Introduction

The incidence of endometrial cancer is rising in developed countries and this uptick is rapid in Japan, with several-fold increases in the past few decades.1 2 The majority of endometrial cancers are diagnosed at an early stage, which carries a favorable prognosis.3 In Japan, >70% of endometrial cancers are stage I disease where survival is generally excellent (5 year overall survival rate of 93.9%).4 Primary surgical therapy with a total hysterectomy, adnexectomy, and additional lymph node evaluation in selected cases is the current standard surgical treatment approach for early disease in Japan.5

Surgical treatment for endometrial cancer continues to evolve over time. Given multiple clinical trials demonstrating comparable oncologic outcomes and improved surgical morbidity with laparoscopic surgery compared with a laparotomy approach,6 7 there has been a paradigm shift in endometrial cancer surgery towards minimally invasive hysterectomy.8 This landscape shift is also occurring in Japan.9 Multiple clinical practice guidelines, including those of the Japan Society of Gynecologic Oncology, adopted laparoscopic hysterectomy as a favored surgical approach for endometrial cancer with apparent uterine-confined disease.5 10 11

One area of active investigation and controversy is the oncologic safety of intrauterine manipulator use during laparoscopic hysterectomy for endometrial cancer.12–16 This surgical device is inserted into the endometrial cavity through the uterine cervix during laparoscopic hysterectomy to facilitate anatomical identification and exposure during surgery. Some surgeons advocate for its use to reduce surgical morbidity, while others voice concerns related to possible iatrogenic tumor spill by insertion of an intrauterine manipulator into the tumor site. This variability in opinion has resulted in a heterogenous and wide range of intrauterine manipulator use (20–90%).15–17

Moreover, results of previous investigations examining the effects of intrauterine manipulator use on pathological findings of hysterectomy specimens as well as survival outcomes in endometrial cancer are mixed,12–16 highlighting the importance and necessity to address this surgical practice. The objective of this study was to examine the association between intrauterine manipulator use and pathological factors and oncologic outcomes in patients with endometrial cancer who underwent laparoscopic hysterectomy in Japan.

MethodsTumor Registry

The tumor registry data platform of the Japan Society of Obstetrics and Gynecology was launched more than two decades ago in 2001, and has been managed and maintained by the Japan Society of Obstetrics and Gynecology Gynecologic Tumor Committee.18 Their data capturing mechanism collects ~70% of new cases of gynecologic malignancy across 477 participating sites throughout the nation.

Cohort Eligibility

This was a preplanned ancillary analysis of a previously organized database for endometrial cancer via the Japan Society of Obstetrics and Gynecology Tumor Registry.9 In this secondary analysis, the study cohort was restricted to endometrial cancer cases that had primary surgical therapy with laparoscopic hysterectomy from January 2015 to December 2017. Data on laparoscopic (including robotic assisted) hysterectomy were available starting in 2015. Among the cases that had primary laparoscopic hysterectomy, additional data were collected through the Japan Society of Obstetrics and Gynecology accredited hospitals, and detailed information pertinent to the laparoscopic hysterectomy was obtained.

Exposure

The eligible cases for analysis were grouped according to the use of an intrauterine manipulator during laparoscopic hysterectomy. Cases with unknown intrauterine manipulator use were excluded. An intrauterine manipulator was defined as a manipulator to which an instrument is connected and inserted into the endometrial cavity. Vaginal colpotomizers (eg, vaginal pipe) were not considered an intrauterine manipulator in this study.

Outcome Measures

The two main outcome measures were postoperative pathological factors and survival outcomes (overall, cause-specific, and disease-free survival) associated with intrauterine manipulator use. Overall survival was defined as the time interval between hysterectomy and death from any cause. Cause-specific survival was defined as the time interval from hysterectomy to endometrial cancer death. Disease-free survival was defined as the time interval from hysterectomy to first recurrence of disease or death. Patients who did not have the measured survival events were censored at the last visit.

