Comparative operative performance of three robotic platforms in benign gynecologic surgery: a single-surgeon retrospective cohort study

From January 2020 to May 2026, 218 patients were included in this study: 109 (50.0%) treated with da Vinci Xi, 41 (18.8%) with Hugo RAS, and 68 (31.2%) with da Vinci SP. Overall, 135 patients underwent hysterectomy and 83 underwent myomectomy. Age differed among groups ( P = .026), as did obesity prevalence ( P = .047) and procedure distribution ( P = .020). Hysterectomy accounted for 70.6% of Xi, 58.5% of Hugo, and 50.0% of SP procedures. Height, weight, and BMI did not differ significantly (Table  1 ). All 218 procedures were completed without conversion to laparoscopy or laparotomy. Table 1 Baseline demographic and clinical characteristics by robotic platform Parameter Xi Hugo SP P value Overall Cohort ( N = 218) Xi ( N   = 109) Hugo ( N   = 41) SP ( N   = 68) Age, years 48.28 (26–78) 45.46 (28–73) 45.10 (29–59) 0.026 Height, cm 159.12 (143.6–170.3) 159.61 (143.7–174.2) 158.58 (149.1–175.7) 0.647 Weight, kg 63.10 (39.5–98.9) 66.73 (45–99.8) 61.61 (42–126) 0.144 BMI, kg/m² 24.89 (17.26–36.62) 26.29 (17.36–37.47) 24.43 (16.54–48.67) 0.153 Obesity, n (%) 12 (11.0) 11 (26.8) 9 (13.2) 0.047 Procedure 0.020  Hysterectomy, n (%) 77 (70.6) 24 (58.5) 34 (50.0)  Myomectomy, n (%) 32 (29.4) 17 (41.5) 34 (50.0) Myomectomy Subgroup ( N  = 83) Xi ( N  = 32) Hugo ( N  = 17) SP ( N  = 34) Number of myomas, n 3.72 (1–9) 2.76 (1–6) 2.35 (1–8) 0.018 Maximum diameter, cm 8.45 (5-13.5) 7.99 (4.6–12) 6.81 (2.5–13) 0.014 Values are mean (range) unless otherwise indicated Baseline demographic and clinical characteristics by robotic platform Values are mean (range) unless otherwise indicated Postoperative length of stay was similar across platforms ( P = .826). Total specimen weight and uterine weight did not differ significantly. Myoma weight differed ( P = .003), with lower mean weight in the SP group. Overall estimated blood loss differed ( P = .009); the difference was significant for myomectomy ( P = .007) but not hysterectomy ( P = .441). Overall operative time differed ( P < .001), concentrated in myomectomy ( P < .001). Console time differed overall ( P < .001) and within both hysterectomy ( P = .002) and myomectomy ( P < .001) (Table  2 ). Table 2 Intraoperative and postoperative outcomes by robotic platform Parameter Xi ( N  = 109) Hugo ( N  = 41) SP ( N  = 68) P value Length of stay, days 3.67 (2–43) 3.34 (2–9) 3.49 (2–16) 0.826 Total specimen weight, g 485.5 (3–3516) 454.4 (5–1975) 371.8 (7–1738) 0.276  Hysterectomy (Uterine weight), g 523.2 (60-3516) 495.1 (77–1975) 553.2 (66–1738) 0.918  Myomectomy (Myoma weight), g 395 (3-1130) 397.1 (5–1040) 190.3 (7–740) 0.003 EBL overall, mL 298.7 (0–3100) 418.9 (3–1900) 159.9 (0–1050) 0.009  Hysterectomy EBL 236.5 (0-2000) 281.8 (3–1800) 160.7 (0–1050) 0.441  Myomectomy EBL 448.3 (15-3100) 612.4 (10–1900) 159.2 (1–1000) 0.007 Operative time, min 339.1 (180–896) 350.7 (218–736) 275.4 (155–575) < 0.001  Hysterectomy OT 332.9 (180–896) 330.1 (234–736) 285.9 (155–575) 0.109  Myomectomy OT 354 (185–565) 379.9 (218–721) 264.9 (160–504) < 0.001 Console time, min 156.0 (30–410) 107.2 (25–370) 99.1 (25–357) < 0.001  Hysterectomy CT 143.0 (30–350) 98.5 (40–202) 99.9 (27–357) 0.002  Myomectomy CT 187.4 (30–410) 119.5 (25–370) 98.3 (25–331) < 0.001 Blood transfusion, U 0.94 (0–24) 1.24 (0–10) 0.38 (0–8) 0.146  Hysterectomy BT 0.95 (0–24) 1 (0–10) 0.59 (0–8) 0.793  Myomectomy BT 0.91 (0–10) 1.59 (0–6) 0.18 (0–2) 0.009 EBL, estimated blood loss; OT, operative time; CT, console time; BT, blood transfusion. Values are mean (range) Intraoperative and postoperative outcomes by robotic platform EBL, estimated blood loss; OT, operative time; CT, console time; BT, blood transfusion. Values are mean (range) After adjustment for age, obesity status, and total specimen weight, the estimated geometric mean console time for hysterectomy was 96.3 min (95% CI, 80.9–115.0) with SP, 95.6 min (78.8–116.0) with Hugo, and 136.5 min (119.1–156.0) with