On Sept. 15th 2016, a directive for the upcoming organized Cervical Cancer Screening Program in Germany was issued. Women from the age of 35 years will be offered co-testing by an HPV test in combination with cytology every three years. Younger women still have annual PAP smears. New requirements for an HPV DNA test in an organized population based screening setting should be as follows – the test shall be able to detect only the 13 HPV types classified by the IARC/WHO as carcinogenic to humans. The clinical sensitivity and specificity of an HPV test for the detection of CIN2+ must not be lower than 95% and 98% of already established and in RCTs used HPV test systems, such as the HC2 (QIAGEN) test, respectively. Furthermore the test-positivity rate in women with NILM should not be higher than that of the HC2 test. Intra- and inter-laboratory reproducibility on different instruments and with different staff should be at least 90%. And because all screening programmes that include HPV testing comprise screening intervals of at least 2-3 years, the longitudinal cumulative incidence rate of CIN2+ and the NPV should not be significantly different from the HC2 test used in European RCTs. In addition, the desired test should have demonstrated convincing performances in pilot trials, should be accepted by the lab and the referring ObGyn and should be cost-effective.
In the past years, we have performed three HPV DNA and RNA test comparisons (Abbott: realtime high risk HPV test, Hologic: CERVISTA, APTIMA) in cross-sectional studies (N= approx. 10,000 each) based on routine screening populations and currently perform one new RNA test comparison (APTIMA; N=9451) in a long-term prospective (5 years) routine screening cohort.
Although all tests revealed a high degree of automatisation, inter- and intra-laboratory reproducibility and non-inferiority in the clinical performance to the gold standard, other features need to be considered as well. Tests like CERVISTA revealed a twice as high positivity rate in women with normal cytology than the HC2 test leading to an increased referral rate in our comparative cross-sectional study. The Abbott HPV test missed two cases of HPV31-positive CIN3. In contrast, the APTIMA RNA-based test consistently shows a comparable clinical sensitivity to HC2 in combination with a higher clinical specificity. This reduces the number of follow up procedures by 23% in a screening program while at the same time keeping a high sensitivity for the detection of CIN2+.
Prior to their introduction into nationwide screening programmes, HPV tests need to be evaluated in pilot studies under real screening conditions.