Automated Cervical Screening and Triage, Based on HPV Testing and Computer-Interpreted Cytology

Area of Science:

• Gynecology
• Oncology
• Pathology

Background:

• Cervical cancer prevention relies on HPV vaccination and screening/treatment of precancer.
• Human papillomavirus (HPV) testing detects precancer but requires triage to identify high-risk individuals and prevent overtreatment.
• Current triage methods often depend on expert cytology, which is scarce in many regions.

Purpose of the Study:

• To develop and validate a fully automated triage method for cervical cancer screening.
• To assess the performance of a cytologic risk score algorithm in predicting cervical precancer among HPV-positive women.

Main Methods:

• A novel cytologic risk score algorithm was designed using computer-scanned liquid-based slide features.
• The algorithm's performance was compared with abnormal cytology in predicting precancer among 1839 HPV-positive women.
• Prospective validation was conducted by comparing the algorithm with cytology results in 243,807 women screened between 2016-2017.

Main Results:

• The automated algorithm demonstrated triage performance comparable to abnormal cytology among HPV-positive women.
• Combined with HPV typing, the automated strategy referred 91.7% of HPV-positive CIN3/AIS cases to colposcopy while deferring 38.4% of women for retesting.
• In prospective validation, predicted risk scores strongly correlated with cytology (P < .001).

Conclusions:

• A completely automated approach can achieve high-quality cervical screening and triage performance.
• Automated technology holds the potential to expand high-quality cervical screening and triage coverage to underserved areas.