Study Covariates

Preoperative factors included age, year of diagnosis, preoperative International Federation of Gynecology and Obstetrics (FIGO) cancer stage, radiographic nodal status, preoperative endometrial biopsy results, regional registry area in Japan, hospital’s surgical volume status per prior analysis,19 and hospital’s accreditation status for the Japan Society of Gynecologic Oncology and the Japan Society of Gynecologic and Obstetric Endoscopy.

Intraoperative factors included intrauterine manipulator use, lymph node evaluation, and performance of additional surgical procedures (eg, fallopian tubal occlusion at the beginning of surgery, surgical specimen retrieval through the vaginally approached colpotomy, and transvaginal upper vaginal cuff creation).20 Postoperative factors included FIGO cancer stage, histology subtypes, lymphovascular space invasion, malignant peritoneal cytology, depth of myometrial tumor invasion, pathological lymph node metastasis, postoperative adjuvant therapy, and survival information, including follow-up period and vital status.

Statistical Analysis

Independent preoperative and intraoperative characteristics associated with intrauterine manipulator use were assessed with a multivariable binary logistic regression model. All preoperative and intraoperative factors were considered in the modeling. Multicollinearity was assessed among the measured study covariates. The effect size for intrauterine manipulator use was expressed as adjusted-odds ratio (aOR) and corresponding 95% confidence interval (CI).

Postoperative pathological factors associated with intrauterine manipulator use were assessed by creating a propensity score matched cohort. Preoperative and intraoperative measured confounders between the two exposure groups were considered to generate a propensity score with a binary logistic regression model. An automated algorithm was used to perform 1-to-1 matching with the optimal caliper width for estimating differences of equal to 0.2 of the standard deviation for the logit of the propensity score (0.04). Final pathological factors and postoperative therapy were evaluated in a multivariable binary logistic regression analysis, expressed with aOR and 95% CI.

Oncologic outcomes related to intrauterine manipulator use were also assessed in the propensity score matched cohort. Point estimates for the 5-year survival rate were computed with the Kaplan–Meier method, and a Cox proportional hazard regression model was fitted to estimate the hazard risk for the intrauterine manipulator group in comparison with the non-manipulator group. To account for the possible mediator effect of pathological factors associated with intrauterine manipulator use on oncologic outcome, the exposure–outcome relationship was further adjusted for the intrauterine manipulator related pathologic factors identified in the prior step analysis (ie, lymphovascular space invasion and malignant peritoneal cytology) in addition to historical prognostic factors (nodal metastasis and postoperative therapy). The survival effect for intrauterine manipulator use compared with non-use was expressed with an adjusted hazard ratio (HR) and corresponding 95% CI.

Various sensitivity analyses were undertaken to assess the robustness of the study findings. First, the exposure–outcome association was examined in several subcohorts of interest according to patient clinicopathological characteristics. Second, cluster modeling according to the stratification pattern of intrauterine manipulator use and its associated pathological factors (ie, lymphovascular space invasion and malignant peritoneal cytology) was performed in an exploratory fashion. Last, anatomical sites of the first recurrence were assessed.

Statistical interpretation followed a two tailed hypothesis, and a p-value <0.05 was considered statistically significant. IBM SPSS Statistics (V.26.0, Armonk, New York, USA) was used for the analysis. This study followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) reporting guidelines to summarize the performance of the cohort study.21

ResultsStudy Population

A total of 3846 patients were examined for analysis (Online Supplemental Figure S1 and Online Supplemental Tables S1–S3), including 1607 (41.8%) patients who had intrauterine manipulator use and 2239 (58.2%) patients who did not. Median age was 56 years (IQR 49–65). Most tumors were stage I disease (96.5%) and grade 1–2 endometrioid histology (81.9%).

Characteristics of Intrauterine Manipulator Use

A preoperative diagnosis of atypical endometrial hyperplasia was associated with intrauterine manipulator use, whereas hysterectomy at Japan Society of Gynecologic and Oncologic Endoscopy accredited centers was less likely to be associated with intrauterine manipulator use (aOR 0.30, 95% CI 0.23 to 0.39). Patients with intrauterine manipulator use were less likely to undergo lymph node evaluation (aOR 0.85, 95% CI 0.73 to 0.99) compared with those in the non-manipulator group (Online Supplemental Table S4).