Xi. Relative to SP, Xi had significantly longer adjusted console time (Tukey-adjusted P = .001), whereas Hugo did not differ from SP ( P > .99) (Table  3 ). Within the myomectomy subgroup, the SP cohort had fewer myomas, a smaller maximum myoma diameter, and lower myoma weight than the Hugo and Xi cohorts (Table  2 ). After additional adjustment for myoma count and maximum diameter, adjusted geometric mean console times were 107.6 min (83.0–139.0) with SP, 97.9 min (72.7–132.0) with Hugo, and 140.6 min (108.8–182.0) with Xi. Neither Hugo ( P = .86) nor Xi ( P = .22) differed significantly from SP (Table  3 ). Thus, the previously observed myomectomy console-time difference was attenuated after accounting for measured fibroid burden. Table 3 Multivariable-adjusted console time for hysterectomy and myomectomy Procedure / Platform Adjusted geometric mean, min 95% CI Adjusted P vs. SP Hysterectomy – SP 96.3 80.9–115.0 Reference Hysterectomy – Hugo 95.6 78.8–116.0 > 0.99 Hysterectomy – Xi 136.5 119.1–156.0 0.001 Myomectomy – SP 107.6 83–139 Reference Myomectomy – Hugo 97.9 72.7–132 0.86 Myomectomy – Xi 140.6 108.8–182 0.22 Hysterectomy model adjusted for age, total specimen weight, and obesity status. Myomectomy model adjusted for age, total specimen weight, number of myomas, maximum myoma diameter, and obesity status. Console time was log-transformed; estimates and 95% CIs were back-transformed. P values were adjusted using Tukey’s method Multivariable-adjusted console time for hysterectomy and myomectomy Hysterectomy model adjusted for age, total specimen weight, and obesity status. Myomectomy model adjusted for age, total specimen weight, number of myomas, maximum myoma diameter, and obesity status. Console time was log-transformed; estimates and 95% CIs were back-transformed. P values were adjusted using Tukey’s method For hysterectomy, back-transformed adjusted estimated blood loss was 76.3 mL (95% CI, 42.6–123.8) with SP, 158.3 mL (96.4–242.2) with Hugo, and 136.8 mL (95.4–188.8) with Xi. Neither Hugo ( P = .07) nor Xi ( P = .07) differed significantly from SP (Table  4 ). In the myomectomy model additionally adjusted for myoma count and maximum diameter, adjusted estimated blood loss was 191.1 mL (109.5–305.6) with SP, 415.8 mL (255.3–632.6) with Hugo, and 272.7 mL (168.2–413.3) with Xi. Hugo remained associated with greater blood loss than SP ( P = .04), whereas Xi did not differ significantly from SP ( P = .45) (Table  4 ). Table 4 Multivariable-adjusted estimated blood loss for hysterectomy and myomectomy Procedure / Platform Back-transformed adjusted estimate, mL 95% CI Adjusted P vs. SP Hysterectomy – SP 76.3 42.6–123.8 Reference Hysterectomy – Hugo 158.3 96.4–242.2 0.07 Hysterectomy – Xi 136.8 95.4–188.8 0.07 Myomectomy – SP 191.1 109.5-305.6 Reference Myomectomy – Hugo 415.8 255.3-632.6 0.04 Myomectomy – Xi 272.7 168.2-413.3 0.45 Hysterectomy model adjusted for age, total specimen weight, and obesity status. Myomectomy model adjusted for age, total specimen weight, number of myomas, maximum myoma diameter, and obesity status. Estimated blood loss was cube-root transformed; estimates and 95% CIs were back-transformed. P values were adjusted using Tukey’s method Multivariable-adjusted estimated blood loss for hysterectomy and myomectomy Hysterectomy model adjusted for age, total specimen weight, and obesity status. Myomectomy model adjusted for age, total specimen weight, number of myomas, maximum myoma diameter, and obesity status. Estimated blood loss was cube-root transformed; estimates and 95% CIs were back-transformed. P values were adjusted using Tukey’s method No statistically significant between-platform differences were observed for reoperation, transfusion, ileus, or fever (Table  5 ). One reoperation occurred in the SP group due to revision of large scarred insertion wound. Non-significant complication comparisons should not be interpreted as evidence of equivalence. Table 5 Perioperative complications and Clavien-Dindo classification by robotic platform Complication Xi ( N  = 109) Hugo ( N  = 41) SP ( N  = 68) P value Reoperation, n (%) 0 (0) 0 (0) 1 (1.5) 0.500 Transfusion, n (%) 25 (22.9) 13 (31.7) 11 (16.2) 0.168 Ileus, n (%) 3 (2.8) 2 (4.9) 4 (5.9) 0.590 Fever, n (%) 18 (16.5)* 4 (9.8) 8 (11.8) 0.232 Clavien-dindo Garde, n (%) 0.24  0 70 (64.22) 24 (58.54) 49 (72.06)  2 39 (35.78) 17 (41.46) 18 (26.47)  3 0 (0) 0 (0) 1 (1.47) Perioperative complications and Clavien-Dindo classification by robotic platform

