Robotic segmental ureterectomy for upper tract urothelial carcinoma: comparable short-term oncological outcomes and improved postoperative recovery compared to open approach

Abstract

Background

Segmental ureterectomy (SU) is an established kidney-sparing strategy for upper tract urothelial carcinoma (UTUC), yet few studies have compared the standard open approach to the emerging robotic modality.

Methods

We performed a clinical outcome and safety evaluation of robotic-assisted SU compared to the traditional open approach. In this retrospective study of 53 patients (40 open, 13 robotic) treated between 2008 and 2025, we assessed perioperative, functional, and oncological parameters.

Results

Regarding safety endpoints, robotic SU demonstrated a favorable profile with no significant difference in postoperative complication rates compared to open surgery (15.4% vs. 26.5%, p = 0.7). While the robotic approach involved longer operative times (351 vs. 285 min, p = 0.01), it resulted in a significantly shorter median hospital stay (4 vs. 12 days, p < 0.001). Utilizing Firth’s penalized Cox regression to account for sparse events, no significant differences were observed in overall survival (HR 1.25, 95% CI 0.13–6.17, p = 0.815) or disease-free survival (HR 1.22, 95% CI 0.31–3.63, p = 0.75). Similarly, secondary endpoints including intraluminal recurrence-free (p = 0.188), intravesical recurrence-free (p = 0.581), and metastasis-free survival (p = 0.997) did not differ significantly. Postoperatively, functional evaluation revealed a trend toward improved renal function in the robotic cohort (+ 22.94 vs. +7.46 mL/min/1.73 m², p = 0.08).

Conclusions

Robotic SU appears feasible and was not associated with an apparent compromise in short-term oncological outcomes during the available follow-up period, while offering the benefits of minimally invasive recovery. Given the limited follow-up and small sample size of the robotic cohort, these results provide preliminary evidence of safety. Further multi-institutional, large-scale studies with extended follow-up are required to establish definitive long-term non-inferiority.

Key points

UTUC, upper tract urothelial carcinoma; SU, segmental ureterectomy; RNU, radical nephroureterectomy.

Background

Upper tract urothelial carcinomas (UTUC) involve ureteral, renal pelvis, and renal calyces lesions, and are considered relatively rare [1]. Unlike bladder cancer, where only 15% to 20% of patients present with invasive tumors, roughly half of all UTUC patients have invasive disease at the time of initial diagnosis, making prompt and effective intervention essential [2, 3].

Historically, radical nephroureterectomy (RNU) with bladder cuff excision has been the gold standard, particularly for high-risk disease [4–6]. However, in the last decade, the therapeutic landscape has shifted toward kidney-sparing surgery (KSS), including segmental ureterectomy (SU). According to the current guidelines, SU is the preferred approach for low-risk disease to mitigate the morbidity and chronic kidney disease associated with RNU [4, 7]. In high-risk patients, SU is increasingly utilized for imperative indications (e.g., solitary kidney, bilateral disease, renal insufficiency) or in selective cases where nephron preservation is prioritized [4, 8]. Recent propensity score-matched analyses indicate that in carefully selected patients, SU offers cancer-specific and overall survival comparable to RNU, with the distinct advantage of preserving estimated glomerular filtration rate (eGFR) [9–14]. Preservation of renal function is vital, as it maintains patient eligibility for adjuvant chemotherapy in locally advanced disease [15, 16].

Motivated by these functional advantages, our center utilizes a highly inclusive criteria for KSS: we offer SU to any patient presenting with localized UTUC amenable to segmental resection, irrespective of tumor location or grade, so strict oncological safety is upheld [4, 5, 7, 8].

While the oncological safety of SU compared to RNU is well-documented, the optimal surgical approach to SU remains a subject of investigation. Several groups have reported on robotic SU; however, these publications are predominantly restricted to small patient samples [17–21], making it difficult to draw firm conclusions about its true comparative value. As robotic surgeries become more widespread, robust data is needed to verify if robotic SU can mirror the proven outcomes of the standard open technique. Accordingly, this study aims to detail our center’s experience by evaluating and comparing the oncological, functional, and perioperative results of robotic SU versus the traditional open method in patients with UTUC.

