Early evaluation of a next‐generation surgical system in robot‐assisted total laparoscopic hysterectomy: A prospective clinical cohort study

Abstract Introduction This study aimed to demonstrate the safe and effective use of the Versius surgical system (CMR Surgical, Cambridge, UK) in robot‐assisted total laparoscopic hysterectomy. This surgical robot was developed iteratively with input from surgeons to improve surgical outcomes and end‐user experience. We report data from the gynecology cohort of an early clinical trial designed in broad alignment with IDEAL‐D (Idea, Development, Exploration, Assessment, Long‐term follow‐up – Devices) stage 2b (Exploration). Material and methods The study is registered in the Indian clinical trials register (CTRI/2019/02/017872). Adult women requiring total hysterectomy who provided informed consent and met the eligibility criteria underwent procedures at one of three hospitals in India. Five surgeons performed robot‐assisted total laparoscopic hysterectomies using the device from March 2019 to September 2020. The primary endpoint was rate of unplanned conversion to conventional laparoscopic or open surgery. Adverse events were adjudicated by an independent clinical events committee using endoscope video recordings and clinical notes. Results In total, 144 women underwent surgery (median age: 44 years [range: 28–78]; median body mass index 25.8 kg/m2 [range: 14.3–47.8]). The rate of unplanned conversion to conventional laparoscopy was 2/144 (1.4%); neither conversion was device related. No surgery was converted to open. In total, 13 adverse events occurred among seven (4.9%) patients, comprising seven serious adverse events and six adverse events. One serious adverse event was deemed device‐related. Two patients were readmitted to hospital within 30 days; both made a full recovery. No patients died within 90 days of surgery. Conclusions The device provides a safe and effective option for total laparoscopic hysterectomy; these findings support its continued implementation in larger patient cohorts and expansion in other major minimal access indications.

The device was developed in broad alignment with the IDEAL-D (Idea, Development, Exploration, Assessment, Long-term follow-up -Devices) recommendations throughout its evolution. [15][16][17][18][19][20] First, its design was based on testing by and feedback from end users to better meet their needs, 16 and usability was demonstrated early in the development process. 15 Several procedures were then successfully performed in cadaver and live porcine studies, 14,17,18 and a purposedesigned training program was validated. 21 A small number of firstin-human minor surgeries were then safely completed. 22 This prospective, early clinical study broadly aligns with stage 2b (Exploration) and aimed to provide a full device safety and performance analysis from a larger cohort of more than 120 women requiring total laparoscopic hysterectomy.

| Surgeons
The participating surgeons were accredited, practicing, high-volume consultant gynecology surgeons with extensive experience in MAS.
Surgeons had no prior experience using other robotic systems and had limited or no experience using this robotic device in humans prior to the clinical trial (two or fewer cases per surgeon). In accordance with the training protocol, all participating surgeons completed approximately 10 h of online training and a minimum 6 h of simulator training and passed the validated 3.5-day training program immediately prior to the start of the study. 21 Surgeons completed procedures at two study sites in Maharashtra, India: the Deenanath Mangeshkar Hospital and Research Center (March 12, 2019, to January 24, 2020) and the HCG Manavata Cancer Center (October 30, 2019, to January 20, 2020). Following a regulatory request for "final finished device" data, an additional 25 cases were performed as part of a bridging study completed at both centers and at an additional third center (because of the COVID-19 pandemic), the Healing Hands Clinic, Pune, Maharashtra, India, between August 24, 2020, and September 04, 2020.
conditions: uncontrolled hypertension and/or diabetes mellitus (blood glucose concentration >200 mg/dl), known presence of regional and/or distant metastases, medical instability prior to surgery, and any obvious contraindications to abdomino-pelvic surgery. Patients were excluded if their physical status was American Society of Anesthesiologists (ASA) class III or higher. 23 In June 2020, this exclusion criterion was changed to ASA class IV or higher to extend eligibility once several procedures had been safely performed. Eligible participants were identified from study hospital surgical lists and approached directly by their surgeon or clinical team. After being provided with relevant study information, patients provided written informed consent to participate in the study, and audio-visual consent was recorded for patients who enrolled at the Deenanath Mangeshkar Hospital and Research Center.