We conducted a retrospective cohort study at a single tertiary academic medical center. The study was approved by the institutional review boards (IRB #114043 and IRB #115002), covering different enrollment periods, and written informed consent was obtained from all participants. Women undergoing robotic hysterectomy or robotic myomectomy for benign indications were eligible. Indications included symptomatic uterine leiomyomas, adenomyosis, abnormal uterine bleeding refractory to medical therapy, and other benign uterine disorders. Patients undergoing concomitant malignant procedures, emergency operations, or conversion to laparotomy before completion of robotic surgery were excluded. Patients were categorized according to the robotic platform used: da Vinci Xi, Hugo RAS, or da Vinci SP. Platform use was not randomized. The Xi and Hugo cohorts partially overlapped in calendar time, and the Xi and SP cohorts also partially overlapped; however, the Hugo and SP cohorts were noncontemporaneous. Platform selection reflected clinical and institutional practice, including platform availability and clinical indication. Figure  1 shows the timeline of implementation of the three platforms. Fig. 1 Timeline and cumulative case distribution of the three robotic platforms across study phases Timeline and cumulative case distribution of the three robotic platforms across study phases All procedures were performed by a single surgeon with established clinical experience using all three robotic platforms. Platform-specific experience was not quantified; therefore, no formal claim was made that learning-curve effects were completely eliminated. Robotic hysterectomy consisted of total hysterectomy with bilateral salpingectomy with and without oophorectomy. Robotic myomectomy was performed with excision of leiomyomas followed by multilayer myometrial reconstruction. Xi and Hugo procedures used conventional multiport approaches. SP procedures were performed through a 2.5- to 3-cm umbilical single-port access without accessory trocar placement. The primary outcome was robotic console time. Secondary outcomes were total operative time, estimated blood loss, total specimen weight, uterine weight, myoma weight, postoperative length of stay, and perioperative complications. Baseline variables included age, height, weight, body mass index, obesity status, and procedure type. For the myomectomy subgroup, the number of myomas enucleated and maximum myoma diameter were additionally recorded as measures of anatomical burden. Continuous variables were summarized as mean (range) and compared using one-way analysis of variance or the Kruskal-Wallis test. Categorical variables were compared using the chi-square or Fisher exact test. Key operative outcomes were examined separately for hysterectomy and myomectomy. Hysterectomy multivariable models adjusted for age, obesity status, and total specimen weight. To address differences in fibroid burden, myomectomy models additionally adjusted for the number of myomas and maximum myoma diameter. Console time was log-transformed and estimated blood loss was cube-root transformed. Adjusted marginal estimates and 95% confidence intervals were back-transformed. Pairwise post hoc comparisons used Tukey’s method. Statistical significance was defined as two-sided P < .05. Analyses were performed using R version 4.0.2. Because the three platforms were not fully contemporaneous, adjusted associations were interpreted as platform-associated differences rather than causal effects.

