For the historical control group, the raw study data of a prospective cohort study on the learning curve of the da Vinci Xi system, published in 2018, were used [8]. This study was approved at Maria Middelares Ghent hospital with the Belgian trial number B300201629629. This study protocol was submitted at ClinicalTrials.gov (NCT02975401).

Inclusion criteria for both studies were almost the same except for excluding patients with a BMI higher than 35 kg/m² in the da Vinci group. For this reason, in the original publication of this cohort one patient was excluded from analysis. Since we could use the raw data from the control group, we included that patient in the current analysis. No patient in the Hugo SUSHI cohort had a BMI higher than 35 kg/m².

All procedures were performed by a single surgeon who, at the start of the SUSHI study, had over 25 years of experience with laparoscopic TAPP and more than 7 years of experience performing r-TAPP using the da Vinci Xi system. The surgeon has been a proctor for robotic surgery for 6 years having proctored 100+ hernia surgeons. At the time of the start of the da Vinci Xi, the surgeon had a 20-year experience with laparoscopic inguinal hernia repair, but no experience with robot assisted surgery.

Preparation before the first procedures with Hugo^TM RAS involved simulator training on the Hugo^TM RAS following dedicated simulator training modules and a 3-day training at the ORSI academy training centre in Melle, Belgium. This involved a day of technology training on Hugo^TM RAS, a day of skills acquisition on a porcine model performing cholecystectomy and a day of hernia specific training on human cadaver. This training at ORSI academy included training a second surgeon as first assist and the nursing team.

The team in the operating room consisted of a first assist and nursing team with no previous clinical experience using Hugo^TM RAS, but an extensive experience with the da Vinci X and da Vinci Xi systems. For the da Vinci group the nursing team had extensive experience with the da Vinci Xi system from urological robot-assisted surgery.

In the Hugo SUSHI cohort, a set-up specialist from Medtronic was present during all 50 cases, whilst in the da Vinci group logistic support from Intuitive was only present during the first 10 surgeries in the operating room.

Patients were operated consecutively using Hugo^TM RAS. The Hugo^TM RAS is different from the da Vinci Xi in that it has an open console and a modular set up, with separate arm carts to be positioned around the patient. There is one arm cart for the camera endoscope and two arm carts for the instruments. The patient was positioned in a dorsal decubitus with a mild Trendelenburg position of 10° and the arms tucked alongside the patient’s body. A 12 mm optical trocar was placed at the supraumbilical position after creation of pneumoperitoneum using a Veress needle with an intra-abdominal pressure of 12 mmHg. Two additional trocars (8 mm) are placed bilaterally in the flank at the level of the umbilicus under direct vision. The distance between the lateral trocars and the umbilical trocar was 8 cm at minimum. After inspection of both groins, a decision was made to perform either a unilateral or a bilateral repair. Mesh and peritoneal closure suture were introduced into the peritoneal cavity. The meshes used in the study were 13 × 15 cm for unilateral hernias and 13 × 28 cm for bilateral hernias (Progrip™ Self-Fixating Mesh, Medtronic, Minneapolis, MN, US). The suture for closure of the peritoneum was a slowly absorbable barbed suture 15 cm in length for each side (V-Loc™ 90, Medtronic, Minneapolis, MN, US). The trocars were docked to the robotic arms after positioning the arm carts using the Hugo^TM RAS CE approved set-up guide for inguinal hernia repair with the endoscope at the umbilicus (Figures 1A,B). Tilt angles were installed correctly, and the docking angles were placed with a maximum allowed deviation of the recommended angle of −5° to +5°. The need for intraoperative adjustments of the angels or the position of the arm carts and robotic arms was documented on the case report form (CRF).

Robotic instruments used were a large needle driver, a bipolar fenestrated grasper and monopolar curved shears. Inguinal hernia repair was performed according to the standard surgical principles, with mesh placement after appropriate preperitoneal dissection and critical view of the myopectineal orifice, as described by Jorge Daes and Edward Felix [9]. The self-fixating mesh was appropriately positioned, and no additional fixation was utilized. After mesh placement, the peritoneum and subsequently the skin were closed.

For the da Vinci group the procedure was described in detail in the previous publication [8]. The surgical technique was standardized in both groups with two exceptions. Firstly, in the da Vinci group the robotic system involved the da Vinci Xi and secondly, bilateral hernias were repaired using two separate meshes. In the Hugo SUSHI cohort one mesh of 13 × 28 cm was used, a technique previously shown to be feasible and safe for bilateral repairs [10].

Primary endpoint of the study was the skin-to-skin operative time defined as the time between first skin incision and the last skin suture. Secondary endpoint was the total operative time defined as the time between arrival of the patient in the operating room and the time the patient left the operating room. The adherence to the official set-up guide and the need for perioperative adjustments were documented. Patient reported outcome was documented using the EuraHS quality of life (QoL) instrument preoperatively and at the follow up clinical control visit at 4 weeks postoperatively [11].

Patient data recorded include: age; BMI; recurrent inguinal hernia; previous abdominal surgeries; unilateral versus bilateral hernia; hernia side in unilateral hernias; planned for day care clinic or overnight admission; preoperative and at 4 weeks postoperatively QoL assessed with the EuraHS QoL score [11]; inguinal hernia classification (medial-lateral-femoral, size 1-2-3) [1]; postoperative complications; length of stay; unplanned admission and overnight stay; complications at follow up at 4 weeks; seroma at 4 weeks, classified according to Morales et al [12].

