Viral Tests Outdoes Pap Smear in Sensitivity

MONTREAL, Oct. 17 -- Viral testing to detect cervical cancer appears to outperform the conventional Pap smear in some respects, researchers here have concluded.

In a randomized trial, a test for the human papillomavirus was significantly more likely (P=0.01) to identify cases of cervical intraepithelial neoplasia (CIN) than a Pap smear, found Marie-Hlne Mayrand, M.D., of McGill University and colleagues.

On the other hand, the viral test was significantly more likely (P

"We're not there yet," Dr. Runowicz said.

Between Sept. 26, 2002 and Feb. 3, 2005, the Canadian researchers randomly assigned 10,154 women, ages 30 to 69, to one of two groups - a "focus on Pap" group, which got a Pap smear followed by the viral test, and a "focus on HPV" group, which got the HPV test first, followed by a Pap smear.

For ethical reasons, all women got both tests, which were performed at the same session, but the design allowed the researchers to compare the two tests head to head, Dr. Mayrand and colleagues said.

The researchers found that the viral test - commercially available and approved by the FDA - correctly identified cervical intraepithelial neoplasia of grade two or three 94.6% of the time, compared to 55.4% for the Pap test.

On the other hand, the specificity of the viral test was 94.1%, compared to 96.8% for Pap testing, they found.

From May 1997 through November 2000, the Swedish researchers randomly assigned 12,527 women ages 32 to 38 to get a Pap test (and form a control group) or a Pap test combined with a test for human papillomavirus using polymerase chain reaction.

They calculated rates of CIN2, CIN3, and cervical cancer at baseline and at subsequent screenings, according to Pontus Naucler, M.D., Ph.D., of Lund University in Malmo, and colleagues.

At baseline, the study found, the proportion of women in the two-test group who had CIN2, CIN3, or cancer was 51% higher than among the women who only got the Pap test.

But at the subsequent screening, following treatment where appropriate, the rate in the two-test group was 42% lower than in the control group.

Some of the differences might have been the result of over-diagnosis at baseline, the researchers said, but not all. "The improved sensitivity of HPV testing is not merely due to over-diagnosis but is attributable, at least in part, to earlier diagnosis of lesions that do not regress," they concluded.

The Swedish study was supported by the Swedish Cancer Society and Europe against Cancer. The researchers reported no potential conflicts.

Dr. Runowicz reported no potential conflicts.