Twenty-five percent of women in the United States are obese and are thus at a 10-fold increased risk for endometrial cancer. The National Institutes of Health divides obesity into three body mass index (BMI) classes: I (30–34.9), II (35–39.9), and III (extreme obesity, 40 or higher) (BMI is calculated as weight (kg)/[height (m)]2).1 In developed countries, obesity strongly influences the increased incidence of endometrial cancer.2–4 For medically fit endometrial cancer patients, the foundation of treatment is comprehensive surgical staging,5,6 including pelvic and aortic lymphadenectomy; this procedure defines the biology of the disease and assigns the need for adjuvant therapy.5,7
Laparoscopic surgical staging often is used for the treatment of endometrial cancer,8 and it recently has been reported in a randomized trial to have fewer grade 2 or higher postoperative complications than staging by laparotomy (Walker JL, Piedmonte M, Spirtos N, Eisenkop SM, Schlaerth JB, Mannel RS, et al. Randomized trial of laparoscopy vs. laparotomy for comprehensive surgical staging of uterine cancer: a Gynecologic Oncology Group study [LAP2]. J Clin Oncol. In press.). However, in this study, the rate of comprehensive staging (including pelvic and aortic lymphadenectomy) was lower in the laparoscopy group (92%) compared with the laparotomy group (96%, P<.001). One could hypothesize that this discrepancy in the rate of comprehensive staging could be even more apparent in obese women with endometrial cancer. Although previous data have demonstrated the feasibility of surgical staging in obese women using laparoscopy or robotic approaches,6,9–14 there have been limited data rigorously comparing the adequacy of surgical staging in this population using minimally invasive techniques. As such, we set out to compare formally the adequacy of staging and outcomes of laparotomy and robotics in obese women with endometrial cancer.
MATERIALS AND METHODS
Institutional review board approval was obtained from the cancer institutional review board of the Ohio State University (Ohio State) and the institutional review board of the University of Alabama at Birmingham (Alabama-Birmingham). A cohort study was performed at Ohio State and Alabama-Birmingham. Eligible patients included all patients with clinical stage I or occult stage II endometrial cancer who underwent robotic or open hysterectomy and lymphadenectomy who were obese (defined as a BMI at least 30).2 The robotic cohort was identified at each institution from the inception of their robotics programs (Ohio State January 2006 and Alabama-Birmingham May 2006 to August 2008). All robotic surgeries were performed on the da Vinci Surgical System (Intuitive Surgical, Sunnydale, CA) using a four-arm robot. The robotic cohort was compared with an obese group of patients undergoing laparotomy for endometrial cancer, matched for surgeon and BMI±2 in a 1:2 fashion (Ohio State June 3, 1998, to January 2006 and Alabama-Birmingham June 24, 2004, to May 2006). When two eligible laparotomy cases could not be identified, a single patient served as a match.
Two faculty surgeons at each institution performed these procedures, with residents or fellows as bedside assistants or console surgeons. In these institutions, it generally is held that comprehensive surgical staging should be performed when feasible, including hysterectomy, salpingo-oophorectomy, and pelvic and aortic lymphadenectomy. The usual boundaries of the lymph node dissections at both institutions adhere to the Gynecologic Oncology Group (GOG) Surgical Procedures Manual for pelvic and aortic lymphadenectomy.15 For the aortic dissection, the superior boundary of the dissection is at least to the duodenum on the right and the inferior mesenteric artery on the left.
All inpatient and outpatient charts were reviewed. Data were collected on the following potential confounders or effect modifiers: age, comorbidities, previous surgeries, preoperative grade, fellow assistance during surgery, and start time of surgery (am or pm). Surgeries were defined as morning cases if they began before 11 am. Operative times for all surgeries were divided into two categories: room time and skin time. Room time was defined as the total length of time (minutes) the patient spent in the operating room. Skin time was defined as the length of the operation (minutes) from the operation start time (from first incision or vaginal instrumentation) to the time all incisions were closed. Extending a robotic port to accommodate removal of a specimen was not considered conversion to laparotomy. All other laparotomy incisions in the robotic cohort were considered conversions.
All operative and postoperative complications were recorded. Operative complications included blood loss requiring transfusion (either intraoperatively or postoperatively) and major vessel, nerve, gastrointestinal, and urinary tract injuries. Postoperative complications were defined as readmissions within 30 days of the procedure, venous thromboembolic events, and cardiac, pulmonary, gastrointestinal, urinary tract, neurologic, and any other significant events.
Wound problems were considered any documentation of seroma, hematoma, wound disruption without infection, or wound infection. Length of stay was defined as the number of nights the patient spent in the hospital. Lymphadenectomy arbitrarily was considered adequate if the total lymph node count was at least 10, pelvic node count was at least six, and paraaortic node count was at least four on the pathology report.
