Robotic-assisted surgery in the management of endometrial cancer

Authors


Professor Robert W. Holloway, Obstetrics and Gynecology, University of Central Florida, Director, Gynecologic Oncology, Florida Hospital Cancer Institute and the Global Robotics Institute, Suite 800, 2501 N. Orange Ave, Orlando, FL 32804, USA. Email: robhollowaymd@gmail.com or sarfraz.ahmad@flhosp.org

Abstract

Advanced laparoscopic procedures for hysterectomy and staging lymphadenectomy were not widely adopted for management of endometrial cancer despite nearly 20 years of improvements in laparoscopic technology. Many surgeons have recently embraced da Vinci robotic-assisted laparoscopy in preference to traditional laparoscopy because of its technological advantages of wristed instrumentation, high-definition 3-D optics, ergonomics and autonomy of camera control; the majority of women with endometrial cancer in the USA now undergo robotic-assisted surgery. The purpose of this article is to review the robotic surgical techniques for hysterectomy, pelvic and aortic lymphadenectomy procedures, and the current comparative literature discussing perioperative outcomes. Additionally, literature that discusses challenges managing obese patients and robotic surgical costs are reviewed. Future multi-institutional, prospective registration studies comparing perioperative outcomes, complications, pain, recovery time, cost and long-term clinical outcomes with open, laparoscopic and robotic procedures will be necessary to completely appreciate the impact of robotic-assisted technology.

Introduction

Uterine corpus malignancies are the most common diseases managed by gynecological oncologists, comprising 6% of all cancers diagnosed in women. Comprehensive surgical staging that includes total hysterectomy, bilateral salpingo-oophorectomy and systematic pelvic and aortic lymphadenectomy is required by the International Federation of Gynecology and Obstetrics (FIGO) for staging.1 Lymph node staging is a critical determinant for risk stratification and treatment planning in the opinion of most gynecological oncologists in the USA. The overall therapeutic value of surgical staging continues to be debated, especially in light of recent European phase III clinical trials that showed no improvement in progression-free or overall survival for surgically staged patients.2,3 However, these trials have been heavily criticized for deficiencies in trial design and conduct calling into question the trials' conclusions and surgical staging remains ‘standard’ in the opinions of most thought leaders in the USA.4,5

Laparoscopy was first proposed for surgically staging patients with endometrial cancer in the early 1990s.6,7 While investigators immediately recognized benefits to minimally invasive surgery in terms of recovery and shortened hospital stay, most surgeons in the USA found this approach difficult for routine clinical use because of increased operating time and a protracted ‘laparoscopic learning curve’. Few centers offered laparoscopic staging for the majority of their patients with endometrial cancer by 2005,8–10 largely relegating this technique to non-obese patients with lower grade malignancies. The da Vinci Surgical System (Intuitive Surgical, Inc., Sunnyvale, CA, USA) was cleared by the U.S. Food and Drug Administration (FDA) in 2005 for gynecological procedures and rapidly gained acceptance by surgeons as an effective device for performing hysterectomy with staging lymphadenectomy in the management of endometrial cancer. It is estimated that more than 50% of such cases were managed with robotic-assisted surgery in 2010, representing a U.S. paradigm shift towards minimally invasive surgery not previously achieved with traditional laparoscopic surgery.

Da Vinci Technology

The da Vinci robotic system has three major components: a three-dimensional vision system, an ergonomic surgeon console and the robotic platform with four operative arms. It is now manufactured as a third generation ‘Si model’ and can be configured with a dual-console design for teaching. After establishing a pneumoperitoneum and placing abdominal laparoscopic ports, the robotic platform is ‘docked’ and the surgeon sits at a console, viewing the pelvis through a high-definition (1080-i) 3-D vision system. The surgeon manipulates cabled ‘endowrist’ instruments that mimic the freedom of human hand and wrist motion. Endowrist technology provides better precision of movement allowing superior operative technique, better exposure and less torquing of the abdominal wall through the operative ports, which results in less post-operative pain. Furthermore, Endowrist technology allows surgical maneuvers that more closely mimic open techniques, allowing those with less advanced laparoscopic skill to progress quickly along the learning curve with difficult tasks such as intracorporeal suturing and knot-tying. In addition to manipulating the uterus, the bedside assistant can suction, grasp, retract and pass sutures through an accessory laparoscopic port. The robotic surgeon controls the camera with clutch maneuvers and the master hand controls, and activates energy sources via foot pedals at the console. Robotic surgeons are seated and much more comfortable than laparoscopic surgeons who stand, often holding two instruments in a somewhat ergonomically contorted fashion, viewing the pelvis on a two-dimensional monitor remote to the patient while communicating camera maneuvers to an assistant. Consequently, fatigue and frustration become less of a limiting factor for the robotic surgeon.

