Comparison of seven screening methods in the diagnosis of bladder cancer / 中华医学杂志(英文版)

ABSTRACT

<p><b>BACKGROUND</b>We compared the validity (evaluated by sensitivity and specificity), reliability (evaluated by reproducibility) and yield (evaluated by predictive value, examining complexity and cost) of individual and combined tests for bladder tumour antigen stat (BTAstat), nuclear matrix protein 22 (NMP22), hyaluronic acid (HA), survivin, CD44v6, vascular endothelial growth factor (VEGF), and voided urine cytology (VUC) in detecting bladder cancer. And at the same time we evaluated the clinical value of these seven detecting methods in the diagnosis of bladder cancer.</p><p><b>METHODS</b>The six markers and VUC were detected in the urine of cancer group (151 patients with bladder cancer) and two control groups (50 patients with benign urological diseases and 50 healthy controls). The sensitivity, specificity, predictive value, reproducibility, examining complexity and checking cost of each marker and combined markers were calculated.</p><p><b>RESULTS</b>There was a significant difference between bladder cancer group and the two control groups. The sensitivity, specificity and positive predictive value were as follows: VUC (36.4%, 100.0%, 100%), BTAstat (76.8%, 87.0%, 89.9%), NMP22 (77.5%, 81.0%, 86.0%), HA (82.8%, 83.0%, 88.0%), survivin (70.2%, 85.0%, 87.6%), CD44v6 (50.3%, 79.0%, 78.4%), and VEGF (68.2%, 93.0%, 93.6%). The highest sensitivities were 91.4% for NMP22 + BTAstat and HA + NMP22, whereas the combined marker with the lowest sensitivity (62.3%) was VUC + CD44v6. The highest specificity was 93.0% for the combined use of VUC + VEGF and HA + CD44v6 had the lowest specificity (73.0%). The most convenient examining method was the detection for BTAstat, the lowest cost was the detection for HA, and the best reproducibility were the detection for BTAstat and VUC.</p><p><b>CONCLUSIONS</b>All the markers have obvious clinical value in diagnosis of bladder cancer. The use of BTAstat + HA or NMP22 + BTAstat are better examining methods in terms of validity, reliability, and yield.</p>