Preclinical evaluation of the newly developed carina platform in robotic-assisted proctectomy with porcine and cadaveric models Authors: Yifei Feng, Junwei Tang, Yi Zhang, Dongsheng Zhang, Wenbin Zhou, Li Li, Yueming Sun Published: 2026 ABSTRACT AbstractRobotic-assisted laparoscopic surgery has made progress in addressing many of the technical challenges associated with conventional laparoscopy. Recently, a new robotic surgical system, Carina Platform, has been developed. The aim of this study is to evaluate the feasibility and safety of Carina for proctectomy surgery in preclinical models. Port and cart positions were initially determined by surgical simulation based on three-dimensional reconstruction. Six pigs, divided equally into acute and chronic groups, underwent robotic-assisted proctectomy using the Carina robotic system. Operative time, device errors and perioperative complications were recorded. All animals were observed postoperatively to assess their mental status, and after four weeks, the chronic pigs were euthanized for necropsy to evaluate the recovery of the anastomotic stoma. The optimal port and patient cart positions were further assessed in three human cadavers. Robotic-assisted proctectomy was successfully completed in all porcine subjects. No device-related intraoperative complications were observed. An autopsy examination revealed that the chronic pigs, which had been followed up for 28 days, had made a favorable recovery. The configurations of ports and patient carts were successfully validated in human cadaver models that could simulate the completion of a rectal resection. The feasibility and safety of the newly developed Carina robotic system for performing proctectomy were successfully demonstrated in porcine and cadaveric models. Further studies are required to validate its clinical application in human patients.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1038/s41598-026-43313-1.Keywords: Carina Platform, Three-dimensional reconstruction, Preclinical models, Robotic-assisted proctectomySubject terms: Rectum, Gastrointestinal models Introduction Laparoscopic surgery is the standard of care across various surgical specialties and is recommended in guidelines1,2. Many surgical departments routinely operate on gastrointestinal perforations, incarcerated hernias and bowel obstruction laparoscopically. Compared to traditional open surgery, laparoscopic surgery offers numerous advantages, including less trauma, faster recovery, and less pain3. However, laparoscopic surgery is technically challenging and usually increases the learning curve to perform a safe and effective procedure, especially when operating within restricted anatomical spaces. These challenges stem from limitations such as restricted instrument motion, a two-dimensional operative view, difficulty in performing precise suturing, and involuntary tremor4–6. Robot-assisted surgery has advanced the capabilities of laparoscopic surgery in more delicate and complex procedures by providing a magnified three-dimensional view, improved dexterity, tremor filtration, improvement of surgeon ergonomics, and great precision of movement7–9. Consequently, robotic surgery has been an integral part of general surgery for over two decades after the clinical introduction of the da Vinci Surgical System (Intuitive Surgical, Inc.). In 2018, there were more general surgery robotic procedures performed in the United States than any other specialty, reaching up to 15.1% of all general surgical operations in the USA10. The da Vinci surgical system is currently pioneering robotic-assisted surgery. However, its high acquisition and maintenance costs make it inaccessible to many healthcare facilities11. In response, a series of new robotic surgical platforms have been developed in China, supported by national policies and funding, aiming to break the market monopoly held by the da Vinci system12–14. The reduction in cost, coupled with the ability to meet clinical requirements, will facilitate the expansion of robotic surgery to a greater number of medical facilities. Nevertheless, prior to the commercialization of robots, it is imperative to conduct preclinical studies to verify their safety and efficacy. The Carina platform (Ronovo Surgical, Shanghai), is a new modular robotic surgery system with independent patient carts. Its patient carts have specific characteristics and flexibility due to the technical design. To assess the feasibility and safety of the novel Carina robotic surgery system, we conducted a two-year experimental study using animal and cadaver models across various surgical disciplines, including urology, gynecology, thoracic, and general surgery. Herein, we present our first experience with the newly developed robotic surgical system-assisted rectal surgery in porcine and cadaveric models. Materials