During the course of the surgery, 8 time points were documented on the CRF: 1) arrival of the patient in the operating room, 2) end of induction of anaesthesia, 3) first skin incision, 4) surgeon into the console, 5) start placement of the mesh, 6) start suturing the peritoneum, 7) last skin suture, 8) patient leaving the operating room. These time points define seven distinct time intervals, T1-T7, in minutes. All data were entered in a REDCap^® database, downloaded to an excel file and double checked with the CRFs for errors before finalising the database and the start of the analysis. The total operative time was the sum of all 7 times intervals (T1+T2+T3+T4+T5+T6+T7). The skin-to-skin operative time was the sum of T3+T4+T5+T6.

A sample size of 50 patients was empirically chosen as being large enough to evaluate the learning curve effect on operative time and small enough to be performed within a reasonable time frame. It also mirrored the sample size of the da Vinci Xi group.

The statistical methodology was selected and implemented by an independent statistician. Patient characteristics were summarized using proportions (%N) or by means and standard deviations (SD). Differences in continuous or dichotomous variables between groups were analysed using the T-test or Fisher’s exact test. Overall operative times were calculated for both study groups. They were also analysed separately for the first 25 and the last 25 patients in each group. The evolution of operative time during the learning curve was graphically illustrated using least square regression lines. Operative times for both study groups were also calculated overall and specifically for unilateral and bilateral inguinal hernia repairs. For the calculation of the EuraHS QoL scores (total and subdomains) the previously described methodology was used [13]. The distributions of these QoL scores were summarized both numerically and graphically, employing median and interquartile ranges (P25-P75), and were compared using the Mann-Whitney U test. A p-value of ≤ 0.05 was considered statistically significant. All analyses were performed with SAS software (release 9.4, Cary, NC, US).

Between September 2023 and December 2024, the first 50 patients operated with the Hugo™ RAS at Sint Vincentius hospital, Deinze (Belgium) were enrolled in this study (Hugo SUSHI cohort). The da Vinci group consists of the first 50 patients who were operated with the da Vinci Xi during a 5-month inclusion period from September 2016 until January 2017 at Maria Middelares hospital, Ghent (Belgium).

Patient demographics, as summarized in Table 1, demonstrated no differences between groups in terms of gender, BMI, hernia recurrence rate and hernia side. The patients in the Hugo SUSHI cohort were older with a mean (SD) of 65.0 (14.2) years, compared to 58.7 (15.3) years in the da Vinci group (p = 0.035).

Table 2 shows the perioperative outcomes of the Hugo SUSHI cohort and da Vinci control group. Two intraoperative complications occurred in the Hugo SUSHI cohort (4%). These complications were not procedure- or robot-related, consisting of one patient vomiting during induction of anaesthesia and one patient with an episode of atrial fibrillation. The setup guide developed for the Hugo RAS system for inguinal hernia repair (as shown in Figure 1) was validated and no adjustments needed to be made perioperatively. There were three instances of robot related issues during the operations. In all three cases, this did not lead to any harm to the patient but only to longer operation times. All errors could be solved during the operation. One error was a cart failure with rebooting taking 25 min, another error a start-up failure causing a 7 minute delay and one error was an instrument failure with no impact on the surgical time.

There is no significant difference in intrahospital complications between both groups. The intrahospital complication in the Hugo SUSHI cohort (n = 3) were postoperative nausea and vomiting, haematuria and urinary retention. The haematuria was not considered procedure-related since the patient presented with kidney stones only days before the procedure and already had a ureteral stent in situ. All five patients in the da Vinci group who experienced an intrahospital complication had urinary retention. Most patients in both cohorts did not stay overnight. In the Hugo SUSHI cohort, 80% of patients were discharged the same day, compared to 68% in the Da Vinci group. Postoperative seroma occurred in 10% of patients in the Hugo SUSHI cohort, compared to 14% in the Da Vinci group.

The detailed outcome data for the individual time blocks for both cohorts is presented in Table 3. All steps took less time in the Hugo SUSHI cohort, except for T6 (closure of the peritoneum).

The outcome data for the total operative time and the skin-to-skin operative time is presented in Table 4. The operative times for unilateral and bilateral hernias separately and for the first 25 patients and the subsequent 25 patients for each group are mentioned. The mean total operative time with Hugo^TM RAS was 85.6 min and for da Vinci Xi 102.7 min. Operative times in the Hugo SUSHI cohort were significantly shorter with a mean difference of 17 min (95% CI +10.0 to +24.0, p < 0.001). This significant difference is not present for the skin-to-skin operative times, with 57.0 min in the Hugo SUSHI cohort and 62.8 min in the da Vinci group and a mean difference of 5.9 min (95% CI -0.9 to +12.6, p = 0.09).

The skin-to-skin operative time for unilateral hernias with Hugo^TM RAS was not significantly different from the operative time with da Vinci Xi. The mean skin-to-skin operative time for bilateral hernias with Hugo^TM RAS was significantly shorter compared to da Vinci Xi, with a mean difference of 17.4 min (95% CI +7.1 to +27.8, p < 0.001). When comparing the first 25 patients of both groups, the mean total operative time and skin-to-skin operative time were significantly shorter in the Hugo SUSHI cohort, with a mean difference of resepectively. The difference in skin-to-skin operative time was no longer present when comparing the second 25 patients of each group.

The evolution of the learning curve of the skin-to-skin operative time over time is graphically depicted for each cohort separately in Figures 2A,B. In the Hugo SUSHI cohort, the skin-to-skin operative time remained stable across cases, indicating no measurable learning curve.

The results of preoperative and postoperative analysis of the EuraHS QoL scores are presented in Table 5. There was no difference between the study groups overall, for the specific domains of the EuraHS QoL scores preoperatively and postoperatively, or for the change between preoperative and postoperative assessments. EuraHS QoL scores significantly improved from preoperative to postoperative assessment across all domains in both groups, which is graphically depicted in Figure 3.