Aims and background: Automated reading of smears is increasingly used in cervical screening, and quality control procedures to check its performance are required. The aim of the present study was to evaluate the usefulness of rapid review of all smears coded as "no further review" (NFR) by the AutoPap system as a quality control procedure.
Methods: From 2002-2005, 153,269 smears were processed by AutoPap at the Centro per lo Studio e la Prevenzione Oncologica of Florence, Italy, and 24,503 (15.9%) were coded as NFR. All NFR smears underwent immediate rapid review by expert readers.
Results: Sixty-five of 118 NFR cases reviewed as ASCUS-R complied to recommended 6 months cytology, and 25 of 30 NFR cases reviewed as ASCUS-SIL or more severe accepted immediate colposcopy. As a result of such assessment, one single case of microinvasive carcinoma was detected in a woman aged 34 years: NFR smear had been reviewed as AGCUS, was followed by repeat AGCUS cytology and repeat negative colposcopy, and finally underwent conization.
Conclusions: NFR report at AutoPap was associated with an extremely high negative predictive value, which might suggest using AutoPap as a primary screening tool, with NFR report prompting a "return to screening" recommendation. Rapid review of NFR smears is not recommended as a quality control measure for the negligible yield of high-grade lesions detected.