Clinical Prediction Model for Anal High-Grade Squamous Intraepithelial Lesions Risk

Area of Science:

• Oncology
• Gastroenterology
• Epidemiology

Background:

• Timely treatment of anal high-grade intraepithelial lesions (HSIL) is crucial for preventing anal cancer progression.
• Current screening tools like anal Pap tests and high-risk human papillomavirus (hrHPV) testing have suboptimal performance, leading to over-referral for high-resolution anoscopy.
• There is a need for improved methods to accurately triage individuals at increased risk for anal cancer.

Purpose of the Study:

• To develop and externally validate a clinical prediction model for identifying histologic anal HSIL.
• To enhance the triage process for high-resolution anoscopy in individuals at high risk for anal cancer.

Main Methods:

• A penalized logistic regression model with elastic net regularization was developed using data from two institutions.
• Candidate predictors included age, sex, HIV status, history of anogenital HPV-related disease, immunosuppressant use, anal cytology, and anal hrHPV status.
• The model underwent internal validation via five-fold cross-validation and external validation in a third institution cohort.

Main Results:

• The derivation dataset included 536 patients (71.3% with HIV), with an HSIL prevalence of 21.1%. The area under the ROC curve (AUC) was 0.80.
• The external validation dataset comprised 242 patients (65.7% with HIV), with an HSIL prevalence of 37.2%. The AUC for the external validation was 0.73.
• The final model incorporated age, sex, anal cytology, anal hrHPV, immunosuppressant use, history of anogenital HPV-associated disease, and two interaction terms.

Conclusions:

• The developed clinical prediction model shows promising performance for identifying histologic anal HSIL.
• The model utilizes objective and easily obtainable factors, facilitating improved triage to high-resolution anoscopy.
• This tool has the potential to optimize the management of individuals at risk for anal cancer.