New steps of robot-assisted radical prostatectomy using the extraperitoneal approach: a propensity-score matched comparison between extraperitoneal and transperitoneal approach in Japanese patients

Robot-assisted radical prostatectomy (RARP) has been used worldwide since it was introduced in 2000 [1, 2]. The 4-armed da Vinci® S surgical system (Intuitive Surgical, Sunnyvale, CA, USA), which is responsible for the widespread use of RARP, has 8-cm-wide arms, and the trocars should be spaced at intervals >8 cm to prevent them from colliding with each other [3]. RARP is commonly performed using the transperitoneal (TP) approach because it offers enough arm distance, a larger working space, and familiar laparoscopic intraperitoneal landmarks [2]. The working space is relatively smaller when using the extraperitoneal (EP) approach. However, the procedure can be performed with little effect on patients with previous intraabdominal surgery or severe obesity [4], and it causes minimal intraabdominal complications [2, 5]. Furthermore, because the steep Trendelenburg position is not required, the EP approach is effective in patients with contraindications to this position [6, 7]. However, the EP approach has several limitations including a small working space and collision of the robotic arms with one another. Within a small operating cavity, the separated prostate impedes the operator’s visibility, and performing the vesicourethral anastomosis becomes difficult. It is challenging to acquire enough space to keep the removed prostate out of the robotic view and avoid collision of the robotic arms in patients who are physically small in particular. The development of new methods is necessary to overcome these difficulties.

Herein, we introduce four new steps for the surgical procedure and investigate the feasibility of using the EP approach with six trocars over a 4-cm distance in a less than 5° Trendelenburg position. We also compare the surgical results and complications of RARP using the EP and TP approaches.

The study protocol was approved by the Tokushukai Group Ethical Committee (approval number: TGE00700–016). Clinical data were gathered starting from the beginning of our experience with RARP. The database included preoperative, operative, and postoperative information.

From August 2012 to August 2016, we retrospectively reviewed the data of 200 consecutive patients who underwent RARP with the EP approach in a flat or 5° Trendelenburg position and 428 consecutive patients who underwent RARP with the TP approach in a steep Trendelenburg position. Our surgical team began performing RARP using the TP approach in May 2011. Prior to that, they had experience performing >600 laparoscopic radical prostatectomies using TP and EP approaches [8‐11]. After we managed 135 cases of RARP with the TP approach, we began performing RARP with the EP approach.

Patients in the EP and TP group were balanced using propensity score matching (Table 1). The patient with the smallest physique in the EP group was 143.6 cm tall and weighed 34.8 kg; he underwent RARP using the EP approach safely. Patient backgrounds such as follow-up duration, age, height, weight, and BMI were not significantly different between the groups (Table 1). Preoperative clinical characteristics such as the mean PSA level, biopsy Gleason score, clinical stage, and previous abdominal surgery, were also not significantly different between the groups (Table 1).

Table 2 shows the operative and postoperative results, as well as complications in both groups. The mean operative time was significantly longer in the EP group than in the TP group (254.5 vs 225.8 min, p < 0.01). The mean robot console time and vesicourethral anastomosis time were not significantly different between the EP and TP group (180.5 vs 175.4 min, p = 0.07; and 30.3 vs 28.0 min, p = 0.09, respectively). The mean volume of estimated blood loss and duration of indwelling urethral catheters were significantly lower in the EP group than in the TP group (139.9 vs 184.9 mL, p < 0.01; and 5.6 vs 7.7 days, p < 0.01, respectively). There were no significant differences between the groups in terms of the positive surgical margin, lymph-node dissection rate, and urinary continence. The rate of postoperative inguinal hernia was significantly lower in the EP group than in the TP group (1.1 vs 7.4%, p < 0.01). There were no cases of conversion to open surgery in the EP group.

Table 3 compares the surgical and pathological results of the 200 patients in the EP group subdivided into two chronological groups of 100 patients. The mean operative time, robot console time, and vesicourethral anastomosis time were significantly shorter in the latter group (257.9 vs 247.2 min, p < 0.01; 184.5 vs 174.1 min, p < 0.01; and 33.4 vs 24.2 min, p < 0.01, respectively). None of the patients developed an indirect inguinal hernia postoperatively after our method was introduced.

Our method of RARP using the EP approach can be safely implemented, even in patients with small physiques. RARP using the EP approach has the benefit of reducing the amount of blood loss and shortening the duration of indwelling urethral catheter use compared with RARP using the TP approach.

We divided the surgical procedure into four steps. First, we arranged six trocars spaced 4 cm apart. In all 200 patients in the EP group, it was possible to perform surgery without revising the trocar placement. This result suggests that it is possible to perform RARP with the EP approach using six trocars spaced 4 cm apart.