The study’s sample size was limited by the number of robotic patients that met the criteria. To detect a 10% difference in adequacy of lymphadenectomy (80% robotic, 90% open), the primary outcome of our study, we would need 180 patients in the robotics arm and 360 patients in the laparotomy arm (two-sided α=0.05, β=0.15, matched 1:2).
Random-effects linear regression was used to test for differences across groups (robotic and laparotomy) for continuous variables that were normally distributed with stable variances across the groups. This method takes into account the variance within and across the strata created when the patients were matched on BMI. For binary variables, random-effects logistic regression was used to test for differences across groups; for categorical variables, ordinal logistic regression with a clustering variable was used to test differences. The data were analyzed using STATA 10.1 (Stata Corporation, College Station, TX) statistical software.
A total of 109 obese patients with endometrial cancer underwent robotic staging and were compared with 191 patients who underwent laparotomy. Of the 109 patients undergoing robotic staging, 70% had BMIs of at least 35. The ranges were as follows: 30% BMI 30–34, 21% BMI 35–39, 24% BMI 40–44, 17% BMI 45–49, and 8% BMI 50 or higher. Likewise, because of matching on BMI, the laparotomy group was similar (28% BMI 30–34, 23% BMI 35–39, 24% BMI 40–44, 16% BMI 45–49, and 9% BMI 50 or higher). The robotic patients were younger (58±10 years compared with 62±12 years, P=.003), more often had at least three comorbidities (43% compared with 26%, P=.05), and less commonly had previous surgery (51% compared with 63%, P=.04) compared with the laparotomy group (Table 1). There was no difference among the entire cohort in preoperative grade, stage, or final histology.
Of 109 patients undergoing surgery with the intent of robotic staging, 92 procedures were completed robotically and 17 (15.6%, 95% confidence interval [CI] 9.5–24.2%,) were converted to laparotomy. When these 92 patients were compared 162 matched controls, 87% compared with 85% (P=.65) had a lymphadenectomy in addition to hysterectomy and 85% compared with 91% (P=.16) had an adequate lymphadenectomy (Table 2). In 90% of robotic cases, an adequate pelvic lymph node dissection only was done, similar to the rate in the open group (95%, P=.16). Likewise, the adequacy of aortic lymphadenectomy was no different between the robotic and laparotomy cohorts (76% compared with 79%, P=.70). The number of lymph nodes retrieved was consistent over the course of the study. In addition, we failed to find a difference in adequacy of staging when the surgery started in the morning compared with the afternoon or related to fellow assistance. However, more fellows participated in robotic surgeries compared with laparotomy surgeries (74% compared with 49%), and fellows or residents sat the console for 39% of robotic procedures.
A univariable analysis comparing the robotic and laparotomy groups demonstrated a potential relationship between adequate staging and at least two or at least three comorbidities, preoperative grade, and history of previous surgery. By multivariable analysis, patients undergoing robotic surgery were less likely to have adequate staging when adjusting for two or more comorbidities, previous surgical history, and preoperative grade (2 or 3) compared with laparotomy patients (odds ratio [OR] 0.22, 95% CI 0.05–0.90).
The mean estimated blood loss was less for robotic procedures (109 mL compared with 394 mL, P<.001), and robotic procedures were associated with a reduced transfusion rate (2% compared with 9%). Patients undergoing robotic procedures were 78% less likely to need a red blood cell transfusion than were patients undergoing laparotomy (OR 0.22, 95% CI 0.05–0.97, P=.046, Fig. 1). Adjusting for comorbidities and history of previous surgery did not alter this finding significantly. The total operative time from skin opening to closure (228±43 minutes compared with 143±47 minutes P<.001) and overall room time (entry to exit, 284±49 minutes compared with 186±51 minutes, P<.001) were longer for robotic staging when compared with laparotomy. The mean length of hospital stay was lower for robotic surgery than for laparotomy—1 night (interquartile range 1–2) compared with three nights (interquartile range 3–4), P<.001.
The overall rate of complications excluding wound complications (11% compared with 27%, OR 0.29, 95% CI 0.13–0.65 P=.003, Table 3) and in particular wound complications (2% compared with 17%, OR 0.10, 95% CI 0.02–0.43, P=.002) was reduced for robotic surgery compared with laparotomy (Fig. 1). After adjusting for two or more comorbidities and previous surgery, the reduction in wound complications was even more apparent (OR 0.06, 95% CI 0.006–0.59, P=.02). However, the overall rate of complications excluding wound complications was not affected by the adjustment (OR 0.35, 95% CI 0.13–0.95, P=.04). The most common complication in the laparotomy cohort was postoperative ileus (10% compared with 1.8% in the robotic group). In the laparotomy group, one patient went into cardiac arrest after pulseless electrical activity and another developed pneumonia, systemic inflammatory response syndrome, and sepsis. Additionally, one patient in this cohort died from suspected aspiration pneumonia. There were no deaths in the robotics group.