Operating Room Set-up, Patient Preparation and Surgical Technique

Patients at our institution are counseled pre-operatively to expect an overnight admission to the hospital following surgery and a less than 5% risk for conversion to laparotomy. A mechanical bowel prep of MiraLAX in G-2 Gatorade preceded by two bisacodyl tablets is prescribed to improve visibility in the pelvis and displacement of the bowel for aortic lymphadenectomy. Trendelenburg positioning to 32 degrees requires special consideration to prevent sliding on the operating table. Washable gel-pads are placed under the sacrum and shoulders, and a tension tape is positioned over towels on the patients' clavicles and shoulders. We avoid using shoulder braces because of the potential risk of brachial nerve injury. It is important to avoid placing the strap-tape below the shoulders over the breasts and chest, which can result in sliding of the patient's torso. Lower placement of the strap below the humeri also limits chest expansion during ventilator breaths, forcing downward movement of the diaphragm that can compromise exposure during infra-renal lymphadenectomy. Other anti-slide alternatives include ‘eggcrate’ foam or the vacuum ‘bean-bag’ mattresses. The arms are padded and tucked along the torso. For obese patients or patients with cardio-pulmonary disease, we recommend testing ventilatory performance while in steep Trendelenburg position prior to prepping and establishing pneumoperitoneum. Pressure-controlled anesthesia is required for ventilation in steep Trendelenburg and neuromuscular blockade must be maintained throughout the procedure. It is helpful to ask the anesthetist to adjust the ventilator with reduced tidal volumes and increased rate (to maintain minute-ventilation) during the infra-renal portion of the aortic lymphadenectomy, when exposure is most difficult. For particularly high-risk patients who may require conversion to laparotomy, we have open instruments available in order to avoid delay should conversion be necessary. Because the arms are tucked away, we recommend patients undergoing lymphadenectomy have two peripheral intravenous cannulae placed prior to tucking the arms for ease of access in the event of major bleeding, although we have had no need for intra-operative transfusion in our five years of experience with almost 2000 patients.

Robotic hysterectomy is greatly facilitated with use of an assistant surgeon's uterine manipulator. Uterine manipulation improves safety, far outweighing any unlikely theoretical risk of dislodging cancer cells from the uterus or cervix in our opinion.11 We prefer the V-CARE disposable uterine manipulator (ConMed, Utica, NY, USA) for its ease of placement and single piece design. Many surgeons prefer their experience from laparoscopic hysterectomy using the Zumi manipulator, KOH ring and a separate pneumo-occluder balloon (Cooper Surgical, Turnbull, CT, USA). Still other surgeons use a metal EEA (U.S. Surgical Corporation, Norwalk, CT, USA) rectal sizer as an obturator to distend the vaginal fornices for colpotomy while using the third da Vinci operating arm for uterine manipulation. Port sites are anesthetized with bupivacaine 0.5% with the camera port placed 23–27 cm above the symphysis pubis, depending on the individual patient's height, torso length, uterine size and need to perform aortic lymphadenectomy. Care must be taken to avoid placing the camera port too close to the uterus or pelvic mass and obscuring the view of the anatomy. The ‘S’ and ‘Si’ da Vinci models have longer instruments than the original standard model, allowing higher placement of ports and easier dissection of aortic nodes and hysterectomy. We find that placing the third operating arm cephalad to the second arm in the left flank at the level of the camera port allows its optimal use as a retractor during aortic node dissection as well as obturator space dissection (Fig. 1).

Figure 1.

Da Vinci port set-up for aortic lymphadenectomy.

