Ipsilateral recurrences after kidney-sparing surgery for upper tract urothelial carcinoma: a systematic review and meta-analysis

Abstract

Objectives

Ipsilateral recurrence is a relevant outcome measure after kidney-sparing surgery (KSS) for urothelial upper tract carcinoma (UTUC). However, studies comparing ipsilateral recurrence after segmental ureterectomy (SU) and endourological treatment (ET) are lacking. This study aimed to quantitatively assess ipsilateral recurrence in ET and SU for UTUC to inform patient counselling and future study design.

Design

Systematic review and meta-analysis.

Setting

A systematic literature search, according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), identified relevant studies from inception to September 2024. Random-effects meta-analysis was conducted to estimate the year-wise and pooled ipsilateral recurrence separately for ET and SU. Heterogeneity was explored by subgroup and meta-regression analyses and risk of bias was assessed using Joanna Briggs Institute Critical Appraisal Checklist as well as funnel plots and Egger’s regression test.

Participants

Patients undergoing ET or SU included in eligible studies reporting ipsilateral recurrence outcomes.

Main outcome measures

Time-to-event-derived ipsilateral recurrence probabilities and crude pooled ipsilateral recurrence proportions.

Results

Seventy-four studies were included, of which 40 reported ipsilateral recurrence after ET and 34 after SU (n=3044). In 11 of these, time-to-event data were available. Time-to-event-derived ipsilateral recurrence probabilities after 4 years were 0.53 (95% CI 0.25 to 0.80) after ET and the corresponding crude ipsilateral recurrence proportion was 0.48 (95% CI 0.39 to 0.57) with considerable heterogeneity (p<0.001 Cochran’s Q test). Time-to-event-derived ipsilateral recurrence probabilities after 4 years were 0.20 (95% CI 0.11 to 0.30) after SU with a crude ipsilateral recurrence proportion of 0.11 (95% CI 0.08 to 0.13) with moderate heterogeneity (p=0.20 Cochran’s Q test). The estimates are derived from separate single-arm studies with differing tumour characteristics, treatment indications, surveillance protocols and other selection mechanisms that cannot be adjusted for. Results are limited by sparse reporting of numbers at risk and censoring, and lack of uniform reporting of outcome-related variables.

Conclusions

Ipsilateral recurrence outcomes following SU were homogenous across studies. Direct comparative evidence evaluating ET versus SU remains lacking. Future studies should emphasise standardised reporting on outcomes and associated variables in KSS, as ipsilateral recurrence is a clinically important endpoint for patient counselling in UTUC.

PROSPERO registration number

CRD42025640780.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Ipsilateral recurrence is a clinically relevant factor when counselling patients for kidney-sparing surgery (KSS) for upper tract urothelial carcinoma (UTUC).

• Studies comparing ipsilateral recurrence after endourological treatment (ET) and segmental ureterectomy (SU) are lacking.

WHAT THIS STUDY ADDS

• This systematic review and meta-analysis found that the 4-year estimated ipsilateral recurrence probability was 0.53 (95% CI 0.25 to 0.80) after ET and 0.20 after SU (95% CI 0.11 to 0.30).

• Ipsilateral recurrence probabilities were not comparable between ET and SU due to selection mechanisms that could not be adjusted for.

• These findings are novel and, to our knowledge, the first attempt to determine ipsilateral recurrences after KSS on a meta-level.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• In this study, high heterogeneity particularly among ET studies, limited the robustness of the results.

• Therefore, the authors present a list of factors to support standardised outcome reporting in future studies, strengthening the evidence base for patient counselling and shared decision-making in KSS for UTUC.

Introduction

Upper tract urothelial carcinoma (UTUC) is a rare diagnosis, with an incidence of one to two per 100 000 individuals, accounting for approximately 5% of urothelial cancers.¹ Radical nephroureterectomy (RNU) with bladder-cuff excision remains the standard treatment, yet in recent years, more interest has been directed to kidney-sparing surgery (KSS).¹ KSS includes endourological treatment (ET) and segmental ureterectomy (SU) and is recommended for low-risk tumours (unifocal, small <2 cm, low-grade) and in high-risk tumours with an imperative indication.¹ Tumour location influences treatment choice, with ureteric tumours being more suitable for SU and renal pelvis tumours for ET. Furthermore, SU can be performed with or without a perioperative lymph node dissection for high-risk or invasive tumours with oncological outcomes comparable to those of RNU.²

Apart from intravesical recurrence, KSS also carries a risk of ipsilateral recurrence and progression requiring further local treatment or RNU. Reported ipsilateral recurrence rates after ET for UTUC range widely from 8% to 90%³ while subsequent RNU rates range from 9.5% to 40%. Although KSS is often reported as one entity, and there is to our knowledge an absence of direct comparative studies between ET and SU, we hypothesise that there are clinically relevant differences in ipsilateral recurrence rates after ET and SU.

