The definitive diagnosis of upper urinary tract lesions often relies o
n providing adequate tissue samples for cytopathologic review. Uretero
pyeloscopy allows direct examination of these lesions and enables tiss
ue sampling under visual control. A retrospective analysis of 11 patie
nts with 13 abnormal upper tracts was performed, comparing sampling te
chniques performed under endoscopic vision (brush biopsy, fine-needle
aspiration [FNA], and forceps biopsy) with standard retrograde cathete
r lavage. Forceps biopsy was the most accurate of the direct vision sa
mpling methods (100%), although it provided adequate tissue only 56% o
f the time. Direct vision brush biopsy and FNA were better than forcep
s biopsy at providing adequate tissue but were less accurate, at 83% a
nd 89%, respectively. Surprisingly, retrograde catheter lavage was sup
erior to direct vision sampling methods. These results may be attribut
able to patient selection bias in that five of seven upper tracts foun
d to have malignancy had multifocal disease. One patient developed a d
istal ureteral stricture from instrumentation that was managed success
fully with an endoureterotomy.