Role of Postoperative Radiation Therapy After Radical Cystectomy for Bladder Cancer: A Systematic Review and Meta-Analysis

Abstract

Purpose

Radical cystectomy with neoadjuvant cisplatin-based chemotherapy is the standard of care for muscle-invasive bladder cancer (MIBC). However, recurrence remains common. Postoperative radiation therapy (PORT) has been explored to improve locoregional control and survival, but evidence remains inconclusive. This systematic review and meta-analysis assessed the role of PORT in MIBC.

Methods and Materials

PubMed, Embase, and Cochrane were searched through March 2025 (CRD42025632052). Eligible studies included randomized controlled trials (RCTs), prospective nonrandomized studies, and comparative observational cohorts of patients with MIBC undergoing radical cystectomy, with or without PORT. Prospective single-arm studies of PORT were included for safety outcomes. Primary outcomes were overall survival (OS) and recurrence-free survival (RFS); secondary outcomes included locoregional recurrence (LRR) and treatment-related toxicities. Random-effects models were used.

Results

Fifteen studies (5 RCTs, 2 prospective nonrandomized, and 8 retrospective cohorts) were included. A total of 19,033 patients (PORT, 1157) were included in the comparative analysis. PORT was associated with improved OS (hazard ratio, 0.84; 95% CI, 0.71-0.98; P = .026), although these results were not maintained in urothelial-only analysis. LRR was markedly reduced (risk ratio, 0.31; 95% CI, 0.15-0.65; P = .002), a finding maintained in urothelial-only and RCT-only analyses. No RFS benefit was observed in the primary analysis, but sensitivity, urothelial-only, and RCT-only analyses demonstrated significant improvements. Grade ≥3 acute gastrointestinal and genitourinary toxicities occurred in 3.3% and 1.8% of patients, respectively; late toxicities in 4.3% and 0.62%.

Conclusions

PORT reduced LRR across all analyses, improved OS in the overall and RCT-only analysis, and showed RFS benefits in sensitivity, urothelial-only, and RCT-only analyses. PORT may benefit selected high-risk patients, and its integration with contemporary radiation and systemic therapies warrants further prospective evaluation.

Introduction

Neoadjuvant cisplatin-based chemotherapy followed by radical cystectomy (RC) remains the standard of care (SOC) for muscle-invasive bladder cancer (MIBC). Recently, the incorporation of immune checkpoint inhibitors in the perioperative setting, as well as novel antibody-drug conjugate combinations, has improved outcomes in selected patients, particularly those with residual disease after neoadjuvant chemotherapy or those ineligible for cisplatin-based regimens.1, 2, 3, 4, 5 However, despite all recent advances in systemic agents, a high risk of recurrence still exists, and small increments have been made to the classic 50% to 60% estimated 5-year overall survival (OS) mark.^3,6,7 Metastatic recurrence is the most common recurrence pattern and is largely responsible for the mortality associated with MIBC. Nevertheless, locoregional recurrences are also frequent, occurring in up to a third of the cases, and are typically associated with morbidity and poor outcomes.⁸ A predictable pattern of local failure has been studied, occurring in iliac and/or obturator nodes in patients with stage pT3 tumors or higher and in cystectomy bed and/or presacral nodal when margins were positive, delineating a rationale for postoperative radiation therapy (PORT) use in locally advanced disease, with the number of lymph nodes compromised and resected also being a prognostic factor to be considered.^9,10

PORT has been explored post-RC in several studies over the last decades, showing benefits in local control, with more inconsistent results in recurrence-free survival (RFS) and OS.11, 12, 13 Moreover, many of such studies have been subject to criticism regarding the population included, with a high proportion of nonurothelial histology, limited use of contemporary systemic therapy, and predominance of non–Western-based populations. Special concerns have been raised regarding the safety of PORT, with older techniques resulting in considerable risk of late gastrointestinal (GI) toxicities.¹¹ Yet, with the advent of modern techniques such as intensity modulated radiation therapy (IMRT), toxicity profiles have improved. Importantly, high-risk patients, such as those with pT4 disease, positive surgical margins, extensive nodal involvement, or those with incomplete lymphadenectomy, remain underrepresented in contemporary perioperative immunotherapy trials, despite being at highest risk of locoregional failure. In this context, and given the potential synergy between radiation therapy and immunotherapy, the role of PORT in the contemporary era using modern radiation therapy techniques warrants reassessment.