Similar to other rare oncological settings with heterogeneous treatment strategies and outcomes, systematic reviews and meta-analyses can synthesise available evidence.^4,5 Thus, this study aims to quantitatively assess ipsilateral recurrence after KSS (ET and SU) using a meta-analysis approach, addressing existing research gaps to better inform future study design and clinical counselling in patients with UTUC.

Methods

This systematic review and meta-analysis was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)⁶ and was registered with PROSPERO (ID CRD42025640780). This study was designed without any formal patient or public involvement; however, it is intended to support shared decision-making in UTUC-treatment counselling.

Data source, screening strategies, inclusion and exclusion criteria

A systematic search for literature was conducted in PubMed, Embase, Cochrane Library, Academic Search Complete, Web of Science and Scopus with specific search strategy (online supplemental material). Records published until 24 September 2024 were searched and retrieved. The studies were imported to Covidence software for screening of titles, abstracts and full texts. Only cross-sectional or cohort studies were included since the research question is to estimate the cumulative incidence of ipsilateral recurrence. Studies were included according to specific Population, Exposure, Comparator, Outcome and Study type, inclusion and exclusion criteria (online supplemental material). Studies were deemed ineligible when the ET consisted of chemoablation or photodynamic therapy or if the approach was percutaneous. Each of the records was independently screened by two reviewers and conflicts were resolved by a third reviewer.

Data extraction, synthesis and meta-analysis

The data extraction was performed by ES and FL and all discrepancies were resolved by JB. Extracted information included countries, authors and year of publication, study population, median age, gender, multifocality, elective or palliative indication, use of surveillance ureteroscopies (URS) during follow-up and duration of follow-up, tumour site, tumour size, clinical tumour grade, previous/synchronous bladder cancer, risk categories based on the European Association of Urology (EAU) risk stratification model, laser type used during ET and adjuvant upper tract instillation together with funding information and conflict of interest. As the included studies did not consistently distinguish between sex and gender, the terms were treated as equivalent. Studies reporting sample size and ipsilateral recurrence were considered for the meta-analysis. Where ipsilateral recurrence was reported as percentage, the number of patients was calculated from the sample size. If the same data sources were used for several studies, only one study was included to minimise duplicate bias as suggested by Wood.⁷ When duplicate publications were identified, the highest quality or most recent publication was included. Progression to RNU and major complications (Clavien-Dindo >2) were ascertained as secondary outcome measures. Separate meta-analyses were conducted for ET and SU. From studies reporting Kaplan-Meier estimates of ipsilateral recurrence-free survival accompanied by risk tables, censoring marks or CIs, this information was extracted.

Quality and bias assessment

The methodological quality of the included studies was assessed using the Joanna Briggs Institute Critical Appraisal Checklist (JBI) adapted to meet the research question and outcome measures of this study.⁸ Quality was assessed based on three domains: (1) selection, (2) comparability and (3) outcomes, consisting of an overall 10-point score with three levels of quality: good (8–10 points), fair (6–7 points) and unsatisfactory (<6 points). Quality was assessed independently by two reviewers and a third made the consensus. The JBI checklist was used to assess methodological quality, including risks related to selection bias and aspects of detection bias. Publications bias at the meta-analysis level was evaluated using funnel plots and Egger’s regression test. Subgroup and sensitivity analyses were performed.

Subgroup and sensitivity analyses

Based on anatomical limitations when performing SU, we performed a subgroup analysis of patients with mid-ureteric or lower-ureteric tumours. Tumours situated in both distal and proximal ureter were not included for subgroup analysis. A second subgroup analysis included patient undergoing elective rather than imperative/palliative KSS, given the impact of preoperative renal function on treatment selection. Sensitivity analyses were also performed based on reported use of URS in the surveillance protocols and study quality.

