An automated quantitative DNA image cytometry system detects abnormal cells in cervical cytology with high sensitivity

Objective To evaluate the performance of an automated DNA ‐image‐cytometry system as a tool to detect cervical carcinoma. Methods Of 384 liquid‐based cervical cytology samples with available biopsy follow‐up were analyzed by both the Imager System and a high‐risk HPV test (Cobas). Results The sensitivity and specificity of Imager System for detecting biopsy proven high‐grade squamous intraepithelial lesion ( HSIL , cervical intraepithelial neoplasia [ CIN] 2‐3) and carcinoma were 89.58% and 56.25%, respectively, compared to 97.22% and 23.33% of HPV test but additional HPV 16/18 genotyping increased the specificity to 69.58%. The sensitivity and specificity of the Imager System for predicting HSIL + ( CIN 2‐3+) lesions among atypical squamous cells of undetermined significance samples were 80.00% and 70.53%, respectively, compared to 100% and 11.58% of HPV test whilst the HPV 16/18 genotyping increased the specificity to 77.89%. Among atypical squamous cells‐cannot exclude HSIL , the sensitivity and specificity of Imager System for predicting HSIL + ( CIN 2‐3+) lesions upon follow up were 82.86% and 33.33%%, respectively, compared to 97.14% and 4.76% of HPV test and the HPV 16/18 genotyping increased the specificity to 19.05%. Among low‐grade squamous intraepithelial lesion cases, the sensitivity and specificity of the Imager System for predicting HSIL + ( CIN 2‐3+) lesions were 66.67% and 35.71%%, respectively, compared to 66.67% and 29.76% of HPV test while HPV 16/18 genotyping increased the specificity to 79.76%. The overall results of imager and high‐risk HPV test agreed in 69.43% (268) of all samples. Conclusions The automated imager system and HPV 16/18 genotyping can enhance the specificity of detecting HSIL + ( CIN 2‐3+) lesions.