The Evolution of the Cervical Cancer Screening: HPV Testing

Abstract & Commentary

By Robert L. Coleman, MD, Associate Professor, University of Texas; M.D. Anderson Cancer Center, Houston, is Associate Editor for OB/GYN Clinical Alert

Dr. Coleman reports no financial relationship to this field of study.

Synopsis: In a community-based, randomized controlled clinical trial comparing conventional Pap testing to HPV testing, the latter was associated with greater sensitivity for high grade cervical intraepithelial neoplasia.

Source: Mayrand MH, et al. Human Papillomavirus DNA versus Papanicolaou Screening Tests for Cervical Cancer. New Engl J Med. 2007; 357:1579-1588.

Human Papillomavirus (HPV) testing has proved efficacious in triaging minimally abnormal cytology to further investigation or surveillance. However, despite its high sensitivity for cervix pathology, its value as a primary screening technology has been formally evaluated in only a limited way. To address the hypothesis that HPV screening is superior to conventional Pap testing, a randomized clinical trial was conducted in 10154 women in 30 clinics from Montreal and St. John's, Canada. Eligible participants included women aged 30 to 69 years, who otherwise were not being followed up for a cervical lesion, lacked a cervix, were pregnant, had a history of cervical cancer or had undergone Pap testing in the previous year. Patients were randomized 1:1 to either a "focus on Pap" or "focus on HPV" screening group. Since the hypothesis consideration was two-tailed, double sampling occurred in all patients; however, the first test performed was directed by randomization. High-grade (grade 2 or higher) cervical intraepithelial neoplasia in tissue was the primary endpoint. The endpoint could be reached by biopsy (liberal-definition) or excision (conservative-definition). The sensitivity for HPV testing for this endpoint was 94.6% compared to 55.4% for conventional Pap testing (p = 0.01); the specificity was 94.1% for HPV testing compared to 96.8% for conventional Pap (p <0.001). Performance was unaffected by sequence. All patients with significant pathology were identified when both tests were used, however, specificity dropped to 92.5%. Combining tests improved specificity but at the cost of sensitivity. As compared to Pap testing, HPV-based screening has greater sensitivity for cervical intraepithelial neoplasia.


Since the introduction of the Pap smear, deaths from cervix cancer in the US, as well as other resource-rich countries, have steadily dropped; the screening test is the most successful ever introduced for a solid malignancy. Its success is dependent on several factors, not the least of which is an index disease that is well suited for screening. However, outside the available, accepted and cheap sampling method, resources required to make this screening effort successful are vast and include not only specialized personnel to interpret the cytology but also an availability to evaluate (colposcopy) and to exact treatment (excision/destruction) for pre-invasive lesions that are suggested or confirmed. The costs are prohibitive for resource-poor situations and the system doesn't address the highest risk groups now facing the medical community—the never- and under-screened population. Since nearly 100% of cervix cancer is associated with HPV infection, testing returns and unambiguous result, and the sampling for HPV in the general population can be accomplished with little medical intervention, attention has turned to evaluation of the next iteration of screening in an effort to break the stalemate of mortality still observed in this and other countries.

The primary objective of this provocative randomized clinical study was to address the potential impact of a screening program based solely on the identification of HPV DNA as compared to conventional cytology. It was performed in a population of average risk women seeking care in the community setting. Given the previously suggested improved sensitivity of HPV testing for high-grade CIN, it was expected that molecular testing would do a better job in representing patients with disease. However, as was also expected, the potential for false positive results would necessarily lead to greater secondary referral. Nevertheless, this front-end "disease" focus over the long run could translate into the need for less repeat testing—this has been suggested by others but was untested in the current study. Additionally, it may be inferred from the current study that secondary cytologic evaluation of HPV positive screens could improve overall screening performance. This is a viable health care question and should be a primary focus of future clinical trials. Other recent efforts to raise the bar on screening performance have focused on the processing, interpretation and automation of cytology-based screening. While these objectives are of great importance, an avenue to address mortality from this preventable cancer where it is of greatest burden is a persistent and formidable challenge.

Other Reading:

Belinson JL, et al. Shanxi Province Cervical Cancer Screening Study II: self-sampling for high-risk human papillomavirus compared to direct sampling for human papillomavirus and liquid based cervical cytology. Int J Gynecol Cancer. 2003;13:819.

Ratnam S, et al. Human papillomavirus testing for primary screening of cervical cancer precursors. Cancer Epidemiol Biomarkers Prev. 2000;945-951.