and methods Carina Platform The Carina Platform consists of four independent patient carts, a surgeon console, and an integration hub (Fig. 1A). The system supports both straight-stick instruments and wristed instruments, allowing surgeons flexible choice of instruments (Fig. 1B). The surgeon console has an immersive, stereoscopic viewing window, two hand manipulators, a set of foot pedals, and an interactive screen that allows the user to achieve an optimal ergonomic position. The integration hub serves as an intermediate unit between the surgeon console and patient carts, undertaking the functions of information transmission, peripheral connections, and system control during operation. Each patient cart is equipped with an individual robotic arm that has eight joints that contribute to ease of movement in the surgical field. Instruments can be installed on any one of the instrument-holding patient carts, while the endoscope can be installed on the scope-holding patient cart. A distinguishing feature of the system is the modular design of the patient cart, which enables independent configurations of different operation spaces (Fig. 1C) and cart distances (Fig. 1D) that combine to optimize accessibility to the surgical field. Monopolar curved scissors, fenestrated bipolar forceps, large needle drivers, monopolar cautery hooks, fenestrated forceps, atraumatic grasping forceps, and visceral grasping forceps were used in this study. The Carina Platform. (A) Patient carts, surgeon console, and integration hub. (B) Wristed and straight-stick instruments. (C) Illustration of operation space settings: -30°and + 30°positions. (D) Illustration of cart distance. Surgical team All surgical procedures were performed exclusively by a lead surgeon with robotic surgery experience supported by a surgical team. The lead surgeon was responsible for carrying out the surgical steps associated with the procedure and for evaluating the system in accordance with the specific objectives of the study. Meanwhile, the assistant surgeon was responsible for carrying out any additional manual tasks, such as inserting and removing sutures and gauze and retracting tissue to assist with the surgery. In addition, personnel present recorded the placements of ports and positions of each patient cart, along with the subsequent outcomes. Before commencement of the study, all users engaged in sufficient training to properly perform surgical procedures using the Carina. Study design Prior to preclinical labs, we used three-dimensional simulation techniques to inform the plan for surgery. This was an observational study with live animal and human cadaver subjects. The objective of the live animal experiments was to evaluate the feasibility and safety of Carina in proctectomy by examining information obtained intraoperatively (including vital signs, intraoperative complications, technical malfunction, maneuverability of the instruments, and instrument handling of live tissue) alongside survival outcomes of a 4-week follow-up period. Thereafter, the primary objective of the human cadaver experiments was to ascertain the accessibility of instrument manipulation in rectal surgery and the ability to simulate the completion of an entire procedure in a realistic human anatomical environment. All procedures were conducted in a simulated operating room in accordance with the standards and protocols typically observed in a clinical setting. The live porcine experiments were performed during July and August 2023 at a third-party preclinical laboratory (Shanghai Harborside Medical Technology Co., LTD) which is certified by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) and specializes in animal studies in the field of medical device development. The experimental protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of Shanghai Harborside (IACUC-2023-074). All animal studies were performed in accordance with the institutional guidelines and complied with the ARRIVE guidelines. These experiments were conducted in accordance with Good Laboratory Practice and were designed to align with the 3R principle, which seeks to reduce the number of animals used in scientific research. Thereafter, human cadaveric studies were conducted with the approval of the Institutional Ethics Committee of the First Affiliated Hospital of Nanjing Medical University. Informed consent was obtained from the next of kin of the cadaver donor for the conduct of this study. All human cadavers were donated with informed consent following the guidelines of the China Organ Donation Administrative Center. Port placements and cart positions We used three-dimensional simulation techniques in Blender (Blender Foundation, The Netherlands) to determine optimal port placements and cart positioning. A three-dimensional human model was placed in the Trendelenburg position, and