Materials and methods

Patient selection

We conducted a retrospective cohort study at a tertiary referral center of adult patients who underwent SU between 2008 and 2025. The patients were identified from institutional electronic health records using a big data software (MDClone, ADAMS Platform, Israel) and verified through chart review.

Inclusion criteria: age ≥ 18 years, pathologically confirmed UTUC (HG\LG) based on diagnostic ureteroscopy, eligible for kidney preservation treatment (localized disease, any tumor size suitable for SU) or imperative indications (e.g., a solitary kidney, acute or chronic renal failure). Patient selection is summarized in Fig. 1.

Fig. 1.

Patient selection flowchart. SU: segmental ureterectomy; RCC: renal cell carcinoma; AML: angiomyolipoma. * Benign causes include urethral stricture, non-functioning kidney, melacoplakia, chronic pyelonephritis

Outcomes

Perioperative outcomes included operative time, estimated blood loss (EBL), blood transfusion, complications (Clavien-Dindo classification), and length of hospital stay.

Functional outcome was evaluated independently for the robotic and open cohorts, defined as the median change in estimated glomerular filtration rate (ΔeGFR) for each patient. The individual ΔeGFR was calculated as the difference between the baseline eGFR (defined as the measurement obtained within 7 days before surgery) and the postoperative eGFR (the most recent measurement available between 3 weeks and 12 months postoperatively).

Primary oncologic endpoints were overall survival (OS) and disease-free survival (DFS). OS was defined as time to death from any cause. DFS was defined as time to first urothelial recurrence (ipsilateral\contralateral UTUC, intravesical recurrence, or distant metastasis) or UTUC-specific death, whichever occurred first. Deaths from other causes and cases with unknown cause were censored at death. Secondary oncologic endpoints included intraluminal upper-tract recurrence (ILRFS) defined as time to first ipsi- or contralateral ureteral recurrence, intravesical recurrence-free survival (IVRFS), defined as time to first bladder recurrence, and metastasis-free survival (MFS), defined as time to first distant metastasis. Patients without an event were censored at last follow-up.

Surgical approach

All robotic procedures were performed by two fellowship-trained surgeons with extensive experience in upper tract reconstruction, open and robotic oncology, ensuring standardization of the surgical technique. The open cohort was comprised of cases performed by five senior urologic-oncology members: notably, 72.5% (29/40) of the open procedures were performed by the same two surgeons responsible for the robotic cohort.

Patient positioning and port placement

Patients were placed in low lithotomy or supine position for distal tumors with operating table adjusted to a maximal Trendelenburg position, or lateral decubitus for mid-ureteral tumors. Pneumoperitoneum was established at 15 mmHg via a Veress needle. A transperitoneal robotic approach was utilized using a standard four-arm configuration (da Vinci^® Xi or Si, Surgical System [Intuitive Surgical, Inc., Sunnyvale, CA, USA]) consisting of three 8-mm robotic trocars and one 12-mm assistant port with AirSeal^® System (CONMED Corporation, Utica, NY, USA).

Dissection and tumor localization

Following adhesiolysis and bowel mobilization, the ureter was identified and encircled. For strictures or tumors without clear external margins, we utilized a hybrid approach: a flexible ureteroscope was advanced retrograde to identify the lesion boundaries under simultaneous robotic and endoscopic visualization. The involved ureteral segment was excised after applying Hem-o-lock clips proximally and distally to the planned resection site, and intraoperative frozen sections confirmed the absence of residual cancer [22].