Pathological Factors Related to Intrauterine Manipulator Use

A total of 1274 patients in the intrauterine manipulator group were compared with 1274 patients in the non-manipulator group in the matched model (Online Supplemental Table S5 and Table 1). Patients in the intrauterine manipulator group were 35% more likely to have lymphovascular space invasion (17.8% vs 13.3%, aOR 1.35 95% CI 1.08 to 1.69) and 77% more likely to have malignant peritoneal cytology (10.8% vs 6.4%, aOR 1.77, 95% CI 1.29 to 2.31) compared with those in the non-manipulator group. Based on these results, lymphovascular space invasion and malignant peritoneal cytology were defined as the possible pathological mediators associated with intrauterine manipulator use in this study. In the clustered model (Online Supplemental Table S6), about one in eight cases were classified as intrauterine manipulator related pathological factor groups (13.2%).

Table 1

Pathological factors associated with intrauterine manipulator use

Oncologic Outcome Associated With Intrauterine Manipulator Use

Median follow-up was 5.0 years. The 5-year overall survival rates were 94.2% for the intrauterine manipulator group and 96.6% for the non-manipulator group (HR 1.64, 95% Cl 1.12 to 2.39, Figure 1A and Online Supplemental Figure S2). After adjusting for intrauterine manipulator related pathological factors (lymphovascular space invasion and malignant peritoneal cytology), nodal metastasis, and postoperative therapy, adjusted HR for overall survival was 1.36 (95% CI 0.93 to 2.00, Figure 2). These pathological mediator associations remained consistent for cause-specific survival and disease-free survival (Figures 1B,C and 2 and Online Supplemental Figure S2)

Figure 1

Survival associated with intrauterine manipulator (IUM). Survival curves for intrauterine manipulator and non-manipulator groups are shown for (A) overall survival (5-year rates 94.2% vs 96.6%), (B) cause-specific survival (95.9% vs 97.9%), and (C) disease-free survival (90.0% vs 92.1%). Log rank test for p-values. Axes were truncated to maximize visualization.

Figure 2

Sensitivity analysis. Effects of intrauterine manipulator (IUM) use on survival outcomes (post-mediator adjusting) were examined for overall survival (OS), cause-specific survival (CSS), and disease-free survival (DFS). *Possible pathological mediator accounted hazard estimates for outcome measures are shown (IUM vs non-IUM). The exposure–outcome association was adjusted for IUM related pathological mediators (lymphovascular space invasion and malignant peritoneal cytology), nodal metastasis, and postoperative therapy. G1–2 endometrioid, grade 1 and 2 endometrioid carcinomas; G3 end, non-end, grade 3 endometrioid carcinoma, serous carcinoma, clear cell carcinoma, and carcinosarcoma.

The clustered analysis per lymphovascular space invasion and malignant peritoneal cytology status had three notable observations to assess the possible pathological mediator effects on survival (Figure 3, Table 2, and Online Supplemental Tables S7–S9). The 5 year recurrence rates for all five anatomical sites differed across the eight clustered groups (all p<0.001; Online Supplemental Table S10). The exposure–outcome association was examined in a total of six subcohorts (Figure 2). These included age <60 years (n=1511), age ≥60 years (n=951), preoperative grades 1–2 endometrioid tumors (n=2012), preoperative grade 3 endometrioid and non-endometrioid tumors (n=134), preoperative FIGO stage I disease (n=2396), and preoperative FIGO stage I, grades 1–2 endometrioid tumors (n=1985). In these subcohort evaluations, the survival association of intrauterine manipulator use was consistent with the whole cohort (all p>0.05).

Figure 3

Overall survival based on intrauterine manipulator (IUM) related pathological factors. Survival curves are shown based on the combination pattern of IUM use and its associated pathological factors (lymphovascular space invasion (LVSI) and malignant peritoneal cytology (MPC)). Eight groups (Grp 1–8) based on the combination of IUM use, LVSI, and MPC were examined (definition in online supplemental Table S6; survival meta-data in Table 2). Red letters indicate IUM use. Log rank test for p-value. X and Y axes were truncated at 60–100% and 0–60 months to maximize visualization.

