Background  Human papillomaviruses (HPV) genotyping is an important method for cervical cancer screening, but the insufficient technician and unqualified lab is not favorable for its application.

  Objective  To evaluate the sensitivity, specificity and overall agreement of an automatic bio-chips in detecting HPV by comparing it with fluorescent quantitative polymerase chain reaction (PCR) method.

  Methods  For clinical samples that undergoing routine fluorescence quantitative PCR testing, a simple random sampling method is used for each type. Then selected samples were tested by automatic bio-chips method, and Kappa test was used to compare the detection results of the two methods for any differences. The inconsistent results were confirmed by direct sequencing to the amplicon of E6/E7 region.

  Results  Totally 124 samples were selected randomly from 4 589 samples, including 50 negative samples and 74 positive samples (99 tests counting by cases of types). If counting by samples, the two methods were completely consistent (Kappa=1.000, P＜0.0001). If counting by types, the sensitivity, specificity, positive predictive value, negative predictive value and overall agreement of automatic bio-chips was 98.0%, 92.6%, 96.0%, 96.2% and 96.1%, respectively (Kappa=0.913, P＜0.0001). The inconsistent results were found in 6 positive samples (1 case each for type 39 and 59, 2 cases each for type 52 and 68), which were all confirmed in duplicate test and direct sequencing. Besides, additional 18 positive cases were found by automatic bio-chips, whose HPV types were not covered by fluorescent quantitative PCR method, thus providing more information for the treatment.

  Conclusion  Automatic bio-chips shows excellent agreement with fluorescent quantitative PCR in HPV genotyping, which has good prospect to be applied in medical institutions with insufficient technician and unqualified PCR lab.