Statistical analysis

Time-to-event-based meta-analysis

In 11 studies (seven ET and four SU), time-to-event data were displayed as Kaplan-Meier estimates of ipsilateral recurrence-free survival, accompanied by either risk tables, censoring marks or CIs. Individual patient data were reconstructed from published Kaplan-Meier curves using the R base graphics ‘locator’ function and the IPDfromKM package, yielding a close approximation of the original data. The method digitises Kaplan-Meier curves and extracts available graphical information (eg, CIs and/or risk tables) to estimate survival probabilities and their SEs. Survival probabilities at prespecified follow-up times (1, 2, 3 and 4 years) were transformed to an approximately normal scale and combined using inverse-weighting under random-effects models. Survival functions were estimated using the Kaplan-Meier method, and the precision of survival probabilities was quantified using Greenwood’s variance estimator. At each event time, the estimator incorporates the number of participants at risk and the number of events to derive an SE for the corresponding Kaplan-Meier survival estimate.⁹ Greenwood-based SEs were used to construct CIs for the survival curves and to obtain study-specific measures of uncertainty for inclusion in the meta-analysis. A graphical explanation of the data extraction process is presented in the online supplemental material.

Crude ipsilateral recurrence proportions meta-analysis

The DerSimonian-Laird inverse variance random-effects model was considered when calculating the pooled estimates¹⁰ instead of a fixed-effect model, since we anticipated that the studies would differ in patient risk profiles, adjuvant instillation protocols and follow-up duration, leading to genuine between-study heterogeneity in ipsilateral recurrence risks. The variance was stabilised using the Freeman-Tukey double arcsine transformation.¹¹

The heterogeneity was examined using the χ² test on Cochran’s Q statistic, which was derived from the H and I² indices.^12–14 Univariate and multivariable meta-regression analyses were performed to identify the source of heterogeneity, using variables available for all studies: reporting of previous bladder cancer, publication year before 2015 (treatment decision based on high-risk and low-risk UTUC first presented in EAU guidelines 2014) and median follow-up time <24 months. The statistical analyses were performed using STATA V.18 with the ‘meta esize’ command for meta-analysis and R V.4.5.0¹⁵ with IPDfromKM package V.0.1.10.¹⁶

Results

Narrative synthesis

In total, 74 studies involving 3044 participants were included in the systematic review as depicted in the PRISMA flow chart (online supplemental material). Out of these 74 studies selected for a narrative review, 40 publications reported outcomes after ET^17–56 and 34 after SU.^2,57–89 However, for the meta-analysis, 33 studies were selected for ET and 28 studies for SU, respectively, to minimise duplication bias. The characteristics of ET studies are summarised in table 1. Sample sizes varied between six and 168 participants. Follow-up durations and ipsilateral recurrence rates showed substantial variability across studies. Out of 33 ET studies included in the meta-analysis, 26 defined the surgical technique and laser type used. Surveillance protocol was defined in 30 of 33 studies, of which 17 reported CT and 30 reported URS as part of the surveillance protocol. Four ET studies reported outcomes separately for low-risk and high-risk disease,^18,30,33,34 of which one study included only low-risk disease³⁰ and one included only high-risk disease.³³ In total, 114 patients treated with ET were categorised as low risk, of which 42 (37%) developed ipsilateral recurrence. Among 84 high-risk ET patients, 35 (42%) developed ipsilateral recurrence. Of the studies included in the meta-analysis, major complications were reported in 21 ET studies among 36/800 (4.5%) patients and progression to RNU was reported in 191/1193 (16%) patients. Multifocality was reported in 307 of 757 (41%) patients treated with ET. Additional characteristics and outcome measures for both ET and SU are given in the online supplemental material.

Table 1. Descriptive characteristics of selected original publications of endourological treatment for upper urinary tract urothelial carcinoma, identified by systematic literature review.