Methods and Materials

A systematic literature review and meta-analysis was conducted using the most recent Cochrane Handbook recommendations as guide. This study followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines (Tables E1 and E2).¹⁴

Literature search and selection process

A systematic literature search was conducted in PubMed, Embase, and Cochrane Central from inception to March 2025. The full search strategy is detailed in Appendix E1. Additional records were identified through manual reference screening. Only studies published in English were included. For trials with multiple reports, the most recent and comprehensive publication was selected, prioritizing full-text articles unless abstracts provided sufficient data. Efficacy outcomes from the Murthy et al¹⁵ 2025 randomized trial were incorporated through targeted updating, as the conference abstract¹⁶ was released after completion of the primary search strategy and provided outcome data not available in the initial full-text publication, which reported only safety results. Two reviewers (RLN and FC) independently screened titles and abstracts; discrepancies were resolved through discussion or third-party review (DVA). The review protocol was registered in PROSPERO (CRD42025632052).

Eligibility criteria

This review included both comparative and single-arm studies, with eligibility criteria tailored to the objectives of each analysis. We used the population, interventions, comparator, outcome, and study design framework to define the eligibility criteria (Table E3).

For the comparative analysis, eligible studies were: (1) randomized controlled trials (RCTs) or comparative observational studies; (2) comparing PORT + SOC versus SOC alone; (3) enrolling patients with localized bladder carcinoma undergoing RC; and (4) reporting OS, RFS, or local disease control.

For the single-arm analysis, eligible studies included prospective cohorts evaluating patients with MIBC treated with PORT following RC, reporting on adverse events or treatment-related toxicities.

Data collection and synthesis

All authors independently extracted data. Discrepancies were resolved by consensus or a third reviewer (DVA).

For survival outcomes, hazard ratios (HRs) and 95% CIs were extracted when available. Multivariable-adjusted estimates were prioritized. When HRs were not reported, Kaplan-Meier curves were reconstructed, and HRs were estimated using the IPDfromKM and coxph R packages.¹⁷ Log-transformed HRs and standard errors were computed for meta-analysis.

For single-arm studies, we extracted adverse event data based on the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE),¹⁸ focusing on grade 2 (G2) and grade ≥3 (G3+) toxicities. The most consistently reported systems, GI and genitourinary (GU), were prioritized. For G3+ events, we captured the proportion of patients experiencing at least one such event, regardless of recurrence.

When studies used alternative grading systems (Radiation Therapy Oncology Group and National Cancer Institute’s Common Toxicity Criteria), we mapped these to CTCAE-equivalent grades to ensure consistency across cohorts. Full mapping methods are detailed in Appendix E2.

Endpoints

Outcomes assessed included OS, RFS, and LRR. Definitions of LRR varied across studies and are summarized in Table E4. Secondary outcomes included adverse events related to radiation therapy, with particular attention to G2 to G5 toxicity when available.

Quality assessment

Risk of bias was assessed using version 2 of the Cochrane Risk of Bias (RoB 2) tool for randomized trials and Risk Of Bias In Non-randomized Studies of Interventions(ROBINS-I) for observational studies reporting OS.^19,20 Two independent reviewers (RLN and HC) conducted the assessments, resolving disagreements by consensus.

We assessed publication bias using funnel plot. For outcomes with at least 7 studies, we performed Egger’s regression test.

Certainty of evidence was graded as very low, low, moderate, or high using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach²¹ by 2 independent reviewers (FCM and LML); disagreements were resolved by consensus.

Statistical analysis

Comparative outcomes

For time-to-event outcomes (OS and RFS), HRs with 95% CIs were pooled. For dichotomous outcomes (LRR), risk ratios (RRs) were used. Heterogeneity was assessed using the Cochran Q test (P < .10) and quantified by the I² statistic (I² >25% indicating moderate/high heterogeneity). Analyses were performed in R (v4.4.1) using a random-effects model (DerSimonian and Laird method).²²

We conducted sensitivity and subgroup analyses to assess robustness. For the main outcomes, a leave-one-out analysis was performed by sequentially excluding each study.

For OS, we selected Ernandez et al²³ and Fischer-Valuck et al²⁴ from among 5 National Cancer Database studies25, 26, 27 based on analytical profile, cohort characteristics, and methodological rigor (see Table E5 for details).

