Objective
The objective of the study was to compare the cost-effectiveness of 3 screening options for endometrial cancer in asymptomatic, postmenopausal women prior to undergoing morcellation in minimally invasive supracervical hysterectomy and minimally invasive sacral colpopexy for the treatment of pelvic organ prolapse.
Study Design
A decision tree model was constructed to compare no screening, endometrial biopsy, and transvaginal ultrasound for asymptomatic, postmenopausal women prior to surgery. Effectiveness was measured by life-years. The incremental cost-effectiveness ratio, defined as the difference in cost between 2 screening options divided by the difference in life-years between the 2 options, was calculated in 2012 US dollars for endometrial biopsy and transvaginal ultrasound, in comparison with no screening.
Results
Using an endometrial cancer prevalence of 0.6% and a 40% risk of upstaging after morcellation, the expected per-patient cost was $8800, $9023, and $9112 over 5 years for no screening, endometrial biopsy, and transvaginal ultrasound, respectively. The expected life-years saved compared with no screening were 0.00108 for endometrial biopsy and 0.00105 for transvaginal ultrasound, ie, 0.39 and 0.38 days, respectively. The estimated incremental cost-effectiveness ratio was $207,348 for endometrial biopsy and $298,038 for transvaginal ultrasound compared with no screening. A sensitivity analysis showed that the prevalence of endometrial cancer and the risk of endometrial cancer upstaging after morcellation had the greatest impact on the cost-effectiveness of screening.
Conclusion
For asymptomatic, postmenopausal women, preoperative endometrial evaluation via endometrial biopsy or transvaginal ultrasound helps improve the preoperative detection of endometrial cancer, but universal screening is not cost effective.
Pelvic organ prolapse (POP) is defined as the descent of the anterior, posterior, or apical vaginal wall. It is a common disorder that has a great impact on a woman’s quality of life, and by 2050, the prevalence of prolapse is expected to nearly double as the elderly population grows. An effective surgical treatment for apical vaginal prolapse is a sacrocolpopexy, in which a graft is attached to the anterior and posterior aspects of the vagina and suspended from the anterior longitudinal ligament of the sacrum. In women with a uterus, many surgeons perform sacrocolpopexy with concurrent supracervical hysterectomy (SCH) because of the increased occurrence of mesh erosion that has been reported after total hysterectomy.
Minimally invasive supracervical hysterectomy and minimally invasive sacrocolpopexy, performed either laparoscopically or robotic assisted, are alternatives to the traditional open approach to treat apical prolapse. Minimally invasive approaches allow for smaller incisions and shorter hospital stays, prompting many surgeons to transition from open approaches to minimally invasive surgeries. In contrast to an open SCH, during a minimally invasive SCH, the uterus is morcellated inside the peritoneal cavity to allow for its removal through a laparoscopic port. Although this is typically done without complication, little is known about the consequences of intraperitoneal morcellation in the presence of an unanticipated uterine malignancy.
Based on the current literature, there is a 0.6% (range, 0.2–1.2%) risk of diagnosing endometrial cancer in asymptomatic postmenopausal women undergoing POP surgery. Although there are few data on the survival of women with endometrial cancer after intraperitoneal uterine morcellation, there is evidence suggesting a worse prognosis in patients who underwent morcellation of previously undiagnosed endometrial cancer. In addition, pathology reports on morcellated specimens can be inaccurate secondary to the destruction of anatomic landmarks.
Preoperative evaluation of the endometrial cavity can be performed in an outpatient setting by either endometrial biopsy or transvaginal ultrasound. The aim of this study was to assess the cost-effectiveness of 3 screening options for endometrial cancer in asymptomatic, postmenopausal women prior to undergoing minimally invasive supracervical hysterectomy with morcellation and sacrocolpopexy for the treatment of POP: no screening, endometrial biopsy, and transvaginal ultrasound.
Materials and Methods
Decision model
This study received internal review board exemption (Human Investigations Committee Protocol number 1201009598), given that only previously published data without specific human subject identifiers were used. We constructed a decision tree model to compare the cost-effectiveness of the 3 preoperative screening options: no screening, endometrial biopsy, and transvaginal ultrasound ( Figure 1 ). The distribution of endometrial cancer stages I, II, and III in asymptomatic women was based on published epidemiological data. Stage IV endometrial cancer was excluded because the likelihood of asymptomatic stage IV endometrial cancer was extremely low. In all cases, we assumed that the tumor histology was adenocarcinoma because this is the most common histology found in endometrial cancer.
