Authors

1. Lockwood, Suzy PhD, MSN, RN, OCN, CHPN
2. Ritzert, Barbara PhD, RN

Abstract

Review question/objective: The objective of this review is to synthesize the evidence from economic evaluations of the resource use and cost effectiveness of serum CA125 testing compared to transvaginal ultrasound screening for ovarian cancer.

The review questions are: What are the estimates of the cost effectiveness of serum CA125 testing in females of any age compared to the use of transvaginal ultrasound testing for ovarian cancer screening? If the evidence presents a wide range of cost effectiveness measures, can any conclusions be drawn regarding the circumstances in which serum CA125 is likely to be more effective and less costly than transvaginal ultrasound as a screening test for ovarian cancer?

The ultimate objective of the review is to provide information that can be used by policy makers and administrators to make evidence-informed recommendations regarding ovarian cancer screening.

Background: Ovarian cancer is the eighth most common female gynecologic cancer in the United States and fifth most common internationally.1,2 While the worldwide lifetime risk of developing ovarian cancer is 1 in 70, it remains the deadliest of all gynecologic cancers, with a five year survival rate for all stages of only 46%.1,2 Data on the incidence of ovarian cancer at an international level is difficult to ascertain due to the limited number of national cancer registries. In 2008, a total of 224,747 new cases of ovarian cancer were reported globally, with higher numbers seen in the more developed countries; particularly Europe, North America, Australia, South America, Asia and Africa.1

Early ovarian cancer causes minimal to no symptoms, consequently most females are diagnosed in an advanced stage. The prognosis for women with ovarian cancer is closely related to the stage at diagnosis. Using the staging system developed by the International Federation of Gynecology and Obstetrics (FIGO), women diagnosed at Stage I have a five year survival rate between 79-87% whereas those that present at Stage III only have a 25-23% survival rate. 3 The fact that only 20% of women present at Stage I versus 58% at Stage III makes these figures all the more significant.3 Effective early detection of ovarian cancer is beneficial not only from a patient wellbeing perspective, but is also important because by identifying cancer at an early stage, high expenditure on treatment can be averted, thereby saving scarce societal resources.

Ovarian cancer is classified into three primary histologic groups: epithelial, sex-cord stromal and germ cell, with epithelial being the most common type. Age is the primary risk factor for ovarian cancer; with 80% of cases occurring in post-menopausal women.2 Significant genetic indicators include BRCA1, BRCA2 and Lynch II syndrome. Women with a history of breast cancer have an increased risk for epithelial ovarian cancer.1,2 Recent research has suggested that risk for ovarian cancer might also be related to the number of ovulatory cycles a woman experiences.4 Lack of clarity about risk factors for ovarian cancer has increased interest in effective screening tools. To support decision making that promotes optimal health outcomes, decision makers need to understand not only what the most effective screening tool is, but also the resource implications associated with different tools and which tool the evidence suggests is most cost effective and in which settings.

Numerous prospective studies throughout the past 25 years have evaluated the role of CA-125 and/or transvaginal sonography for routine screening in both pre- and postmenopausal women. Historically, studies have rendered both positive and negative endorsements of CA-125 and/or transvaginal ultrasound as cost effective strategies to improve early detection and diagnosis for ovarian cancer.1,5,6 Although the recently published multicenter Prostate, Lung, Colorectal, and Ovarian (PLCO) cancer screening randomized trial failed to demonstrate any benefit in early diagnosis and overall survival rates between control and screening intervention groups,5 three other large-scale multicenter trials demonstrated statistically significant improvements in stage at diagnosis and two of the three studies reported improvement in survival rates as a result of screening protocols including serum analysis of CA-125 and/or transvaginal sonography.6,7

The pivotal outcome in the development of an effective screening strategy for ovarian cancer is identification of malignancy in early stages. One of the most challenging factors precluding early diagnosis of ovarian cancer is the non-specificity of symptoms or the asymptomatic nature of women in the early stages of the disease. Typically, symptoms do not present until late in disease progression and women only present when substantive symptoms have arisen, such as increased abdominal girth, bowel blockage and urinary difficulty. Despite advances in chemotherapeutic regimens, disease recurrence is very common, frequently associated with a dismal prognosis and little chance of curative treatment.1,5,8

The ideal screening protocol for ovarian cancer described by Jacobs8 demonstrates high sensitivity and specificity, with a minimum Positive Predictive Value (PPV) of 10%, and Negative Predictive Value (NPV) of 100%. In other words, a positive test is reliable only 10% of the time and a negative test result should be reliable 100% of the time. Due to the high prevalence of ovarian malignancy diagnosed at an advanced stage of disease and the staggering mortality rates associated with same, the NPV becomes extremely important.

