Objective: To evaluate the effectiveness of the PAPNET system in identifying microbial organisms.
Study design: Two hundred forty-nine previously screened cervicovaginal smears with a diagnosis of an infectious agent were rescreened by PAPNET system. The images were reviewed on a high-resolution monitor and the presence or absence of coccobacilli, trichomonads, and Candida and herpes organisms was noted.
Results: Of 99 cervicovaginal smears with excessive coccobacilli, 16 were missed by PAPNET (percentage of accuracy, 84). Only 60% of Candida organisms (33 of 55) and 77% of trichomonads (42 of 54) were identified by PAPNET. Herpesvirus was diagnosed correctly on PAPNET in 24 of 41 cases. Two or more infectious agents were present in 14 cases, 5 of which were correctly identified by PAPNET. The accuracy of PAPNET in identifying infectious organism when measured against manual screening was 70%, with a high false negative rate (30%).
Conclusion: The PAPNET system identifies coccobacilli, trichomonads and Candida organisms in decreasing order of frequency, missing a high percentage of Candida organisms and trichomonads (40% and 33%, respectively), the most common causative agents in female genital tract infections.