We also included 4 nonoverlapping studies by Zaghloul et al11, 12, 13^,28 (1992, 2007, 2018, and 2024). The 1992¹¹ and 2007²⁸ trials focused on squamous cell carcinoma (SCC) and adenocarcinoma, respectively, while the 2018¹² and 2024¹³ RCTs enrolled urothelial carcinoma patients and employed modern radiation therapy techniques.

Subgroup analyses evaluated potential effect modifiers, including histologic subtype (urothelial) and study design (RCTs only).

Single-arm safety analysis

Safety outcomes were pooled using random-effects single-arm meta-analyses. Detailed transformation methods are described in Appendix E2.

Results

Study selection and characteristics

The initial search yielded 1028 records. After removing duplicates and screening titles and abstracts, 41 studies underwent full-text review (Fig. 1). Ultimately, 15 studies met the eligibility criteria and were included in the meta-analysis. Of these, 13 were comparative studies (4 RCTs, 1 prospective nonrandomized study, and 8 retrospective cohort studies), and 6 contained prospective data eligible for the single-arm portion evaluating PORT (References E1-E17).

Figure 1.

Preferred Reporting Items for Systematic reviews and Meta-Analyses flow diagram of study selection.

Abbreviation: SOC = standard of care.

In the comparative arm, a total of 19,033 patients were included across all studies, of whom 1157 (6.1%) received PORT after RC.

The majority of patients had urothelial carcinoma (87.7%), while other frequently reported histologies were SCC and adenocarcinoma. PORT doses ranged from 34.2 to 60 Gy.

Median follow-up times varied across studies. Patient and study characteristics are summarized in Table 111, 12, 13^,15,23,24,28, 29, 30, 31, 32, 33, 34 and Table 2.^{12,13,15,30,35,36}

Table 1.

Characteristics of studies included in the comparative analysis of postoperative radiation therapy (PORT) versus no radiation therapy (non-PORT) in muscle-invasive or locally advanced bladder cancer

