Clinical Experience with the ThinPrep Imager System

ACTA CYTOLOGICA 481 C we change the way we screen cervical cytology?1 This has been the focus of the cytopathology profession, from individuals to entire laboratories, in approaching implementation of the ThinPrep Imager System (TIS) (Cytyc Corp., Boxborough, Massachusetts, U.S.A.) automated interactive Pap smear screening device. The cytopathology community was reassured by the multicenter clinical trial that initially demonstrated equivalence to manual screening with 10,742 ThinPrep Pap tests analyzed by the TIS. In addition, increased sensitivity for the detection of atypical squamous cells (ASC+) and increases in cytotechnologist productivity were shown with the TIS.2 The cytopathology community believed that the technology was robust and would offer many advantages for our laboratories, as evidenced by the fact that since Food and Drug Administration approval in June 2003, the TIS has been installed in over 200 facilities in the United States and at least 6 have been installed in Europe. Many laboratories have indeed changed the way they screen cervical cytology. The profession must now try to answer the question of whether we have improved the cervical cancer screening process, thereby leading to lower false negative rates. In this issue of Acta Cytologica, Bolger et al report a comparison of manual screening vs. TIS screening, in the first such publication outside of the United States.3 Their results are favorable for TIS screening, showing at least equivalence to manual screening using the British Society for Clinical Cytology (BSCC) guidelines for gynecologic cytology evaluation. There are some differences between their study and the multicenter clinical trial, which are important to note. One of the major differences is in the evaluation of their first 6,000 cases. The authors choose to look at only the 22 fields of view (FOV) selected by the TIS in the Review Scope for the first set of 6,000 cases. They stated that this was for purposes of obtaining experience with the system in spite of going through a training process recommended by the TIS manufacturer. Not surprisingly, a lower sensitivity and specificity was observed in the first set of 6,000 cases when only the selected 22 FOV were examined in the Review Scope compared to manual screening of the full slide. The value of the data in the first set of cases is that they demonstrate that a full screen of the slide must be performed if there are any abnormalities in the initial 22 FOV. For the remainder of this editorial, I will refer to the results in only the second set of 6,000 cases since these cases were evaluated according to manufacturer’s recommendation. In the second set of 6,000 cases, screening was performed according the Cytyc Corp. recommendations, which require that a full screen of the slide be performed if there are any abnormalities observed in the initial 22 FOV. The sensitivity (92.84%) and specificity (98.62%) of the TIS in this second set was statistically significantly improved over those in the first set evaluated with the TIS. However, in the second set the manual screening sensitivity (94.42%) and specificity (98.77%) also improved and there was no statistically significant difference between the types of screening. The authors thought the improvement in manual screening in the second set was due in part to familiarity with the darker ThinPrep Pap stain. However, it is also important to note that the cytotechnologists in this study performed the manual screen after the slides had been imaged by the TIS and evaluated on the Review Scope. There may have been a tendency toward closer scrutiny of the slides by the cytotechnologists when performing the manual screen in order to attempt to “beat the machine.” In the second set of slides there were 46 false negative cases in the primary screening group (0.76%). A close look at the false negative values revealed 60 cases in the TIS group. The majority of the false negative cases in the primary screening and TIS groups were from the borderline nuclear abnormality (BNA) and CIN 1 categories and only 1 was from CIN 3 category. A closer examination of the false negative cases in the TIS set highlights a valuable contribution of the TIS in evaluating the etiology of false negative results. In 27 of the TIS false negative cases, Clinical Experience with the ThinPrep Imager System EDITORIAL ACTA