The Positive Predictive Value (PPV) of a screening test (the number of diagnostic procedures necessary to diagnose one case of ovarian cancer) is determined by the sensitivity and specificity of the test as well as the disease prevalence. Unfortunately, when the prevalence of preclinical disease is low, as is the case with ovarian cancer, the Positive Predictive Value of a given screening test will be low, even in a screening test that demonstrates high sensitivity and specificity.9,10 For example, to achieve the benchmark of a PPV of 10%, specificity would need to reach a level of 99.6%, with sensitivity greater than 75% to detect early stage disease.11

Transvaginal ultrasound and the serum biomarker CA-125 are two of the standard methodologies used to detect ovarian cancer, in addition to physical pelvic examinations and more discriminating imaging studies used to confirm a preliminary suspicion of an ovarian neoplasm (CT, MRI, and PET scans). Transvaginal ultrasound has been established as an independent screening modality with demonstrated reliability in the measurement and evaluation of ovarian size, the detection of small masses and characterization of complex tumors with solid or papillary projections suggesting neoplasia.12-14

CA-125 is a tumor antigen originally used in a serial measurement protocol to monitor patients diagnosed with epithelial ovarian carcinoma;15 however, it has now been established as a screening biomarker for ovarian cancer utilized as a single test, serial analysis, or in combination with transvaginal ultrasonography for multimodal screening.16 Although other serum biomarkers to detect ovarian cancer have been studied during the past decade, CA-125 remains the most robust biomarker in sensitivity and specificity.17,18 The traditional threshold designating a positive level is 35 U/mL; however, recent screening trials have suggested changes in CA-125 levels over time in the context of an individual's age-specific risk may increase the sensitivity and specificity for early diagnosis, reducing false positive results.5,19,20

CA-125 is elevated (> 35 U/ml) in 85% of women with advanced stages of ovarian cancer, 50% of women with early-stage disease and 6% of women demonstrating no evidence of disease.21 The biomarker CA-125 is not produced exclusively by ovarian cancer cells, nor is it specific only to ovarian tissue. Elevations in CA-125 may be a result of non-malignant conditions such as ovarian cysts, endometriosis, pregnancy, as well as normal fluctuations occurring on a cyclic basis in premenopausal women, hence the difficulty in establishing specificity for CA-125 alone in the detection of ovarian cancer.

Olivier, et al.,22 conducted a prospective study of 132 women over a six year period in the Netherlands who were considered to be high risk for developing ovarian cancer (breast cancer 1 and breast cancer 2 germ line mutation carriers, hereditary breast and ovarian cancer families, and breast cancer patients from a hereditary breast cancer family) by means of pelvic examination, serum CA-125 sampling and transvaginal ultrasonography. The majority of the abnormal findings normalized within one to three months; however, the study did conclude the sensitivity, specificity, and PPV of CA125 and TVS are increased in women with risk factors for ovarian cancer above that of values from normal population studies.

Four major ovarian screening trials published within the past five years have studied the use of CA-125 and transvaginal ultrasound to determine if there is a shift between diagnosis from late stage to early stage, and a documented improvement in survival as a result of screening. Two trials have been completed with reported data analysis,20,23 one trial has reported partial analysis of data6 and one trial has reported diagnostic trends (stage at diagnosis) with pending analysis of the influence of screening on survival.7 With the exception of the PLCO trial (USA), all three other screening trials have reported a significant shift in diagnosis to earlier stages with screening.

The PLCO trial did not demonstrate a statistically significant difference between the stages of diagnosis in the control group versus the screened group, nor did the trial demonstrate an improvement in overall mortality between groups.5,16 Some have speculated that the lack of improved findings in this trial may be due to the lack of uniform treatment guidelines that were used to treat women with a screening abnormality.

The multicenter Japanese screening trial reported a significant shift in detection of ovarian cancers from Stages III and IV to Stages I and II; however, they have not reported analysis of mortality data for the intervention and control group.6 The University of Kentucky Ovarian Cancer Screening Trial reported both a shift in diagnosis from late stage to early stage as a result of their screening protocol, in addition to an improvement in the five-year survival rate.23

Preliminary results from the UKCTOCS trial also reported a shift toward earlier diagnosis in Stage I and II associated with the two intervention arms (TVS alone and CA-125 + TVS), however mortality analyses are still pending.24 Speculation regarding differences in protocols to interpret CA125 and ultrasound imaging, in addition to well defined treatment interventions based on screening findings may account for differences in the study results between the UKCTOCS trial and the PLCO trial. The protocol and treatment differences between the trials may provide valuable definitive evidence to demonstrate benefits in early diagnosis, resulting in improved morbidity and mortality, ultimately tipping the cost effectiveness analysis toward the positive side of the balance in the evaluation of CA-125 and/or transvaginal ultrasonography.

It is in this context that the importance of understanding estimates of the relative cost effectiveness of two well established screening methodologies, the CA-125 serum biomarker and/or transvaginal ultrasonography for the early detection of ovarian cancer should be understood.

No systematic review examining the cost effectiveness of using transvaginal ultrasonography compared to the serum tumor maker cancer antigen CA-125 for screening in females of any age for ovarian cancer has been identified by initial searches of the Cochrane Library of Systematic Reviews, JBI Library of Reviews, EMBASE, CINAHL, or PubMed/MEDLINE. The initial search did however, identify a couple of economic evaluations on the resource use/cost compared to the benefits/effects of transvaginal ultrasonography compared to serum CA-125 screening for ovarian cancer.