DNA testing for the two most common types of cancer-causing HPV detects the most risky cervical lesions better than liquid-based cytology, data from the ATHENA trial indicate.

Detecting HPV-16 and HPV-18, the types responsible for about 70% of invasive cervical cancers, predicted high-grade cervical intraepithelial neoplasia with a 92% sensitivity compared with cytology’s 53.3% (P<0.0001), Philip E. Castle, PhD, of the American Society for Clinical Pathology in Washington, D.C., and colleagues determined.

Furthermore, combining DNA testing with liquid-based cytology offered little advantage over the HPV-test alone, the researchers reported in Lancet Oncology.

Continue Reading

“HPV testing with individual genotyping for HPV16, HPV18 or both could provide a more efficient strategy for cervical cancer screening than do existing programs based on cytology,” the researchers wrote.

The American Society for Obstetrics and Gynecology currently recommends liquid-based cytology as the screening method of choice, with DNA testing only utilized in cases of abnormal or unclear cytology. But this may soon change, as ATHENA findings along with those from several other longitudinal studies, suggest that stand-alone HPV DNA testing may be a better option in high-income countries.

In an accompanying editorial, Guglielmo Ronco, MD, PhD, of the Centre of Cancer Prevention in Turin, Italy, and colleagues wrote that the increased sensitivity observed with DNA testing could enable earlier detection of precancerous lesions, “which should provide additional (though ill-defined) protection against cervical cancer.”

The arm of the ATHENA trial that compared HPV screening methods involved 41,955 U.S. women, aged 25 or older, who had cervical specimens collected during routine screening and analyzed with three types of tests: liquid-based cytology, first-generation HPV DNA assays (Amplicor HPV test, Roche; Linear Array HPV genotyping test, Roche) a second-generation cobas HPV test with individual HPV-16 and HPV-18 detection.

The cobas test detected HPV among 4,275 women (10%), whereas 2,617 women (6%) had abnormal cytology.

Although the analysis did not directly compare cytology with cobas HPV testing, 92% of women tested positive for either HPV-16, HPV-18 or both vs. 53.3% with atypical squamous cells of undetermined significance on cytology. These results indicate that HPV DNA testing had a 38.7% absolute advantage in sensitivity (P<0.0001), but a 16% lower specificity (P<0.0001).

Combining cobas HPV testing with cytology only slightly improved sensitivity for cervical intraepithelial neoplasia grade 3 (CIN3) to 96.7%, and improved negative predictive value from 99.5% with DNA testing to 99.7% with co-testing.

However, these slight improvements lead to a substantial 35.2% increase in the number of women who screened positive requiring colposcopy, the researchers noted.

When the HPV DNA test was used as a triage to determine if women who screened positive for any type of HPV had either HPV-16 or HPV-18, sensitivity (59.5% vs. 52.8%, P=0.11) and positive-predictive value (15.5% vs. 14.1%, P=0.20) was equal to detection of atypical squamous cells of undetermined significance or worse on cytology for detecting CIN3 lesions.

Among HPV-positive women, detection of HPV-16, HPV-18, or low-grade squamous intraepithelial lesion or worse cytology was more sensitive than atypical squamous cells of undetermined significance or worse on cytology (P<0.0001). The same was true for high-risk HPV or high-grade squamous intraepithelial lesion or worse cytology (P=0.0011).

“In view of our observations, use of HPV testing as a the primary screening test to rule out disease, and use of a specific test, like liquid-based cytology, to decide which women need immediate colposcopy, seems to be a rational approach,” the researchers concluded.

The researchers were unable to determine whether one screening method reduced cancer incidence over the other due to the cross-sectional design of the study.  “Future studies will need to assess the comparative performance and cost-effectiveness of the different cervical cancer strategies to identify best practices,” they added.

Castle PE et al. Lancet Oncol. 2011;12:880-890.

Ronco G et al. Lancet Oncol. 2011; 12: 831–832.