Study	Country	Inclusion years	Sample size	Follow-up (months) (mean/median)	Individuals with ipsilateral recurrence	Quality
Baboudjian et al17	Spain	2015–2021	60	-/36	18	Good
Basile et al†18	Spain	2015–2021	109	-/47	30	Good
Boorjian et al19	USA	1993–2003	38	37/-	26	Medium
Bozzini et al20	Italy	2015–2016	47	12/-	9	Low
Chen and Bagley21	USA	1989–1998	23	35/-	15	Medium
Cornu et al22	France	2003–2007	35	-/30	21	Good
Daneshmand et al23	USA	1987–2001	26	-/31	23	Low
Defidio et al24	Italy	2005–2018	101	29/-	50	Good
Elliott et al25	USA	1983–1994	44	60/-	17	Good
Fajkovic et al26	Austria	1996–2012	20	20/-	5	Good
Fanari et al*27	Italy	1996–2009	105	72	32	Low
Figaroa et al28	The Netherlands	2010–2020	71	-/50	59	Good
Grasso et al29	USA	1996–2011	66	52/-	51	Medium
Hoffman et al†30	Israel	2000–2010	25	-/26	9	Good
Hsieh et al†31	Taiwan	2012–2018	34	25/-	15	Medium
Iborra et al32	Spain	1984–2001	23	85/-	7	Low
Jimie et al33	UK	2016–2020	30	-/30	14	Medium
Johnson et al34	USA	1995–2005	35	32/-	24	Medium
Keeley et al35	USA	1985–1995	38	-/26	8	Low
Lucas et al36	USA	1990–2005	27	-/54	13	Low
Matsuoka et al37	Japan	–	27	-/-	10	Low
Mugiya et al38	Japan	1997–2002	6	13/-	2	Low
Mugiya et al39	Japan	1997–2004	7	32	5	Low
Musi et al40	Italy	2012–2016	42	-/26	11	Medium
Proietti et al41	Italy	2021–2022	28	6/-	5	Medium
Raymundo et al42	USA	1999–2009	21	18/-	4	Good
Rouprêt et al43	France	1990–2004	27	-/51	4	Medium
Sanguedolce et al44	Spain, Italy	2015–2019	47	-/24	13	Good
Scotland et al46	USA	1994–2011	63	-/44	40	Good
Scotland et al45	USA	1994–2011	168	66/-	125	Good
Shenhar et al†47	Israel	2000–2018	24	60/-	22	Good
Shvero et al48	Israel	2014–2019	59	/22	46	Good
Tada et al†49	Japan	2003–2007	15	/26	5	Medium
Thompson et al†50	USA	1983–2004	83	-/55	46	Medium
Thompson et al51	USA	1988–2003	22	/59	11	Medium
Villa et al52	France	2003–2015	92	52	70	Good
Wen et al53	China	2013–2017	32	-/-	7	Low
Yamada et al54	Japan	2000–2002	5	10/-	0	Low
Yamada et al55	Japan	2000–2003	7	68/78	1	Medium
Yoshida et al†*56	Japan	2017–2019	10	24	2	Low

^*Own calculation of ipsilateral recurrence number based on Kaplan-Meier curves or percentages.

^†Time-to-event data used in combined Kaplan-Meier curve.

NR, not reported.

The characteristics of SU studies are summarised in table 2. Sample sizes varied between three and 161 participants. Out of the 28 SU studies included in the meta-analysis, seven studies reported use of URS in the surveillance protocol. Four SU studies included high-risk disease only,^65,70,75,81 with ipsilateral recurrence reported in 20/139 (14%) patients. One SU study with 10 low-risk patients⁷⁴ reported one ipsilateral recurrence. Major complications (Clavien-Dindo >2) were reported in 12/28 of the SU studies among 39/599 (6.5%) patients and progression to RNU was reported in 58/664 (8.7%) patients. Only 53 of 693 (7.6%) patients subjected to SU had multifocal tumours. Additional characteristics and outcome measures for both ET and SU are given in the online supplemental material.

Table 2. Descriptive characteristics of selected original publications of segmental ureterectomy for upper urinary tract urothelial carcinoma, identified by systematic literature review.

