Background: HPV-testing has been introduced as the next paradigm in diagnostic cervical cancer prevention. Although cervical cancer and the cancer precursors CIN2 and CIN3 are rare in a population, the prevalence of HPV is high. It is expected that 70-80 % of all women will have a cervical HPV infection during their lifetime.

Objectives: The objective of this work is to focus on the paramount issues of primary screening algorithms.

Methods: The literature describing key statistical properties of different HPV-tests has been evaluated with emphasis on head-to-head studies. Sensitivity, specificity, PPV and NPV are known for different tests. These statistical properties govern how HPV-tests may be implemented in primary screening.

Results: HPV-DNA tests have a high sensitivity, but a low specificity. Suggested primary screening algorithms typically suggest the use of cytology to triage the high number of HPV-DNA positive women to reduce the number of women referred to colposcopy and biopsy. However, this approach reduces the initially high sensitivity of DNA-tests due to the much lower sensitivity of cytology. On the other hand technologies like the HPV mRNA test PreTect HPV-Proofer have a high sensitivity (although lower than HPV-DNA tests) but at the same time also a high specificity. This high specificity enables a primary screening algorithm without any triage step. The initial sensitivity of PreTect HPV-Proofer will then be equivalent to the algorithmic sensitivity. The literature also indicates that PreTect HPV-Proofer has a higher sensitivity than cytology.

Conclusions: To be able to exploit the HPV-test sensitivity in a primary screening algorithm triaging positives should be avoided. The most sensitive primary screening algorithm is expected to be obtained using a test not needing a subsequent triage step resulting in an algorithmic sensitivity equal to the test sensitivity.