the height of the pneumoperitoneum was simulated to achieve a level as close as possible to that of the actual human body. According to the established laparoscopic techniques and the surgical preferences of the surgical team, four ports (one camera port and three instrument ports) were mapped onto the three-dimensional human model (Fig. 2A), while the points and trajectories where the instruments need to be passed during the proctectomy procedure were marked on the three-dimensional model. In the reachability simulation of the proctectomy procedure, the set range of surgical operation areas include rectal region, sigmoid region, descending colon region, splenic flexure. A distinctive trajectory was established for each specific operation area within the surgical simulation. The spatial coordinates of the points along the trajectory are employed as the target point of the instrument’s motion. Using the patient cart coordinates and the instrument target coordinates, we can perform inverse kinematics analyses to calculate the initial positions and orientation of the cart by using MATLAB software as follows: (1) Each cart is placed in a fixed position, extending a distance of 1.5 m each in x and y directions as the search area; (2) The range of motion in which the cart can reach and be maneuvered is restricted by the fixed point of the incision port; (3) The trajectory information is utilized as instrument target point input, with the solution to the inverse kinematics of the cart determined at 5-cm intervals within the specified search area. If the arm has a solution for all trajectory points, the position is marked with a blue dot. Conversely, if there is no solution for at least one point, the position is marked with a red dot. Subsequently, the optimal locations for each cart within the operating room were identified based on the given operating room layout and surgical team preferences (Fig. 2B). All the carts were then subjected to collision analysis in accordance with the surgical trajectories using three-dimensional model calculation software (PyBullet) to establish optimal positioning of the patient carts (Fig. 2C). Finally, the findings were utilized as an informational output to develop a dedicated procedure card for routine clinical use. Planning of the incisions and the placement of patient carts through three-dimensional simulation techniques. (A) The white point represents the port incision, while the red tracings depict the respective target trajectory. R: for right instrument; C: for camera; L: for left instrument; 4: for 4th instrument. (B) For each cart, the blue curvature within the green box represents the location designated for the cart placement. (C) Layout of the patient carts. Porcine experiments The anatomy and physiology of the abdominal cavities of pigs are similar to those of humans, and their body size is also conducive to laparoscopic surgical operations. Animal testing was conducted on 6 healthy pigs, obtained from Changzhou Beiwang Biotechnology Co., Ltd. (Jiangsu, China), which were divided randomly into acute group (N = 3) and chronic group (N = 3). Prior to the surgical procedures, the porcine subjects were evaluated to determine their preoperative health status. This evaluation included a physical examination, blood chemistry analysis, and quarantine. Each subject underwent preoperative fasting for 12 h. After being anesthetized, the subjects received tracheal intubation and were placed in Trendelenburg position, after which the surgical field was thoroughly sterilized. The procedure commenced with the insertion of a Veress needle below the umbilicus, which was followed by the creation of pneumoperitoneum. Following the insertion of the camera port in the corresponding location, three robotic trocars and an additional assistant trocar were placed in the abdomen under laparoscopic vision to ensure that no intra-abdominal organs were injured during port placement (Fig. 3A). Positioning the patient carts around the surgical table involved three steps: (1) deploying each robotic arm using buttons on the patient cart to achieve a configuration where the cart distance is appropriate for the procedure; (2) placing each patient cart at the location indicated on the procedure card with the robotic arm pointing at the corresponding port site; and (3) locking the patient cart in position (Fig. 3B). After docking Carina to the trocars, the endoscope and the robotic instruments were inserted. Port placements and patient cart positions in porcine subjects. (A) Port positions tested in porcine model.12 mm for assistant port. 10 mm for camera port. 8 mm for right instrument cart assembled with monopolar electric hook (or bending shear) and large needle holding forceps generally. 8 mm for left instrument cart assembled with bipolar window pliers. 