Guidelines for anal cytology--to make cytological diagnosis and follow up much more reliable.
Scholefield JH. Johnson J. Hitchcock A. Kocjan G. Smith JH. Smith PA. Ferryman S. Byass P.
Department of Surgery, University Hospital, UK.
Anal intraepithelial neoplasia is a difficult diagnostic and management problem, particularly when it occurs in women with synchronous or metachronous genital intraepithelial neoplasia. Diagnosis and follow up by colposcopy is too specialized for widespread use, and although anal cytology has been used before it has been thought of as too inconsistent for practical application. This study standardized collection of specimens and investigated interobserver variation. The aim of the study was to determine whether observers could reliably distinguish high grade anal intraepithelial neoplasia from other conditions. Standardized collection of anal preparations was achieved in the host centre. A meeting of experienced cytopathologists was convened to agree guidelines for anal cytology. These guidelines were sent to the panel of six observers who were subsequently circulated with 30 cytopathological preparations in random order and asked to report them all. The results were collected and processed centrally. Four individuals were in complete agreement about those preparations which were inadequate for reporting, but two others had a lower threshold for rejecting preparations as inadequate. There was agreement between the observers in over 95% of cases in distinguishing high grade intraepithelial neoplasia from other cytological conditions. Kappa values range from 0.66 to 1.00. This study demonstrates that the provision of guidelines for the interpretation of anal cytopathological preparations can result in a high degree of interobserver agreement about the clinically important distinction between high grade anal intraepithelial neoplasia and other conditions. Anal cytology is a more useful technique for diagnosis and follow up of 'at risk' individuals than has previously been suggested, and should be utilized more widely in this group of patients.
A comparative study of three variations of the learning vector quantizer in the discrimination of benign from malignant gastric cells.
Karakitsos P. Ioakim-Liossi A. Pouliakis A. Botsoli-Stergiou EM. Tzivras M. Archimandritis A. Kyrkou K.
Department of Clinical Cytology and Cytogenetics, Laiko Hospital, Athens, Greece.
A prospective study was undertaken to investigate the potential value of morphometry and artificial neural networks (ANN) for the discrimination of benign and malignant gastric lesions. Two thousand five hundred cells from 23 cases of cancer, 19 cases of gastritis and 58 cases of ulcer were selected as a training set, and an additional 8524 cells from an equal number of cases of cancer, gastritis and ulcer were used as a test set. Images of routine processed gastric smears stained by the Papanicolaou technique were processed by a custom image analysis system. The application of the learning vector quantization (LVQ) classifier enabled correct classification of > 97% of benign cells and > 95% of malignant cells, obtaining an overall accuracy of > 97%. This study presents the capabilities of ANN, and also indicates that ANN and image morphometry may offer useful information on the potential of malignancy in gastric cells.