Reported September 09, 2008

TURIN, Italy, Sept. 9 -- The specificity of DNA tests for human papillomavirus (HPV) can be improved by the use of a protein that is over-expressed in cervical cancer cells, researchers here said.

Testing for HPV has been shown to be more likely to pick up cases of high-grade cervical intraepithelial neoplasia (CIN) than conventional cytology, according to Guglielmo Ronco, M.D., of the Center for Cancer Prevention, and colleagues.

But the molecular method also gives more false positives, increasing the number of unneeded referrals for colposcopy, Dr. Ronco and colleagues reported online in Lancet Oncology.

In a large randomized trial in Italy -- the New Technologies for Cervical Cancer screening study (NTCC) -- the HPV test was 63% more sensitive than conventional cytology, Dr. Ronco and colleagues said, but it was less specific for CIN because many infections regress without developing into high-grade lesions.

To improve the specificity, they considered the cyclin-dependent kinase inhibitor p16-INK4A (p16), which is considered to be a marker of HPV infection, and also of activated viral oncogenes and of virus-induced deregulation of the cell cycle.

In a sub-study of the NTCC, the researchers looked at women in the HPV-test arm who tested positive for the virus and who also had cells tested for p16 by immunostaining.

All women who tested positive for HPV were referred for colposcopy and slightly less than half had a colposcopy directed biopsy, the researchers said. All told, 508 of the 1,137 women in the sub-study had a biopsy and 629 did not.

The endpoint of the analysis was sensitivity and specificity of p16 triage for CIN2 or higher, the researchers said, but they also calculated the relative sensitivity and relative referral to colposcopy, compared with conventional cytology, that would have been obtained.

 

Dr. Ronco and colleagues found:

* Using the p16 triage in HPV-positive women of any age accurately detected CIN2 or higher in 81 of the 92 women with such lesions, for a sensitivity of 88%.
* The specificity, on the other hand, was 61%, correctly identifying 633 of 1,045 women as not having CIN2 or higher.
* Among women 35 to 60, the relative sensitivity, compared with conventional cytology, was 1.53 (with a 95% confidence interval from 1.15 to 2.02).
* In that age group, the relative rate of referral for colposcopy was 1.08 (with a 95% confidence interval from 0.96 to 1.21).
* In women ages 25 through 34, the p16 triage was three times as sensitive as conventional cytology, with a relative sensitivity 3.01 (and with a 95% confidence interval from 1.82 to 5.17).
* At the same time, the relative referral rate was again not different from conventional cytology at 1.15 (with a 95% confidence interval from 0.96 to 1.37).

"Our data show that in HPV-positive women, p16-INK4A over-expression is strongly associated with the presence of histologically confirmed CIN2+, suggesting that it actually is a marker of progression," Dr. Ronco said.

"This study supports the application of triage by P16INK4A immunostaining in HPV-positive women," he added.

The researchers cautioned that the study was preformed in women who had not been given prophylactic HPV vaccination and are not immediately applicable to vaccinated women.

The study was supported by the European Union, the Italian Ministry of Health, the Regional Health Administrations of Piemonte, Tuscany, Veneto, and Emilia-Romagna, and the Public Health Agency of Lazio Region. Dr. Ronco reported being paid for participation in advisory boards by GenProbe, whose products are not involved in the current study.

Primary source: Lancet Oncology