Cost effectiveness of a test to detect metastases for endometrial cancer.

OBJECTIVE

To estimate the potential cost-effectiveness of a hypothetical test to screen for lymph node metastases in women with newly diagnosed, apparent early stage endometrial cancer.

METHODS

A decision model was constructed to inform a choice between the following strategies: (1) Usual care, in which the probability of undergoing full surgical staging (29%) is based on literature review; (2) Non-invasive diagnostic testing for metastasis (Testing), in which patients with abnormal test results undergo full surgical staging; (3) 100% referral, in which all patients are referred for full surgical staging. Survival was modeled using Surveillance Epidemiology and End Results (SEER) database. Base case diagnostic test characteristic estimates (sensitivity 0.90, specificity 0.90) were varied for sensitivity analysis. Cost of the diagnostic test was set at $500 and varied; costs of treatment for endometrial cancer (surgery, adjuvant therapies, diagnosis of recurrence, salvage therapies and palliative care) were incorporated.

RESULTS

Usual care was the least expensive strategy, while Testing was more expensive and more effective, with an incremental cost-effectiveness ratio (ICER) of $18,785 per year of life saved (YLS) compared to Usual care. 100% referral was the most expensive and most effective strategy, with an ICER of $35,358 per YLS compared to Testing. Results are relatively sensitive to variation in test characteristics and the cost of the diagnostic test but insensitive to cost of treatment and probability of adjuvant therapies. Testing remains cost-effective compared to Usual care unless the usual rate of referral to a Gynecologic Oncologist for full staging exceeds 90%.

CONCLUSIONS

Given the current low rates of full surgical staging and/or referral to a Gynecologic Oncologist, a diagnostic test to detect nodal metastasis for endometrial cancer has potential to be cost-effective when compared to usual care. Testing is also potentially cost-effective compared to 100% referral at very high test sensitivities and at the lower range of test costs.