Reconstruction was dictated by the defect’s size and anatomical location:

Ureteroureterostomy (End-to-End)

For proximal or mid-ureteral defects (n = 3, 23%), the ureteral ends were spatulated. Reconstruction was performed over a double-J stent (6–7 Fr) using either a continuous running suture (4 − 0 barbed or 4 − 0 PDS) or interrupted absorbable sutures (4 − 0 Vicryl) to ensure a watertight, tension-free anastomosis. In cases requiring additional length, proximal and distal mobilization were used.

Distal ureterectomy with ureteroneocystostomy

For distal tumors, the ureter was excised with a bladder cuff (n = 10, 76.9%). The bladder defect was closed in two layers. The ureter was reimplanted into a new cystotomy at the bladder dome over a double-J stent (6–7 Fr) using either a continuous running suture (4 − 0 barbed or 4 − 0 PDS) or interrupted absorbable sutures (4 − 0 Vicryl).

Psoas hitch

In cases with significant defects requiring tension relief, the bladder was extensively mobilized by dividing the median umbilical ligaments and dissecting the retropubic space. The posterior bladder wall was anchored to the ipsilateral psoas tendon using interrupted 2 − 0 absorbable sutures. The anastomosis was completed with a detrusorrhaphy to provide an anti-reflux mechanism.

All repairs were validated for watertightness via a 200–300 mL retrograde bladder fill, followed by closed suction drain (Jackson-Pratt) placement in the pelvis or retroperitoneum and standard fascial and skin closure.

Statistical analysis

We reported normally distributed continuous data as means with standard deviations (SD), whereas non-parametric data were expressed as medians with interquartile ranges (IQRs). Categorical variables were summarized as proportions. The independent sample t-test for normally distributed data was used to compare continuous variables between groups. The Mann-Whitney U test was used for abnormally distributed data. Categorical variables were evaluated using the Pearson chi-squared test or Fisher’s exact test, as appropriate. The 95% confidence intervals (CI) for the difference between medians were calculated using the Hodges-Lehmann estimator.

Missing data was handled using complete-case analysis. Variables with missing data constituted less than 3.8% of the overall dataset, and the specific denominators for these variables are explicitly denoted in the baseline characteristics table to ensure transparency.

Time-to-event outcomes (OS, DFS, ILRFS, IVRFS, and MFS) were analyzed using the Kaplan-Meier method, with differences between survival curves assessed via the log-rank test. To estimate the association between the surgical approach and survival outcomes, Cox proportional hazards regression models were utilized. The proportional hazards assumption was evaluated visually for survival endpoints using Schoenfeld (partial) residuals. Where event frequencies permitted (OS and DFS), this assumption was formally verified by incorporating a time-dependent covariate interaction term into the models. Given the small sample size in the robotic cohort (n = 13) and the sparse number of oncological events for secondary endpoints, which limits the stability of standard Cox regression and formal time-dependent testing, we performed a secondary sensitivity analysis utilizing Firth’s penalized Cox regression. This method is specifically designed to reduce small-sample bias and handle rare events or complete separation in survival data.

All tests were two-sided (p < 0.05). While primary descriptive and comparative statistics were performed using IBM SPSS Statistics (v29.1), all Firth’s penalized regression models were executed using the “readxl”, “survival” and “coxphf” packages in R (version 4.5.3).

Ethics

The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines and was approved by the institutional review board of Sheba Medical Center (SMC-4146; last update 4 June 2025).

Results

The analysis included 53 patients with a mean age 73 ± 10.3 years of them 68% male. The open group comprised 40 patients, and the robotic 13 patients. Demographics, medical background and tumor characteristics were similar, including baseline median eGFR (open 48.8 vs. robotic 37.89 mL/min/1.73 m², p = 0.71). Patient characteristics are summarized in Table 1.

Table 1.