A zero-degree camera lens is used for most cases; however, a 30-degree down lens can provide better visualization of the upper anatomy for perirenal dissection and omentectomy if ports were inadvertently placed too low. Mono-polar scissors are used in the right operative arm, a fenestrated bipolar grasper is placed in the left arm and a double-fenestrated grasper/retractor is preferred for the third operative arm because of its length as a retractor. For controlling vascular pedicles, we prefer the da Vinci fenestrated bipolar grasper, which avoids the cost of an additional laparoscopic instrument and energy source, and is a more efficient grasper than the Maryland bipolar grasper. Once the uterus is delivered per vagina, a sterile lap pad inside a glove is used as a vaginal occluder during cuff closure. A long-acting suture such as 0-PDS or a 2-0 barbed suture used in a running, non-locking fashion is our preferred closure technique and seems to avoid cuff healing issues that were previously recognized with polyglycolic acid suture closures.12

Performance of aortic lymphadenectomy to the left renal vein is possible with da Vinci, but can be limited by difficult exposure in morbidly obese patients, especially those with short stature. A Ray-Tec sponge is placed during port placement as the omentum and small bowel is gently folded into the left upper quadrant. It may be necessary to laparoscopically lyse omental adhesions in patients having undergone prior surgical procedures in order that the omentum and transverse colon are optimally displaced. The peritoneum along the right common iliac artery is opened, the ureter and ovarian vessels are identified and retracted laterally by the bedside assistant with a laparoscopic suction cannula. Common iliac lymph nodes are dissected free using scissors and gentle monopolar cautery, taking care to avoid injury to the vein and genitofemoral nerve running below the artery along the psoas muscle. Maximum exposure to the aortic recess below the duodenum is created with a laparoscopic bedside assistant grasper on the small bowel peritoneum that effectively ‘tents’ away the small bowel, along with the double-fenestrated grasper in the ‘third’ arm reflecting the duodenum. In order to maintain exposure, dissection of lymph nodes from the vena cava and aorta is best performed in a cranial-to-caudal direction because the patient is in the Trendelenburg position. Caution must be exercised to avoid vascular injury by identifying the insertion of the right gonadal artery, gonadal vein and inferior mesenteric artery (IMA). Minor vena caval bleeding can be easily managed with pressure and Fibrillar (absorbable hemostatic cellulose; Ethicon, Inc., Somerville, NJ, USA). Care must be taken to avoid lifting the nodal bundle off the vena cava, sheering the anterior perforating ‘fellow's’ veins and instead dissecting around the veins using bipolar and monopolar cautery carefully. The third arm is then moved to the left, retracting the duodenum above the left renal vein and exposing the left gonadal vein. Lymph nodes from the left renal vein to the IMA are dissected free of the descending colon mesentery and aorta with blunt dissection and short bursts of monopolar cautery. The aortic lymphadenectomy is completed by dissecting below the IMA with the double fenestrated grasper retracting the ureter laterally down to the external iliac artery. We do not find it necessary to sacrifice the IMA and believe that major lymphatic trunks should be bipolar cauterized or clipped when identified to minimize the risk of lymphoceles or chylous ascites.

Florida Hospital Adoption of da Vinci Surgery for Endometrial Cancer

Prior to the FDA gynecology indication for da Vinci, urologists at our institution were performing robotic prostatectomy on a four-arm ‘standard’ model. Our three gynecological oncology surgeons preferentially selected younger, less obese patients with lower grade endometrial cancers during their learning curves, which was apparent in our initial comparative analysis of robotic and open procedures from 2006.13 A da Vinci ‘S’ model was purchased in 2006, allowing more surgeon access to robotic equipment and improved surgical capabilities with the ‘S’ model design changes. Subsequently, surgical performance for the group was examined and, through the tumor board review, the surgical approach was standardized such that during the second half of the 100-case experience, lymph node yields and operative times approached parity for three surgeons.14 The total caseload of endometrial cancer cases and the percentage of cases managed with da Vinci increased during the next four years (Fig. 2). An examination of surgery during the fourth year revealed that the majority of laparotomy cases had either stage IV disease or uteromegaly above 14 cm, and obesity (mean body mass index [BMI] 50 kg/m2) accounted for only 10% of the laparotomies. For robotic cases, age, BMI and lymph node yields increased significantly from year 1 to year 4, and the mean operative time decreased significantly (Table 1).15 By the fourth year, operative times declined to a mean 145 min (approximately 60% cases undergoing complete aortic dissection) and the mean lymph node yields were 29 in total for those undergoing a complete dissection. Our current practice is to stage the majority of patients with a visible grade 1–2 lesion with at least a pelvic lymphadenectomy, and all patients with high grade tumors (endometrioid grade 2/3, papillary serous or clear cell histologies), bulky (≥2 cm) or deeply invasive (≥50%) tumors undergo complete aortic lymphadenectomy to the renal vein bilaterally when possible.

