ABSTRACT
We evaluated image cytometric deoxyribonucleic acid ploidy analysis of urine sediments as an adjunct to visual cytology in the detection of urothelial cell carcinoma. Both procedures were performed on 384 specimens obtained from voided urine (78%), bladder washings or by catheterization. Of these specimens 235 came from patients with active urothelial cell carcinoma, 23 from patients with currently inactive urothelial cell carcinoma and 126 from patients with benign disorders of the urinary tract. Visual cytology classified 171 specimens as positive for urothelial cell carcinoma, 143 as negative and 70 as atypical but diagnostically ambiguous. Sensitivity was 72.3% (170 of 235) and specificity was 99.3% (148 of 149, with atypical and negative specimens grouped together). The slides were then restained for ploidy analysis. Each of the deoxyribonucleic acid histograms of the 170 true positive and the 129 true negative specimens was described by a set of order statistics. These sets were entered into a multivariate discriminant analysis. The discriminant function obtained was then applied to the 70 ambiguous specimens. As a result 27 of these specimens, all from patients with active urothelial cell carcinoma, were reclassified as positive. Sensitivity was increased to 83.8% (197 of 235), with no loss of specificity. Image cytometry is a useful adjunct to visual cytology for the detection of bladder cancer in urine sediments.