Study	Design	Comparison	RT technique; dose	Follow-up RT vs SOC	Number of patients, total N (RT n vs SOC n)	Female, n (%) RT vs SOC	Age, y RT vs SOC	Lymph node involvement (positive), n (%) RT vs SOC	Surgical margin positive, n (%) RT vs SOC	Histology n (%) urothelial, n (%) others RT vs SOC	pT n (%) of the total study sample size
Zaghloul11, 1992	RCT	PORT CF vs non-PORT	2D-RT; 50 Gy/25 F/5 wk	Median, 69 mo*	161 (78 vs 83)	13 (16.7%) vs 17 (20.5%)	Mean, 48 ± 9; Mean, 47 ± 8, 50± 9†	21 (27%) vs 19 (22.8%)	N/A	14 (17.9%), 64 (82.1%) vs 19 (22.6%), 65 (77.4%)‡	pT3, 130 (80.7%), pT4, 31 (19.2%)
Karakiewicz29, 2006	Retrospective	PORT vs non-PORT	N/A	Median, 24.9 (range, 0.1-183.4)*	728 (34 vs 694)	129 (17.7%)*	Mean, 64.5, median, 66 (range, 33.1-89.2)*	N/A	N/A	34 (100%), 0 (0%) vs 694 (100%), 0 (0%)	pT0, 56 (7.7%), pTis, 92 (12.6%), pTa, 23 (3.2%), pT1, 94 (12.9%), pT2, 163 (22.4%), pT3, 215 (29.5%), pT4, 85 (11.7%)
Zaghloul28, 2007	Retrospective	PORT vs non-PORT	2D-CRT; 50 Gy/25 F/5 wk	Median, 47 mo (range, 0.1-108)*	216 (82 vs 134)	66 (30.6%)*	Mean, 51.0 ± 10.6 (range, 24-80)*	N/A	N/A	0 (0%), 82 (100%) vs 0 (0%), 134 (100%)	pT1, 7 (3.3%); pT2, 56 (25.9%); pT3, 116 (53.7%); pT4, 37 (17.1%)
Abdel Raheem30, 2011	Prospective	ART vs non-ART	2D-CRT; 50 Gy/25 F/5 wk	Mean, 16 ± 10 mo*	71 (38 vs 33)	14 (36.8%) vs 10 (30.3%)	<50: 22 (57.9%) vs 15 (45.5%) ≥50: 16 (42.1%) vs 18 (54.5%)	3 (7.9%) vs 7 (21.2%)	N/A	0 (0%), 38 (100%) vs 0 (0%), 33 (100%)	pT2, 61 (85.9%), pT3, 7 (9.9%), pT4, 3 (4.2%)
Bayoumi31, 2014	Retrospective	PORT vs non-PORT	3D-CRT; Mean, 49 ± 6 Gy (range, 42-60)	Median, 47 mo (range, 17-77)*	170 (92 vs 78)	22 (23.9%) vs 49 (62.8%)	Mean, 56 ± 7 vs 60 ±13	40 (43.4%) vs 18 (23%)	N/A	60 (65.2%), 32 (35.5%) vs 44 (56.4%), 34 (43.6%)	pT3, 114 (67%); pT4, 56 (33%)
Tuanquin32, 2016	Retrospective	PORT vs non-PORT	IMRT; dose range, 34.2-58 Gy	N/A	63 (10 vs 53)	N/A	N/A	N/A	6 (60%) vs 3 (6%)	7 (70%), 3 (30%) vs 49 (92%), 4 (8%)	pT3/pT4, 63 (100%)
Pouessel33, 2017	Retrospective	PORT + AC vs non-PORT + AC	3D-CRT; median, 45 Gy	Median, 4.2 y (range, 0.4-11.0)*	226 (13 vs 213)	37 (16.4%)*	Median, 62.4 (range, 35-82)*	N/A	N/A	13 (100%), 0 (0%) vs 213 (100%), 0 (0%)	pT0, 7 (3.1%), pTa-pTis, 3 (1.3%), pT1, 5 (2.2%), pT2, 31 (13.7%), pT3, 115 (50.9%), pT4, 65 (28.8%)
Zaghloul12, 2018	RCT	CCRT vs AC	3D-CRT; 45 Gy/1.5 Gy twice-daily F/3 wk	Median, 24 mo vs 27 mo	120 (75 vs 45)	15 (20.0%) vs 8 (17.8%)	Median, 52 vs 55	35 (46.7%) vs 17 (37.8%)	0 (0%) vs 0 (0%)	41 (54.7%), 34 (45.3%) vs 23 (51.1%), 22 (48.9%)	pT2, 11 (9.1%); pT3, 90 (75%); pT4, 19 (15.8%)
Fischer-Valuck24, 2019	Retrospective	PORT vs non-PORT	EBRT; median, 50.4 Gy (IQR, 45-55.8)	Median, 18.6 mo vs 18.8 mo	15,124 (512 vs 14,612)	193 (37.7%) vs 4029 (27.6%)	Mean, 64.6; median, 65 (range, 22‐90) vs mean, 68.0, median; 69 (range, 32‐90)	254 (49.6%) vs 6934 (47.5%)	263 (51.4%) vs 2905 (19.9%)	389 (76%), 123 (24%) vs 12,972 (88.8%), 1640 (11.2%)	pT3, 9948 (65.7%), pT4, 5176 (34.2%)
Ernandez23, 2022	Retrospective	AC + PORT vs AC	N/A; ≥45 Gy	Median, 22.5 mo (15.0-31.0) vs 26.0 mo (13.0-53.0)	1684 (66 vs 1618)	27 (41%) vs 387 (24%)	Median, 65 (IQR, 59-70) vs 65 (IQR, 58-70)	41 (62.1%) vs 1060 (65.5%)	29 (44%) vs 269 (17%)	66 (100%), 0 (0%) vs 1618 (100%), 0 (0%)	pT1/T2, 249 (15%), pT3, 987 (59%), pT4, 448 (27%)
Zaghloul13, 2024	RCT	PORT vs non-PORT	IMRT; 50 Gy/25 F	43.4 mo (IQR, 26.4-52.6)*	122 (62 vs 60)	7 (11%) vs 5 (8.3%)	Mean, 58 (IQR, 54-65) vs 64 (IQR, 54-67)	25 (40.3%) vs 15 (25%)	0 (0%) vs 0 (0%)	62 (100%), 0 (0%) vs 60 (100%), 0 (0%)	pT2, 21 (17.2%), pT3, 77 (63.1%), pT4, 24 (19.6%)
Marcq34, 2025	Retrospective	PORT vs non-PORT	N/A	Median, 51 mo (range, 44-62)*	195 (18 vs 137)	64 (32.8%)*	Median, 65 (IQR, 59-71)§	18 (100%) vs 137 (100%)	51 (26.2%)§	18(100%), 0(0%) vs 137(100%), 0(0%)	ypT0, 10 (5.13%), ypT1, 11 (5.64%), ypT2, 29 (14.87%), ypT3, 79 (40.51%), ypT4, 61 (31.28%), ypTis, 5 (2.56%)§
Murthy15, 2025	RCT	PORT vs non-PORT	IMRT; 50.4 Gy/28 F	Median, 23 mo	153 (77 vs 78)	N/A	N/A	23 (29.8%)§	N/A	77(100%), 0(0%) vs 78 (100%), 0(0%)	pT3/T4, 95 (62%)