| Study design
Patients completed pre-operative screening, then underwent total hysterectomy on Day 1 (the surgical procedure steps are provided in Table S1). Perioperative care was uniform across all patients unless adverse events (AEs) occurred. Following surgery, patients were discharged from hospital when deemed safe by the operating surgeon and postoperative care team and were followed-up on postoperative Day 30 (±2 days) and Day 90 (±7 days) via telephone or clinic visit ( Figure S1). Patients in the bridging study were followed up to at least Day 30 (±2 days).

| Device set-up and OR layout
The device consisted of a surgeon console, instrument bedside units (two or three instrument bedside units can be used according to procedure type and surgeon preference; three were used in this study), and one visualization bedside unit ( Figure 1A+B). Energy instruments, including a Monopolar Hook and Bipolar Maryland Grasper, were used during the procedures. The most frequently used port positioning and OR layouts are illustrated in Figure 1C+D. The camera port was positioned up to 2 cm above the umbilicus on the midline, with a 5 mm robotic port on the right and left mid-clavicular line, at the level of the umbilicus. A 5-10 mm assistant port was positioned below the umbilicus at the midline. For women with high BMI (≥30 kg/m 2 ), the camera port was positioned below the umbilicus.

F I G U R E 1
Overview of the device, port positioning, and operating room layout. Schematic representation of the setup of the device (A) and real-world image of the device setup (B); adapted from Haig et al. 15 Common port positioning (C) with corresponding BSU positions (D); adapted from Kelkar et al. 22 The assistant port was for nonrobotic laparoscopic instruments. Umbilicus is where the ML crosses the SUL. Aux: auxiliary monitor; BSU: bedside unit; Console: surgeon console; Endo: endoscope; Instr: instrument; MCL: mid-clavicular line; ML: midline; SUL: supine-umbilical line

| Study procedures and evaluations
The primary endpoint was the rate of unplanned conversion of robot-assisted procedures to conventional laparoscopic or open surgery. Secondary endpoints included operative time (from first incision to skin closure), estimated intra-operative blood loss, intraoperative complications, return to the OR within 24 h, length of hospital stay, hospital readmission within 30 days, postoperative complications through 30 and 90 days, and mortality rate at 90 days.
Uterine weight was measured postoperatively at a histopathology laboratory. As a standard protocol, length of stay was pre-emptively extended to 4 days for some patients living a long distance from the study site because of transportation limitations if readmission was required. included all medical occurrences that were life-threatening or led to death, required hospital admission, prolonged hospitalization, or resulted in persistent disability or permanent damage. Any other AE the CEC deemed "medically significant" was also classified as an SAE.

| Adverse events
Prior to the adjudication meetings, each CEC member was provided with detailed information on each AE/SAE, as recorded by the operating surgeon. For SAEs classified by the surgeon, additional information, including ethics committee notifications, follow-up reports, and summary operative and recovery notes, were provided.
The CEC then systematically reviewed each AE regarding expectedness (expected/unexpected), seriousness (AE/SAE), and relatedness to the device (not related/possibly related/probably related/ related) until consensus. Expectedness was determined based on whether the complication is typically listed on the patient consent form for hysterectomy and in view of the comorbidities present for each patient. If required, the committee had access to endoscope video recordings for all surgeries to aid their review. CEC members "upgraded" the event classification in uncertain cases.

| Statistical analyses
A sample size of at least 120 women was determined sufficient to estimate conversion rates with satisfactory accuracy, with confidence intervals of an appropriate size; alpha was set at 0.05, and the conversion rate was assumed to be 1.8% based upon a literature search. The inclusion of patients in the bridging study further increased the sample size. Unless otherwise stated, continuous data summaries were used to present the number of observations, the median, and range. Data were collected in SAS format, and all analyses were performed using SAS version 9.4.

| Patient disposition and baseline characteristics
Of the 154 consenting women who were screened, 145 were eligible for the study and 144 proceeded to robot-assisted MAS ( Figure 2).
One patient underwent total hysterectomy and cholecystectomy in the same surgery. Although the two procedures were successfully completed using the device without complications or readmission, this patient was excluded from the analysis presented herein as only gynecology procedures were included. A total of 144 women underwent robot-assisted total hysterectomy, of whom two also underwent robot-assisted bilateral salpingo-oophorectomies. The median age was 44 years (range: 28-78), and the median BMI was 25.8 kg/m 2 (range: adenomyosis (n = 13; 9.0%), and uterine prolapse (n = 9; 6.3%; Table S2).
In total, 67 women (46.5%) had previously undergone abdominal/pelvic surgeries, one of which was in the previous 12 months. Median uterine weight (n = 41) was 170 g (range 50-600). The demographic and clinical characteristics of the cohort are summarized in Table 1.

