Implementation of the Versius robotic surgical system for thoracic surgery: first clinical evaluation of feasibility and performance

Abstract

OBJECTIVES

The aim of this study is to demonstrate the ability of the Versius surgical system to successfully and safely complete a range of thoracic procedures aligned with Stage 2a (Development) of the Idea, Development, Exploration, Assessment and Long-term follow-up framework for surgical innovation.

METHODS

This prospective study included the first 30 consecutive patients who underwent robotic surgery with Versius by 2 surgeons without prior robotic experience between 1 April 2023 and 30 December 2023 [25 lung resections (wedge, segmentectomy and lobectomy) and 5 thymectomies]. There were no specific predetermined selection criteria for each case. The primary outcome was safe completion of the procedure without unplanned conversion. Secondary outcomes included intraoperative and postoperative complications, intraoperative device-related outcomes and pathology results.

RESULTS

Twenty-eight (93.3%) cases were completed without conversion. Both conversions were to thoracoscopy, one due to a ‘console alarm’ and the other due to pulmonary artery bleeding. In lung resections, median console time was 103 (90–129) min. Five (20%) patients experienced postoperative complications, most frequent was persistent air leak (16%). Median length-of-stay was 3 (2–4) days. Neither readmissions nor mortality was observed. In thymectomies, no intraoperative or postoperative complications, readmissions, reinterventions or mortality were observed. Median console time was 77 (75–89) min and median length of stay was 1 (1–1) day.

CONCLUSIONS

This phase 2a IDEAL-D study confirms lung resections and thymectomies are feasible with the use of Versius system, laying the foundation for larger phase 2b and 3 clinical studies within the IDEAL-D framework.

Open in new tabDownload slide

INTRODUCTION

The use of robotic-assisted thoracic surgery (RATS) is making headway worldwide [1]. Multiple studies have analysed RATS outcomes in comparison with open and video-assisted thoracic surgery (VATS). Although RATS has shown improved results over to thoracotomy [2, 3], there is not enough evidence to determine its superiority to thoracoscopy yet. However, some studies have started to signal promising results: increased rates of lymph node harvested [4, 5], shorter hospital stays [2, 6], decreased conversion rate [2] and fewer intraoperative blood loss [6].

Benefits of RATS for surgeons are evident: fully ‘wristed’ instruments, high-resolution magnified and binocular three-dimensional (3D) view, amplified surgical precision and dexterity, physical strain reduction and more ergonomics, and potential teaching superiority [7]. As a result, more thoracic surgeons have adopted this approach to operate lung surgery and mediastinal tumour resections, and they are succeeding in performing complex cases as well, such as anatomical resections after neoadjuvant treatment or bronchovascular sleeve resections, with the use of RATS.

Several robotic systems are being currently implemented, including the Versius surgical system (Cambridge Medical Robotics [CMR] Surgical Ltd, Cambridge, UK), which incorporated surgeons feedback in its development [8]. This relatively new robotic system has a modular design, with an open console and 3 or 4 individual bedside units (BSUs) that control 5 mm wristed robotic instruments (not docked to the trocars). The design enables flexibility in port placement facilitating its transfer and set-up, and an easy conversion to thoracoscopy when necessary. The open console, which has solely hand controllers, can be adjusted to either a standing or sitting operating position and helps to maintain a communicative environment within the operating room. The endoscope camera provides surgeons with 3D high-definition video feedback [9].

The IDEAL-D (Idea, Development, Exploration, Assessment, Long-term follow-up—Devices) framework has been established as a guidance for improving the evidence base from research during each step of surgical innovation [10]. A preclinical study stage 0/1 of the IDEAL-D framework was previously conducted, demonstrating proof of concept for the use of Versius in thoracic surgery [11]. As some of the earliest adopters of this technology in thoracic surgery, the aim of this study is to present a prospective evaluation of Versius in an IDEAL Stage 2a study (Development), focusing on performance, technical details and feasibility of the thoracic procedures in live humans.

MATERIALS AND METHODS

Ethical statement

The present study was approved by the Institutional Review Board and Hospital Ethics Committee (code 244/23). Informed consent was obtained by all patients to undergo surgery, as well as prospective data collection in a specific registry.