Baseline characteristics of patients underwent open versus robotic segmental ureterectomy

Parameter	Open n = 40	Robotic n = 13	P value	95% CI of the Difference
Age at surgery, mean (± SD)	71 + 10.9	70.9 + 8.73	0.98	-6.6-6.79
Gender, male, number (%)	26 (65)	10 (76.9)	0.52
Distal ureteral tumor, number (%)	23 (58)	10 (76.9)	0.33
Tumor size, median mm, (IQR)	25 (18.8–31.3)	32.5 (27.8–47.5)	0.22	-23.3-6.1
Laterality, left, number (%)	25 (67.5)	11 (84.6)	0.3
Medical Background, (n = 51) *, number (%)
DM	6 (15.8)	5 (38.5)	0.12
HTN	21 (55.3)	5 (38.5)	0.35
IHD	8 (21)	0 (0)	0.09
COPD	4 (10.5)	0 (0)	0.56
Pre-operation parameters**
Creatinine, mg/mL, median (IQR)	1.2 (1-1.7)	1.39 (0.87–1.8)	0.98	-0.28-0.29
eGFR (mL/min/1.73m2), median (IQR)	48.72 (29.4–57.1)	37.89 (28.7–77.1)	0.71	-11.46-16.5
Hemoglobin, mg\mL, median (IQR)	12.39 (10.9–13.5)	12.56 (11.3–15.6)	0.22	-2.3-0.58
Number of URS performed before SU, mean (± SD)	1.66 + 2.3	2.76 + 1.8	0.17	-2.7-0.5
Neoadjuvant chemotherapy, n (%)	0 (0%)	0 (0%)	1
Bladder UC (n = 51) *, number (%)
History of bladder UC before SU	15/251 (60)	2/31 (66.7)	1
NMIBC needed intravesical treatments before SU (intermediate and high-risk groups) ^	13/15 (86.7)	1/2 (50)	0.07

mm Millimeter, DM Diabetes mellitus, HTN Hypertension, IHD Ischemic heart disease, COPD Chronic obstructive pulmonary disease, eGFR Estimated glomerular filtration rate, URS Ureteroscopy, SU Segmental ureterectomy, POD Post operation day, UC Urothelial cancer, NMIBC Non-muscle invasive bladder cancer

*Where denominators vary due to missing data, counts (n) and percentages reflect only patients with available data for that specific variable

** Measured within 7 days prior to SU

¹ Represents the proportion of patients with a history of bladder UC whose diagnosis occurred prior to SU

^Intravesical treatments include BCG, Mitomycin, Synergo®

As expected, the total follow-up time was longer among patients who were treated in the open approach 45.5 months (IQR 17.5-115.5) vs. 11 in the robotic group (IQR 6–27), p < 0.001 (95% CI 26.3–64.3).

Operative parameters significantly differed by operating time, which was longer with robotics (351 vs. 285 min, p = 0.01), while length of hospital stay was shorter in the robotic group (4 vs. 12 days, p < 0.001). Postoperative complication rates were similar (26.5% open vs. 15.4% robotic, p = 0.7). Long-term renal function (assessed at 6–12 months) increased in the robotic cohort (37.89 mL/min/1.73 m² to 60.83 mL/min/1.73 m², p = 0.055), with a trend toward greater ΔeGFR improvement (+ 22.94 vs. +7.46 mL/min/1.73 m², p = 0.08). Comparison between operational and functional parameters is shown in Table 2.

Table 2.

Operative and functional outcomes between open and robotic SU

	Open (n = 40)	Robotic (n = 13)	P value	95% CI
Intraoperative complications, number (%)	0 (0)	0 (0)	1
EBL, mL, median (IQR)	100 (50–300)	150 (50–200)	0.24	-100 to 150
Operative time, min, median (IQR)	285 (200–330)	351 (290–365)	0.01	-160 to -20
Post-operative complications, number (%)
Any grade of Clavien-Dindo classification (I-V)	9 (26.5)	2 (15.4)	0.7
Clavien-Dindo ≥ 3	1 (2.9)	2 (15.4)	0.2
Need for blood transfusion^, n (%)	2 (6)	0 (0)	0.6
Hospital stays, days median (IQR)	12 (9.5–12)	4 (2–7)	< 0.001	2 to 7
△eGFR, mL/min/1.73 m2 median (IQR)	7.46 (-10-21)	22.94 (19.2–41)	0.08	-36.8 to 6.4
The need for subsequent RNU, n (%)	5 (12.5)	1 (7.7)	1