| Intra-operative endpoints
Overall, 142/144 women (98.6%) underwent a successful total laparoscopic hysterectomy using the device (Figure 2). No procedure was converted to open surgery. Two (1.4%) surgeries were completed using conventional laparoscopy. One of the conversions was secondary to a bladder injury, which occurred at the time of anterior colpotomy ( Table 2). This was sutured by a urologist who had no training on the device and therefore used conventional laparoscopy.
The CEC determined that this was a recognized (expected) complication of that stage of the procedure and was not related to the patient lost more than 500 ml of blood during their surgery, and blood loss of less than 5 ml was noted in 18 patients (12.5%; Figure 3B). No patient required a blood transfusion during or after surgery.

| Postoperative endpoints
The median length of hospital stay, from admission to discharge, was 3 days (range: 1-9), and the median length of postoperative stay was F I G U R E 2 Study CONSORT diagram. a One patient did not return to the site for screening. b Based on routine endoscopic examination, surgeons decided that alternative more familiar approaches (three conventional laparoscopic hysterectomies and one open surgery) were more appropriate for four women who had highly complex anatomies and/or disease states, such as an enlarged uterus and fundal fibroid, and multiple extensive adhesions involving the uterus, sigmoid colon, rectum, ovaries, and ureter. c One patient underwent total hysterectomy and cholecystectomy in the same surgery, both of which were successfully completed using the device without complications or adverse events. d Includes two patients who also underwent bilateral salpingo-oophorectomies. e One patient had a urinary bladder injury at the time of anterior colpotomy, which was sutured with conventional laparoscopy by a urologist. SAE, serious adverse event

| AEs
Of the 144 women who underwent total hysterectomy, there were 13 AEs in seven patients (4.9%), comprising seven SAEs in five patients and six non-serious AEs in two patients ( Table 2).

| DISCUSS ION
This early clinical study demonstrated the safe and effective use of the device for major gynecology surgery in women with varied diagnoses and a range of BMIs and uterine weights. The rate of unplanned con- shorter. In any case, patient safety is of paramount importance and was the key outcome of this current study.
Alongside its sister general surgery cohort study, 33 this study supports the implementation of the device in more patients and in a greater range of abdominal and pelvic surgeries. Such expansion will continue to follow the IDEAL-D recommendations, proceeding to stage 3 (Assessment), with the explicit aims of demonstrating middleand long-term clinical outcomes and cost effectiveness. We envisage that continued use and expansion of robot-assisted MAS will improve surgical outcomes for patients, with fewer intra-and postoperative complications than with conventional MAS and open surgery. Robotic assistance in major surgeries may also reduce the overall length of hospital stay, leading to higher case throughput and surgeon availability.
We hope that the evolution of robotic surgical system designs will facilitate wider access of MAS to surgeons, with a shorter learning curve and less challenging operating techniques than conventional instruments.
Further, by removing the need for awkward and static positioning while operating, robotic devices may play an important part in alleviating the physical burden on surgeons, potentially extending surgical careers.
Results pertaining to length of surgery are important, and further clinical series will demonstrate mature use of the device FIGURE 3 Operative time, intra-operative blood loss, and length of hospital stay. Operative time from first incision to skin closure (A), estimated intra-operative blood loss (B), length of hospital stay from day procedure performed to patient discharge from hospital (C). For A and C, middle vertical lines represent the medians, left and right box edges represent the first and third quartiles, and lower and upper whiskers extend to the respective lowest and highest values. a Includes patients with estimated blood loss recorded as <100 ml. b Includes patients with estimated blood loss recorded as <500 ml in relation to these perioperative outcomes. The data collected in this clinical study have been entered into a Versius surgical registry created in alignment with IDEAL-D stage 4 (Long-term study) to enable monitoring of rare events, longer-term outcomes, and quality assurance. The registry will enable continual collection of real-world data to evaluate ongoing patient safety, a crucial aspect for medical device vigilance and postmarket surveillance.