Prediction of Treatment Failure After Excisional Treatment of Cervical Precancer: A Systematic Review and Meta-analysis.

Abstract

OBJECTIVE To evaluate the diagnostic accuracy and clinical utility of posttreatment tests to predict treatment failure after excisional treatment of cervical intraepithelial neoplasia grade 2 or worse (CIN 2+). DATA SOURCES Electronic databases (EMBASE, PubMed MEDLINE) were searched for studies published from January 1975 to August 2024 assessing the occurrence of treatment failure in women who underwent excisional treatment for histologically confirmed CIN 2+ lesion. METHODS OF STUDY SELECTION Previously published meta-analyses were extended and updated. A total of 1,802 studies were reviewed. Studies that assessed the diagnostic accuracy of the margin status, cytologic testing, combination of cytology and high-risk human papillomavirus (HPV), or combination of margin status and high-risk HPV compared with high-risk HPV testing were included. The primary outcome was treatment failure (residual or recurrent CIN 2+) and the absolute and relative diagnostic accuracy to predict this outcome. Studies with at least 18 months of follow-up were included. TABULATION, INTEGRATION, AND RESULTS Forty-six studies and 20,385 women were included in the analysis. Treatment failure occurred in 6.6% of patients. The pooled sensitivity and specificity of high-risk HPV testing were 86.8% and 80.5%, respectively. Cytology had a sensitivity of 70.8% and a specificity of 85.7%, pooled from 34 studies. Compared with high-risk HPV testing in the same studies, cytology was 6.5% more specific (95% CI, 1.024-1.108) but 21.3% less sensitive (95% CI, 0.702-0.882). Assessment of the margin status was 39.9% less sensitive (95% CI, 0.532-0.678) but similarly specific (95% CI, 0.970-1.069) to high-risk HPV testing in 29 studies, with a pooled sensitivity and specificity of 48.9% and 82.5%, respectively. Co-testing with cytology and high-risk HPV was similarly sensitive (95% CI, 0.992-1.061) but 10.5% less specific (95% CI, 0.850-0.944) compared with high-risk HPV testing in 16 studies, with a pooled sensitivity and specificity of 94.7% and 69.9%, respectively. The pooled sensitivity and specificity of co-testing with margin status and high-risk HPV were 96.9% and 55.7%, respectively, 6.6% more sensitive (95% CI, 1.021-1.114) but 26.7% less specific (95% CI, 0.637-0.844) than high-risk HPV testing in eight studies. Involved resection margins, abnormal cytology, and a positive high-risk HPV test result were associated with a failure risk of 16.1%, 29.0%, and 26.1%, respectively. Women with negative margins, normal cytology, and a negative high-risk HPV test result had a failure risk of 3.6%, 2.3%, and 0.9%, respectively. The risk of treatment failure was highest for women with involved margins and a positive high-risk HPV test result (45.3%) and lowest for women with negative margins and a negative high-risk HPV test result (0.3%). Abnormal cytology and a positive high-risk HPV test result increased the risk of treatment failure to 42%, whereas normal cytology and a negative high-risk HPV test result decreased the risk to 0.6%. CONCLUSION High-risk HPV testing is highly sensitive and specific to predict residual or recurrent CIN 2+ after excisional treatment. Cytologic testing and assessment of the margin status are slightly more and similarly specific, respectively, but both are significantly less sensitive. High-risk HPV testing can sufficiently inform the posttreatment management based on the posttest risk of treatment failure, unlike both cytology and assessment of the margin status. High-risk HPV testing alone performs similarly to co-testing with cytology or margin status. Posttreatment high-risk HPV testing is therefore an accurate predictor of treatment failure in women treated for CIN 2+.