EBL Estimated blood loss, mL Milliliter, min Minute, IQR Interquartile range, RNU Radical nephroureterectomy

^ Blood transfusion was classified as Clavien-Dindo grade II

△eGFR represents the median of individual patient-level changes, calculated as eGFR_postOP minus eGFR_baseline

eGFR_postOP = postoperative value at ≥3 weeks or the latest measurement within 6-12 months postoperatively

eGFR_baseline = preoperative value within 7 days before surgery

Postoperative pathological characteristics were comparable between the open and robotic cohorts, Table 3. No significant differences were observed in the distribution of pathological T-stage (p = 0.69) or tumor grade (p = 1). High-grade disease was present in 63.2% of the open cohort and 61.5% of the robotic cohort. Similarly, nodal status (pN) was balanced between groups (p = 0.47), with positive lymph nodes identified in one robotic case (pT3N1M0) and two open cases. Ten patients (7 open, 3 robotic) demonstrated no evidence of residual malignancy (pT0) following initial endoscopic resection. Surgical margins were negative in all robotic cases (13/13, 100%), while positive margins occurred in 10.8% of the open group (p = 0.56). One patient needed a subsequent RNU (compared to 5 patients in the open approach), within 19 months due to disease recurrence.

Table 3.

Postoperative pathological characteristics

Characteristic	Open Cohort n = 40 *	Robotic Cohort n = 13	p-value
Pathological T Stage, n (%) ** 0.69
pT0***	7/35 (20)	3 (23.1)
pTa / pT1	14/35 (40)	5 (38.5)
pT2	6/35 (17.1)	2 (15.4)
pT3	8/35 (22.8)	3 (23.1)
Tumor Grade, n (%) 1.0
Low Grade	9 (22.5)	2 (15.4)
High Grade	24 (60)	8 (61.5)
Lymph Node Status, n (%) 0.47
pN0 (Negative)	15/35 (42.8)	6 (46.2)
pN+ (Positive)	3/35 (8.6)	1 (7.7)
pNx (Not performed)	17/35 (48.6)	6 (46.2)
Surgical Margins, n (%) 0.56
Negative (R0)	35/39 (90)	13 (100)
Positive (R1/R2)	4/39 (10.2)	0 (0)

*Where denominators vary due to missing data, values are presented as n/N, with the denominator (N) denoting the number of patients with available data for that specific variable

**The analysis excluded cases with pT0 or missing pathological data

***The pT0 category is included solely to provide a complete descriptive overview of the cohort. pT0 cases were excluded from the formal comparative analyses

Primary oncological endpoints: Survival analysis using the Kaplan-Meier method demonstrated comparable oncological outcomes with no significant difference in OS (log-rank p = 0.927). While the open approach was associated with a median survival of 135 months, the median survival for the robotic group was not reached (mean estimate: 126.9 months, 95% CI: 96.4-157.4 for open vs. 34.2 months, 95% CI: 27.3–41). Standard Cox regression showed no association between the surgical approach and OS (HR 0.9, 95% CI 0.11–7.71, p = 0.927). Given the wide CIs, a post hoc sensitivity analysis using Firth’s penalized regression was performed, which confirmed these findings (HR 1.25, 95% CI 0.13–6.17, p = 0.815). DFS did not differ between the groups on Kaplan-Meier analysis (log-rank p = 0.898). In the open group, the median DFS was 55 months (95% CI 25-85.1), while the median DFS in the robotic group was not reached (mean estimate: 65.1 months, 95% 47.2–82.9 open vs. 30.8 months, 95% CI 17.9–43.6). Standard Cox regression demonstrated no significant association between surgical approach and DFS (HR 0.49, 95% CI 0.11–2.14, p = 0.343). Firth’s penalized regression was performed, which confirmed these findings (HR 1.22, 95% CI 0.31–3.63, p = 0.75).