Study	Country	Inclusion years	Sample size	Follow-up (months) (mean/median)	Individuals with ipsilateral recurrence	Quality
Abrate et al58	Italy	2003–2013	84	-/23	8	Good
Abrate et al57	Italy	2003–2013	26	-/27	4	Good
Abrate et al59	Italy	2003–2013	27	-/23	0	Medium
Anderstrom et al60	Sweden	1963–1978	21	-/83	1	Low
Bin et al61	USA	2000–2010	17	6/-	2	Medium
Campi et al62	Italy, France	2015–2018	15	-/21	3	Good
Chen et al63	Taiwan	1993–2003	12	-/23	7	Medium
Dalpiaz et al*64	Austria	1984–2011	49	-/51	2	Good
Ditonno et al65	Multinational	2015–2022	58	30/-	7	Good
Dragicevic et al66	Serbia	–	9	67/-	1	Low
Fang et al67	China	2003–2016	53	51/-	8	Good
Fukushima et al68	Japan	1977–	43	-/50	3	Good
Giannarini et al69	Switzerland	1974–2004	19	-/58	2	Medium
Huang et al*70	China	2011–2016	24	22/-	4	Good
Hung et al71	Taiwan	2004–2010	35	48/-	5	Medium
Johnson and Babaian72	USA	1971–1979	6	44/-	1	Low
Kato et al73	Japan	–	12	49/-	2	Good
Koterazawa et al74	Japan	2017–2022	10	-/33	1	Good
Lee et al76	Taiwan	2006–2014	19	28/-	2	Good
Lee et al75	South Korea	2011–2020	46	40/-	9	Good
Liu et al77	China	2014–2019	54	-/-	11	Medium
Masson-Lecomte et al*78	France	2010–2020	155	60/-	28	Good
McClain et al79	USA	2008–2012	5	33/-	0	Medium
Murayama et al80	Japan	1982–1992	3	30/-	1	Low
Nasu et al81	Japan	2008–	11	35/35	0	Medium
Palagonia et al82	Belgium	2013–2017	11	-/26	1	Medium
Rouprêt et al83	France USA	2001–2005	6	32/-	1	Medium
Saini et al84	USA	2008–2021	15	-/41	2	Medium
Seisen et al*†2	Multinational	2004–2013	134	60/-	19	Medium
Silberstein et al85	USA	1994–2009	33	-/50	3	Good
Simonato et al86	Italy	–	73	-/87	2	Medium
Wang et al87	Taiwan	–	161	-/44	25	Good
Jia et al88	China	2000–2014	40	64/-	3	Good
Zoretic and Gonzales89	USA	1961–1968	16	56/-	1	Low

^*Time-to-event data used in combined Kaplan-Meier curve.

^†Own calculation of ipsilateral recurrence number from the percentages.

NR, not reported.

Time-to-event-based meta-analysis

We were able to retrieve time-to-event information from seven ET studies. After ET, time-to-event derived ipsilateral recurrence probabilities increased with longer follow-up, from 0.36 after 1 year to 0.46 after 2 years, 0.49 after 3 years and 0.53 after 4 years. The degree of heterogeneity also rose over time (increasing I² values), indicating greater variability in ipsilateral recurrence estimates at later follow-up. Not all studies contributed data to every time point, so the number of included studies differed across the 1-year, 2-year, 3-year and 4-year pooled estimates (figure 1). Time-to-event data were accessible in four SU studies. After SU, time-to-event-derived ipsilateral recurrence probabilities were increasing from 0.06 after 1 year to 0.12 after 2 years, 0.17 after 3 years and 0.20 after 4 years. Heterogeneity was generally modest, suggesting similar ipsilateral recurrence estimates across studies at each follow-up time. As with ET, not all studies reported outcomes at every time point, and the pooled estimates after each year were based on the subset of studies with available follow-up data (figure 2). The extracted individual patient data set for ET and SU studies combined in common survival analyses and a compact validation table for each reconstructed study are presented in the online supplemental material.

Figure 1. Forest plot of year-wise ipsilateral recurrence rates after endourological treatment for upper tract urothelial carcinoma. REML, restricted maximum likelihood.

Figure 2. Forest plot of year-wise ipsilateral recurrence rates after segmental ureterectomy for upper tract urothelial carcinoma. REML, restricted maximum likelihood.

Meta-analysis of crude ipsilateral recurrence proportions

The forest plot displays a crude pooled ipsilateral recurrence proportion of 0.48 (95% CI 0.39 to 0.57) after ET (figure 3), with high heterogeneity among studies (I²=91%; p<0.001, based on Cochran’s Q test), based on 774 ipsilateral recurrences among 1481 patients. After SU, the corresponding proportion was 0.11 (95% CI 0.08 to 0.13), with low heterogeneity (I²=29%; p=0.20, based on Cochran’s Q test) (figure 4), based on 151 ipsilateral recurrences among 1175 patients. Individual study proportions ranged from zero⁸¹ to 0.33,⁸⁰ although most estimates clustered between 0.03 and 0.20. Notably, larger studies such as Masson-Lecomte et al (n=155)⁷⁸ and Wang et al (n=161)⁸⁷ demonstrated proportions of 0.18 and 0.16, respectively, and with narrower CIs. No significant correlation between crude pooled ipsilateral recurrence proportions and mean/median follow-up after ET or SU was observed (online supplemental material).

