Table 1.

Parameters for assessing the clinical performance of HPV tests in cervical cancer screening.

Terminology	Characteristics to measure
Sensitivity1	The proportion (expressed as percent) of patients with a certain disease (e.g. CIN III or invasive cervical cancer) who have a POSITIVE HPV test. E.g., a sensitivity of 95% for CIN III means that 5 of 100 CIN III cases will be NEGATIVE (false-negative) and will be missed by this test. Sensitivity is affected mainly by the test itself and to some extent by the nature of samples submitted for testing.
Specificity2	The proportion of subjects (expressed as percent) without a certain disease (e.g. CIN II or higher severity) who have a NEGATIVE HPV test. E.g., a specificity of 70% means that if 100 women without CIN II or higher severity participate in the screening, 30 will be POSITIVE (false-positive) by the test. Specificity is affected mainly by the test itself and to some extent by the nature of samples submitted for testing.
Positive predictive value	The chance (expressed as percent) that a POSITIVE HPV test result indicates the presence of a certain disease (e.g. invasive cervical cancer). In the context of cervical screening, this predictive value can be further divided into predictive value for the current situation or in the next 5 or 10 years. The positive predictive value of a test depends mainly on the prevalence of the disease in the target population. Cancer screening tests, including HPV, are bound to have low positive predictive value when applied to the general population.
Negative predictive value	The chance (expressed as percent) that a NEGATIVE HPV test result excludes the presence of a certain disease (e.g. CIN II or higher severity). Under the context of cervical screening, this predictive value can be further divided into predictive value for current situation or in the next 5 or 10 years. The negative predictive value of a test depends mainly on the prevalence of the disease in the target population.

¹The definition presented here refers to “clinical” sensitivity. From the laboratory performance point of view, the concept of analytical sensitivity is often used, which means the lowest amount of analyte required in the specimen to generate a positive result, (also referred to as detection limit, e.g. 100 copies of viral DNA). For the HPV test, increasing the analytical sensitivity may not always increase the clinical sensitivity, but usually specificity is lost.

²From the laboratory performance point of view, “specificity” means cross-reaction with similar targets. This is more relevant for assays intended for HPV typing. e.g. an assay with poor specificity may misidentify an HPV 58–positive sample as HPV 33.