Abstract
Human papillomavirus (HPV) genotype classification relies on DNA sequence similarity to reference (prototype) sequences. Most HPV assays used for cervical cancer screening were clinically validated against European HPV prototypes. However, the impact of HPV sequence polymorphisms on test performance remains unexplored. We evaluated whether sequence variation in E6/E7 relative to HPV prototypes affects test performance by analyzing cervicovaginal samples from 990 women enrolled (2019-2022) across Belgium, Portugal, Brazil and Ecuador. Samples were tested using cobas, Anyplex, and next-generation sequencing (Ampliseq/Ion Torrent targeting E6/E7). Sequence variation was defined as the proportion of single nucleotide polymorphisms across E6/E7 relative to reference sequence. Sequence variation was, on average, higher in HPV-negative than HPV-positive samples for HPV16 (0.46% vs 0.13%) and HPV18 (0.44% vs 0.37%) using cobas. Similar patterns were observed with Anyplex (HPV16: 0.78% vs 0.13%, HPV33: 0.66% vs 0.40%, HPV58: 0.79% vs 0.53, and HPV66: 1.14% vs 0.25%). For HPV45, sequence variation was, on average, higher in HPV-positive than HPV-negative samples when tested with Anyplex (0.87% vs 0.43%). For HPV types 18, 31, 35, 39, 51, 52, 56, 59 and 68, the mean sequence variation was similar between HPV-negative and HPV-positive samples using Anyplex. Our findings show that sequence variation relative to prototypes may impact test performance.