Figure 3. Forest plot of crude ipsilateral recurrence rates after endourological treatment for upper tract urothelial carcinoma. REML, restricted maximum likelihood.

Figure 4. Forest plot of crude ipsilateral recurrence rates after segmental ureterectomy for upper tract urothelial carcinoma. REML, restricted maximum likelihood.

Publication bias

The assessment of publication bias using funnel plots for both studies for ET and SU revealed largely symmetrical distributions around the pooled effect size, with no substantial asymmetry observed in the plots for both the time-to-event meta-analysis and crude proportions meta-analysis (online supplemental material). This visual inspection suggests an absence of significant publication bias in the included studies for any of the treatment modalities. Funnel plots for ET after the fourth year and for SU after the second, third and fourth years each included only three studies and were therefore not interpreted.⁹⁰ The funnel plots for ET should be interpreted with caution, as substantial between-study heterogeneity limits the reliability of visual inferences regarding small-study effects and publication bias. The Egger test was insignificant for all the meta-analyses, indicating no evidence of publication bias.

Meta-regression analysis

Multivariable meta-regression analysis was performed to explore potential sources of heterogeneity among ET studies. Reporting of previous bladder cancer was significantly associated with the outcome (β=0.41, 95% CI 0.07 to 0.74, p=0.016) explaining 21% of the between-study variance. In contrast, publication year after 2015 and follow-up duration longer than 24 months were not significantly associated with the outcome. Other factors potentially contributing to heterogeneity, such as laser type used, use of URS during follow-up, elective versus imperative indication or tumour characteristics were not reported consistently, and therefore could not be included in the meta-regression model. In the online supplemental material, we detail the rationale for the selection of covariates for the meta-regression and present the corresponding results.

Subgroup analyses

The forest plots based on the subgroup analyses are presented in the online supplemental material. Subgroup analyses of ET demonstrated notable differences in outcomes based on clinical context. For cases involving middle or distal ureteric locations, the pooled proportion of the outcome of interest was 0.48 (95% CI 0.34 to 0.62), with moderate heterogeneity (I²=45%; p=0.09, based on Cochran’s Q test). In contrast, for elective procedures, the pooled proportion was higher at 0.53 (95% CI 0.36 to 0.70), although with substantial heterogeneity (I²=87%; p<0.001, based on Cochran’s Q test) (online supplemental material). The observed variation between studies was more pronounced in the elective subgroup, as reflected by the higher heterogeneity estimate.

For SU patients with middle or distal ureteric tumours, the pooled proportion of the outcome was 0.10 (95% CI 0.7 to 0.13), with moderate heterogeneity (I²=31%; p=0.15, based on Cochran’s Q test) (online supplemental material). Similarly, in the elective intervention subgroup of SU, the pooled proportion was also 0.10 (95% CI 0.6 to 0.14), with moderate heterogeneity (I²=44%; p=0.07, based on Cochran’s Q test). Moderate heterogeneity suggests reasonable consistency in results across studies within each subgroup (online supplemental material).

Sensitivity analyses

The 30 ET studies that reported use of URS in the surveillance protocols had a pooled proportion ipsilateral recurrence of 0.50 (95% CI 0.40 to 0.59), while three ET studies with no report of URS during follow-up had a corresponding proportion of 0.28 (95% CI 0.20 to 0.35). Seven SU studies reporting use of URS in the surveillance protocols had a pooled proportion ipsilateral recurrence of 0.12 (95% CI 0.08 to 0.16) compared with 0.10 (95% CI 0.07 to 0.14) in the 21 SU studies where use of URS during follow-up was not reported (online supplemental material). The study found that ‘good’ quality studies had a pooled proportion of 0.58 (95% CI 0.42 to 0.73) ipsilateral recurrence, while ‘medium’-quality and ‘low’-quality studies had a proportion of 0.45 (95% CI 0.32 to 0.59) and 0.37 (95% CI 0.23 to 0.53), respectively for the ET studies (online supplemental material). For SU studies, ‘good’-quality studies had a pooled outcome proportion of 0.12 (95% CI 0.09 to 0.15), ‘medium’-quality studies reported 0.10 (95% CI 0.04 to 0.16) and ‘low’-quality studies had a pooled outcome proportion of 0.7 (95% CI 0.00 to 0.17).

