Karin Sundström

PURPOSE Detection of human papillomavirus (HPV) by polymerase chain reaction in invasive cervical cancer is strongly associated with prognosis but previous studies have not considered sequencing efforts. We aimed to assess the association when also including comprehensive analysis of HPV infection by deep sequencing and a longer follow-up period. MATERIALS AND METHODS We subjected all 392 of 2,845 invasive cervical cancer cases that were polymerase chain reaction–negative for HPV to RNA sequencing on the NovaSeq 6000 platform (Illumina) and identified an additional 169 cases as HPV-positive. We followed all women from date of diagnosis to December 31, 2016, emigration, or death, whichever occurred first. The main outcome was all-cause mortality by December 31, 2016. We calculated 5-year cumulative relative survival ratios compared with the female general population and used Poisson regression to estimate excess hazard ratios of all-cause mortality by infection with any of the 13 most oncogenic (high-risk [hr]) HPV types in the tumor. All models were adjusted for age, time since diagnosis, stage, histology, and education level. RESULTS The 5-year cumulative relative survival ratio was 0.45 (95% CI, 0.39 to 0.51) in the hrHPV-negative group, and 0.74 (95% CI, 0.72 to 0.75) in the hrHPV-positive group. This translated to a statistically significantly 43% lower excess mortality in the hrHPV-positive group compared with the hrHPV-negative (corresponding to an excess hazard ratio 0.57; 95% CI, 0.48 to 0.69). There was no association between HPV risk group, clade, or number of HPV infections and prognosis. CONCLUSION hrHPV status is a strong determinant of cervical cancer prognosis over 15 years after diagnosis, above and beyond other established factors.