Cause of ovarian cancer

This statement summarizes the U.S. Preventive Services Task Force (USPSTF) recommendation on screening for ovarian cancer and the supporting scientific evidence, and updates the 1996 recommendations contained in the Guide to Clinical Preventive Services, 2d ed. (1) In 1996, the USPSTF recommended against routine screening for ovarian cancer (D recommendation). (1) Since then, the USPSTF criteria to rate the strength of the evidence have changed. (2) Therefore, this recommendation statement has been updated and revised based on the current USPSTF methodology and rating of the strength of the evidence. Explanations of the current taskforce ratings and strength of overall evidence are given in Tables 1 and 2, respectively.

The complete information on which this statement is based, including evidence tables and references, is available in the brief evidence update (3) on this topic on the USPSTF Web site (http://www.preventiveservices. ahrq.gov). The recommendation statement and brief evidence update also are available in print from the Agency for Healthcare Research and Quality Publications Clearinghouse (telephone: 1-800-358-9295; e-mail: ahrqpubs@ahrq.gov). The recommendation also is posted on the Web site of the National Guideline Clearinghouse (http://www.guideline.gov).

This recommendation statement was first published in Ann Fam Med 2004;2:260-2.

Summary of Recommendation

The USPSTF recommends against routine screening for ovarian cancer. D recommendation.

The USPSTF found fair evidence that screening with serum CA-125 level or transvaginal ultrasound can detect ovarian cancer at an earlier stage than it can be detected in the absence of screening; however, the USPSTF found fair evidence that earlier detection would likely have a small effect, at best, on mortality from ovarian cancer. Because of the low prevalence of ovarian cancer and the invasive nature of diagnostic testing after a positive screening test, there is fair evidence that screening could likely lead to important harms. The USPSTF concluded that the potential harms outweigh the potential benefits.

Clinical Considerations

* There is no existing evidence that any screening test, including CA-125, ultrasonography, and pelvic examination, reduces mortality from ovarian cancer. Furthermore, existing evidence that screening can detect early-stage ovarian cancer is insufficient to indicate that this earlier diagnosis will reduce mortality.

* Because there is a low incidence of ovarian cancer in the general population (age-adjusted incidence of 17 per 100,000 women), screening for ovarian cancer is likely to have a relatively low yield. The great majority of women with a positive screening test will not have ovarian cancer (i.e., they will have a false-positive result). In women at average risk, the positive predictive value of an abnormal screening test is, at best, approximately 2 percent (i.e., 98 percent of women with positive test results will not have ovarian cancer).

* The positive predictive value of an initially positive screening test would be more favorable for women at higher risk. For example, the lifetime probability of ovarian cancer increases from about 1.6 percent in a 35-year-old woman without a family history of ovarian

cancer to about 5 percent if she has one relative and 7 percent if she has two relatives with ovarian cancer. If ongoing clinical trials show that screening has a beneficial effect on mortality rates, then women at higher risk are likely to experience the greatest benefit.

Discussion

Ovarian cancer is the fifth leading cause of cancer death among women in the United States, accounting for an estimated 25,400 new cases and 14,300 deaths in 2003. (4) Several risk factors are associated with ovarian cancer. Family history increases the risk for ovarian cancer: having one first- or second-degree relative with ovarian cancer increases risk by about threefold. (5) Carriers of the BRCA1 or BRCA2 gene mutations also are at increased risk. (6) The risk for developing ovarian cancer is reduced with oral contraceptive use and pregnancy of any duration. (7) Some studies have shown that postmenopausal women taking estrogen may be at increased risk for developing ovarian cancer. (8,9)

Most women with ovarian cancer have nonlocalized disease at the time of diagnosis. (4) A randomized controlled trial (RCT) using multimodal screening (CA-125 screening, followed by ultrasonography in patients with abnormally elevated levels) reported that 50 percent of patients with ovarian cancer in the screened group were in stage I, compared with only 5 percent in the control group. (3,10) This difference was not statistically significant. Two large cohort studies using transvaginal ultrasonographic screening reported that 59 to 65 percent of ovarian cancers were diagnosed in stage I. (11,12) However, there is no evidence that detecting earlier-stage tumors through screening leads to a decrease in ovarian cancer-specific mortality.

Establishing the true sensitivity of CA-125 or ultrasonography is limited by several factors. The studies assessing the accuracy of screening tests have used different thresholds to define an elevated CA-125 level and different lengths of clinical follow-up, and have included small numbers of patients. In women at average risk for ovarian cancer, using thresholds of 30 U per mL or 35 U per mL, the one-year follow-up sensitivity of CA-125 screening, followed by ultrasonography, has been reported to be about 80 percent; the specificity is nearly 100 percent. (13-15) However, using a similar CA-125 threshold for women at high risk for ovarian cancer, the sensitivity would be reduced to 50 percent. The estimated sensitivity of annual transvaginal ultrasonography at one-year follow-up is 88 percent (95 percent confidence interval [CI], 47 to 100 percent); the specificity is estimated to range from 97 to 99 percent. (16) There is conflicting evidence as to whether adding color Doppler imaging to ultrasonographic screening can reduce the rate of false-positive test results. (17,18) There are few data available to determine the sensitivity and specificity of successive rounds of screening.

There is a significant potential for harms associated with screening for ovarian cancer, although there are few data to assess the magnitude of harms from screening, such as needless surgery or increased anxiety. A study by the British Health Technology Assessment program estimated that screening a hypothetical cohort of 10,000 women aged 50 to 64 for ovarian cancer, using either annual CA-125 or twice-yearly transvaginal ultrasonography (assuming specificities of 97 and 93 percent, respectively), would result in 300 women (using CA-125) or 350 women (using ultrasonography) who do not have ovarian cancer being recalled each year for further assessment, resulting in potential distress and anxiety in otherwise healthy women. (16) Of these, 20 (using CA-125) or 65 (using ultrasonography) women without ovarian cancer would undergo surgery each year. For women at average risk for ovarian cancer, the positive predictive value of an abnormal screening test is, at best, approximately 2 percent. On the other hand, the potential benefits of screening (based on this model's optimistic assumption that earlier treatment leads to a 40 percent mortality reduction) would yield a maximum of four additional cancers detected per year and would result in 1.5 additional five-year survivors for each year of screening.

Although no RCT of screening for ovarian cancer with mortality outcomes in the general population has yet been complete& at least three such RCTs are currently in progress: the United Kingdom Collaborative Trial of Ovarian Cancer Screening; the National Institutes of Health Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial; and the European Randomized Trial of Ovarian Cancer Screening. (19-21)

Recommendations of Others