Abstract
Bladder cancer attacked approx 50,500 Americans in 1995 and killed about 11,200 (1). Bladder cancer appears to develop along two mam tracks: a deeply mvasive, high-grade form that rapidly becomes life-threatening, and a much less dangerous low-grade form (2–5). Although low-grade tumors are usually cured readily, by simple resection if detected early or by Bacille Calmette-Guerin (BCG) therapy in the case of multiple tumors, the detection of lowgrade tumors is pressing because approx 15% of patients with these tumors progress to dangerous disease (6) Given this tendency to progress, even though approx 70% of bladder cancers are low grade on inriral diagnosis, the number of deaths caused by bladder cancer is almost equally divided between those with aggressive disease upon presentation and those who progress from lowgrade disease. Thus, the ability to detect a group at high risk for progression, or to detect progression early, is crucial to decreasing the death toll from bladder cancer, particularly if detection could be based on noninvasive techniques that quantitate biochemical changes in exfoliated cells found in urine Conventional cytologic methods have poor sensitivity to low-grade tumors (2,3), though the addition of DNA ploidy by image analysis, which only detects the limited class of low-grade tumors with aberrant ploidy or bladders with field disease, improves the sensitivity 15–20% compared to Papanicolaou cytology (7–9).