The analysis was conducted from a health care system’s perspective (secondary to unavailability of data on patients’ productivity loss and transportation costs) with a 5 year time frame. Effectiveness was measured by life-years, which reflect the number of years a patient is expected to live over the 5 years (calculated based on the expected probability of mortality in each month).
The decision tree model accounted for the sensitivity and specificity of screening using endometrial biopsy or transvaginal ultrasound, prevalence of endometrial cancer by stage, differences in surgical management of patients with and without endometrial cancer, procedure-specific mesh erosion rates, adverse impact of intraperitoneal morcellation on endometrial cancer prognosis by stage, and stage-specific relative survival rates as well as costs associated with each health state and event.
Assumptions about subsequent management following the different screening tools were based on American Congress of Obstetricians and Gynecologists committee opinions, the literature, and typical standards of care. If a woman underwent a transvaginal ultrasound with a thickened endometrial stripe (≥5 mm), she was then further evaluated using endometrial biopsy ( Figure 1 ). Women found to have cancer on endometrial biopsy underwent preoperative consultation with gynecologic oncology followed by a change in surgical plan at the time of primary surgery (ie, total laparoscopic hysterectomy, bilateral salpingo-oophorectomy [BSO], lymph node dissection, and concomitant minimally invasive sacrocolpopexy).
Women incidentally found to have endometrial cancer on their postoperative pathological specimens would have first undergone a minimally invasive supracervical hysterectomy with intraperitoneal morcellation and sacrocolpopexy, followed by a postoperative referral to gynecologic oncology for consultation and a subsequent second surgery for staging of endometrial cancer including trachelectomy, BSO, and lymph node dissection. Mesh erosion rates varied, depending on what type of procedure was performed ( Table 1 ). Women suffering from mesh erosion were assumed to undergo surgical mesh revision.
| Parameter | Base value | Minimum | Maximum | Sources |
|---|---|---|---|---|
| Probability of study population having endometrial cancer | 0.6% | 0.2% | 1.2% | |
| Transvaginal ultrasound sensitivity (≤5 mm) | 97% | 94% | 98% | |
| Transvaginal ultrasound specificity (≤5 mm) | 55% | 53% | 57% | |
| Endometrial biopsy sensitivity | 94% | 84% | 99% | |
| Endometrial biopsy specificity | 99% | 99% | 100% | |
| Probability of stage I endometrial cancer at index surgery | 74.75% | 74.23% | 75.27% | |
| Probability of stage II endometrial cancer at index surgery | 13.71% | 12.37% | 15.05% | |
| Probability of stage III endometrial cancer at index surgery | 11.54% | 9.68% | 13.40% | |
| Probability of stage V endometrial cancer at index surgery | 0.00% | — | — | Assumption |
| Probability of mesh erosion following minimally invasive total laparoscopic hysterectomy, BSO, lymph node dissection, minimally invasive sacrocolpopexy | 8.10% | 2.30% | 14.50% | |
| Probability of mesh erosion following minimally invasive SCH/sacrocolpopexy | 2.50% | 0% | 2.70% | |
| Probability of mesh erosion following minimally invasive sacrocolpopexy, trachelectomy, BSO, lymph node dissection | 2.50% | 0% | 14.50% | Author’s assumption |
| Proportion of morcellated patients who were upstaged | 40.00% | 0% | 100% | |
| Probability of surviving 5 years if patient has stage I endometrial cancer | 96.20% | 95.80% | 96.90% | |
| Probability of surviving 5 years if patient has stage II endometrial cancer | 81.40% | 78.80% | 83.60% | |
| Probability of surviving 5 years if patient has stage III endometrial cancer | 62.20% | 60.10% | 64.30% | |
| Probability of surviving 5 years if patient has stage IV endometrial cancer | 18.50% | 16.60% | 20.50% | |
| Probability of surviving 5 years if no endometrial cancer | 100% | — | — | Author’s assumption |
Parameter estimates
A comprehensive literature review was performed to identify the best estimates and plausible ranges for the value of each input parameter ( Table 1 ). In the base case analysis, it was assumed that the clinical prognosis of endometrial cancer would be worsened by 1 stage (eg, stage I becomes stage II) for 40% of cases in which the uterus was morcellated intraperitoneally. This estimate was based on the only study currently available by Einstein et al, which showed a 40% risk of cancer upstaging in patients who underwent morcellation of previously undiagnosed endometrial cancer.