Table 2

Point estimates for overall survival

DiscussionSummary of Main Results

There were two key findings in this study. First, intrauterine manipulator use at the time of laparoscopic hysterectomy for endometrial cancer was associated with an increased risk of prognostic factors for endometrial cancer: lymphovascular space invasion and malignant peritoneal cytology. Second, the results of survival data suggested a possible mediator effect of intrauterine manipulator-introduced pathological factors on endometrial cancer survival.

Results in the Context of Published LiteratureIntrauterine Manipulator Use and Lymphovascular Space Invasion

Lymphovascular space invasion is a well established pathological factor that is associated with decreased survival in endometrial cancer, including low risk groups.22–24 Whether intrauterine manipulator use leads to mechanical tumor cell shedding into the lymphatic and/or vascular capillary space in the myometrium has been debated in recent decades.25 The results of this study were similar to the 2021 Spanish multicenter study that demonstrated an increased risk of lymphovascular space invasion with intrauterine manipulator use versus non-use during laparoscopic hysterectomy (24.6% vs 11.7%).12

In contrast, a 2023 Italian meta-analysis summariezed the results of nine studies assessing lymphovascular space invasion.16 In a pooled analysis, the incidences of lymphovascular space invasion were 19.5% versus 13.8% for the intrauterine manipulator and non-manipulator groups, respectively (OR 1.11, 95% CI 0.67 to 1.90). One explanation for these heterogenous findings across past and current investigations may be related to differences in study populations such that distribution of cancer stage and histology type differed among these studies.12–14 16 25–28 As the baseline incidence of lymphovascular space invasion differs based on these two tumor factors, an analysis accounting for cancer stage and histology type would be of value.29 An exploratory analysis suggested that the survival effect of intrauterine manipulator may be larger among studies with lower lymphovascular space invasion rates.29 Other speculation may be that lymphovascular space invasion related to intrauterine manipulator may represent indirect evidence of excess manipulation of the intrauterine manipulator that may result in tumor spill.

Intrauterine Manipulator Use and Malignant Peritoneal Cytology

Malignant peritoneal cytology is another prognostic factor in endometrial cancer, including both early and advanced stages as well as endometrioid and non-endometrioid tumors.30–32 A 2011 meta-analysis suggested iatrogenic retrograde transtubal tumor spread via hysteroscopy use as a possible etiology of malignant peritoneal cytology.33 With regards to laparoscopic hysterectomy for endometrial cancer, the association of intrauterine manipulator use and malignant peritoneal cytology has been an area of active interest.14 16 34 35 A 2022 prospective trial of 124 laparoscopic hysterectomies for endometrial cancer showed that possible iatrogenic introduction of malignant peritoneal cytology with an intrauterine manipulator was seen in 8% of surgeries,35 whereas a 2021 randomized controlled trial reported that none of 78 patients with intrauterine manipulator use had malignant peritoneal cytology.14

More recently, a 2023 meta-analysis of six studies examined the incidence of malignant peritoneal cytology related to intrauterine manipulator use.16 In a pooled analysis (n=227 for intrauterine manipulator vs n=375 for non-manipulator), the incidences of malignant peritoneal cytology were 9.3% versus 5.3% (OR 1.66, 95% CI 0.67 to 4.16). The CI of the study widened compared with the prior 2022 meta-analysis (low–high range, 3.49 vs 2.12),15 implying variable findings across the studies. Taken together, more robust data with a large sample size to examine the effect of intrauterine manipulator use on malignant peritoneal cytology is useful given the scarcity of previous studies.