Study design

The study included 30 consecutive patients who received robotic thoracic surgery using the CMR Versius Surgical Robot System at a single institution between April 2023 and December 2023. Patients underwent lung resection (wedge, anatomical segmentectomy or lobectomy) with systematic lymphadenectomy or thymectomy. There were no specific predetermined selection criteria for individual cases.

Preoperative data was gathered from each patient, such as baseline demographics, comorbidities and the American Society of Anaesthesiologists Physical Status Classification System (ASA).

The primary outcome was the safe completion of the procedure without unplanned conversion to VATS or open surgery. Secondary outcomes included intraoperative complications and complications occurring during the hospital stay or within 90 days after discharge (graded according to Clavien-Dindo classification) [12]. Additional secondary end-points were set-up/docking time for BSUs and robotic console time, console alarm, readmissions to hospital and reoperations within 30 or 90 days, length of stay and 30-day mortality. Set-up time was defined as the time elapsed from the first port training to the last one. Docking time was described as the time elapsed from the first skin incision to the end of port training. Console time refers to the time required by the surgeon to complete a specific surgical procedure in the console. Console alarm refers to a system alert in which the console can no longer be used for surgery and all BSUs connected to the console disengage. Pathology results collected entailed histologic type, pathological size and pathologic stage (according to the 8th IASLC TNM staging in cases of lung cancer and Masaoka staging for thymic tumours).

Surgical team

The team comprised of 2 surgeons who participated alternatively as lead surgeon or bedside surgeon. Both had extensive experience in VATS (>50 cases per annum), but neither in RATS (there was no robotic thoracic surgery program at the hospital at that time) nor with the Versius. All surgeons participated in a tailored 3.5-day training (including 2 days of wet-lab with cadaver) with 2 nurses, who attended the first cases. The study began immediately after completing the 3-day training. The first thymectomy and the second lobectomy were proctored by an external surgeon.

Operating room layout and surgical technique

Thymectomy

Thymectomies or mediastinal tumour resections were performed with 2 instrument BSUs and 1 visualization BSU. Patient preferred position was the same as in VATS anterior mediastinal tumour resections. BSU locations and port placement in left-side thymectomies are shown in Fig. 1A and B. Arms were given a Z-shape (Fig. 1C). An 11-mm camera port and two 5-mm working ports were inserted. Instruments used were a fenestrated grasper and a bipolar Maryland grasper (Fig. 1D). This procedure was performed under insufflation of CO². A 30-degree face-down endoscope was used. The decision of approaching from the right or left side was determined by the tumour location. An external ultrasonic energy device could be used to section thymic veins, which was inserted through one of the 5 mm ports after removing the robotic instrument. All patients underwent an intercostal block at the beginning of the operation. No changes to system setup, instruments and operative technique were necessary during the study.

Lung resection

Robotic anatomical and non-anatomical lung resections with lymph node dissection were performed utilizing 2 instrument BSUs and 1 visualization BSU. The patient was positioned as in VATS procedures, except for the angulation, which was placed lower towards the patient’s hip. Lung resection approach was configured and decided based on surgeons’ experience in VATS and training in the laboratory. The recommended BSU positions are detailed in Fig. 2A. Instruments BSU position can slightly vary depending on the lobe of resection and the anatomy of the patient, and it may need to be repositioned during the procedure if the instrument cannot reach a specific location in the operative field.

An 11-mm camera port was set at the 8th intercostal space as well as a 5-mm working posterior port (Fig. 2B). An anterior 3-cm minithoracotomy was performed at the 5th intercostal space and protected with a soft tissue plastic retractor (Fig. 2B). For upper lobectomies, the camera port was placed anterior to the posterior axillary line, in order to avoid clashes with the patient’s hip. Our port placement differed from the one described in the previous published phase 0/1 study [11], since the surgeons preferred to minimize the number of incisions. At the beginning of the series, arms were given a C-shape. During the study, instrument arms were switched to Z-shape (Fig. 2C), as it decreased the necessity of repositioning BSUs.

