Validation of cobas 4800 HPV assay in SurePath Papanicolaou specimens for cervical cancer screening

Abstract

The cobas (Roche Diagnostics, Indianapolis, IN) HPV assay was approved by the US Food and Drug Administration for human papillomavirus (HPV) testing in SurePath (Becton Dickinson, Franklin Lakes, NJ) Papanicolaou specimens for cervical cancer prevention. To validate the cobas HPV assay in SurePath specimens in our institution, we compared its accuracy and clinical efficacy to that of the Cervista (Hologic, Marlborough, MA) HPV HR assay. This study used 138 Papanicolaou (Pap) cytology specimens collected in SurePath preservative fluid at our institution in 2018. After Pap cytology testing, the residual specimens were split for testing with the cobas and Cervista assays. Polymerase chain reaction (PCR)-based HPV testing (GP5+/GP6+) was performed on specimens with discrepant results. Clinical follow-up data were reviewed. The cobas HPV and Cervista HPV HR assays showed good concordance (89.1%), with a kappa value of 0.78 (95% CI: 0.675-0.885). Fifteen specimens showed discrepant results between the 2 assays. Of 7 cases with cobas+/Cervista- results, 5 (71%) were confirmed positive by PCR. Of 8 cases with cobas-/Cervista+ results, 4 (50%) were confirmed positive by PCR. cobas HPV and Cervista HPV HR showed the same HPV-positive rate in cases of pathologically diagnosed ASC-H, LSIL, or HSIL. The sensitivities and specificities for detecting high-risk HPV of cobas HPV (93.7%, 97.3%) and Cervista HPV HR (92.1%, 94.7%) were comparable. The cobas HPV assay had false-negative results in 4 cases (5.2%) including 1 false-negative case that failed to predict CIN3. The cobas HPV assay is valid in SurePath Pap cytology specimens for cervical cancer screening but has limitations of false-negative results with clinical implications.