Which abnormal ovarian findings can be followed by serial TVUS?

Ovarian cancer causes more deaths than any other cancer affecting the female reproductive system.¹ One reason it’s so deadly: It usually isn’t detected until it has reached an advanced stage. No clear-cut symptoms point definitively to ovarian malignancy, and no feasible screening strategy has been found to increase detection at an early stage.

Among the strategies that have been utilized to detect ovarian cancer are bimanual examination of the adnexae (primarily in postmenopausal women), measurement of cancer antigen (CA) 125, and transvaginal ultrasonography (TVUS) of the ovaries. The last two strategies sometimes are combined in high-risk women.

TVUS can highlight ovarian abnormalities and provide information about their structure. The question then becomes which abnormalities are likely to resolve without treatment, and which should be scrutinized more closely. In this study, Pavlik and colleagues reviewed TVUS findings from 39,337 women enrolled in the University of Kentucky Ovarian Cancer Screening Program, which involved 221,576 baseline and interval TVUS scans.

Details of the study
Women in this study were screened with annual TVUS scans between 1987 and 2002. The population included:

• asymptomatic women aged 50 or older
• asymptomatic women over age 25 who had a first- or second-degree relative with documented ovarian cancer.

The initial TVUS scan was normal in almost 90% of women, and only about 10% subsequently experienced an abnormal scan. About half (46.7%) of the ovarian abnormalities identified via TVUS were found on the very first scan. Of these, 63.2% resolved during follow-up with no treatment.

Approximately 80% of women had no abnormal TVUS findings at any time during the observation period. This is notable because participants had a high risk for ovarian cancer by virtue of advanced age or family history.

TVUS abnormalities had a higher prevalence in premenopausal women (35%) than in postmenopausal women (17%; P<.001). The incidence of ovarian cysts also was significantly higher among premenopausal women (15.3% vs 8.2%; P<.001). These differences are to be expected, owing to the functional nature of premenopausal ovaries in regard to folliculogenesis, ovulation, and endometriosis.

Positive predictive values ranged from 15.3% to 24.7%
Over the 25 years covered by this study, our understanding of the malignant potential of various ovarian masses has evolved considerably. We have long known that unilocular cysts are extremely unlikely to be malignant, but now we are aware that even septated cysts are unlikely to represent cancer.

As for the success of this ovarian cancer-screening program, which identified 85 true malignancies and 472 nonmalignancies in surgical specimens, it had an overall positive predictive value of 15.3%. After January 1, 2008, however, when serial observation expanded to include septated cysts (because published data confirmed these masses to have low malignant potential), positive predictive value improved to 24.7%.

Pavlik and colleagues also discussed findings from the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial, which relied on a single TVUS abnormality to trigger a recommendation for surgery, with a positive predictive value of only 5.1%.²

Most cancers were diagnosed at an early stage
Of the invasive epithelial cancers identified in this study, the stage distribution at diagnosis was:

• Stage 1: 45%
• Stage 2: 23%
• Stage 3: 32%
• Stage 4: None.

This finding is notable, given statistics from the “real world,” where about 80% of ovarian cancers are diagnosed at Stage 3 or Stage 4.

Among benign findings that were managed surgically, 47% were serous cystadenomas, 13% were hemorrhagic cysts, 9% were fibromas, thecomas, or Brenner tumors, and the rest were fairly equally divided between hydrosalpinx or paratubal cysts; endometriomas; and mucinous cystadenomas, leiomyomas, and cystic teratomas.

What this evidence means for practice

In general, unilocular or septate cysts can be followed every 6 months by TVUS. Although more complex tumors may resolve spontaneously, they should be followed with serial TVUS, with caution, at intervals of 6 weeks to 3 months. The findings of each scan should determine the subsequent course of action, which could involve further monitoring or surgical extirpation.

Regrettably, this study did not utilize color flow Doppler imaging. Because malignant tumors are rich in neovascularity, and the vessels laid down by such tumors often lack a normal media layer, they often exhibit very low resistance to flow. Although neovascularity is not a perfect diagnostic indicator of malignancy, the presence of abundant blood flow and low resistance can raise the index of suspicion. In my opinion, color flow Doppler should be incorporated into ultrasonographic evaluation of potential ovarian malignancies.

—Steven R. Goldstein, MD

Tell us what you think, at rbarbieri@frontlinemedcom.com. Please include your name and city and state.