8 mm for the 4th instrument cart assembled with long viscera grasping forceps (no energy). (B) Patient cart positions tested in porcine model. After all the preparations were completed for robotic proctectomy, console operation commenced under stereoscopic vision (Fig. 4A-D). During the operation, the inferior mesenteric artery (IMA) and vein (IMV) were exposed and isolated, while a dissection in the mesorectal plane was performed. The sigmoid colon was pulled ventrally to expose the origin of the IMA. The coated IMA surface was cut with a bipolar electric hook and the Toldt gap was entered. The loose tissue around IMA was bluntly separated to expose the vascular sheath. The main IMA trunk (1–2 cm from the aorta) was clipped with hemolock and dissected. The IMV was found at the lower edge of the pancreas (often adjacent to the ligament of Treitz). The tissue surrounding the IMV was slowly dissociated using the bipolar electric hook, and the IMV was disconnected after hemolock clampling. Then, a subsequent lower rectal transection was performed using an endoscopic linear cutter- stapler (EC-60, Ethicon Endo-Surgery). Following the withdrawal of the robotic system, the rectum was pulled out of the pelvic cavity via a longitudinal incision and part of the proximal rectum was resected, which was more than 3 cm in length (Fig. 4E-F). The rectum was anastomosed using the double-stapling technique; the anvil of a circular stapling device (CDH29, Ethicon Endo-Surgery) was placed into the rectal stump and secured via purse-string suturing. After placing the bowel back into the abdomen and closure of the incision, the pneumoperitoneum was maintained and the robotic system was reconnected. A circular stapler was then introduced through the anus and an end-to-end anastomosis of the rectum was performed to restore intestinal continuity. Key steps of proctectomy surgery in porcine model from acute group. (A) Exposure of the mesorectum. (B) Complete exposure of the rectum. (C) Closure of the distal rectal segment using a straight-line stapler. (D) End-to-end anastomosis of the rectum. (E) Specimen of resected rectum. (F) Abdominal incisions sutured. For each procedure, the surgeon performed rectal resections on the rectum and lower part of the colon. All instruments available were utilized for all operations on live tissues to test the robotic system’s ability to perform common surgical maneuvers, such as isolation, division, grasping, dissection, vessel coagulation and suturing on living tissue. During each procedure, blood loss, technical malfunctions (usually include collision between robotic arms or between robotic arms and assistants, control failure, error reporting, out-of-sync laparoscopic images, system breakdown, etc.), intraoperative complications, and surgeon feedback regarding the system and instruments in use were collected. After completion of each procedure, the acute subjects were observed for at least one hour, during which time their mental status were monitored, and then humanely euthanized for anatomical autopsy evaluation. The chronic subjects were transferred to the observation facility where they were observed and fed for 28 days while any adverse events were recorded. The chronic pigs were observed for vital signs at least once daily by a veterinarian according to the following parameters: diet, coat condition, feces, urine, mental status, voluntary activity, and respiration. Furthermore, a series of hematological and plasma biochemical tests were conducted at the following time points: three days post-surgery, 14 days post-surgery and 28 days post-surgery. Following a four-week period, the chronic subjects were weighed and euthanized. The pigs were anaesthetized with intramuscular injections of zoletil 50 (5–10 mg/kg) and propofol injectable emulsion (1–8 mg/kg), and subsequently euthanized with a 10% potassium chloride solution at a dosage of 1 ml/kg. Afterwards, they underwent a detailed examination of the surgical sites and an assessment of the success of organ removal and the integrity of the anastomosis. Cadaver experiments The human cadavers in these experiments were fresh-frozen, having been rapidly frozen after death according to standard policy15, and thawed at room temperature prior to the commencement of the study. The cadavers selected for this study were chosen to represent a wide range of body mass indices (BMI), and none had history of abdominal surgery. After all the preparations were completed, the procedures were performed in the following steps16: sigmoid retraction, isolation and division of the inferior mesenteric artery, dissection of the left colon, splenic flexure mobilization, total mesorectal excision, rectal isolation and transection, and rectal suturing to conclude. For these steps, the robotic instruments were used to reach extremely caudal, lateral, ventral, or dorsal surgical