[429] UroVysion
(UroV) 
False Positives: True False Positives or Anticipatory True Positives? The University of Pittsburgh Experience
M Youschak, K Cieply, C Sherer, K Cumbie, J Salvatore, D Zynger, A Parwani, R Dhir, L Kiss, S Bastacky. University of Pittsburgh, Pittsburgh, PA
Background: UroV was developed to increase sensitivity for detecting urothelial neoplasia (UN) over urinary cytology (UC) alone. We have occasionally encountered pts with UroV+/UC-, who later developed a positive UC / bx. The aim of this study is to investigate characteristics of UroV+/ UC- cases.
Design: We retrospectively reviewed 667 UC cases assessed by UroV (Vysis; Downer's Grove, IL) from 9/20/04 to 3/31/07. Subjects include: 1) pts under surveillance for UN with any UC dx and 2) pts screened for UN with an atypical / suspicious / positive UC. UC cases were classified as (+) (suspicious / positive) and (-) (negative / atypical). UroV was prepared from residual urine and scored manually by a Nikon fluorescence microscope. A + UroV required 4 double trisomic cells (chrom 3, 7, 17) or 12 cells with homozygous 9p21 deletion. Cases with UroV+ cells below threshold were considered inconclusive. We identified UC/UroV discordant cases, and reviewed UC slides for UC-/UroV+ cases. Minimum f/u was 18 mo.
Results:
| Category | # cases | Trisomy (# cases) | 9p21 loss (# cases) | # cells counted (median) |
| UroV (-)/UC(-) | 448 | 0 | 0 | 180 |
| UroV(-)/UC(+) | 64 | 0 | 0 | 63 |
| UroV(+)/UC(+) | 127 | 93 | 34 | 87 |
| UroV(+)/UC(-) | 12 | 11 | 1 | 19 |
| UroV (inconclusive)/UC(+) | 7 | N/A | N/A | 208 |
| UroV (inconclusive)/UC(-) | 9 | N/A | N/A | 147 |
76 (11%) cases were discordant: UroV+/UC- (n=12; 2%); UroV-/ UC+ (n=;9%). For the UroV+/UC- cases, 3 had suboptimal UC (low cellularity/degeneration/obscuring). 10 cases had rare to many urothelial cells with high N:C ratios, but absent or mild atypia. 11 (92%) and 1 (8%) were trisomic and 9p21 deleted. Nine trisomic cases had
1 cell with concurrent chromosome 3/7/17 trisomy. Number of total cells counted to reach threshold number of UroV+ cells was 4-50 (median 19). Interval from UroV+ to + UC/ bx was 1 - 210 days (median 32 days). Follow-up UN was high-grade in 8/12 (67%) cases; for UC+/UroV- cases, UN was high-grade in 26/64 (41%) cases.Conclusions: 1) Discrepant UC/UroV occurred in 11% of cases. These cases were UC+/UroV-, suggesting lower UroV sensitivity for UN than UC. 2) UroV+ confers a high predictive value for UN in UC- cases. 3) Most UroV+/UC- pts are trisomic, similar to UroV+/UC+ pts. 4) Unlike UroV inconclusive pts, UroV+/UC- anticipatory (+) pts required only a small number of cells to be counted to reach threshold number of UroV+ cells. 5) Most UroV+/UC- anticipatory (+) pts have high-grade UN.
Category: Cytopathology
Monday, March 9, 2009 2:45 PM
Platform Session: Section F, Monday Afternoon