Urine cytology in the evaluation of upper tract urothelial lesions.

Approximately 10% ofall urothelial tumors occur in the upper tracts, and many arise in patients with a history of bladder cancer. In most instances diagnosis ofupper tract neoplasms is straightforward and is based on positive radiological or endoscopic findings. Diagnostic difficulties occur when urine cytology is reportedly abnormal but endoscopic and radiological studies reveal no apparent lesions in either the bladder or the upper tracts. Cytologic specimens from the upper tracts are usually obtained by direct catheterization, washings through a catheter or brushings. The sensitivity ofthese methods in detecting urothelial carcinoma depends on the grade ofthe neoplasm. The sensitivity ofurine cytology is good for high grade lesions (up to 76%) because the cells ofthese lesions are discohesive and tend to exfoliate readily. Low grade urothelial neoplasms are less likely to shed diagnostic cells, contributing to the incidence off alse-negative results (greater than 50%). The accuracy ofdiagnosis in urine cytology also varies with the grade ofneoplasm. High grade neoplasms tend to shed abundant cells, singly and in small clusters. Nuclei oftumor cells are typically large, hyperchromatic, pleomorphic and markedly variable in size. Such nuclei have irregular contours, are often notched or grooved and may demonstrate the presence ofnucleoli. Nuclear chromatin is coarse, and occasional mitotic figures may be observed. In some cases there may be background cellular debris suggestive oftumor necrosis. With this constellation off indings, the diagnosis of malignancy is relatively easy and accurate in more than 90% ofcases. Figure 1 illustrates many ofthe cytologic findings of exfoliated carcinoma cells from a case of urothelial carcinoma ofthe upper ureter. Figure 2 is a histological section from the nephroureterectomy specimen in this case, which shows carcinoma in situ that has extended into underlying von Brunn’s nests. Nephroureterectomy was performed after 3 consecutive positive washings from this renal unit. Ureteroscopy and retrograde pyelograms on 2 occasions failed to disclose any identifiable lesion. The diagnostic dilemmas created by low grade urothelial neoplasms of the upper tract are much more difficult to resolve. As noted, such lesions may not shed evaluable cells. Furthermore, the differential diagnosis of cells derived from upper tract sources by washing or brushing includes “instrumentation effect” (dislodgement of normal urothelium by catheters or ureteroscopes) and a host ofinf ectious and inflammatory processes, including urothelial atypia induced by urinary calculi. Cells derived from low grade urothelial neoplasms tend to be arranged in cohesive papillary clusters with a fairly uniform cell size, modestly increased nuclearto-cytoplasmic ratio, and minimal to moderate nuclear irregularity. Unfortunately, these same characteristics are frequently observed in reactive (nonneoplastic) urothelium. Consequently, the value ofurine cytology in assessing upper tract lesions, notably those that are not high grade, has traditionally been questionable. In recent years adjunctive diagnostic techniques, such as immunocytologic staining or fluorescence in situ hybridization, have been used for evaluating the presence or absence ofneoplastic cells in urine. Use ofthese techniques in diagnostically equivocal cases may prove to be beneficial.