Robotic surgical staging is feasible in obese patients with endometrial cancer and may result in similar surgical outcomes as patients staged by laparotomy, with fewer complications and shorter length of hospital stay. The potential advantages of minimally invasive surgery for patients with endometrial cancer are well-documented and include decreased length of hospital stay and postoperative complications as well as improvement in short-term quality of life.16,17 As a group, obese patients with endometrial cancer are most likely to benefit from a successful minimally invasive surgery because of their higher rate of comorbidities and postoperative complications.10,13 Despite the potential advantages compared with laparotomy, only a minority of patients with endometrial cancer undergo minimally invasive, comprehensive surgical staging procedures.18 The laparoscopic approach is difficult, with a long learning curve for most surgeons; these challenges are magnified in obese patients.
Our study shows that there is not sufficient evidence to suggest that the adequacy of surgical staging is different between robotics and laparotomy; however, because of the sample size, our study is not powered to detect such a difference. Although we discovered a potential decrease in the adequacy of robotic staging when adjusting for two or more comorbidities, previous surgical history, and preoperative grade (2 or 3) compared with laparotomy patients, this may be an artifact of modeling. Deleting one of these variables from the regression model leads to a loss of this finding.
Others have demonstrated the ability to stage obese patients comprehensively by laparotomy or with minimally invasive techniques using both laparoscopy and robotics. Pavelka et al19 studied obese women with endometrial cancer and found that, with increasing BMI, the rate of aortic lymphadenectomy declined. However, when lymphadenectomy was performed, the lymph node yields were similar among overweight, obese, and morbidly obese women. In a retrospective matched cohort study of obese patients undergoing laparotomy or laparoscopy for endometrial cancer, Scribner et al6 found a shorter length of stay, fewer postoperative fevers, and lower rates of ileus with laparoscopy. Importantly, lymph node counts were comparable between the two groups—18 pelvic, five common iliac, and six–seven aortic lymph nodes.6 Similarly, Eisenhauer et al9 reviewed institutional data on obese endometrial cancer patients (BMI 35 or higher) planning to undergo either laparotomy (with or without panniculectomy) or laparoscopy. The laparotomy group without panniculectomy and the laparoscopy group had similar rates of completing pelvic (40% compared with 45%) and aortic (36% compared with 32%) lymphadenectomies and similar median node counts (pelvic 15 compared with 12, aortic six compared with four).9
However, the GOG LAP2 study (GOG#2222), a randomized prospective trial of laparoscopy compared with laparotomy for endometrial cancer, was unable to demonstrate equivalency for comprehensive laparoscopic staging. Only 94% of the laparoscopic cohort underwent aortic nodal sampling, compared with 97% in the laparotomy group (P<.001) (Walker JL. In press.). Recently, comparisons of robotic surgery with laparoscopy in obese women with endometrial cancer have been reported.11 In this study, Gehrig et al11 report shorter operative times and lower estimated blood loss in the robotics cohort. However, the lymph node yield was identical in obese women regardless of whether they were staged robotically or laparoscopically.11
Despite encouraging data that the adequacy of staging in obese women with endometrial cancer is similar whether performed by laparotomy or robotics, the current study has several limitations. The retrospective nature of data collection induces bias in selection and ascertainment. It is possible that obese patients undergoing robotic surgery are preselected by the surgeon to undergo robotics preferentially rather than open surgery based on a variety of factors. By eliminating cases performed by laparotomy after the initiation of an institution’s robotic surgery program, we attempted to minimize the effect of selecting out the least favorable robotic candidates for laparotomy. Although laparoscopy for endometrial cancer staging was being performed at both institutions during the timeframe in which cases were selected for study inclusion, the laparoscopic cases were chosen carefully, with most patients having BMIs less than 30.20 Thus, our design would not allow a comparison laparoscopy arm.
We attempted to control for confounders by including open cases only for those surgeons who also perform robotics, matching for BMI and collecting data on comorbidities, previous surgeries, and preoperative grade. Ascertainment bias is possible in data collection because the investigators who collected the data were not blinded to the type of surgery that was performed. It is also possible that wound and other complications were underrepresented in this study because both participating institutions see patients from a wide geographic area in which local physicians are involved in emergent and urgent postoperative issues, sometimes without reporting back to the surgeon. However, this reporting bias is equally likely in both cohorts. Additionally, because of the study design, unknown potential confounders are a factor.
