Robot-assisted radial forearm free flap harvesting: a propensity score-matched case–control study

Although some surgeons prefer anterolateral thigh and latissimus dorsi flap for soft tissue reconstruction in the head and neck area because it minimizes donor site complications, the radial forearm flap remains the workhorse for soft tissue reconstruction due to its reliability. To reduce donor site morbidity, the authors developed a novel technique for radial forearm flap harvesting using a robotic device. 42 radial forearm free flap reconstruction cases were studied, consisting of 31 conventional and 11 robot-assisted cases. 1:1 propensity score matching was done according to age, sex, previous and postoperative radiation therapy history and method used for vein anastomosis. There was no significant difference in flap outcome, which was 100% vitality in the robot-assisted group and 90.9% vitality in the conventional group. The robot-assisted group showed significantly longer mean harvesting time than did the conventional group, being 107.2 min and 67.0 min, respectively. Robot-assisted radial forearm flap harvesting can reduce donor site complications by minimizing incision. When more surgical experience is gained under appropriate case selection, we expect our robot-assisted method will yield a harvesting time similar to that of the conventional method and thus become more reliable and feasible.


Introduction
Radial forearm free flap has been the workhorse flap in head and neck reconstruction since Yang et al. reported it in 1981 [1]. It is a fasciocutaneous flap with a major vascular trunk based on the radial artery. Abundant blood flow can be supplied to the skin paddle of the flap via several septocutaneous perforators, allowing the harvesting of large skin paddles [2]. Radial forearm free flap has definite advantages in head and neck reconstruction due to its superior flexibility and pliability. Moreover, the large diameter and long pedicle length of the vessel make vascular anastomosis easier and more reliable.
Despite this reliability and versatility, radial forearm free flap conveys shortcomings such as a long linear scar with a wide area of altered skin texture, and possible complications such as altered sensation and decreased range of motion [3,4]. Various attempts have been made to minimize such shortcomings. While full or split-thickness skin graft is widely used to cover skin paddle defects, attempts with local flap and autologous fat transplantation have been tried, with favorable results [5][6][7]. Endoscopic-assisted pedicle dissection has been utilized to reduce linear scarring for harvesting vascular pedicle [8].
Due to its reliability and usefulness, robot-assisted surgery in the head and neck area has recently come into wide use not only in tumor ablation but also in flap harvesting [9,10]. Compared to endoscopic-assisted flap harvesting, robot-assisted surgery allows 3-dimensional and higherresolution viewing. Various instruments, including electrocautery devices, provide greater flexibility and thus more precise manipulation [11].
To our knowledge, there have been no trials or reports on robot-assisted radial forearm free flap harvesting. One study reported that endoscopic-assisted radial forearm free flap harvesting reduced the linear scar from pedicle dissection, yielding more aesthetic results and reducing donor site discomfort [8]. Robot-assisted surgery, being more technically advanced than endoscopic-assisted surgery, lowers the entry barrier to radial forearm free flap harvesting and entails less donor site morbidity due to the reduction in linear scarring. The aim of our study is to compare cases of robot-assisted harvesting with conventional harvesting of radial forearm free flap under case-control using propensity score matching to compare the safety and efficacy of each approach.

Material and methods
Case selection 42 cases of radial forearm free flap reconstruction done in a single institute by a single surgeon between October 2018 and July 2021 were selected. 31 cases of flap harvesting were done conventionally and 11 cases were robot-assisted. Patients were informed about the merits and disadvantages of each method and given a choice of surgical procedure. To minimize the effects of other factors on flap vitality aside from the method of surgery, 1:1 propensity score matching was done. Age, sex, previous and postoperative radiation therapy history and method used for vein anastomosis were used as the criteria. Finally, 22 cases, 11 from each group, were included ( Table 1).

Evaluation of a surgical method
Flap survival rates were compared to assess the overall outcome of surgery. Harvesting times were compared to evaluate the difficulty and efficacy of surgery. Statistical analysis was done by IBM SPSS Statistics for Windows, version 26 (IBM Corp., Armonk, N.Y., USA).

Robot-assisted harvesting procedure
A DaVinci Si robot was used for robot-assisted radial forearm free flap harvesting ( Fig. 1). To minimize postoperative bleeding, harvesting was done without a tourniquet. After the flap was designed, the radial artery path was identified by palpation of the pulse. Next, flexor carpi radialis and palmaris longus tendons, which must be preserved for hand movement, were marked. A skin paddle was designed according to the defect size, and the vascular pedicle, including the cephalic vein, was dissected as usual. After raising the skin paddle, an additional 2 cm of skin was dissected without a conventional linear incision to make space for the retractor and robotic devices. When the retractor was successfully positioned, careful dissection and ligation of the pedicle were done with robotic devices and the flap was completely harvested.
After flap harvesting, essential insetting with a key suture was done. Under microscopy, end-to-end artery anastomosis was done with 9-0 nylon and end-to-end vein anastomosis was done with a 9-0 nylon suture or GEM coupler (Synovis Micro Companies Alliance, Inc, Birmingham, AL, USA). Donor site repair was done with a split-thickness skin graft from the thigh with dermatome and negative pressure wound therapy was applied to protect and accelerate donor site healing.

