The mean age (± SD) of the patients was 61.37 ± 11.9 years. The patient data and their relationships to tumor morphology are summarized in Table 1.
The chi-square test showed that the coral-like group and clump-like group had significant differences in surgical methods, histologic grade, depth of invasion, and lymph node metastasis.
The Spearman correlation test showed that bladder tumor morphology was moderately correlated with invasion depth (ρ = 0.492, p < 0.001) and invasion status (ρ = 0.467, p < 0.001), whereas muscle-invasive status (ρ = 0.36, p < 0.001), surgical methods (ρ = 0.213, p < 0.001), histologic grade (ρ = 0.321, p < 0.001), and lymph node metastasis (ρ = 0.167, p = 0.001) were weakly correlated, and the remaining variables were not correlated (Table 2).
We used binary logistic regression to analyze the hazard ratios between invasive status and various factors, and the results suggest that tumor morphology was associated with invasive status (HR = 8.27; 95% CI: 4.3–15.79, p < 0.001), partial cystectomy (HR = 10.65; 95% CI: 2.74–41.37, p = 0.001) and radical cystectomy (HR = 15.75; 95% CI: 7.93–33.98, p < 0.001), as shown in Table 3.
Among the 386 patients, 371 survived (median follow-up duration 45 months, interquartile range 29–60 months). In total, 350 patients did not experience recurrence (median follow-up duration 43 months, interquartile range 22–58 months). Disease progression occurred in 36 patients, 20 of whom survived.
Kaplan–Meier analysis produced the following results: different bladder tumor morphologies (coral-like and clump-like) were not associated with OS, and tumor morphology was not associated with OS (log-rank p = 0.206) or PFS (log-rank p = 0.250), as shown in Fig. 2.
In univariable analyses, tumor morphology was not associated with OS (HR = 1.9; 95% CI: 0.69–5.26, p < 0.214). However, coral-like morphology was associated with better PFS (HR = 0.63; 95% CI: 029–1.39, p = 0.225). In terms of the surgical type, radical cystectomy was significantly associated with inferior OS (HR = 16.49; 95% CI: 3.69–73.69, p < 0.001) compared to TURBT and not associated with PFS (HR = 1.48; 95% CI: 0.73–2.99, p = 0.274). Outer invasive depth was significantly associated with both inferior OS (HR = 78.49; 95% CI: 9.16–672.83, p < 0.001) and PFS (HR = 4.57; 95% CI: 1.7–12.33, p < 0.002). Perivesical invasive depth was associated with inferior OS (HR = 34.27; 95% CI: 3.83–306.93, p = 0.002) but was not associated with PFS (HR = 1.93; 95% CI: 0.65–5.7, p < 0.235). Lymph node metastasis status was significantly associated with inferior OS (HR = 13.46; 95% CI: 4.27–42.45, p < 0.001) and PFS (HR = 5.25; 95% CI: 1.86–14.88, p = 0.005).
We performed multivariate analysis, which revealed that tumor morphology was not a significant independent factor for OS (p = 0.337) but was associated with PFS (p = 0.026). Outer invasive depth was an independent factor that was significantly associated with both inferior OS (HR = 53.74; 95% CI: 3.24–892.21, p = 0.005) and PFS (HR = 7.73; 95% CI: 2.06–29.06, p < 0.002). In the multivariate analysis, lymph node metastasis status remained significant for OS (HR = 6.36; 95% CI: 1.33–30.44, p = 0.021) and PFS (HR = 8.07; 95% CI: 1.9–34.27, p = 0.005), as indicated in the NCCN and EAU guidelines. However, age, sex, tumor size, surgical type and histologic grade were not significant for OS and PFS in the multivariable analysis (Table 4).