Intrauterine Manipulator Use and Endometrial Cancer Survival

The results of the survival analysis in this study are nuanced in that they both support and discourage the use of an intrauterine manipulator during laparoscopic hysterectomy for endometrial cancer, as was discussed by several investigators.12–14 In support of intrauterine manipulator use, several studies have reported no association between intrauterine manipulator use and endometrial cancer survival.12 13 A 2023 meta-analysis of three studies reported a pooled HR of 1.07 (95% CI 0.65 to 1.76) although their follow-up was relatively limited (median 3.2–3.8 years).16

Based on the observations in this study, two hypotheses can be derived. First, if intrauterine manipulator use does not introduce lymphovascular space invasion and/or malignant peritoneal cytology, it may not have a survival effect. Second, if these adverse pathological factors are already present, addition of an intrauterine manipulator may not impact oncologic outcomes. As surgeons would not know these histopathological data until after hysterectomy, extra care must be taken when introducing this instrument into the tumor site within the uterus.

There are also data discouraging intrauterine manipulator use. The 2023 meta-analysis also shows an adverse survival effect of intrauterine manipulator use on disease recurrence with a marginal lower boundary of CI (five studies, HR 1.52, 95% CI 0.99 to 2.33).16 Data observed here and by others suggest a possible link between intrauterine manipulator use and adverse endometrial cancer outcomes, which may be mediated by intrauterine manipulator introduced adverse pathological factors.12 This type of iatrogenic tumor spill and decreased survival was previously reported in ovarian cancer surgery in the setting of intraoperative capsule rupture,36 including the laparoscopic approach.37 However, as causality was not assessable in a retrospective study, prudent and careful interpretation is necessary for the our study results.

Strengths and Weaknesses

A larger sample size and adequate follow-up compared with previous investigations enhanced the interpretation of our study findings.16 Key limitations of our study include unmeasured confounding bias with lack of information related to details of intrauterine manipulator use (the decision making process, surgeon’s criteria, intrauterine manipulator type, timing of insertion, and uterine perforation), peritoneal cytology sampling timing (before or after intrauterine manipulator use), lymphovascular space invasion data (extent (focal or extensive) and location (inner or outer half)), surgeon parameters (specialty and experience), preoperative hysteroscopy use, past history of tubal ligation or salpingectomy, patient body habitus, gross tumor contamination to surgical field at colpotomy, tumor / uterine size, and use of robotic platform.

The majority of our study population had low-grade, early stage endometrial cancer, and therefore the generalizability of the study results in high-grade or advanced stage endometrial cancer remains unknown. Data on molecular classification and use of sentinel lymph node biopsy with ultrastaging were not available in this study.

Implications for Practice and Future Research

This hypothesis generating observation warrants further investigation in a prospective setting to assess the oncologic effect of intrauterine manipulator use during laparoscopic hysterectomy for endometrial cancer, as was endorsed by other investigators.16 Such a trial would need to take the aforementioned unmeasured confounders into account, especially the factors related to intraoperative tumor spill with intrauterine manipulator use.

A 2021 US survey study reported that a large number of gynecologic oncologists have experienced tumor spill during laparoscopic hysterectomy for endometrial cancer, such as uterine perforation with intrauterine manipulator or tumor spill during colpotomy.38 Oncologic outcomes related to intrauterine manipulator use could be related to these spill events, but this has not been evaluated in endometrial cancer surgery to date. In the absence of more data, careful patient selection for intrauterine manipulator use and its manipulation during surgery are highly recommended (level 3 evidence, Online Supplemental Table S11). Developing surgical techniques to replace the need for an intrauterine manipulator would also be necessary.

Conclusion

These real world data in Japan suggest that intrauterine manipulator use during laparoscopic hysterectomy for early stage, low-grade endometrial cancer may be associated with an increased risk of lymphovascular space invasion and malignant peritoneal cytology. Further investigation in a prospective setting is warranted to assess the possible mediator effects of intrauterine manipulator use on survival.

Data availability statement

No data are available. This study used the Japan Society of Obstetrics and Gynecology database. The Japan Society of Obstetrics and Gynecology Committee forbids the transfer, rent, or sale of the data to any third party without prior approval. For inquiries about access to the data used for this study, Japan Society of Obstetrics and Gynecology can be contacted at +81-03-5206-1982.