The proportional hazards assumption was formally confirmed to be met for both primary endpoints, showing no significant interaction between surgical approach and time for OS (p = 0.748) or DFS (p = 0.946).

To address the potential bias introduced by patients with no residual malignancy in the surgical specimen, a sensitivity analysis was performed excluding 3 pT0 and 7 pT0 cases in robotic and open cohort, respectively. Utilizing Firth’s penalized Cox regression, the robotic approach continued to demonstrate comparable short-term outcomes to the open cohort for both OS (HR: 3.9, 95% CI: 0.33–32.94, p = 0.24) or DFS (HR: 1.71, 95% CI: 0.43–5.35, p = 0.41).

Kaplan-Meier curves are shown in Fig. 2.

Fig. 2.

Kaplan-Meier estimates (a) Overall Survival and (b) Disease-Free Survival. a OS was similar between groups (log-rank p = 0.927). The median survival of the open group was 135 months. Mean estimate: 126.9 months for open vs. 34.2 months for robotics. b DFS was similar between groups (log-rank p = 0.89). The median DFS of the open group was 55 months. Mean estimate: 65.1 months for open vs. 30.8 months for robotics

Secondary oncological endpoints showed no significant differences between the open and robotic cohorts. Specifically, ILRFS (p = 0.211), IVRFS (p = 0.803) and MFS (p = 0.754), were comparable across groups. The Firth’s regression confirmed that the surgical approach was not a significant predictor for intraluminal recurrence (HR 3.25, 95% CI 0.52–17.35, p = 0.188), intravesical recurrence (HR 1.75, 95% CI 0.17–10.15, p = 0.581), or metastasis (HR 1, 95% CI 0.1–4.7, p = 0.997).While median recurrence intervals were reached in the open group for intraluminal (103.8 months) and intravesical recurrence (130 months), medians were not reached in the robotic arm for any endpoint. Survival analysis endpoints are presented in Table 4.

Table 4.

Survival analysis of open vs. robotic segmental ureterectomy

Survival Endpoint	Events, n (%)		Log-Rank (p value)	Hazard Ratio (95% CI) *	Firth’s Regression (p value)	Hazard Ratio (95% CI)**
	Open n = 40	Robotic n = 13
Primary Endpoints
Overall Survival	11 (27.5)	1 (7.7)	0.927	0.9 (0.11–7.71)	0.815	1.25 (0.13–6.17)
Disease-Free Survival	22 (55)	3 (23.1)	0.898	0.49 (0.11–2.14)	0.750	1.22 (0.31–3.63)
Secondary Endpoints
Intraluminal Recurrence free survival	11 (27.5)	2 (15.4)	0.211	3.04 (0.49–18.81)	0.188	3.25 (0.52–17.35)
Intravesical Recurrence free survival	8 (20)	1 (7.7)	0.803	1.33 (0.14–2.43)	0.581	1.75 (0.17–10.15)
Metastasis-Free Survival 1	9 (23.7)	1 (8.3)	0.754	0.71 (0.09–5.92)	0.997	1 (0.1–4.7)

Hazard Ratios for all endpoints are calculated using the open approach as the reference category. For primary endpoints, both standard Cox and Firth’s penalized regression are provided for sensitivity. For secondary endpoints, Firth’s penalized Cox regression was utilized primarily to address the small sample size and absence of events in the robotic cohort, ensuring stable hazard ratio estimates

CI Confidence Interval

¹Metastasis-free survival data was available for 50 patients (Open: n = 38, Robotic: n = 12); percentages for this endpoint are calculated based on these available cases

* Derived from standard Cox regression

** Derived from Firth’s penalized Cox proportional hazards regression

Discussion

SU has become the nephron-sparing procedure of choice for low-risk UTUC over the past decade. By relying on intraoperative frozen sections to guarantee tumor-free margins, SU yields adequate tissue for accurate pathological grading and staging while successfully preserving the involved kidney [22]. Although RNU remains the reference standard for high-risk disease, accumulating real-world evidence supports SU as a reasonable alternative in selected high-risk patients, particularly when the functional consequences of nephron loss are significant [5–14, 23].