In this single-surgeon retrospective cohort comparing three robotic platforms for benign hysterectomy and myomectomy, platform-associated differences varied by procedure and by adjustment for case complexity. For hysterectomy, da Vinci Xi had a longer adjusted console time than SP, whereas Hugo and SP were similar. In myomectomy, the initial console-time difference between Xi and SP was attenuated and no longer statistically significant after incorporating myoma count and maximum diameter. Estimated blood-loss findings were procedure-specific: no significant adjusted difference from SP was demonstrated for hysterectomy, whereas Hugo was associated with greater adjusted blood loss than SP during myomectomy. Myomectomy requires enucleation, hemostasis, and multilayer reconstruction, during which instrument articulation, arm positioning, and internal or external clashing may affect workflow. SP deploys articulating instruments through a single multichannel cannula without external clashing and minimized internal clashing, whereas Hugo uses independently positioned arm carts with possible external clashing during the procedure. Because clashing and ergonomic events were not prospectively measured, we present these mechanisms as plausible contextual considerations rather than demonstrated explanations for the observed outcomes. These findings argue against interpreting robotic platform performance as a single hierarchy. Published comparative studies report heterogeneous results: SP has shown similar, shorter, or longer operative metrics than Xi depending on the cohort [ 19 – 26 ], while Hugo-versus-Xi comparisons have also produced context-dependent differences [ 17 , 18 ]. Multiplatform hysterectomy studies similarly suggest that performance differences may depend on the specific systems and setting [ 27 , 28 ]. In the present study, console time showed a reproducible pattern across both procedures, whereas blood-loss findings varied by procedure. A distinctive feature of this study is the single-surgeon setting, which reduces heterogeneity attributable to differences in operative technique. However, surgeon experience was not quantitatively tracked per platform, and learning-curve analyses demonstrate that platform-specific workflow metrics continue to evolve over time [ 9 , 21 ]. Crucially, surgical competencies and console familiarity gained from extensive multiport robotic experience (such as with the da Vinci Xi) are highly transferable to single-port platforms (da Vinci SP) during hysterectomy, which can significantly accelerate procedural proficiency and optimize console efficiency [ 29 ]. Similarly, initial institutional outcomes and the learning curve during Hugo RAS implementation are strongly modulated by the surgeon’s prior robotic experience, as surgeons with established multiport backgrounds consistently achieve shorter docking and console times compared to robotic-naïve operators [ 30 ]. The temporal structure is another critical consideration: Xi overlapped partly with Hugo and partly with SP, whereas Hugo and SP were noncontemporaneous. Changes in case selection, operating-room workflow, or cumulative team experience over time may therefore have contributed to the observed differences. The procedure-stratified findings are clinically relevant. The crude difference in total operative time was concentrated in myomectomy; however, the adjusted console-time contrast between Xi and SP was no longer statistically significant after accounting for myoma count and maximum diameter. This attenuation suggests that measured fibroid burden contributed to the crude platform difference. By contrast, the association between Hugo and greater estimated blood loss relative to SP persisted after the expanded adjustment. Myomectomy involves tumor enucleation, hemostasis, and multilayer reconstruction, and platform architecture may influence these tasks, but the observational design does not establish a platform-specific mechanism or causal advantage. Several limitations warrant emphasis. This was a retrospective, nonrandomized