Examining the Evidence
Is ovarian Ca screening effective in postmenopausal women?
Ovarian cancer is relatively rare, but late diagnosis leads to a higher death rate than the death rate observed with other gynecologic malignancies. Detection at an early stage (I or II) is associated with higher survival than detection at a later stage (III or IV).
Earlier studies have suggested that ovarian cancer screening with the serum biomarker CA 125 and transvaginal ultrasonography (TVS) may help the clinician diagnose ovarian cancer at an earlier stage.
British study finds screening to be effective
In the United Kingdom Collaborative Trial of Ovarian Cancer Screening, postmenopausal women 50 to 74 years old were randomized to one of the following:
- no screening (n=101,359)
- annual assessment of CA 125 using a proprietary risk-of-cancer algorithm, with TVS as a second-line test (multimodal screening, or MMS) (n=50,078)
- annual TVS in a 2:1:1 ratio (n=48,230). (A scan was considered abnormal when one or both ovaries had complex morphology or a simple cyst exceeded 60 cm3 in size.)
The mean age at screening was 60.6 years, and 96.5% of participants were white.
Women who underwent MMS were triaged according to their estimated risk of developing ovarian cancer, based on the CA 125 level and age-specific risk estimates. Women in the group with the lowest risk of ovarian cancer continued annual assessment of CA 125, whereas those at highest risk underwent repeat measurement of CA 125, followed by TVS if the repeat assay suggested elevated risk. If TVS findings were also abnormal, the patient underwent clinical evaluation.
In the MMS and TVS groups, 8.7% and 12.0% of subjects, respectively, underwent clinical assessment, and surgery was performed in 0.2% and 1.8%, respectively. A primary ovarian or tubal cancer was diagnosed in 42 and 45 women in the MMS and TVS groups, respectively. In addition, 8 (MMS) and 20 (TVS) borderline malignancies were identified.
Overall, 48.3% of invasive cancers were diagnosed during stage I or II, with no difference in the distribution of stages between MMS and TVS groups. The positive predictive value (PPV) of MMS and TVS for primary invasive epithelial and tubal cancers was 35.1% and 2.8%, respectively. The ratio of surgery to case of invasive ovarian cancer was 2.9:1 (MMS) and 35:1 (TVS).
In the US study, women underwent both CA 125 assessment and TVS
The US study involved four annual rounds of screening in women 55 to 74 years old, who underwent both CA 125 measurement and TVS imaging in each round. A CA 125 level of 35 U/mL or above, or ovarian volume greater than 10 cm3 (or detection of an ovarian cyst with complex morphology) was considered abnormal.
Of 34,621 women randomized to screening, 30,630 underwent at least one screen during the four rounds. Almost two thirds of participants were 55 to 64 years old, and almost 89% were non-Hispanic white.
The percentage of women who had at least one positive screen decreased over the four rounds of screening, from 5.8% in year 1 to 4.9%, 4.6%, and 4.5% in years 2 through 4, respectively. In each round of screening, TVS was more likely to be positive than was CA 125 measurement (e.g., 4.6% vs 1.4% in the first round).
Of the 28,746 women who underwent the initial (prevalence) screen, 1,675 (5.8%) had positive findings, 566 (1.97% of those who were screened) underwent surgery, and 27 neoplasms were detected (0.06% of those who were screened), including 18 ovarian or primary peritoneal invasive cancers. Nine borderline tumors were also identified.
The PPV of a positive screen in the first round was 1.1%. Of the 18 invasive cancers identified in this round, 16.7% were diagnosed during stage I or II, and 83.4% were identified during stage III or IV. The ratio of surgery to cases of invasive ovarian cancer was 31.4:1.
Why did the trials have different findings?
The findings of the UK study suggest that an MMS strategy of CA 125 assessment, followed selectively by TVS, can detect early-stage ovarian cancer with an acceptable PPV and ratio of surgery to case of invasive cancer.
In contrast, the US findings are discouraging because of the low PPV, the large percentage of malignancies detected at an advanced stage, and the high ratio of surgery to cases of invasive cancer.
Although the studies had different primary outcomes—ovarian and tubal cancers in the UK and ovarian and primary peritoneal cancers in the US—a majority of invasive malignancies detected in both studies were ovarian.
In the US study, the poor performance of screening may be due, in part, to universal rather than selective use of TVS; that modality generates substantially more false positives than does CA 125. The US study also defined an ovarian abnormality more broadly (volume greater than 10 cm3) than the British study did (volume greater than 60 cm3), which may have lowered the PPV in the US trial.
As investigators in the UK continue to follow participants and report their findings, we will learn more about the value of MMS, including its impact on cancer mortality, possibly as soon as 2014.
WHAT THIS EVIDENCE MEANS FOR PRACTICE
Until we have further data on the value of ovarian cancer screening in postmenopausal women, we should continue the current practice of screening only symptomatic or very-high-risk women for ovarian cancer.—ANDREW M. KAUNITZ, MD