Discussion

In this systematic review and meta-analysis, seven studies with available time-to-event data corresponded to an estimated 4-year ipsilateral recurrence rate of 0.53 (95% CI 0.25 to 0.80) after ET for UTUC. The crude pooled ipsilateral recurrence proportion after ET in 33 available studies was 0.48 (95% CI 0.39 to 0.57); however, the reliability of these findings is substantially limited by a marked heterogeneity across ET studies. Similar results were observed in a subgroup analysis restricted to tumours in the distal or middle ureter (48%). Four studies with available time-to-event data corresponded to an estimated 4-year ipsilateral recurrence rate of 0.20 (95% CI 0.11 to 0.30) after SU for UTUC, and the crude pooled ipsilateral recurrence proportion was 0.11 (95% CI 0.08 to 0.13) with modest heterogeneity. A subgroup analysis in SU studies reporting ipsilateral recurrence separately for patients with tumours in the distal or middle ureter showed a similar crude pooled ipsilateral recurrence proportion (10%).

Previous meta-analyses have suggested that KSS results in similar cancer-specific survival (CSS) as RNU² including a meta-analysis investigating CSS separately after ET and SU.^91,92 Another recent meta-analysis also reported similar CSS, overall survival, recurrence-free survival and intravesical recurrence-free survival outcomes for KSS compared with RNU.⁹³ The common strategy to merge different KSS techniques for comparison with RNU might have advantages for this rare tumour form, yet this might have contributed to a paucity of reliable data on separate outcomes for ET and SU.⁹⁴ Specifically, since CSS generally is acceptable after KSS, ipsilateral recurrence-free survival between ET and SU seems to be the most relevant survival outcome to be discussed with patients in a shared decision-making when considering KSS.¹

There are several selection mechanisms that could influence ipsilateral recurrence after KSS. In a subgroup analysis of studies where ipsilateral recurrences were reported separately for patients with elective indications, similar crude pooled ipsilateral recurrence proportions following both ET and SU were observed, although with high heterogeneity. Furthermore, a second subgroup analysis of seven ET studies (n=102) reporting ipsilateral recurrence outcomes separated by tumour location^{19,31,33,35,37,49,55} showed similar crude pooled ipsilateral recurrence proportions for tumours in the distal or middle ureter compared with the main analysis. Interestingly, the heterogeneity between these seven studies was lower compared with the main analysis. In the setting of SU, it is not known whether SU with termino-terminal ureteric anastomosis in more proximal ureteric tumour locations had similar ipsilateral recurrence rates compared with SU with neocystostomy.⁵⁸

The current meta-analysis also included patients with high-risk tumours since SU is an established treatment option also in this context¹ and due to the fact that some ET series include high-risk tumours as experimental treatment or indicated for imperative reasons. Nevertheless, this allows for imbalances and selection mechanisms in the study populations treated with ET and SU that might affect ipsilateral recurrence rates. Such differences could not be adjusted for, as outcomes were not reported for low-risk and high-risk tumours separately, aside from a few exceptions.

In the studies included in the current meta-analysis, the proportion of high-grade tumours was 228 of 1104 (21%) in patients treated with ET and 514 of 864 (59%) in patients subjected to SU. Multifocality was reported in 307 of 757 (41%) patients treated with ET, whereas only 53 of 693 (7.6%) patients subjected to SU had multifocal tumours. The discrepancy may partly reflect differences in assessing multifocality by endoscopy versus CT urography, that is, not all patients subjected to SU have been subjected to a diagnostic ureteroscopy. However, multifocality has been associated with ipsilateral recurrence after KSS.⁷⁸ The imbalance in the proportions of high-grade and multifocal tumours and between ET and SU limits the possibility of comparing ipsilateral recurrence estimates in this meta-analysis.

A potential explanatory factor for the differences in ipsilateral recurrence estimates observed after KSS is how the recurrences were defined. In sensitivity analyses, the use of URS in the surveillance protocol after ET and SU gave similar crude pooled ipsilateral recurrence proportions (50% and 12%) compared with the meta-analyses (48% and 11%). For ET studies, the discrepancy between pooled recurrence proportions in studies that reported use of URS during follow-up (50%) and studies with no report of using URS during follow-up (28%) suggests that different follow-up protocols and methods of detecting recurrences may systematically bias recurrence estimates.