Figure 2.

Semi-annual growth in robotic-assisted hysterectomy cases for endometrial cancer (EC) at Florida Hospital during the period July 2006 to June 2010.

Table 1.  Yearly comparison of demographic, clinicopathological and perioperative data for patients with endometrial cancer treated with robotic-assisted laparoscopic hysterectomy
FactorsYear 1Year 2Year 3Year 4P-value (Years 1 vs 4)
  • In the entire four-year cases series, 221/391 (56.5%) of cases underwent complete pelvic and aortic lymph node dissection. BMI, body mass index; FIGO, International Federation of Gynecology and Obstetrics; NS, not significant.

Number of cases (n)4279108162 
Age (years)57.4 ± 10.762.3 ± 9.662.9 ± 9.763.8 ± 10.6<0.001
BMI (kg/m2)28.2 ± 7.130.2 ± 6.331.0 ± 7.131.7 ± 7.40.006
Grade 130 (71.4%)44 (55.7%)66 (61.1%)88 (54.3%)NS
FIGO stage I36 (85.7%)56 (70.9%)86 (79.6%)126 (77.8%)NS
Operative time (min)173 ± 56164 ± 40166 ± 43148 ± 34<0.001
Estimated blood loss (mL)94 ± 5986 ± 5781 ± 5479 ± 59NS
Post-operative transfusion rate0002 (1.2%)NS
Length of stay (days)0.98 ± 0.410.98 ± 0.231.06 ± 0.501.36 ± 1.81NS
Pelvic lymph nodes12.5 ± 8.314.6 ± 7.618.1 ± 9.218.9 ± 8.0<0.001
Aortic lymph nodes6.6 ± 5.08.3 ± 4.09.1 ± 4.911.0 ± 6.7<0.001
Number of patients with positive lymph nodes2 (4.8%)9 (11.4%)9 (8.3%)12 (7.4%)NS

Comparative Studies

In the six years since the FDA clearance of da Vinci for gynecological surgery, several retrospective case series and comparative studies (either laparoscopy or laparotomy) for endometrial cancer have been published. There have been no prospective randomized trials completed, but two meta-analyses of observational studies were published simultaneously in 2010.16,17 In an analysis of eight comparative studies that included at least 25 robotic cases (Table 2),13,18–24 Gaia et al. identified 1591 patients with endometrial cancer (589 robotic, 396 laparoscopic and 606 laparotomy cases).16 Robotic and laparoscopic cases appeared to be similar, with the exception of reduced blood loss with the robotic cases (P = 0.001), but with no difference in transfusion rates (OR 0.47, CI 0.10–2.19, P = 0.22). For the robotic–laparotomy comparison, the transfusion rates approached significance in favor of robotic cases (OR 0.25, CI 0.05–1.16, P = 0.06). Operative times for robotic and laparoscopy cases were similar, but both were greater than laparotomy (P < 0.005). Lymph node yields were not statistically different for any surgical method. Conversion to laparotomy was 9.9% for laparoscopy compared to 4.9% for robotic, which approached statistical significance (P = 0.06). Reza et al. also analyzed gynecological surgery publications, including those for endometrial cancer, and reported that robotic surgery was associated with reduced blood loss (76 mL, P = 0.03) and fewer transfusions (OR 0.24, CI 0.09 to 0.64) than conventional laparoscopy.17 In their analysis, conversion to laparotomy was less for robotic surgery (OR 0.43, CI 0.21 to 0.85) and the overall risk of complications was not different for robotic and laparoscopic surgery. Thus, two systematic reviews of comparative studies suggest that, for the management of endometrial cancer, surgeons may expect less blood loss, fewer transfusions, fewer laparotomy conversions, similar operative times and similar overall complications for robotic compared to laparoscopic hysterectomy and lymphadenectomy. However, it is always important to appreciate the limitations of meta-analyses of retrospective observational studies and that their conclusions are not nearly so robust as those from meta-analyses of prospective randomized trials due to the inherent selection biases encountered in retrospective studies.