targets. The success of the procedure was evaluated based on the ability of the robotic system to perform each step in a satisfactory manner, with adequate surgical access and without major robotic arm collisions. Statistical analysis All statistical analyses were performed using SPSS 26.0. Descriptive statistics were used to summarize the key variables of the study. Continuous variables were summarized using mean and standard deviation to provide distribution within the data. T-tests were used for between-group comparisons between the acute and chronic groups and for within-group comparisons in the chronic group. Result Porcine experiments Six robotic-assisted proctectomy procedures were performed in porcine models. Preoperative weight was not statistically different between the acute and chronic groups (45.62 ± 3.88 vs. 48.45 ± 3.09, P value = 0.38). The porcine models of chronic group showed significant increase in several blood measurement parameters at 3 days and 2 weeks after surgery, including white blood cell count (26.09 ± 0.59 vs. 14.96 ± 1.40, P < 0.05), neutrophil count (11.29 ± 0.78 vs. 4.71 ± 1.02, P < 0.05) and Monocyte count (12.48 ± 0.40 vs. 9.40 ± 0.23, P < 0.05) and so on, as shown in Table 1. Table 1The baseline and blood measurement parameters of porcine models.Acute group (N = 3)Chronic group (N = 3)P valuePreoPreoPosto-3dPosto-2wPredWeight (kg)45.62 ± 3.8848.45 ± 3.09--50.25 ± 1.670.38Red blood cell count(M/µL)6.66 ± 0.345.98 ± 0.216.25 ± 0.547.29 ± 1.106.57 ± 1.020.04Hemoglobin(g/dL)11.70 ± 0.5010.60 ± 0.5010.90 ± 1.0013.00 ± 1.40*11.40 ± 1.400.05Hematocrit(%)36.10 ± 2.5032.40 ± 1.7032.90 ± 3.7041.20 ± 6.7035.80 ± 5.500.10Mean corpuscular volume(fL)54.20 ± 1.5054.10 ± 1.1052.60 ± 1.5056.40 ± 0.70*54.40 ± 0.300.93Mean corpuscular hemoglobin(pg)17.60 ± 0.2017.70 ± 0.2017.40 ± 0.3017.90 ± 1.0017.40 ± 0.500.57Mean corpuscular hemoglobin concentration(g/dL)32.50 ± 1.0032.80 ± 0.3033.10 ± 0.6031.70 ± 2.1032.00 ± 0.900.64Reticulocyte count(K/µL)24.90 ± 5.1021.60 ± 5.9035.20 ± 6.1066.80 ± 36.9034.10 ± 14.000.50Platelet count(K/µL)377.00 ± 113.00297.00 ± 34.00343.00 ± 67.00254.00 ± 27.00342.00 ± 52.000.31White blood cell count(K/µL)22.71 ± 5.3914.96 ± 1.4017.55 ± 6.2321.14 ± 3.9926.09 ± 0.59*0.07Neutrophil count(K/µL)7.09 ± 1.444.71 ± 1.025.43 ± 5.026.47 ± 1.6011.29 ± 0.78*0.08Lymphocyte count(K/µL)13.25 ± 4.649.40 ± 0.2310.45 ± 1.9212.74 ± 2.1212.48 ± 0.40*0.22Monocyte count(K/µL)1.13 ± 0.180.77 ± 0.101.08 ± 0.231.39 ± 0.431.30 ± 0.12*0.04Eosinophil count(K/µL)1.13 ± 0.490.08 ± 0.070.57 ± 0.610.51 ± 0.11*0.99 ± 0.52*0.02Basophil count(K/µL)0.11 ± 0.080.01 ± 0.000.01 ± 0.010.03 ± 0.01*0.03 ± 0.020.10Albumin(g/dL)3.00 ± 0.102.80 ± 0.102.80 ± 0.103.00 ± 0.102.50 ± 0.10*0.07Alkaline phosphatase(U/L)118.00 ± 5.00116.00 ± 16.0075.00 ± 4.0090.00 ± 3.0072.00 ± 31.000.85Alanine aminotransferase(U/L)80.00 ± 22.0068.00 ± 2.0077.00 ± 6.0048.00 ± 8.00*38.00 ± 6.00*0.40Creatinine(mg/dL)1.40 ± 0.201.60 ± 0.201.30 ± 0.201.80 ± 0.201.70 ± 0.100.29Globulin(g/dL)3.70 ± 0.204.10 ± 0.104.10 ± 0.205.00 ± 0.50*5.10 ± 0.50*0.04Blood glucose(mg/dL)129.00 ± 50.0067.00 ± 22.0087.00 ± 8.0081.00 ± 11.0077.00 ± 6.000.12Total protein(g/dL)6.70 ± 0.206.90 ± 0.207.00 ± 0.208.00 ± 0.60*7.60 ± 0.500.29All data was shown as mean ± standard deviation.Preo: Pre-operation; Posto-3d: 3-day post-operation; Posto-2w: 2-week post-operation; Pred: Pre-dissect.P value means the Preo measurement of acute group comparing with that of chronic group. * means P value < 0.05 for Posto measurements comparing with Preo data in chronic group. The baseline and blood measurement parameters of porcine models. All data was shown as mean ± standard deviation. Preo: Pre-operation; Posto-3d: 3-day post-operation; Posto-2w: 2-week post-operation; Pred: Pre-dissect. P value means the Preo measurement of acute group comparing with that of chronic group. * means P value < 0.05 for Posto measurements comparing with Preo data in chronic group. All animals maintained normal vital signs throughout the procedures without intraoperative complication and serious malfunction of the robot system and surgical instruments. There was no obvious statistical difference for operation time and blood volume between acute group and chronic group, as shown in Table 2. Three of six were euthanized immediately after the operations (acute group). Autopsy indicated no significant abnormalities in the resected areas or in the rectal anastomosis with a satisfactory blood supply (Fig. 5A-B). The remaining 3 subjects would be observed postoperatively for 28 days (chronic group). Postoperative clinical observations of the subjects revealed no evidence of illness or distress, and all subjects gained body weight normally. The results of hematological and plasma biochemical indices 28 days after surgery were within normal range. Autopsy confirmed that each surgical