Background and purpose: Risk of invasive cervical carcinoma (ICC) and its precancerous lesions following high risk human papillomavirus (hrHPV) infection may vary according to HPV types and geographic regions. Analyzing HPV-type distribution in cervical samples from local women aged 30 years and older

this study aimed to identify HPV types with higher risk of developing CIN2+

and to compare diagnostic performance for CIN2+ using these types and type combinations. Methods: Cervical samples with histology follow-up from patients of a tertiary cancer center in Shanghai were collected for HPV genotyping by PCR-RDB. The risk associations of HPV types with CIN2+ were estimated by logistic regression analysis

and ROC curves were plotted for diagnostic performance evaluation. Results: A total number of 413 specimens were obtained

including 38 CIN1

184 CIN2/3

126 ICC patients and 65 negative control people. The 4 most common HPV types in CIN2+ were HPV16

58

33 and 18

in descending order. And only HPV16 (P0.000 1)

58 (P=0.002)

33 (P=0.015) were significantly associated with CIN2+ lesions.Besides

the area under the ROC curve of the HPV16/18/33/58 test scored statistically higher than the HPV16/18 test did (P=0.006 6). Conclusion: A combined test of HPV16/18/33/58 may offer better performance for detecting CIN2+lesions in our geographic region.