Objectives. To determine what practitioners could reasonably expect from urinary cytology (UC) consultations by analyzing their role in our clinical practice.

Methods. Reports of 227 consecutive interpretations on 130 patients collected during a 13-month period were correlated with all available follow-up information.

Results. In our practice, UC consultations can predict the presence of bladder carcinoma in nearly 90% of patients, the major mitigating factor being the absence of tumor cells in specimens from patients harboring bladder neoplasms. High-grade neoplasms are most reliably detected with UC. Very low grade neoplasms are difficult to detect, primarily because the cells of these lesions lack features of carcinoma. Significant interobserver variation did not occur.

Conclusions. Pathology consultations based on UC can be associated with high diagnostic yields provided that certain factors are understood. These include that (a) the clinical import of diagnostic terms may vary among observers and should be mutually understood in individual practice settings; (b) UC specimens may not contain tumor cells even when patients have bladder cancer; (c) correlative information may be difficult to obtain and thus seem to inflate calculations for false-positive results; (d) interobserver variation can be reduced if cytopathologists use limited and uniform diagnostic terms; and (e) a high level of diagnostic expertise from cytopathologists should be expected.