Data include design, radiation therapy regimen, number of patients, age, lymph node and margin status, histology, and pathologic staging.

Abbreviations: 2D-RT = 2-dimensional radiation therapy; 3D-CRT = 3-dimensional conformal radiation therapy; AC = adjuvant chemotherapy; CCRT =concurrent chemoradiotherapy; EBRT = external beam radiation therapy; F, fractions; IMRT = intensity modulated radiation therapy; N/A = not available; PORT = postoperative radiation therapy; pT = pathologic T stage; RCT = randomized controlled trial; RT = radiation therapy; SOC = standard of care.

^⁎Data for the total study population.

^†Age of the Cystectomy 1981-1984, Cystectomy 1984-1988, and Postoperative Radiation Therapy Conventional Fractionation subpopulation.

^‡RT Urothelial, others, undifferentiated versus SOC urothelial, others, undifferentiated.

^§Data from the total study population, with pathologic staging at Radical Cystectomy, with all patients receiving neoadjuvant chemotherapy.

Table 2.

Characteristics of studies reporting on acute and/or late toxicity after postoperative radiation therapy (PORT)

Study	Design	RT technique; dose	Follow-up	Patients in acute analysis	Patients in late analysis	Female, n (%)	Age, y	Lymph node involvement (positive), n (%)	Surgical margin positive, n (%)	Histology n (%) urothelial, n (%) others	pT, n (%)
Abdel Raheem30, 2011	Prospective	3D-CRT; 50 Gy/25 F/5 wk	Mean, 16 ± 10 mo*	38	38	14 (36.8%)	<50: 22 (57.9%), ≥50: 16 (42.1%)	3 (7.9%)	N/A	0 (0%), 38 (100%)	pT2, 34 (89.4%), pT3, 2 (5.2%), pT4, 2 (5.2%)
El-Monim35, 2013	RCT	2D-RT; 50 Gy/25 F/5 wk	Median, 32 mo (range, 0-69)*	-	45	15 (30%)†	Mean, 55.6 ± 9.7†	9 (20%)†	N/A†	0 (0%), 50 (100%)†	pT2, 5 (10%), pT3, 43 (86%), pT4, 2 (4%)†
Zaghloul12, 2018	RCT	3D-CRT; 45 Gy/1.5 Gy twice-daily F/3 wk	Median, 24 mo (range, 5-127)	-	75	15 (20.0%)	Median, 52	35 (46.7%)	0 (0%)	41 (54.7%), 34 (45.3%)	pT2, 6 (8%), pT3, 61 (83.1%), pT4, 8 (10.7%)
Fonteyne36, 2022	Prospective	IMRT; 50 Gy/25 F	Median, 18 mo (range, 1-72)†	72/17‡	51/11‡	28 (25%)†	70 (34-87)†	47 (65%)†	14 (19%)†	60 (83%), 12 (17%)†	pT≤2, 16 (22%), pT3, 35 (49%), pT4, 21 (29%) 2
Zaghloul13, 2024	RCT	IMRT; 50 Gy/25 F	Median, 42.7 mo*	62	62	7 (11%)	Mean, 58 (IQR, 54-65)	25 (40.3%)	0 (0%)	62 (100%), 0 (0%)	pT2, 7 (11%), pT3, 43 (69%), pT4, 12 (19%)
Murthy15, 2025	RCT	IMRT; 50.4 Gy/28 F	Median, 27 mo§	63	47	8 (10.4%)†	Median, 58 (IQR, 50-64)†	23 (29.8%)†	4 (5.2%)†	77 (100%), 0 (0%)†	pT0, 9 (11.7%), pT1, 12 (16.9%), pT2, 11 (14.3%), pT3, 26 (33.8%), pT4, 19 (24.7%)†

Data include radiation therapy regimen, median follow-up, number of patients analyzed for acute and late toxicities, age, lymph node and margin status, histology, and pathologic staging.