A critical technical challenge in robotic ureterectomy is accurately identifying the tumor’s margins without tactile feedback. Recent innovations, such as those by Alhusseinawi et al., have proposed novel endoscopic dye-marking (Black Eye™ Ink) to enhance surgical precision and margin safety [20]. To address this in our study, deploying a flexible ureteroscope intraoperatively provided real-time luminal visualization, consistently securing negative surgical margins across all robotic cases. This confirms our concurrent endoscopic strategy as a highly effective counterpart to emerging dye-based localization strategies.

Historically, the open approach served as the exclusive standard of care for SU until 2007, when the first case of a robot-assisted laparoscopic distal ureterectomy with psoas hitch was reported [24]. Subsequent studies emphasized the versatility of the robotic platform, establishing its feasibility not only for urothelial carcinoma but also for benign strictures, fistulas, and other non-malignant etiologies [25–27]. Based on this foundation, subsequent descriptive series established the technical feasibility of the robotic approach: Palagonia et al. reported favorable functional outcomes in a single-center series utilizing the psoas hitch technique [17], while Saini et al. reinforced these findings with long-term oncological data, observing durable recurrence-free survival in a similar reconstructive cohort [19]. However, these studies were limited by small sample sizes and the lack of a direct comparator group.

More recently, multicenter efforts have begun to define the role of robotics in high-risk disease. Ditonno et al. (ROBUUST 2.0) compared robotic distal ureterectomy against robotic RNU, demonstrating that the kidney-sparing approach is oncologically safe even for high-grade tumors [21]. Similarly, Campi et al. confirmed that robotic SU offers significantly better preservation of renal function than robotic RNU without compromising cancer control [18].

Building upon these findings, our analysis seeks to move beyond surgical feasibility to provide a comparative assessment of surgical modalities. While Ditonno et al. validated the extent of resection (segmental vs. radical) [21], our study addresses the critical question of surgical modality (robotic vs. open). By directly assessing our robotic results against a long-term open surgical cohort, we found that the robotic platform demonstrated comparable short-term oncological efficacy to the traditional gold standard. This comparability extended to organ preservation, as the need for subsequent RNU due to recurrence was similar across both surgical groups (p = 1). Furthermore, we observed similar intraoperative and postoperative complication rates (p = 0.7), suggesting a comparable safety profile between the two modalities. However, robotic cases exhibited an advantage in postoperative recovery, evidenced by a threefold reduction in median hospital stay (4 vs. 12 days, p < 0.001).

To evaluate the potential for selection bias introduced by prior endoscopic resections, we performed a sensitivity analysis excluding cases with no residual malignancy (pT0) in both groups. Despite the reduced sample size, exclusion of pT0 cases did not materially change the direction of the findings, although the analysis remained limited by small numbers and wide CIs. While these results are constrained by the small number of subjects and limited follow-up, they suggest that the preliminary oncological profile of the robotic approach is not driven solely by a lower-risk subpopulation.

When analyzing functional outcomes, the observed trend toward greater renal recovery following robotic intervention may be attributed to the inherent technical properties of the platform. Precise multi-planar articulation and high-definition visualization facilitate a non-traumatic dissection, preserving the critical ureteral adventitial blood supply and ensuring a tension-free reconstructive repair [28]. This precision is particularly vital for distal ureterectomy, which requires meticulous ‘en bloc’ resection and formal cystotomy closure [28, 29]. Crucially, as 72.5% of the open procedures in our series were performed by the same two expert surgeons responsible for the robotic cohort, the observed trend may reflect a combination of platform-related technical factors, surgeon experience, relief of obstruction, and perioperative selection factors. The impact of chronic obstruction was particularly notable in the robotic group, where nearly 50% of patients required preoperative drainage due to tumor burden or ureteral strictures prior to definitive surgical repair.