In conclusion, our data suggest that robotic surgery is feasible in obese women with endometrial cancer and may result in comparable rates of surgical staging compared with laparotomy. In this group of patients, far fewer complications and a decreased length of stay (at the expense of a longer procedure) are seen. However, because of the numerous factors that determine the surgical approach for a particular individual patient and the biases introduced by our study design, we are cautious to generalize our results. Continued investigation of robotics in endometrial cancer is needed, with particular focus on factors that affect the adequacy of comprehensive staging and the cost-effectiveness and long-term outcomes of this procedure, especially in the population of obese women at the highest risk for morbidity and mortality.
1. National Institutes of Health. National Heart, and Blood Institute. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: the evidence report. Available at: http://www.nhlbi.nih.gov/guidelines/obesity/ob_gdlns.pdf
. Retrieved March 25, 2009.
2. Bergström A, Pisani PM, Tenet V, Wolk A, Adami HO. Overweight as an avoidable cause of cancer in Europe [published erratum appears in Int J Cancer 2001;92:927]. Int J Cancer 2001;91:421–30.
3. Kaaks R, Lukanova A, Kurzer MS. Obesity, endogenous hormones, and endometrial cancer risk: a synthetic review. Cancer Epidemiol Biomarkers Prev 2002;11:1531–43.
5. management of endometrial cancer. ACOG Practice Bulletin No. 65. American College of Obstetricians and Gynecologists. Obstet Gynecol 2005;106:413–25.
6. Scribner DR, Walker JL, Johnson GA, McMeekin DS, Gold MA, Mannel RS. Laparoscopic pelvic and paraaortic lymph node dissection in the obese. Gynecol Oncol 84:426–30.
7. Straughn JM, Huh WK, Kelly FJ, Leath CA 3rd, Kleinberg MJ, Hyde J Jr, et al. Conservative management of stage I endometrial carcinoma after surgical staging. Gynecol Oncol 2002;84:194–200.
8. Barakat RR, Lev G, Hummer AJ, Sonoda Y, Chi DS, Alektiar KM, et al. Twelve-year experience in the management of endometrial cancer: A change in surgical and postoperative radiation approaches. Gynecol Oncol 2007;105:150–6.
9. Eisenhauer EL, Wyppych KA, Mehrara BJ, Lawson C, Chi DS, Barakat RR, et al. Comparing surgical outcomes in obese women undergoing laparotomy, laparoscopy or laparotomy with panniculectomy for the staging of uterine malignancy. Ann Surg Oncol 2007;14:2384–91.
10. Tozzi R, Malur S, Koehler C, Schneider A. Analysis of morbidity in patients with endometrial cancer: is there a commitment to offer laparoscopy? Gynecol Oncol 2005;97:4–9.
11. Gehrig PA, Cantrell LA, Shafer A, Abaid LN, Mendivil A, Boggess JF. What is the optimal minimally invasive surgical procedure for endometrial cancer staging in the obese and morbidly obese woman? Gynecol Oncol 2008;111:41–5.
12. Kadar N. Laparoscopic pelvic lymphadenectomy in obese women with gynecologic malignancies. J Am Assoc Gynecol 1995;2:163–7.
13. Abu-Rustum NR. CO2 pneumoperitoneum or the Bookwalter: choose your access and exposure. Gynecol Oncol 2005;97:1–3.
14. Seamon LG, Cohn DE, Richardson DL, Valmadre S, Carlson MJ, Phillips GS, et al. Robotic hysterectomy and pelvic-aortic lymphadenectomy for endometrial cancer. Obstet Gynecol 2008;112:1207–13.
15. Whitney CW. Gynecologic Oncology Group surgical procedures manual. 9th ed. Philadelphia (PA): Gynecologic Oncology Group. 2005. p. 6–15.
16. Kueck AS, Gossner G, Burke WM, Reynolds RK. Laparoscopic technology for the treatment of endometrial cancer. Int J Gynaecol Obstet 2006;93:176–81.
17. Spirtos NM, Schlaerth JB, Gross GM, Spirtos TW, Schlaerth AC, Ballon SC. Cost and quality-of-life analyses of surgery for early endometrial cancer: laparotomy versus laparoscopy. Am J Obstet Gynecol 1996;174:1795–9.
18. Naumann RW, Coleman RL. The use of adjuvant radiation therapy in early endometrial cancer by members of the Society of Gynecologic Oncologists in 2005. Gynecol Oncol 2007;105:7–12.
19. Pavelka JC, Ben-Shachar I, Fowler JM, Ramirez NC, Copeland LJ, Eaton LA, et al. Morbid obesity and endometrial cancer: surgical, clinical, and pathologic outcomes in surgically managed patients. Gynecol Oncol 2004;95:588–92.
20. Seamon LG, Cohn DL, Henretta MS, et al. Minimally invasive comprehensive surgical staging for endometrial cancer: robotics or laparoscopy? Gynecol Oncol 2009;113:36–41.