Gerhard Renê, Teixeira Sofia, Gaspar da Rocha Adriana, Schmitt Fernando
Institute of Molecular Pathology and Immunology of Porto University, Porto, Portugal.
Diagn Cytopathol. 2013 Sep;41(9):793-8. doi: 10.1002/dc.22958. Epub 2013 Feb 26.
Telecytology has been used for education, training, and consultation. Cytological studies from gynecological, nongynecological and fine-needle aspiration cytology (FNAC) specimens (including studies of thyroid FNAC) analyzed the diagnostic accuracy and reproducibility of telecytology-based predominantly on static digital images. The aim of this study was to evaluate the diagnostic reproducibility of virtual cytology by measuring intraobserver and interobserver agreements among two cytopathologists, using the Bethesda System for Reporting Thyroid Cytopathology (BSRTC) nomenclature. 502 glass slides from 222 cases of thyroid FNAC were retrieved and scanned by a high-resolution scanner generating whole slides images (virtual cytology). Conventional and virtual cytology were analyzed by a skilled cytopathologist and the intraobserver agreement rate was 77.5% with the corresponding κ value of 0.54, suggesting a moderate agreement between both methods. A second cytopathologist analyzed the same slides only by virtual cytology and the interobserver agreement rate was 80.2% with the corresponding κ value of 0.57, suggesting a moderate agreement between both cytopathologists. The virtual cytology resulted in a higher proportion of aspirates classified as nondiagnostic (20.3 and 14.9% for the first and second cytopathologist, respectively) as compared to conventional cytology (8.1%). Regarding specific diagnostic categories as defined by the BSRTC nomenclature, the follicular lesion of undetermined significance category presented the lowest concordance rates, corresponding to 5.9% intraobserver agreement and no (0.0%) interobserver agreement. We suggest that virtual cytology can be an alternative to conventional cytology in assessment of thyroid FNAC specimens, but nondiagnostic aspirates obtained by virtual cytology should be reassessed by conventional cytology.
远程细胞学已用于教育、培训和会诊。对妇科、非妇科及细针穿刺抽吸细胞学（FNAC）标本（包括甲状腺FNAC研究）的细胞学研究，主要基于静态数字图像分析了远程细胞学的诊断准确性和可重复性。本研究的目的是使用甲状腺细胞病理学报告的贝塞斯达系统（BSRTC）命名法，通过测量两名细胞病理学家之间的观察者内和观察者间一致性，评估虚拟细胞学的诊断可重复性。从222例甲状腺FNAC病例中获取502张载玻片，并用高分辨率扫描仪进行扫描以生成全玻片图像（虚拟细胞学）。由一名熟练的细胞病理学家分析传统细胞学和虚拟细胞学，观察者内一致性率为77.5%，相应的κ值为0.54，表明两种方法之间有中度一致性。另一名细胞病理学家仅通过虚拟细胞学分析相同的玻片，观察者间一致性率为80.2%，相应的κ值为0.57，表明两名细胞病理学家之间有中度一致性。与传统细胞学（8.1%）相比，虚拟细胞学导致分类为非诊断性的抽吸物比例更高（第一名和第二名细胞病理学家分别为20.3%和14.9%）。关于BSRTC命名法定义的特定诊断类别，意义未明的滤泡性病变类别一致性率最低，观察者内一致性为5.9%，观察者间一致性为零（0.0%）。我们建议，在评估甲状腺FNAC标本时，虚拟细胞学可作为传统细胞学的替代方法，但通过虚拟细胞学获得的非诊断性抽吸物应通过传统细胞学重新评估。