Second, we created the EP space and guided the trocar with a new technique. Normally, the trocar is placed using finger assistance or laparoscopy to confirm the tip of the trocar so that it does not damage the surrounding structures [24]. However, when the location of trocar placement cannot be sufficiently exposed because of adhesions, it is difficult to appropriately confirm the tip of the trocar. Therefore, we used the laparoscopic dissector to penetrate from the EP space to outside of the body, to avoid the adhesive site, and to guide the trocar. This technique is a useful way to place the trocar into a narrow space, and it can be used during robotic surgery as well as any laparoscopic surgery.

Third, we developed a new technique to hold the separated prostate in the EP space. A 14-gauge cannula can serve as a tract for the thread of the pouch. During robotic or laparoscopic pyeloplasty, a 14-gauge cannula is used as a tract for the insertion of a ureteral stent [25, 26]. Thus, using an intravenous indwelling needle as a small tract through the abdominal wall can be helpful during minimally invasive surgery.

Fourth, the technique of preventing a postoperative inguinal hernia was simple and effective. This technique was originally performed during retropubic radical prostatectomy [21]. Before we introduced this technique, a postoperative inguinal hernia developed in 2 (10%) patients. However, none of the patients developed postoperative inguinal hernia after introducing this technique. The average follow-up duration in our study was over 27 months. The average interval between prostatectomy and postoperative inguinal hernia diagnosis was reported to be 10.6 months [21]. Thus, the follow-up duration we used in our study was sufficient to determine the effect of this technique.

Urethral anastomosis stenosis was observed in 1.1% of patients in the EP group and 2.6% of patients in the TP group. The incidence of strictures of the vesico-urethral anastomosis after radical prostatectomy has been reported to range from 0.5 to 32%, with most occurring within 5 months of radical prostatectomy [27‐29]. Thus, the rate of anastomosis stenosis in our study was lower than those in the literature and the follow-up duration of our study was sufficient to evaluate the rate of anastomosis stenosis after RARP.

Symptomatic lymphocele occurred in 0% of patients in the EP group and 0.5% of patients in the TP group. We removed drains on postoperative day 2 after confirming a decrease in the drain fluid. The duration of pelvic drainage has been reported to influence the rate of symptomatic lymphocele [30]. The group whose drains were removed on postoperative day 1 exhibited higher rates of symptomatic lymphocele than patients whose drains were removed on postoperative day 7 and patients without drainage [30]. In general, lymphatic fluid cannot be absorbed by the peritoneal surface in the EP approach. Accordingly, the rate of lymphocele in the EP approach is greater than that in the TP approach. During the postoperative course of our study, we removed the pelvic drain after the amount of drain fluid decreased. Thus, the rate of lymphocele was low compared with that in the literature [31].

Our four new steps were demonstrated to be successful for performing RARP using the EP approach safely. The positive surgical margin and the time of operation, robot console, and vesicourethral anastomosis were compatible to those in the literature [32]. In particular, the amount of the blood loss and the incidence of the postoperative complications, such as inguinal hernia, anastomosis stenosis, and lymphocele, were low compared with those in the literature [31, 32].

The robot console time and vesicourethral anastomosis time of the EP approach became shorter after 100 cases, and they became shorter than those of the TP approach. However, the total operative time of the EP approach was still longer after 100 cases than that of the TP approach. There is a possibility that extra time was needed to create the EP space. As the operative space is relatively smaller in the EP approach, bleeding should be carefully controlled to maintain the robotic view. Vesicourethral anastomosis was performed after completing hemostasis. The values of blood loss and duration of indwelling urethral catheter use were significantly lower from the beginning of the EP approach than those of the TP approach. We first started RARP using the TP approach, and we overcame the learning curve of 50–100 cases before starting to perform the EP approach [12, 13]. However, 100 cases of the EP approach were needed to overcome the learning curve of EP approach.

We acknowledge that our study had strengths and weaknesses. One strength of our study is that we created a propensity-score matched comparison, which had a balanced covariate preoperative factor profiling, reducing the possibility of cofounding. Although our study was retrospectively designed, the propensity-score matched comparison might reduce selection bias. There were a few additional limitations to our study. First, the sample size might be relatively small for a study of this nature. Second, the model for propensity score matching may not include potentially relevant factors that are simply unknown at this time. Third, erectile data evaluation was not performed. Fourth, the operation was performed by the same surgical team but not by a single surgeon, which might have had an impact on the outcomes of the operation and the postoperative course.

RARP with the EP approach using four new steps was safely performed regardless of patient physique and medical history. Our method of RARP with the EP approach can reduce the amount of perioperative blood loss, the duration of indwelling urethral catheters, and the incidence of postoperative inguinal hernia development. In addition to RARP using the TP approach, the EP approach is useful and it should be learned and mastered for all types of patients regardless of their medical history, if possible.