Abbreviations: D-RT = 2-dimensional radiation therapy; 3D-CRT = 3-dimensional conformal radiation therapy; F = fractions; IMRT = intensity modulated radiation therapy; N/A = not available; PORT = postoperative radiation therapy; pT = pathologic T stage; RCT = randomized controlled trial; RT = radiation therapy.

^⁎Value for the total N of the study.

^†Data for the total N with PORT.

^‡Patients analyzed for gastrointestinal/patients analyzed for genitourinary.

^§Median follow-up for the 104 patients evaluated for late toxicity.

Pooled analysis of comparative studies

OS

Nine comparative studies, including 17,728 patients (887 patients in the PORT + SOC group and 16,841 in the SOC group), reported OS outcomes. PORT was associated with a significant OS benefit compared with SOC alone (HR, 0.84; 95% CI, 0.71-0.98; P = .026; I² = 18.6%) (Fig. 2a), with consistent results on leave-one-out analysis (Fig. E1). Restriction to RCTs confirmed significant benefit (HR, 0.73; 95% CI, 0.53-0.98; P = .039; I² = 0%) (Fig. 2b). In urothelial-only analyses, OS did not differ between groups (HR, 0.85; 95% CI, 0.64-1.12; P = .248; I² = 38.4%) (Fig. E2), including when further limiting the analysis to RCTs (HR, 0.77; 95% CI, 0.54-1.10; P = .148; I² = 0%) (Fig. E3). Funnel plot inspection suggested mild asymmetry (Fig. E4); Egger’s test was not significant (P = .1072).

Figure 2.

(a) Forest plot of overall survival. (b) Overall survival in randomized controlled trials only.

Abbreviations: HR = hazard ratio; IV = inverse variance; PORT = postoperative radiation therapy; SOC = standard of care.

RFS

Ten comparative studies including 1884 patients (535 in the PORT + SOC group and 1349 in the SOC group) reported RFS outcomes. The overall pooled analysis showed no significant difference between PORT and SOC (HR, 0.82; 95% CI, 0.48-1.41; P = .482), with substantial heterogeneity (I² = 93.7%) (Fig. 3a). Exclusion of a single influential study reduced heterogeneity and revealed a significant RFS benefit with PORT (HR, 0.63; 95% CI, 0.52-0.78; I² = 35.6%) (Fig. E5). Restriction to RCTs confirmed significant benefit (HR, 0.54; 95% CI, 0.42-0.70; P < .001; I² = 0%) (Fig. 3b). In urothelial-only analyses excluding the outlier, PORT remained associated with improved RFS (HR, 0.64; 95% CI, 0.46-0.91; P = .011; I² = 0%) (Fig. E6), with consistent results limiting the analysis to RCTs (HR, 0.61; 95% CI, 0.41-0.91; P = .016; I² = 0%) (Fig. E7).

Figure 3.

(a) Forest plot of recurrence-free survival. (b) Recurrence-free survival in randomized controlled trials only.

Abbreviations: HR = hazard ratio; IV = inverse variance; PORT = postoperative radiation therapy; SOC = standard of care.

Karakiewicz et al²⁹ was identified as an outlier under the random-effects model, consistent with its influence on both effect magnitude and heterogeneity. Funnel plot inspection revealed no major asymmetry except for outlier (Fig. E8); Egger’s test was not significant (P = .8426).

LRR

Seven comparative studies including 1005 patients (476 in the PORT + SOC group and 529 in the SOC group) reported LRR outcomes. PORT significantly reduced LRR compared with SOC alone (RR, 0.31; 95% CI, 0.15-0.65; P = .002; I² = 80.8%) (Fig. 4a). This benefit was consistent in leave-one-out analyses despite persistently high heterogeneity (Fig. E9). Restriction to RCTs confirmed significant reduction in LRR (RR, 0.22; 95% CI, 0.10-0.47; P < .001; I² = 57.3%) (Fig. 4b). In urothelial-only RCTs, PORT remained associated with lower LRR (RR, 0.39; 95% CI, 0.22-0.68; P = .001; I² = 0%) (Fig. E10). Funnel plot inspection suggested asymmetry (Fig. E11); Egger’s test was not significant (P = .1781).