Table 2.  List of comparative studies in recent meta-analyses of robotic-assisted hysterectomy for endometrial cancer
  • Articles cited in Gaia et al.;16

  • articles cited in Reza et al.17

Bell et al., 200818
Boggess et al., 200819
Cardenas-Goicoechea et al., 201020
DeNardis et al., 200813
Gehrig et al., 200821
Seamon et al., 200922
Seamon et al., 200923
Veljovich et al., 200824

Managing Obese Patients with Endometrial Cancer

Obesity is associated with a 10-fold increased risk for endometrial cancer and is often mentioned as a ‘limiting factor’ concerning patient selection for either laparoscopic or robotic procedures. Obese patients have an increased risk of laparotomy conversion and less complete lymph nodal dissection.25 Maintenance of exposure during aortic lymph node dissection and adequate ventilation with requirements for steep Trendelenburg positioning can be challenging for both the surgeon and anesthesiologist. Gehrig et al. reported that robotic surgery was preferable to laparoscopy for treatment of endometrial cancer in 36 obese and 13 morbidly patients because of shorter operative times, reduced blood loss and hospital stay, and increased lymph node retrieval.21 Ninety-two percent of the robotic cases and 84% of the laparoscopic cases completed both a pelvic and aortic node dissection. While robotic lymph node yields were greater than those for laparoscopic cases, node counts for morbidly obese (BMI > 40 kg/m2) patients were not greater than laparoscopy, indicating that robotic aortic lymphadenectomy may still have some limitations for this group of difficult patients.

Seamon et al. reported that, of 105 patients with endometrial cancer at The Ohio State University, 13 (12.4%) with a BMI ranging from 47 to 58 and grade 1 cancer did not undergo complete staging: six patients underwent pelvic lymphadenectomy without aortic dissection and seven had no lymphatic staging.23 There were also 12.4% conversions to laparotomy. The mean conversion BMI was 40 ± 7 kg/m2 compared to 34 ± 9 kg/m2 for those completed robotically. The feasibility for completing robotic aortic lymphadenectomy was 67% and 35% for a BMI 45 and 50 kg/m2, respectively.

Most recently, Subramaniam et al. compared outcomes for a group of obese patients with endometrial cancer undergoing robotic (n = 73, BMI 39.8 kg/m2) and open hysterectomy (n = 104, BMI 41.9 kg/m2).26 Lymphadenectomy (total mean node count 8.0) was performed in 66% of robotic cases and conversion to laparotomy was 11%. Transfusions, hospital stay and complications were all improved for robotic-assisted surgery. In our experience, height is also very important when considering patients with an elevated BMI. Short stature and the ‘apple’ shape body habitus (in contrast to a tall patient with ‘pear’ shaped habitus) predicts a significant difficulty with aortic dissection for patients with a BMI in excess of 40 kg/m2, but pelvic lymphadenectomy can almost always be accomplished.

Robotic Staging for Patients with Intermediate-risk Endometrial Cancer Following Simple Hysterectomy

The gynecological oncologist is often asked to consult for women who have been diagnosed with uterine malignancy following a simple hysterectomy for complex endometrial hyperplasia or post-menopausal bleeding. Depending on the uterine pathology, a staging lymphadenectomy and complete removal of the broad ligaments and adnexae are desirable. Standard laparoscopic staging has been described with less morbidity than laparotomy in this setting,6 and may result in less delay of adjuvant therapy. Laparoscopic re-staging was successful in 80% of patients in a Gynecologic Oncology Group study.27 Our center has performed 26 robotic staging operations for such patients who were referred after hysterectomy and had a finding of occult uterine malignancy. Despite adhesions, there have been no conversions to laparotomy. Similar to our experience with robotic hysterectomy for endometrial cancer, patients undergoing robotic-assisted post-hysterectomy staging also benefit, with less perioperative morbidity than patients undergoing a second staging laparotomy (unpublished results).

