[Clinical value of p16INK4a immunocytochemistry in cervical cancer screening].

Objective: To evaluate the value of p16INK4a detected by p16INK4a immunostaining as a new generation of cervical cytology for primary screening and secondary screening in population-based cervical cancer screening, and in improving cytological diagnosis. Methods: Between 2016 and 2018, 5 747 non-pregnant women aged 25-65 years with sexual history were recruited and underwent cervical cancer screening via high-risk (HR)-HPV/liquid-based cytological test (LCT) test in Shenzhen and surrounding areas. All slides were immuno-stained using p16INK4a technology, among them, 902 cases were offered p16INK4a detection during primary screening, and the remaining 4 845 cases were called-back by the virtue of abnormal HR-HPV and LCT results for p16INK4a staining. Participants with complete LCT examination, HR-HPV test, p16INK4a staining and histopathological examination results were included in this study. The performance of p16INK4a in primary and secondary screening, and in assisting cytology to detect high grade squamous intraepithelial lesion [HSIL, including cervical intraepithelial neoplasia (CIN) Ⅱ or Ⅲ] or worse [HSIL (CIN Ⅱ)+ or HSIL (CIN Ⅲ)+] were analyzed. Results: (1) One-thousand and ninety-seven cases with complete data of p16INK4a and histology were included. Pathological diagnosis: 995 cases of normal cervix, 37 cases of low grade squamous intraepithelial lesion (LSIL), 64 cases of HSIL and one case of cervical cancer were found. Among them, 65 cases of HSIL (CIN Ⅱ)+ and 34 cases of HSIL (CIN Ⅲ)+ were detected. The positive rate of p16INK4a in HSIL (CIN Ⅱ)+ was higher than that in CINⅠ or normal pathology (89.2% vs 10.2%; P<0.01). (2) p16INK4a as primary screening for HSIL (CIN Ⅱ)+ or HSIL (CIN Ⅲ)+ was equally sensitive to primary HR-HPV screening (89.2% vs 95.4%, 94.1% vs 94.1%; P>0.05), but more specific than HR-HPV screening (89.8% vs 82.5%, 87.7% vs 80.2%; P<0.05). p16INK4a was equally sensitive and similarly specific to cytology (≥LSIL; P>0.05). (3) The specificity of LCT adjunctive p16INK4a for detecting HSIL (CIN Ⅱ)+ or HSIL (CIN Ⅲ)+ were higher than that of LCT alone or adjunctive HR-HPV (P<0.01), while the sensitivity were similar (P>0.05). (4) p16INK4a staining as secondary screening: p16INK4a was significantly more specific (94.1% vs 89.7%, 91.9% vs 87.4%; P<0.01) and comparably sensitive (84.6% vs 90.8%, 88.2% vs 91.2%; P>0.05) to cytology for triaging primary HR-HPV screening. HPV 16/18 to colposcopy and triage other HR-HPV with p16INK4a was equally sensitive (88.2% vs 94.1%; P=0.500) and more specific (88.3% vs 83.0%; P<0.01) than HPV 16/18 to colposcopy and triage other HR-HPV with LCT≥ atypical squamous cells of undetermined significance (ASCUS), and the referral rate decreased (14.0% vs 19.4%; P=0.005). Conclusions: For primary screening, p16INK4a is equally specific to cytology and equally sensitive to HR-HPV screening. p16INK4a alone could be an efficient triage after primary HR-HPV screening. In addition, p16INK4a immunostaining could be used as an ancillary tool to cervical cytological diagnosis, and improves its accuracy in cervical cancer screening.