Figure 4.

(a) Forest plot of locoregional recurrence. (b) Locoregional recurrence in randomized controlled trials only.

Abbreviations: MH = Mantel-Haenszel; PORT = postoperative radiation therapy; RR = risk ratio; SOC = standard of care.

Safety analysis from single-arm studies

In single-arm analyses of prospective cohorts, PORT demonstrated a favorable safety profile. G3+ treatment-related toxicities were uncommon, with pooled rates of acute GI and GU adverse events of 3.31% (95% CI, 0.6%-7.6%; I² = 47.3%) and 1.8% (95% CI, 0.0%-14.9%; I² = 78.9%), respectively, and late GI and GU events of 4.3% (95% CI, 1.8-7.7; I² = 34.8%) and 0.62% (95% CI, 0.00-5.03; I² = 60.3%). G2 GI toxicity was more frequent, with pooled rates of 47.8% (95% CI, 27.0-68.5; I² = 86.9%) in acute setting and 17.1% (95% CI, 2-41.1; I² = 93.9%) for late events, generally reflecting manageable symptoms. Late G2 GU toxicity had a pooled rate of 8.5% (95% CI, 1.4-38.7; I² = 86.3%).

Subgroup analyses by radiation therapy technique showed comparable rates of G3+ late GI toxicity, with no significant differences across modalities, whereas G2 GI toxicity showed greater variability across cohorts. Detailed stratified analyses are provided in Appendix E3 and Figs. E12-E25. Overall, these findings support the acceptable tolerability of PORT with contemporary techniques.

Risk of bias assessment and GRADE assessment

The risk of bias assessments using the RoB2 and ROBINS-I tools are summarized in Tables E6 and E7. Nine studies were rated as having a serious overall risk of bias, primarily due to confounding and participant selection issues; 3 studies were rated with moderate risk of bias; only Zaghloul et al¹³ was rated as having low risk across all domains. Tuanquin et al³⁰ and Murthy et al¹⁶ were abstracts and had limited data.

The GRADE assessment of the quality of evidence was rated as low for OS and RFS and very low for LRR (Table E8).

Discussion

In this systematic review and meta-analysis evaluating PORT after RC for MIBC, 4 key findings emerged: (1) an important benefit in LRR reduction was demonstrated; (2) RFS improvement was not observed considering all studies; however, a significant benefit was detected in sensitivity and subgroup analyses; (3) a significant OS benefit was seen in the pooled analysis and remained statistically significant in the RCT-only subset, but not when restricted to urothelial histology alone; and (4) overall, PORT had a favorable safety profile.

Our findings build on prior heterogeneous reports of PORT in MIBC. Early single‑institution series and small RCTs demonstrated improved local control with PORT but inconsistent effects on survival.11, 12, 13

In our meta-analysis, PORT reduced LRR, with a 69% relative risk reduction in pelvic failure. This reinforces the capacity of PORT to improve local control.³⁷ However, the clinical significance of this endpoint remains debated, as disease progression in MIBC is often driven by distant relapse,³⁸ which may limit the impact of local interventions on OS. Interpretation of this endpoint should also consider that definitions of LRR were not fully standardized across studies, which may influence cross-study comparability.

Regarding RFS, the pooled analysis was not statistically significant. This result appears to be largely driven by the Karakiewicz et al²⁹ study, in which the estimated effect may be influenced by confounding by indication, given that patients receiving radiation therapy likely represented a higher-risk subgroup. To further explore heterogeneity, we performed leave-one-out sensitivity analyses, which identified this study as a major contributor. Its exclusion substantially reduced heterogeneity and resulted in a statistically significant RFS benefit associated with PORT. This pattern was consistent in analyses restricted to RCTs and urothelial-only cohorts, suggesting that the overall negative result may be driven by this outlier rather than reflecting a true lack of treatment effect.

For OS, we observed a significant benefit in the overall pooled analysis, which was sustained in RCT-only analyses but not in urothelial-only cohorts. This limited signal may reflect several factors. First is the predominance of metastatic relapse in MIBC,³⁸ which may attenuate the impact of improved locoregional control on survival. Second, variability in systemic therapy across eras and regions likely confounds the independent effect of PORT. The observed benefit may be partly influenced by historical cohorts enriched for nonurothelial histologies or treated before the widespread adoption of modern systemic therapies.