This procedure was not performed under insufflation of CO². A 30-degree face-down endoscope was used. A bipolar Maryland grasper was inserted in the right hand (without trocar through the utility incision) and a fenestrated grasper was inserted in the left hand (through the posterior 5 mm trocar) (Fig. 2D). An external endostapler was necessary to carry out the procedure. The bedside surgeon introduced the endostapler through the utility incision, as well as suction and assisted traction when necessary. An ultrasonic energy device or an endoclip applier could be required to dissect and ligate small pulmonary veins or arteries. All patients underwent an intercostal block at the beginning of the operation.

Statistical analysis

For descriptive analysis, continuous variables were tested for normal distribution (Shapiro-Wilk test). Normal distributed variables were reported as mean and standard deviation, while non-normal as median and interquartile range (IQR). Categorical variables were stated as absolute (count) and relative (percentage) frequencies. All statistical analyses were performed using STATA/IC 16.1 (STATACorp, Texas, USA).

RESULTS

Patients’ baseline characteristics are shown in Table 1. Twenty-five patients underwent a lung resection and 5 a thymectomy, of these, 28 (93.3%) were safely completed without conversions. There were no conversions to open surgery.

Lung resection

Twenty patients (80%) underwent a lobectomy, 3 (12%) patients an anatomical segmentectomy and 2 (8%) a wedge resection. All the procedures were accompanied by a systematic mediastinal lymphadenectomy, including wedge resections. None of the patients had undergone neoadjuvant treatment. There were 22 (88%) patients without pleural adhesions, 2 (8%) with firm adhesions and 1 (4%) with soft adhesions. Lobectomies performed are summarized in Table 2. Anatomical segmentectomies performed were 2 segments 6 and 1 segment 1.

Regarding the intraoperative outcomes, 2 controlled conversions to VATS were experienced: one due to pulmonary artery bleeding during a right upper lobectomy and the other due to a console alarm during a S6 segmentectomy. In 5 procedures (20%), an external energy device was used to seal a pulmonary vein or artery. In the lobectomy group, 10 (50%) patients had complete or almost complete fissure, 8 (40%) patients had incomplete fissure and 2 (10%) patients had absent fissure. Median set-up and docking time were 9 (8–12) and 20 (17–20) min. Median console time was 103 (90–129) min. Set-up, docking, and console times tendencies are shown in Fig. 3A and B. Median console time in lobectomies was 104 (90–138) min. One (4%) console alarm was encountered.

Four (16%) patients experienced prolonged air leak, 1 (4%) a haemothorax requiring reintervention and postoperative blood transfusion, 1 (4%) a urinary tract infection, 1 (4%) a pneumonia and 1 (4%) a congestive heart failure decompensation. Thus, only 1 (4%) patient encounter a complication Clavien Dindo grade III or more. In the case of the haemothorax reintervention, a mild bleeding from a vascularized adhesion was found. Median hospitalization was 3 (2–4) days. There were no readmissions and no mortality within 30 and 90 days.

Lung cancer was diagnosed in 22 patients (88%), pulmonary metastasis in 1 patient (4%) and a benign tumour in 2 patients (8%). All cases had negative surgical margins. Mean size of the primary lesion was 1.99 ± 1.18 centimetres. Median number of total, mediastinal and hilar nodes harvested in anatomical resections were 9 (7–11.5), 5 (4–7), and 4 (3–5), respectively. Lung cancer staging and histology are detailed in Table 3.

Thymectomy

Five patients (16.67%) underwent a robotic total thymectomy. None of the patients had myasthenia gravis. No pleural adhesions were observed in this group. Neither intraoperative complications nor conversions were registered. In 1 case (20%), an external energy device was used to seal thymic veins. Median set-up, docking, and console times were 8.5 (7.5–10.5), 20 (17–20) and 77 (75–89) min, respectively (Fig. 3C and D). No console alarms were experienced.

There were no postoperative complications in this group. Median length of stay was 1 (1–1) day. There were no reinterventions, readmissions and no mortality within 30 and 90 days.

Four (80%) patients were diagnosed with thymoma (3 Masaoka Stage IIa and 1 Masaoka Stage I) and 1 (20%) patient with a thymic cyst. All cases had negative surgical margins. Mean size of the primary lesion was 3.42 ± 1.68 centimetres.