study; platform selection may have been influenced by availability and case selection. Baseline differences existed in age, obesity, procedure distribution, and myoma burden. Although the myomectomy models incorporated myoma count and maximum diameter, residual confounding may remain from fibroid location, FIGO type, depth, surgical difficulty, and other unmeasured characteristics. Hugo and SP were noncontemporaneous, introducing potential temporal confounding and operating-room team maturation effects. Furthermore, because the platforms were introduced sequentially, potential cross-platform skill transfer from earlier multiport experience to later-adopted systems cannot be completely ruled out [ 29 , 30 ]. The study was not designed to establish equivalence for uncommon complications. Finally, findings reflect the practice of one surgeon at one institution and may not generalize. In conclusion, among patients undergoing benign robotic hysterectomy or myomectomy in a single-surgeon practice, platform-associated differences depended on procedure and case-mix adjustment. Da Vinci Xi had a longer adjusted console time than SP for hysterectomy, but the corresponding myomectomy difference was no longer significant after accounting for myoma burden. In myomectomy, Hugo remained associated with greater adjusted blood loss than SP. These findings support procedure-specific interpretation of robotic-platform performance and warrant confirmation in larger, contemporaneous, multicenter studies with more detailed measures of surgical complexity.

Robotic-assisted surgery is an established component of minimally invasive gynecologic surgery, including hysterectomy and myomectomy [ 1 , 2 ]. Contemporary robotic systems differ substantially in architecture, port configuration, arm design, and instrument deployment [ 3 – 8 ]. These differences may influence operative workflow and efficiency; however, comparisons across studies are complicated by surgeon experience, institutional practice, patient selection, procedural complexity, and platform-specific learning effects [ 5 , 8 – 10 ]. The da Vinci Xi is an established multiport platform, whereas the Hugo RAS system uses a modular multi-arm configuration with independent arm carts and an open console [ 3 , 4 ], and the da Vinci SP system deploys articulating instruments through a single multichannel cannula [ 11 – 13 ]. Both newer platforms have been increasingly evaluated in gynecologic surgery [ 4 – 8 , 10 – 16 ]. Direct pairwise comparisons with the da Vinci Xi have reported heterogeneous results in operative time, console time, estimated blood loss, and perioperative outcomes [ 17 – 26 ]. Prospective multiplatform comparisons have also begun to emerge [ 27 ]. Collectively, these studies suggest that comparative performance may depend on the specific platform, procedure, and clinical setting. Our group previously compared Hugo RAS with da Vinci Xi across gynecologic procedures [ 17 ]. The present study extends this experience by incorporating the da Vinci SP platform and focusing on benign hysterectomy and myomectomy and there will be difference of the present three-platform, benign-procedure cohort from the earlier heterogeneous Xi–Hugo comparison. To our knowledge, this is the first study to directly compare these three platforms within a single gynecologic surgical cohort. The platforms were not used during completely overlapping calendar periods: Xi overlapped partly with Hugo and partly with SP, whereas Hugo and SP did not overlap. The study should therefore be interpreted as a comparative evaluation across phases of a single institutional robotic program rather than a fully contemporaneous three-arm comparison. All procedures were performed by the same surgeon with established clinical experience using all three platforms, thereby reducing inter-surgeon variation. We aimed to compare operative efficiency and perioperative outcomes, with console time as the primary outcome and procedure-stratified analyses for both hysterectomy and myomectomy.