Most included studies predated contemporary perioperative immunotherapy, which is now part of SOC for selected high-risk patients,1, 2, 3 and may therefore limit the applicability of these findings to current practice. This reflects a shift toward improved systemic disease control, with immunotherapy likely reducing distant recurrence, whereas its impact on locoregional control remains less well defined. Accordingly, the incremental benefit of PORT observed in historical cohorts may not be directly transferable to modern settings. Nevertheless, PORT may still have a role in selected high-risk patients, particularly those with adverse local pathologic features,³⁹ although its incremental benefit in the contemporary context remains uncertain and warrants prospective evaluation.

In addition, histologic subtype represents an important source of heterogeneity and confounding. Several included studies, particularly some of the RCTs, enrolled substantial proportions of patients with nonurothelial histologies such as SCC and adenocarcinoma, which are less common in contemporary Western populations. Nonurothelial histologies may exhibit distinct patterns of spread and potentially different sensitivities to locoregional therapies. In particular, SCC of the bladder is arguably biologically more radiosensitive than urothelial carcinomas, with greater proliferative activity and stronger dependence on locoregional spread, a pattern consistently observed across SCC-predominant malignancies such as cervical⁴⁰ and head and neck cancers.⁴¹ Consistent with this, the pooled LRR analysis restricted to RCTs showed substantial heterogeneity (I² = 57.3%), largely driven by older Egyptian trials with high proportions of nonurothelial bladder cancers,^11,12,28 demonstrated a more pronounced benefit from PORT, compared with more recent urothelial-only RCTs.^13,16 When analyses were confined to urothelial-only populations, the LRR benefit of PORT was maintained but attenuated (RR shifting from 0.31 to 0.39), supporting the notion that nonurothelial cohorts may overestimate the magnitude of benefit. A similar pattern was observed for OS, with no clear benefit in urothelial-only analyses; the RCT-only urothelial subgroup included only 2 studies (n = 275), limiting statistical power. These findings suggest that histologic composition may influence the magnitude of benefit observed with PORT. However, given the limited number of studies and the potential for residual confounding, these observations should be interpreted with caution and considered hypothesis-generating. Future prospective studies are needed to better define the role of PORT across histologic subtypes, particularly in the context of contemporary systemic therapies.

Safety analyses demonstrated low rates of G3+ GI and GU toxicity with contemporary radiation therapy techniques, with events consistently uncommon across studies. Subgroup analyses showed no significant differences in severe toxicity across radiation therapy modalities. In contrast, G2 GI toxicity exhibited substantial variability across cohorts, with higher reported rates in some IMRT series. However, funnel plot asymmetry suggests potential reporting bias, likely reflecting differences in toxicity assessment practices, as more recent IMRT studies tend to use systematic CTCAE-based reporting, whereas earlier cohorts may have underreported lower-grade events.

Limitations

Several limitations must be acknowledged. First, substantial heterogeneity across studies, reflecting differences in study design, patient populations, PORT techniques, and systemic therapy use, may have influenced pooled estimates. Second, reliance on HRs extracted from Kaplan-Meier curves introduces potential estimation error. Third, most included cohorts predated contemporary systemic therapies, limiting applicability to modern practice.⁴² Finally, the certainty of evidence was low to very low, reflecting risk of bias, confounding in observational studies, and indirectness associated with nonurothelial studies conducted outside contemporary Western settings. This is particularly relevant given the weight of National Cancer Database–based analyses in our study, which remain susceptible to selection bias and unmeasured confounding, even when propensity score-matched, potentially influencing the observed treatment effects. To mitigate these limitations, we performed sensitivity and subgroup analyses, including restriction to RCTs, which demonstrated findings with lower heterogeneity.

Conclusion

In conclusion, PORT after RC for MIBC was associated with a statistically significant OS benefit and improved locoregional control. However, OS benefit was not sustained in analyses limited to urothelial carcinoma, warranting cautious interpretation. No significant RFS improvement was observed in the overall analysis, although benefit emerged in sensitivity analyses, including leave-one-out, urothelial-only, and RCT-only analyses. Overall, the benefit of PORT appears context-dependent and may be more evident in selected high-risk patients. Future prospective studies incorporating contemporary systemic therapies and risk stratification are needed to define the role of PORT in multimodal management of MIBC.