TY - JOUR
T1 - Rapid Review of Liquid-Based Smears as a Quality Control Measure
AU - Henderson, Sheryl
AU - Stevens, Mark
AU - Walker, Todd
N1 - Imported on 12 Apr 2017 - DigiTool details were: Journal title (773t) = Diagnostic Cytopathology. ISSNs: 8755-1039;
PY - 2004
Y1 - 2004
N2 - Monolayers with a paucity of diagnostic cells and those displaying subtle nuclear atypia were often overlooked.The objective of this study was to investigate the effectiveness of a standardized method of rapid review (RR) of monolayer preparations for the identification of abnormalities, the presence of an endocervical component and infectious agents. A total of 200 ThinPrep (Cytyc, Boxborough, MA) slides representing the spectrum of abnormalities commonly encountered in cervical/vaginal cytologic specimens was retrieved from archive. The study set comprised 129 cases within normal limits (WNL); 36 low-grade epithelial abnormalities (LGEA); 28 high-grade epithelial abnormalities (HGEA), including 2 endocervical adenocarcinomas in situ (AIS) and 7 carcinomas. Eighteen false negative (FN) cases were also included for study. Originally missed on initial review, these cases were found to be abnormal on quality control review (17 LGEA; 1 AIS). Commonly encountered infectious agents were represented and included Candida albicans, Trichomonas vaginalis, herpes simplex virus, and Actinomyces. The slides were reviewed using a standardized method of RR (turret technique, for 60 sec) by three experienced screeners masked to the original reference diagnosis. Median sensitivity for LGEA was 70% (range, 67-72%); HGEA, 69% (range, 54-80%); and FN, 65% (range, 56-78%). Specificity remained high, median specificity for LGEA was 95%; HGEA, 97%; and FN, 100%. There was no significant overcalling of any diagnostic category. The chi-square test at P < 0.05 showed no significant difference between RR and full manual rescreen of the ThinPrep smears in this study. While no statistical difference was proven, the sensitivity measurements for all categories of abnormality were moderate due to the high proportion of atypical cases included into the study set. Abnormalities on the monolayer preparations frequently displayed fewer, smaller groups of disaggregated cells with rounded cytoplasmic outlines that were difficult to discern on RR. Interobserver variation was noted.
AB - Monolayers with a paucity of diagnostic cells and those displaying subtle nuclear atypia were often overlooked.The objective of this study was to investigate the effectiveness of a standardized method of rapid review (RR) of monolayer preparations for the identification of abnormalities, the presence of an endocervical component and infectious agents. A total of 200 ThinPrep (Cytyc, Boxborough, MA) slides representing the spectrum of abnormalities commonly encountered in cervical/vaginal cytologic specimens was retrieved from archive. The study set comprised 129 cases within normal limits (WNL); 36 low-grade epithelial abnormalities (LGEA); 28 high-grade epithelial abnormalities (HGEA), including 2 endocervical adenocarcinomas in situ (AIS) and 7 carcinomas. Eighteen false negative (FN) cases were also included for study. Originally missed on initial review, these cases were found to be abnormal on quality control review (17 LGEA; 1 AIS). Commonly encountered infectious agents were represented and included Candida albicans, Trichomonas vaginalis, herpes simplex virus, and Actinomyces. The slides were reviewed using a standardized method of RR (turret technique, for 60 sec) by three experienced screeners masked to the original reference diagnosis. Median sensitivity for LGEA was 70% (range, 67-72%); HGEA, 69% (range, 54-80%); and FN, 65% (range, 56-78%). Specificity remained high, median specificity for LGEA was 95%; HGEA, 97%; and FN, 100%. There was no significant overcalling of any diagnostic category. The chi-square test at P < 0.05 showed no significant difference between RR and full manual rescreen of the ThinPrep smears in this study. While no statistical difference was proven, the sensitivity measurements for all categories of abnormality were moderate due to the high proportion of atypical cases included into the study set. Abnormalities on the monolayer preparations frequently displayed fewer, smaller groups of disaggregated cells with rounded cytoplasmic outlines that were difficult to discern on RR. Interobserver variation was noted.
U2 - 10.1002/dc.20091
DO - 10.1002/dc.20091
M3 - Article
SN - 8755-1039
VL - 31
SP - 141
EP - 146
JO - Diagnostic Cytopathology
JF - Diagnostic Cytopathology
IS - 3
ER -