Demographics
Total of 22 cases, consisting of 11 cases of robot-assisted harvesting and 11 cases of conventional harvesting with propensity score matching, were reviewed. Among these, the tongue was the most common defect site at 45.5% and 54.5%, respectively. Due to the case-control with propensity score matching, the demographics of the cases generally showed no statistical difference although in robot-assisted harvesting cases, the mean age was 47.2, lower than that of conventional cases, 55.7 (Table 2).

Flap parameters
To compare flaps, mean surface area, artery and vein used in neck, and veins of the flap were studied. There was no statistical difference in mean surface area between the robot and conventional groups. In both groups, the superior thyroid artery was the most common artery used in the neck, being 54.5% and 63.6% in robot and conventional groups, respectively. In both groups, the facial vein was most frequently used in the neck, being 90.9% and 81.8% in robot and conventional groups. Harvested flaps in both groups included vena comitans. The portion of cephalic vein harvested was 54.5% in the robot group and 63.6% in the conventional group, with no statistical difference (Table 3).

Outcome comparison
To compare the outcome and reliability of robot-assisted and conventional harvesting, the flap survival rate was evaluated and found to be 100% in the robot group and 90.9% in the conventional group, showing no statistical difference (Table 3).
Harvesting time comparison was done to evaluate the efficacy of robot-assisted harvesting. Mean harvesting time was   (Table 3).

Postoperative healing
During postoperative follow-up, robot cases showed more satisfying aesthetic results due to linear scar reduction (Fig. 2).

Discussion
Some surgeons prefer anterolateral thigh and latissimus dorsi flaps over radial forearm flap for soft tissue reconstruction in the head and neck area because donor site repair can be done by primary closure [12]. In addition, it is thought that younger patients who are sensitive to the aesthetic outcome will prefer an anterolateral thigh flap because the scar can be hidden [13]. In our study, however, some patients who were sensitive to cosmetic results preferred robot-assisted radial forearm flap to the anterolateral thigh flap, presumably because the scar of the anterolateral thigh flap is much longer than that of the radial forearm flap and due to the thigh asymmetry caused by large volumetric defect after anterolateral thigh flap. This may also account for the significantly younger mean age of the robot-assisted radial forearm flap group. Many studies reported methods of repairing donor site after radial forearm flap. Usually split thickness and fullthickness skin grafts are used for repair [14,15]. Recently, local flap or allogenic dermal matrix is being used because skin grafts require an additional donor site such as the thigh or abdomen. Potet et al. reported primary closure of the donor site by a local flap ("keystone flap"), [6] and Nam et al. reported using an allogenic dermal matrix to repair the donor site [16]. Recently, endoscopic-assisted radial forearm flap harvesting has been tried, Van Kouwenberg et al. reporting that it reduced linear scar length. This resulted in decreased donor site complications compared to conventional cases, such as numbness, graft loss, hypertrophic scarring, delayed wound healing, tendon exposure and complex regional pain syndrome. Moreover, based on patient surveys after surgery, cosmetic outcomes were better. Finally, functional scoring of such factors as overall discomfort, strength, flexibility, and cold sensitivity was higher [8].
In our study, the mean flap survival rate was 100% in the robot group and 90.9% in the conventional group. The mean survival rate was higher in the robot group but without statistical difference. Mean flap harvesting time was 107.2 min in the robot group and 67 min in the conventional group. Mean flap harvesting time was higher in the robot group, with a statistical difference.
Flap harvesting times for all robot-assisted cases were reviewed ( Table 4). As experience increased, harvesting time showed a decreasing tendency. 3 of 4 recent cases showed 80-90 min, 10 to 20 min longer than the mean time for the conventional group, within the standard deviation range. There is also a learning curve for most kinds of surgeries. We expect robotic flap harvesting time to decrease to a level similar to that of the conventional group after the prolonged experience.
Robot-assisted surgery is technically more advanced than endoscopic-assisted surgery. Robotic devices provide magnified 3-dimensional images with a much higher resolution compared to endoscopic devices. While performing endoscopic-assisted surgery, a surgeon has to use long-shank hand instruments, which require caution due to their high sensitivity. However, robotic arms can freely connect and disconnect several robotic instruments such as Metzenbaum scissors, vascular clamps, and energy devices, allowing a surgeon to manipulate multiple instruments simultaneously. Furthermore, robotic arms are flexible and precise, allowing accurate manipulation free from physical barriers. These technical advances increase the ease and accuracy of some kinds of surgery compared to the use of endoscopic devices.
Ease of flap harvesting, low rates of vascular anatomical variation, and adequate length and diameter of pedicle are the main advantages of radial forearm free flap. Adequate subcutaneous tissue thickness in the forearm area provides better pliability, yielding good functional results after reconstruction. In case sensory reconstruction is needed, lateral

Conclusion
Based on our study and earlier literature findings, robotassisted radial forearm free flap harvesting shows better aesthetic outcome by avoiding long linear scar for pedicle dissection. Compared to endoscopic-assisted radial forearm free flap harvesting, technical robotic advances such as higher resolution, 3-dimensional viewing and precise device manipulation can allow surgeons to do more easily dissect pedicle without linear incisions. When used appropriately, robot-assisted radial forearm free flap harvesting will increase patient satisfaction with aesthetics and functional outcome without any difference in surgical outcome such as flap survival rate.