Robotic-assisted Surgical Cost Comparisons

Everyone caring for women with endometrial cancer understands the benefits of minimally invasive surgery, such as less post-operative pain, fewer wound complications, shorter hospital stays and a quicker return to normal activities. Robotic-assisted surgery has extended these benefits to a much larger proportion of patients with endometrial cancer because of the inherent enabling features of the da Vinci technology. The healthcare benefits of minimally invasive surgery come at a cost in terms of equipment, maintenance, disposables, operating time and training, and are at least partially offset by the reduced costs of hospital bed-days, management of complications and lost wages from a longer convalescence expected with laparotomy cases.

Bell et al. first discussed surgical costs associated with robotic hysterectomy and lymphadenectomy for endometrial cancer in a retrospective comparison of 40 robotic, 30 laparoscopic and 40 open cases from a single institution.18 Direct and indirect hospital costs, as well as estimates of lost household productivity, were calculated for each surgical method. Both laparoscopy and robotic surgery were less costly than laparotomy in this analysis and hospital days were identified as a significant factor in the cost of care.

Three separate cost models were used to compare robotic, laparoscopic and open surgery by Barnett et al.28 In their ‘societal model’, analogous to the Bell et al. article,18 laparoscopy was least expensive at $10 128 per case. Robotic surgery and open surgery costs exceeded laparoscopy by $1348 and $2719, respectively. The authors proposed that reducing disposable costs for the robot to $1046 from the average reported at $2394 would make it equal to laparoscopy. This analysis makes several assumptions about surgical procedure volume (institution-specific variables) that could significantly influence the amortized per-case costs of the da Vinci system and laparoscopic equipment. High volume centers can reduce the cost gap between robotics and laparoscopy because the per-case cost of the da Vinci system is reduced and minimizing disposable instruments reduces costs with all procedures.

A recent study by Jonsdottir et al. described a dramatic change in hysterectomy patterns of care and associated costs in a large tertiary-care hospital.29 A total of 1054 hysterectomies from 2006 were compared to 1079 from 2009, approximately 40% of which were oncology cases. Abdominal hysterectomy fell from 64.7% to 35.8%, laparoscopic hysterectomy increased from 17.7% to 46%, and robotic hysterectomy grew from 2 to 6% of total cases. During this three-year period, the total composite cost of care (which included estimates on ‘societal costs’) increased by 4.1% and abdominal hysterectomy costs decreased by 5%. Laparoscopic costs increased by 32.7%, while robotic costs decreased by 31%. Although robotic hysterectomy was the most expensive form of hysterectomy in 2006, it was the least expensive in 2009 ($1325 less than laparoscopy). In a related study from this institution, these authors previously identified operative time as a significant predictor of surgical costs. Jonsdottir et al. hypothesized that during their three-year study, robotic surgeons and their operating room teams improved efficiencies with reductions in operative and turnover times29 (and perhaps conversely the transition from laparotomy to laparoscopy resulted in longer operative times). This study illustrates the importance of analyzing costs over time for institutions, as efficiencies may be gained or even lost as the patterns of care change and surgeons progress through their separate procedural ‘learning curves’.

Conclusions

Surgical patterns of care for women with endometrial cancer have changed dramatically since Childers et al. first described laparoscopic hysterectomy and lymphadenectomy in 1993.6,7 Currently, the majority of patients in the USA with endometrial cancer undergo robotic-assisted laparoscopic surgery. The da Vinci robotic system has been described as an ‘enabling device’, helping surgeons cross the bridge to minimally-invasive surgery with techniques that mirror the open technique, while assisting the experienced laparoscopic surgeon to push beyond the procedural barriers of standard laparoscopy. There are no prospective randomized clinical trials comparing open, laparoscopic and robotic surgery for endometrial cancer. The bulk of retrospective case series and two meta-analyses indicate similarities with laparoscopy, with the exception of reduced blood loss and fewer laparotomy conversions. Robotic and traditional laparoscopic surgery have better outcomes than laparotomy in terms of post-operative pain, recovery time and complications, including blood loss, transfusion and wound infection. Recent cost analyses indicate that efficiencies gained with robotic surgical experience may translate into significant operating room cost reductions, such that the widely held belief that robotic surgery is ‘too expensive’ may not be true for many institutions.

Disclosure

The authors have no conflict of interest related to this article. Professor Holloway is reimbursed for conducting advanced robotic training programs for gynecological oncologists through Intuitive Surgical, Inc. (Sunnyvale, CA, USA).

Ancillary