The challenges of defining sample adequacy in an era of HPV based cervical screening.

BACKGROUND: The implementation of Human Papillomavirus based cervical screening continues apace on a global scale. Understanding the basis and burden of inadequate or invalid samples is important to ensure confidence in high quality laboratory results and inform the development of new technologies. Here we present population based data from Scotland and Denmark which detail the extent of invalid samples for HPV detection in both clinician-taken and self-taken samples. As a comparator we report on the rate of inadequate cytology preparations in both countries. METHODS: The proportion of samples with an invalid HPV test result was calculated by retrospective analysis of routine laboratory data associated with cervical screening programmes in the two countries. Two assays were in use for the programmes at the time (the Abbott RealTime High Risk HPV assay and the BD Onclarity); both have internal endogenous controls for human genes. In addition, acellular cytology samples were reported through a prospective audit (Scotland) and National quality reporting (Denmark). RESULTS: In total, 89,418 clinician samples and 14,677 self-taken samples were assessed. We observed low rates of invalid HPV tests in clinician taken samples (0.05-0.10 %), irrespective of sample collection media (ThinPrep or SurePath), HPV test system/endogenous control type or clinical indication for testing (primary screening, triage or test of cure). For self-taken samples, the number of invalid samples was 0.18 %. Complete absence of sample material (acellular) in clinician taken samples were observed at a level of 1 in approximately 16.5 thousand. CONCLUSIONS: Clinician and self-taken samples appear robust specimens for HPV testing and acellular samples are very rare. Efforts to develop endogenous controls for HPV assays that provide greater insight into true sample adequacy for cervical disease detection, beyond measuring the presence of human cells, will be welcome.