site demonstrated successful clinical outcome without any unusual occurrences, such as infection, anastomotic leakage, bleeding or hematoma (Fig. 5C-D). Table 2The records of acute and chronic groups during surgery.NumberAcute groupChronic groupP value23P33223P33323P334X ± s23P33523P33623P337X ± sOperation time (min)133126107122 ± 13.45132897297.67 ± 30.920.28Blood volume (ml)5555 ± 05555 ± 0-Intraoperative adverse eventsNoneNoneNone-NoneNoneNone--Serious malfunction of the robot system and surgical instrumentsNoneNoneNone-NoneNoneNone--Operation time was defineded as the time from the skin incision of the animal to the completion of skin suture at the end of surgery. The records of acute and chronic groups during surgery. Operation time was defineded as the time from the skin incision of the animal to the completion of skin suture at the end of surgery. Autopsy findings. Acute group: (A) No abnormalities were observed in the rectum of acute subjects - no evidence of bleeding, anastomotic leakage, or suture looseness. (B) A local view of the rectal anastomosis in an acute subject. Chronic group: (C) The rectal area of chronic subjects exhibited favorable healing trajectory with no discernible anomalies, such as infection, anastomotic leak, hemorrhage or hematoma. (D) A local view of the rectal anastomosis in a chronic subject. Several minor technical errors were observed and immediately resolved intraoperatively, including occasional robotic arm collisions, slight tremor in the robotic arm, docking error of endoscope arm, loading failure of endoscope and needle driver, unresponsiveness of fenestrated bipolar forceps and monopolar curved scissors (Table 3). None of the technical issues resulted in conversion to laparoscopic or open procedure nor compromised animal safety. Table 3Summary of the minimal intraoperative technical errors and troubleshooting.Technical errors type (Frequency)Acute groupChronic groupCadaver groupDescription of solutions23P33223P33323P334Range23P33523P33623P337Range23C00123C00223C003RangeSlight tremor of robotic arm310(0, 3)240(0, 4)210(0, 2)Increase depth of instrument insertionDocking error of endoscope arm to the trocar010(0, 1)000-001(0, 1)Fixing the cart position first and docking secondLoading failure of endoscope001(0, 1)000-010(0, 1)Adjust adapterLoading failure of needle driver000-100(0, 1)000-Remove and reload instrumentFenestrated bipolar forceps and monopolar curved scissors stop responding100(0, 1)020(0, 2)000-Remove and replace both instrumentsTip cover of the monopolar curved scissors breaks000-010(0, 1)000-Replace tip coverOccasional collisions and restricted movement of robotic arms011(0, 1)001(0, 1)110(0, 1)Adjust operation space of robotic arms Summary of the minimal intraoperative technical errors and troubleshooting. Surgeon feedback was also collected to evaluate coagulation efficiency and other technical characteristics, such as grip and traction forces. The robotic surgical instruments used in the operation were found to be satisfactory by the attending surgeons. Cadaver experiments The human cadavers in these experiments presented body mass indices (BMIs) ranging from 22 to 30 kg/m2 (mean: 26.0 kg/m2). Low anterior resection (LAR) was performed in each of 3 cadaveric models to validate the capabilities of the Carina Platform. All cadaveric models were tilted to the right (5–10°) and placed in the Trendelenburg position (10–20°) for this operation. Three different port and corresponding patient cart positions were utilized in these procedures (Fig. 6). During the first two procedures, a few differences were identified in the port and patient cart positions. The first procedure (female, age of 61 years old, height 165 cm, BMI 22) could be completed, but there was an extra-corporeal clash between the left arm and the 4th arm when the left arm isolated the sigmoid colon while the 4th arm lifted the uterus. To address this issue, the camera port placement and the positioning of the left cart were adjusted slightly in the second procedure (female, age of 65 years old, height 163 cm, BMI 30) to accommodate for the increased BMI and to avoid arm collisions. In the third procedure (male, age of 70 years old, height 175 cm, BMI 26), the initial placements of ports and patient carts were optimized to address the operational challenges encountered during the preceding procedures, including arm collisions, limited surgical reach and interference between robotic arms and the table and/or cadaver arm. In addition to the right instrument port placement at the McBurney point, the remaining active ports were all positioned 2 cm above the umbilicus arranged in a horizontal line, with a minimum 8 cm between each port. To facilitate the surgical assistant’s completion of supporting tasks, such as the trans-anal extraction of the specimen, without presenting