From 1991 to 2010, the total number of uterine corpus cancer increased by 5.7-fold in Taiwan. In addition, the annual age-specific rate nearly doubled during 2001 to 2010 when compared with 1991 to 2000 [1]. In Taiwan, there has been a noticeable increase in the number of women adopting a Western-style diet and not having children in recent years, and thus changes in reproductive behavior and an increased rate of obesity may partially be responsible for the increase in endometrial cancer.

The standard treatment for endometrial cancer is staging surgery with total hysterectomy, bilateral salpingo-oophorectomy and pelvic/para-aortic lymph node dissection followed by tailored adjuvant therapy. Surgery is traditionally performed via laparotomy. Since Childers and Surwit first described laparoscopic surgical staging for early endometrial cancer in 1992, many subsequent studies have shown that this approach is an effective alternative to open surgery with a much faster recovery and fewer complications [2-5]. However, these studies lack well-designed randomization, and most of them are retrospective in nature. In 2009, Walker et al. [6] published the initial results of a large randomized controlled trial (LAP2) by the Gynecologic Oncology Group (GOG). With a longer follow-up period, they concluded that laparoscopic staging surgery is an acceptable alternative for patients with presumed early-stage endometrial cancer for better short-term benefits including shorter hospital stay, fewer moderate-to-severe postoperative adverse events, and improved body image. They also demonstrated that this approach improved the patients’ quality of life, and, more importantly, did not compromise overall survival (OS) compared to those treated with laparotomy [7]. In a subsequent randomized controlled trial (Laparoscopic Approach to Cancer of the Endometrium, LACE), the results also demonstrated equivalent disease-free survival at 4.5 years and no difference in OS [8].

In Taiwan, there was only one previous study described the results of head-to-head comparison between laparoscopic and laparotomy surgery for endometrial cancer with limited case number. We started to treat some patients with early endometrial cancer laparoscopically at our department since 2007. Therefore, in this study, we aimed to evaluate whether this surgical approach could be the preferred procedure for these patients compared with conventional open surgery at our institute. We used propensity score matching (PSM) analysis to eliminate the imbalance between groups and reduce the effects of confounding to achieve a random effect in this observational study. Furthermore, we attempted to analyze factors predicting prognosis.

In the present study, we demonstrated the safety and feasibility of laparoscopic surgery for the management of presumed early-stage endometrial cancer. In addition, this surgical approach provided better peri- and post-operative outcomes, and, most importantly, did not compromise survival outcomes. The major factors predicting poor survival for clinical early stage disease were high-grade tumor, non-endometrioid histology, age $>$60, and CA125 $>$35 U/mL but not surgical type. Although our results are similar to previous reports, there are some issues needed to be addressed.

The question of whether to perform pelvic lymphadenectomy routinely has been debated. Similar to LAP2 trial, we performed pelvic lymphadenectomy in all patients during the study period. However in LACE trial, only half of patients had lymphadenectomy. When we looked at the pathological findings, 23% of ours and 12% of LAP2 patients had deep myometrial invasion while none in LACE trial [6,8]. Furthermore, as high as 70% of patients in LACE trial had disease limited to endometrium [8]. That’s why surgeons in LACE trial aborted lymphadenectomy in some patients. Caution should still be made because it is difficult to preoperatively identify these low-risk patients based on gross observation and frozen section results. Studies had demonstrated the uncontrollable variables of change in grade and depth of myometrial invasion on final pathology [9].

Regarding para-aortic lymph node (PAN) retrieval, we performed this procedure in about 70% of patients while over 90% of patients in LAP2 trial had PAN lymphadenectomy. Previous studies have reported that para-aortic lymphadenectomy did not improve clinical outcomes, because the presence of PAN metastasis indicates systemic disease [10-12]. In a study from Northern Taiwan, reported by Chu et al. [5], PAN lymphadenectomy was performed in only 2.8% and 13.6% of patients in the laparoscopic and laparotomic group, respectively. Another single arm study from Taiwan reported by Lee et al. [13], 14.3% of patients had received PAN lymphadenectomy. However, the 5-year DFS and OS rates in both studies were not inferior to the LAP2 result. The potential risks of routine para-aortic lymphadenectomy include a considerably longer operative time, greater blood loss, and higher rate of post-operative ileus [14,15]. Therefore, the NCCN (National Comprehensive Cancer Network) panel has changed their recommendations on PAN lymphadenectomy since 2014. They recommend such a procedure for selective high-risk situations, including those with positive pelvic nodes [16] or high-risk histologic features [17].