Given the paucity of published data on upstaging after morcellation, a sensitivity analysis was performed by varying the assumption of cancer upstaging from 0% to 100%. The 2012 Surveillance, Epidemiology, and End Results (SEER) data were used to determine the endometrial cancer specific relative survival rates by stage. For patients without endometrial cancer, the probability of suffering endometrial cancer specific death over the 5 year period was assumed to be zero.
Estimates of cost parameters for screening services, surgical procedures, pathology services, and physician visits were based on the 2012 Medicare fee schedule, which are developed to reflect the costs of providing medical services to Medicare patients and widely used in economic evaluations to measure the cost of health care services ( Table 2 ). Both hospital facility fees and physician professional fees were included in these cost estimates.
| Cost | Base value | Minimum | Maximum | Sources |
|---|---|---|---|---|
| Transvaginal ultrasound | $129.67 | $64.84 | $194.51 | Medicare 2012 fee schedule |
| Endometrial biopsy | $272.87 | $136.44 | $409.31 | Medicare 2012 fee schedule |
| Gynecological oncology consultation | $257.37 | $128.69 | $386.06 | Medicare 2012 fee schedule |
| Minimally invasive total laparoscopic hysterectomy and sacrocolpopexy, BSO, lymph node dissection | $9245.63 | $6471.94 | $12,019.32 | Medicare 2012 fee schedule |
| Minimally invasive SCH and sacrocolpopexy without pathology finding of endometrial carcinoma | $8398.59 | $5879.01 | $10,918.17 | Medicare 2012 fee schedule |
| Minimally invasive SCH and sacrocolpopexy with pathology finding of endometrial carcinoma | $8501.38 | $5950.97 | $11,051.79 | Medicare 2012 fee schedule |
| Trachelectomy, BSO, lymph node dissection | $8887.56 | $6221.29 | $11,553.83 | Medicare 2012 fee schedule |
| Mesh revision | $6746.83 | $4722.778 | $8770.88 | Medicare 2012 fee schedule |
| 5 year cost of endometrial cancer (if stage I and survived) | $22,066.81 | $20,990.12 | $23,145.11 | |
| 5 year cost of endometrial cancer (if stage I and died) | $39,227.53 | $37,377.51 | $41,079.77 | |
| 5 year cost of endometrial cancer (if stage II and survived) | $22,559.53 | $21,459.85 | $23,660.24 | |
| 5 year cost of endometrial cancer (if stage II and died) | $37,927.33 | $36,140.74 | $39,716.04 | |
| 5 year cost of endometrial cancer (if stage III and survived) | $41,509.42 | $38,022.63 | $44,996.22 | |
| 5 year cost of endometrial cancer (if stage III and died) | $59,959.11 | $54,799.94 | $65,118.29 | |
| 5 year cost of endometrial cancer (if stage IV and survived) | $72,011.56 | $62,135.46 | $81,887.58 | |
| 5 year cost of endometrial cancer (if stage IV and died) | $56,519.04 | $50,497.55 | $62,539.64 | |
| 5 year cost of endometrial cancer (if no endometrial cancer) | $0.00 | — | — | Assumption |
Subsequent cost of care for endometrial cancer over the 5 year period was estimated based on survival data from SEER and published estimates of costs associated with different phases of care (eg, initial year after diagnosis, continuing phase, last year of life) and cancer stage. When minimum and maximum values were not available for parameters from the literature, minimum and maximum values were assumed to be ±50% of their base case estimates. The medical care component of the Consumer Price Index was used to convert all cost estimates to 2012 US dollars. Future costs and life-years were discounted to values at the time of the index surgery using a 3% annual discount rate.
Data analysis
The incremental cost-effectiveness ratio, defined as the difference in costs between 2 screening options divided by the difference in life-years between the 2 options, was calculated for endometrial biopsy in comparison with no screening and for transvaginal ultrasound in comparison with no screening. The incremental cost-effectiveness ratio indicates the additional costs associated with endometrial biopsy and transvaginal ultrasound, respectively, as compared with no screening, for each additional life-year gained.
One-way sensitivity analyses were used to assess the effects of varying the values of each input parameter on the cost-effectiveness of the screening options. A 2-way sensitivity analysis was conducted varying the prevalence of endometrial cancer in an asymptomatic postmenopausal population along with the probability of cancer upstaging after intraperitoneal morcellation of a uterus with endometrial cancer. DecisionTools Suite software (Palisade Corp, Ithaca, NY) was used to construct the decision tree model and analyze the data.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