additional difficulties, the scope-holding patient cart was relocated from between the cadaver legs to above the right of the cadaver head. All planned steps in the third procedure were successfully completed. No robotic instruments collided inside the abdomen, and there were no collisions between the robotic arms and the assistant outside the abdomen. Port and patient cart positions tested in each cadaveric model for LAR. (A) Port placements. (B) Patient cart and surgical assistant positions. Discussion A considerable number of studies have utilized porcine subjects to assess the safety and efficacy of robot-assisted surgery, demonstrating the feasibility of efficiently performing these surgeries in pigs17–19. To investigate the feasibility and the security of the Carina Platform, we performed robotic-assisted proctectomy surgeries on six healthy pigs. All procedures were successfully completed encountering only minor technical issues. It was observed that arm collisions were a common occurrence during surgery on these porcine subjects. However, these collisions were quickly resolved intraoperatively due to the system’s modularity and ease of reconfigurability. No significant perioperative complications were observed and the estimated blood loss was deemed acceptable. Given the proximity of the robotic instruments to major blood vessels during surgery, even a small error, such as an instrument malfunction, could result in catastrophic complications. In this study, the absence of significant complications highlights the effective coagulation capability of the system’s robotic instruments, while the preservation of the integrity of adjacent organs and neighboring tissues indicates the safety of the robotic instruments’ grip and traction forces. These results provide a useful simulation of the system performance expected in live humans and demonstrate the feasibility of using three-dimensional simulation techniques to inform port placement and patient cart positioning. While porcine subjects present similarly to the tissue handling, dissection, surgical plane identification, and control of bleeding in live human surgery, there are meaningful anatomical differences between pigs and humans. Firstly, we found that the shape of the abdominal cavity and pelvic cavity in pigs is different from that in humans, and the operating space is longer and narrower than that in humans. Then, porcine intestines were more curved and stacked, and the anatomical relationship with the surrounding tissues was more complex. Therefore, when we performed rectal resection in porcine model, we stipulated the operation was successful if no less than 5 cm of the arena was resected. Actually, the procedure performed on the porcine subjects in our study may be considered straightforward; therefore, they may not fully reflect the applicability of the tested port placements and patient cart positioning on to live patients. The utilization of cadaveric models affords a more substantial degree of anatomical relevance and realism to live human patients than pigs20. Consequently, Carina was evaluated on human cadaveric models. The cadaver experiments demonstrate that Carina is suitable for robotic-assisted proctectomy surgery. In robotic-assisted laparoscopic procedures, the typical port placement configuration involves three or four ports for the robotic arms and one or two assistant ports21–23. Accordingly, sufficient space must be allocated between each port to ensure freedom of motion for all robotic arms while allowing the bedside assistant to operate effectively. Consequently, an operating surgeon is frequently unable to transfer their preferred conventional laparoscopic port configuration to robotic-assisted procedures. As with our initial two procedures conducted on cadaveric models, collision occurred between the robotic arms, and the surgical assistant lacked sufficient access and range to undertake a series of auxiliary operations. The experience of these two cadaveric experiments suggested that the surface marker (the umbilicus) should not be the only reference for setting port, and the spatial personalized differences of the human abdominal and pelvic cavity should also be taken into account. We believe that taking the distance between the umbilicus and the pubic symphysis as the original reference for ports placement would be more conducive to perfect port-seeting at one time. This is also an important empirical hint for clinical practice. In contrast, after adjustment of port placements and patient cart positions, specially we changed the position of the endoscope cart from the inside of the legs to the right of the cadaver head, no extracorporeal arm collision was observed during the third procedure. This outcome illustrates the advantage of the modular design of the independent patient carts, which allows for a