognostic factors in patients with a single lymph node metastasis at time of radical cystectomy. BJU Int. 2013;111(1):74–84. [DOI] [PubMed] [Google Scholar]
- 34.Mitra AP, Pagliarulo V, Yang D, Waldman FM, Datar RH, Skinner DG, Groshen S, Cote RJ. Generation of a concise gene panel for outcome prediction in urinary bladder cancer. J Clin Oncol. 2009;27(24):3929–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Zheng J, Kong J, Wu S, Li Y, Cai J, Yu H, Xie W, Qin H, Wu Z, Huang J, et al. Development of a noninvasive tool to preoperatively evaluate the muscular invasiveness of bladder cancer using a radiomics approach. Cancer. 2019;125(24):4388–98. [DOI] [PubMed] [Google Scholar]
- 36.Mowatt G, Zhu S, Kilonzo M, Boachie C, Fraser C, Griffiths TR, N’Dow J, Nabi G, Cook J, Vale L. Systematic review of the clinical effectiveness and cost-effectiveness of photodynamic diagnosis and urine biomarkers (FISH, immunocyt, NMP22) and cytology for the detection and follow-up of bladder cancer. Health Technol Assess. 2010;14(4):1–331. iii-iv. [DOI] [PubMed] [Google Scholar]
- 37.Gayed BA, Seideman C, Lotan Y. Cost-effectiveness of fluorescence in situ hybridization in patients with atypical cytology for the detection of urothelial carcinoma. J Urol. 2013;190(4):1181–6. [DOI] [PubMed] [Google Scholar]
- 38.Stordal B, Peters G, Davey R. Similar chromosomal changes in cisplatin and oxaliplatin-resistant sublines of the H69 SCLC cell line are not associated with platinum resistance. Genes Chromosomes Cancer. 2006;45(12):1094–105. [DOI] [PubMed] [Google Scholar]
- 39.Eidelman Y, Salnikov I, Slanina S, Andreev S. Chromosome folding promotes intrachromosomal aberrations under Radiation- and Nuclease-Induced DNA breakage. Int J Mol Sci 2021, 22(22).12186. [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All the data generated or analyzed during this study are included in this published article and its supplementary information files.