DISCUSSION

This is the first publication that we know of describing the implementation of Versius in real patients for thoracic surgery. Our cohort includes Spain’s first Versius robot-assisted lung resection and thymectomy surgeries demonstrating that surgeons were able to successfully perform multiple robot-assisted procedures safely, with few surgical difficulties even without previous robotic experience with other systems. Out of the 30 procedures reported, 28 were successfully completed without conversion to another surgical modality. In the 2 conversions, surgery was effectively accomplished by VATS without needing to convert to thoracotomy. One of the converted surgeries was device-related, as a result of a console alarm which hindered the surgeons from continuing operating. Hence, only 1 conversion (3.3% or 4% if only lung resections are considered) could be associated to the technique. Robotic conversion rate in the literature ranges from 0.8% [7] to 8.6% [13] in lung resections and from 0% to 7% in thymectomies [14]. Bearing this in mind, the conversion rate in this study can be deemed acceptable.

Median console time was 103 and 77 min in lung resections and thymectomies, accordingly. A decreased tendency in median console times throughout the study was not observed since a wide range of procedures were performed within the cohort, from a wedge to a right upper lobectomy with a suture closure of the bronchus stump. Operating time in lung procedures were similar or shorter compared to those previously reported with other robotic systems with standard 4/5 arm approaches [15, 16], and to 3-incision approaches [17] (the latter are more similar to our method). Nevertheless, it is important to recognize that these results are preliminary, and caution must be exercised due to the study’s limited sample size. Operating time in the lobectomy subgroup and in mediastinal procedures were also shorter than those identified in previous initial reports with other robotic platforms [18, 19]. This fact may be attributed to extensive experience of both surgeons in VATS and to the ‘hybrid technique’ they adopted in lung resections. Contrary to console time, a decrease tendency in set-up and docking times was achieved as the team gained knowledge about Versius components functioning. Set-up and docking times were comparable to those obtained in the early use of other robotic systems for these procedure [20, 21]. However, this small series was not intended to evaluate learning curves. Further work is required to establish the surgical learning curve for the Versius Surgical System in different thoracic surgery procedures in line with IDEAL-D Stage 2b and 3.

Postoperatively, 20% of patients experienced a complication in lung procedures, and 25% in the lobectomy subgroup. This low morbidity rate is similar to robotic lung resection outcomes reported by Cerfolio et al. [22] and to lobectomy results of Veronesi et al. [23] initial experience study. Lengths of hospital stay in the lobectomy subgroup were comparable to patients undergoing either VATS [24] or robotic procedures (with other devices) [22, 23].

Thymectomy outcomes (no complications and median hospitalization of 1 day) are considered remarkable compared to early reports of other platforms in the literature [19, 22]. Again, this could be due to the unit’s previous experience with VATS mediastinal resections.

Studies on lymph nodes harvested in robotic thoracic lung resections suggest that robotic-assisted techniques can yield an equal or even higher number of lymph nodes compared to open surgery or VATS. In our study, the median of total lymph nodes harvested in anatomical lung resections was 9 nodes, which is slightly lower than in previously published multicentre randomized trials, such as RAVAL⁵ (10 nodes) and RVlob trial [4] (11 nodes). However, it is necessary to highlight that the sample size in our study is much smaller and the team’s limited experience in robotic surgery. Overtime, the number of lymph nodes removed tends to increase alongside surgeon robotic expertise.

The design of the Versius robotic system has shown both advantages and disadvantages. Commonly encountered technical challenges were the collisions of the arms, which improved over the course of the 30 cases as a result of surgeons progress and procedure repetition. Although the collisions could increase surgical time, they did not cause any accidents, conversions or hinder surgery completion. Moreover, the inability to reach with the tip of the instruments certain points of the operative field, particularly those distant from the anatomical target, has been observed (e.g. when performing a 4R lymphadenectomy having set up the BSUs for a lower lobectomy). Collins et al. [25] reported similar challenges in colorectal surgery as well. This may entail the need to reconfigure the set-up and reposition some BSUs, thereby briefly increasing the surgery time. Inherent to working with a new system, the absence of some commonly used instruments such as endograspers or advanced energy devices was noted. To address this drawback, we adopted an ‘hybrid robotic surgery’ technique, incorporating a utility incision to manually introduce these instruments. The development of advance energy devices, as well as thoracic specific instruments, would benefit performance and wider adoption of the system.