wide spectrum of port placements and enhanced flexibility of robotic arm positioning. Furthermore, Carina’s configurability, wristed instruments, and high-resolution stereoscopic viewing enable adequate surgical access and reach in the pelvis, even in human cadavers with high BMI. Patients with challenging anatomy, such as a narrow male pelvis and obese habitus, or low rectal tumors, may potentially benefit from the port and patient cart configurations validated in our cadaver experiments. Since the introduction of the da Vinci for rectal surgery, several studies have demonstrated the safety and feasibility of robotic-assisted rectal surgery24–28. However, it should be noted that the robotic approach is not a simple one. The experience of the surgical team might have contributed to these results, as all users in our study were trained in the use of Carina and had prior experience performing robotic procedures. Therefore, it is important to train the team efficiently by standardizing surgical techniques and workflows and clearly defining the role of each member of the operating room staff. In addition, compared with the Da Vinci, the split robot has higher flexibility and selectivity in terms of the number of carts and the position arrangement during the operation. This will not only be conducive to carry flexible cost control, but also provide doctors with more space during surgery. Conclusion In conclusion, the Carina Platform, a newly developed robotic surgery system, has preliminarily demonstrated its technical feasibility in robotic-assisted proctectomy. The development and validation of Carina to perform rectal surgery were completed via the utilization of three-dimensional simulation, live porcine subjects, and cadaveric models. Other specialized procedures have been performed in similar developmental sequence. However, preclinical studies cannot provide the full experience of live procedures. As such, the necessity for further clinical trials in live human patients is evident, and we are actively working to promote such trials to bridge the gap between preclinical and clinical settings. In fact, a clinical trial has been initiated, and the feasibility of the Carina Platform to perform rectal surgery has been demonstrated in a case series29,30. Limitation There were some limitations in this study: 1) Although it was reasonable to choose porcine models to verify the safety and efficacy of robotic systems and surgical instruments in performing rectal resection, there were significant anatomical differences between species, so the experience obtained from the porcine models cannot be directly translated to clinical application. The experimental porcine models were healthy, so the surgical effect could not be evaluated on the margin.2) Our experience with cadavers was limited to only three subjects and lack of cadaver studies with BMI above 35. On the one hand, there were significant physical differences between cadaver models and live human subjects, including stiffness and discoloration of cadaver tissues, poor processing fidelity, and lack of bleeding, which made it difficult for us to obtain real clinical experience. There was also a lack of study experience with extremely obese patients. Therefore, this would be an important direction for further research. 3) Due to the limited sample size of animal experiments (N = 6) and cadaver experiments (N = 3) in this study, the surgical experience gained based on this study is finitude as well as its guiding significance for clinical application. In the future, we will verify various surgical procedures on more animals and cadavers. 4) The robotic system used in this study had not yet developed suitable stapler tools, which also greatly limited the application scenarios of robots in surgery. We will consider developing suitable staplers in the next generation of product to make up for the lack of function. Despite these limitations, our positive results provide good simulation of system performance in humans and encourage us to bring it to clinical application in the future. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Author contributions Conceptualization: SY; methodology: FY; material preparation, data collection and analysis: TJ, ZY and ZD; writing-original draft preparation and revise: ZW and LL. All authors read and approved the final manuscript. Data availability All data generated or analysed during this study are included in this published article. Declarations Competing interests The authors declare no competing interests. Ethics approval This study was approved by the Institutional Animal Care and Use Committee (IACUC) of Shanghai Harborside (IACUC-2023-074). Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Data Availability Statement All data generated or analysed during this study are included in this published article.