This study is also limited by a substantial heterogeneity, particularly among the ET studies; therefore, the pooled recurrence proportions of ET should be interpreted as descriptive summaries rather than robust, generalisable estimates. The inconsistent reporting of variables such as tumour grade, localisation, multifocality, imperative indication for surgery, follow-up protocols, adjuvant instillations and laser type limited the ability of the meta-regression to explain more than a small proportion of the heterogeneity observed among ET studies. The finding that ‘reporting of previous bladder cancer’ was associated with the outcome likely reflects a marker of study quality.

Direct comparisons between ET and SU are not feasible due to only one available head-to-head study.² Additionally, with only one study with another comparator than RNU (where ipsilateral recurrence is not possible), that is, ET with and without photodynamic diagnosis,⁵⁶ formal risk of bias assessment (ROBINS-I) is not suitable and consequently the JBI scale was applied. Similarly, given that only 11 studies reported numbers at risk and information about censoring, time-to-event data on ipsilateral recurrence was restricted to this subset of studies. The included studies were predominantly retrospective single-arm series and therefore susceptible to several sources of bias, which limits this meta-analysis. Selection bias may have occurred because patients undergoing SU were selected based on distal tumour location or single tumour status. Residual confounding by factors such as tumour stage, grade, multifocality and treatment indication could not be adequately accounted for in most studies. Outcome ascertainment may have varied because recurrence definitions and follow-up completeness were inconsistent across studies. Moreover, differences in surveillance protocols, particularly the use of routine URS, may have introduced detection bias, with more intensive surveillance increasing the likelihood of identifying ipsilateral recurrence. Consequently, the pooled estimates should be interpreted with caution.

Excluding publications in different languages and seven confirmed or assumed duplicate studies is a limitation. All duplicate publications were excluded except for two ET studies, sharing approximately 50% of their data, which were retained because they reported distinct outcomes.^30,47 Larger Surveillance, Epidemiology and End Results (SEER)-based KSS series for UTUC⁹⁵ did not meet the Population, Exposure, Comparator, Outcome criteria because ipsilateral recurrence data were unavailable. Consequently, the analysis is largely limited to smaller, methodologically heterogeneous studies.

In the absence of previous meta-analyses investigating ipsilateral recurrence after KSS from both ET and SU, this study adds information on an outcome relevant to counselling patients with UTUC in the middle and distal ureter. The findings also provide a foundation for future cost-effectiveness studies as well as underpinning the need for a randomised trial comparing ET and SU and developing predictive tools for ipsilateral recurrence-free survival. The proportion of high-grade (Clavien-Dindo >2) complications reported was slightly higher after SU compared with after ET (7% and 5%, respectively). This must be taken into account together with the need for second-look ureteroscopy in the algorithm for ET and higher proportion of progression to RNU after ET when counselling patients. Furthermore, in the wake of expanding indications for minimally invasive robotic-assisted surgery, robotic-assisted SU is emerging as a putative comparator to ET.⁸² Additionally, considerations should include the morbidity, costs and quality-of-life implications associated with the more invasive nature of SU versus the potential need for repeated ET procedures.

All future KSS studies are encouraged to report ipsilateral recurrence as a separate outcome together with numbers at risk with time intervals. Furthermore, information on elective or imperative indication for surgery, previous bladder cancer, surgical approach (laser type, ureteral sheath, percutaneous), tumour grade, localisation, multifocality, EAU risk category, follow-up protocols (including second-look URS) and adjuvant instillations are encouraged in future KSS studies. Given that cancer-specific survival outcomes are comparable across different KSS strategies, differences in risk of ipsilateral upper tract recurrence among patients suitable for both ET and SU represent a key endpoint in future KSS studies to improve the scientific basis for patient counselling.

Conclusions

In a systematic review and meta-analysis of ipsilateral recurrence after KSS for UTUC using available single-arm studies, we found homogenous outcomes following SU, whereas substantial heterogeneity in the literature made robust quantitative estimation of outcomes after ET impossible. Future studies should emphasise standardised reporting on KSS outcomes and associated variables, as ipsilateral recurrence is a clinically important endpoint when counselling patients with UTUC. Cost-effectiveness analyses and randomised comparisons between treatment modalities for patients with ureteric UTUC also remain unmet needs.

The funding sources had no role in the study design, data analyses, interpretation of the results or writing of the manuscript.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.