Given these limitations, several advantages should be highlighted, including the use of 5 mm trocars to introduce the instruments, which, coupled with the perception of low force exerted by the instruments in the intercostal space, may lead to less postoperative pain. Results from pain and quality of life questionnaires are currently being collected to analyse these outcomes with more patients in the future. On the other hand, the 3D images provided by Versius allow for more precise dissection of thoracic structures. Surgeons involved in the study stated that although the procedures have been carried out in a very similar manner to how they were performed by VATS, the robot provides better ergonomics and greater ease in dissecting vascular structures and suturing facilitated by ‘fully wristed’ instruments.

Previously, IDEAL-D Phase 2 and 3 studies have been published in colorectal [22], urological [26] and cholecystectomy procedures [27], but none in thoracic surgery. However, there is a published prospective multicentre registry of 2083 procedures of multiple specialities that includes 19 thoracic cases [28]. Their outcomes are hardly comparable to those in this study due to the different surgeries included and surgeons previous RATS experience. We consider that our study provides a larger sample size and more technical details and evaluates a wider range of variables to obtain more robust results regarding efficacy and safety.

Nonetheless, several limitations need to be considered. As mentioned above, surgeons lacked previous experience in robotic surgery, therefore, the patients enrolled in the initial cases may represent a highly selective cohort of low-risk individuals. This is a fact present in other early experience reports of a robotic system, to the extent that this patient selection is recommended in the implementation of a robotic program [18]. In addition, the indicated lack of robotic experience enabled us to compare our outcomes with an alternative robotic platform early series reports by pioneering groups, which has since become widely acknowledged and utilized in contemporary practice. Lastly, this study remains a descriptive observation by a single institution with a limited sample size, but this is consistent with the nature of IDEAL-D phase 2a studies.

CONCLUSION

This is the first report to demonstrate the use of the Versius robot platform in thoracic surgery in live humans. Our results demonstrate that the adoption and implementation of Versius is safe and feasible for lung resections and thymectomies and can be used in such procedures with remarkable outcomes, regardless of a surgeon’s prior robotic experience. This study broadly aligns with IDEAL-D Stage 2a (Development) and supports the implementation of Versius on a wider scale. Following the IDEAL-D recommendations, further studies are planned.

ACKNOWLEDGEMENTS

The authors thank CMR Surgical implementation specialists Miguel Nuñez and Juan Carlo Solís for their continuous support in the operating room during the procedures, and Sofia V. Clayton and Cristina Urdaneta for language editing.

FUNDING

None declared.

Conflict of interest: none declared.

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request to the corresponding author.

Author contributions

Sara Fra: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Validation; Visualization; Writing—original draft; Writing—review and editing. Usue Caballero-Silva: Conceptualization; Data curation; Investigation; Methodology; Validation; Writing—review and editing. Alberto Cabañero-Sánchez: Methodology; Supervision; Validation; Visualization; Writing—review and editing. Gemma María Muñoz-Molina: Methodology; Supervision; Validation; Visualization. Cristina Cavestany García-Matres: Conceptualization; Investigation; Visualization. Jose Deymar Lozano-Ayala: Conceptualization; Investigation; Visualization. Luis Lomanto-Navarro: Conceptualization; Investigation; Visualization. Elena Vílchez-Pernias: Conceptualization; Investigation; Visualization. Nicolás Moreno-Mata: Conceptualization; Investigation; Methodology; Project administration; Supervision; Validation; Visualization; Writing—review and editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Alain Bernard and the other anonymous reviewers for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• BSU

  Bed-side units

 
• CMR

  Cambridge medical robotics

 
• IDEAL-D

  Idea, Development, Exploration, Assessment, Long-term follow-up—Devices

 
• RATS

  Robotic-assisted thoracic surgery

 
• VATS

  Video-assisted thoracoscopic surgery