It is important to acknowledge that the current literature regarding robotic SU is heavily restricted by small cohorts and significant heterogeneity in surgical techniques, even within individual series. To the best of our knowledge, we have detailed the available studies addressing this topic. Hence, drawing definitive conclusions based on these comparisons is challenging and underscores the need for larger multi-institutional studies to better define the oncological and functional outcomes of robotic SU.

In our study, the imbalance in sample sizes and follow-up timelines is directly driven by our growing robotics practice, which has accounted for 80% of SU volume since 2022. Consequently, the robotic arm exhibits a high rate of censoring (92.3%) with median survival not yet reached, contrasting sharply with the mature 135-month median observed in the open group. The limited sample size and the imbalance in follow up duration are major limitations. These discrepancies inherently influence survival interpretation: while our Kaplan-Meier and Firth’s penalized regression analyses show no significant difference in survival endpoints, these findings must be viewed as preliminary evidence of short-term oncological safety. The limited number of events in the robotic arm increases the risk of a Type II error, where subtle long-term differences in recurrence patterns may remain undetected particularly for recurrence-free and OS analysis. Consequently, our results support the safety and feasibility of the robotic approach, but long-term equivalence remains to be established through larger, multi-institutional cohorts with extended follow-up.

Another key limitation is the chronological discrepancy between cohorts; robotic cases were performed later, coinciding with the institutional adoption of Enhanced Recovery After Surgery (ERAS) protocols [30]. Thus, the reduced hospital stay (4 vs. 12 days) may be partially influenced by evolved perioperative care rather than the surgical platform alone. However, evidence suggests a synergistic effect; Muhandes et al. demonstrated that while ERAS reduces length of stay across all modalities, the most substantial reduction occurs in robotic cohorts [31]. This suggests the robotic approach may be uniquely protective against morbidity, and may have contributed to the shorter hospitalization alongside ERAS and other temporal changes in perioperative care. Despite these limitations and its retrospective design, this single-center study ensures a high degree of consistency in surgical standardization as well as uniform pathological review and longitudinal data collection. This homogeneity strengthens the internal validity of the comparative analysis and minimizes the variability often encountered in multi-institutional registry studies.

Conclusions

Robotic SU appears feasible and was not associated with an apparent compromise in short-term oncological outcomes during the available follow-up period. The robotic approach was associated with significantly reduced hospital stays and a trend towards improved functional renal recovery. However, these findings should be viewed as preliminary evidence of safety. Future large-scale, multi-institutional studies with extended follow-up are necessary to establish long-term oncological non-inferiority.

Abbreviations

ILRFS

Intraluminal recurrence-free survival

IVRFS

Intravesical recurrence-free survival

KSS

Kidney sparing surgery

MFS

Metastasis-free survival

RNU

Radical nephroureterectomy

SU

Segmental ureterectomy

UTUC

Upper tract urothelial carcinoma

Authors’ contributions

All authors contributed to the study conception and design. Validation, Z.A.D, E.F, O.P, A.S and N.K; formal analysis and investigation, R.L; data curation, Z.A.D, E.F, O.P, A.S, N.K, O.C, M.F; writing - original draft preparation R.L; writing - review and editing, Z.A.D, B.R; visualization, R.L; supervision, Z.A.D. All authors have read and agreed to the published version of the manuscript. All authors commented on previous versions of the manuscript. All authors read and approved of the final manuscript.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines and was approved by the institutional review board of Sheba Medical Center (SMC-4146; last update 4 June 2025). Consent was waived per local regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.