Another interesting issue is the survival outcomes between laparoscopic and laparotomic groups. Most studies have shown equivalent DFS and OS rates between patients undergoing different surgical approaches. Initially, we found better DFS and OS rates in the laparoscopic group, the survival benefits disappeared after a PSM analysis to balance the baseline clinicopathological characteristics such as age, CA125, and histologic type and grade. An unfavorable histology such as serous or clear cell and high-grade endometrioid carcinoma tend to enhance extra-uterine spreading in the early stage of the disease [18,19]. The study reported by Chu et al. showed a better survival than ours could be explained by excluding non-dometrioid histology in their study. In addition, we previously found that a pretreatment CA125 level of more than 40 U/mL was a risk factor for lymph node metastasis [20]. For these reasons we prefer not to perform laparoscopic surgery for these high-risk patients at our hospital, and this may have led to the higher mean level of CA125 and higher proportion of unfavorable histology and grade, and also higher rate of postoperative adjuvant therapy in the laparotomy group. Our multivariate analysis confirmed that the type of surgery did not have any impacts on DFS and OS independently after adjusting for histology, grade, CA125 level, FIGO stage and age.

It is generally believed that operating on obese patients can be challenging especially when a new surgical approach is introduced. Furthermore, the results of LAP2 trial showed that risk of conversion to laparotomy increased with increasing BMI. Based on these findings, we strictly selected our candidates for laparoscopic approach surgery in our study resulting in a low mean BMI of 24.9 kg/m${}^2$, and fortunately none had a conversion to laparotomy. Even though the mean BMI was higher in our laparotomic group, the value was still much lower than that of LAP2 and LACE trials (28–33 kg/m${}^2$). The exact reason is not clear but may be due to the difference of obese prevalence between Asian and White ethnicity [21]. In LAP2 and LACE trials, almost 90% of patients enrolled were White race. In a Japanese study investigating different surgical approaches for early endometrial cancer showed a median BMI of 23–24 kg/m${}^2$ in 120 cases, which was much close to our patients [22].

Last but not least, the recurrence in our study occurred mostly in the local site, including vagina and pelvis, and distant lung metastasis took the second spot for both laparoscopic and laparotomic groups. These results were similar to the LAP2 study, but different from the Chu et al. which showed more lung metastasis in laparoscopic group. According to the explanation of Chu et al., this difference might be related to the inconsistent criteria of receiving postoperative adjuvant brachytherapy between two different types of surgery. Because of the fear of vaginal stump recurrence in the laparoscopic group, more had brachytherapy and no had received chemotherapy although not significant. The real causative reason contributing to this different recurrent pattern in the Chu et al. maybe not clear, but one supposed mechanism is that increased intra-abdominal pressure during laparoscopic surgery may push tumor cells into the lymphovascular space and might cause distant spread of the tumor sequentially [23]. In our study, we had the same criteria of postoperative adjuvant brachytherapy for both groups. However, more multiple recurrent sites in laparotomic group were noted, this could be explained by more unfavorable histology or high-grade lesions.

Although the patients enrolled were not assigned randomly to different surgical approaches which was a limitation of our study, we performed a PSM and successfully to eliminate a certain confounding factors. When randomized trial was limited by the objective conditions, PSM analysis could be applied to reduce selection bias. Another limitation of our study is a single institutional design; however, only 2 surgeons involved in laparoscopic surgery carried the strengths of unique operative procedures and maintaining sufficient experience during the period of patients’ accrual.

This study demonstrated the feasibility of laparoscopic surgery in clinical early stage endometrial cancer patients with the benefits of shorter hospital stay and less blood loss. The survival outcomes were comparable to a laparotomic approach. Factors associated with survival were high-grade tumor, non-endometrioid histology, age $>$60, and CA125 $>$35 U/mL. In selected patients and under experienced surgeons, laparoscopic surgery can be performed safely considering both short-term and long-term outcomes.

HL, FTK, and YCO, design the study; HCF, CHW, CCC, and CCT, analysis or interpretation; FTK, CHW, YJC, CCT, and CCC, literature search; HL and YJC, writing manuscript; HL, HCF, and YCO, critical review. All authors read and approved the final manuscript.

All clinical investigations are conducted according to the Declaration of Helsinki principles. The present study was approved by the Institutional Review Board of Chang Gung Memorial Hospital (approval number: 105-4364C), and the requirement of written informed consent was waived.

This work was partly supported by the Chang Gung Memorial Hospital Research Grants CZRPG 880011, CDRPG 880012, and CDRPG 880013. We appreciated Miss Hsin-Yi Chien and Chih-Yun Lin from the Biostatistics Center of Kaohsiung Chang Gung Memorial Hospital for their assistant in statistic analysis.

This study has not received external funding.

The authors declare no conflict of interest.