Abstract
BACKGROUND: Whole slide imaging (WSI) adoption has been slower in cytopathology due, in part, to challenges in multifocal plane scanning on 3‐dimensional cell clusters. ThinPrep and other liquid‐based preparations may alleviate the issue by reducing clusters in a concentrated area. This study investigates the use of Z‐stacked images for diagnostic assessment and the experience of evaluating urine ThinPrep WSI. METHODS: Thirty ThinPrep urine cases of high‐grade urothelial carcinoma (n = 22) and cases of negative for high‐grade urothelial carcinoma (n = 8) were included. Slides were scanned at 40× magnification without Z‐stack and with Z‐stack at 3 layers, 1 μm each. Six cytopathologists and 1 cytotechnologist evaluated the cases in 2 rounds with a 2‐week wash‐out period in a blinded manner. A Cohen’s Kappa (CK) calculated concordance rates. A survey after each round evaluated participant experience. RESULTS: CK with the original report ranged from 0.606 to 1.0 ( P < .05) without Z‐stack and 0.533 to 1.0 ( P < .05) with Z‐stack both indicating substantial‐to‐perfect concordance. For both rounds, interobserver CK was moderate‐to‐perfect (0.417‐1.0, P < .05). Intraobserver CK was 0.697‐1.0 ( P < 0.05), indicating substantial to perfect concordance. The average scan time and file size for slides without Z‐stack and with Z‐stack are 6.27 minute/0.827 GB and 14.06 minute/2.650 GB, respectively. Surveys demonstrated a range in comfort and use with slightly more favorable opinions for Z‐stacked cases. CONCLUSIONS: Z‐stack images provide minimal diagnostic benefit for urine ThinPrep WSI. In addition, Z‐stacked urine WSI does not justify the prolonged scan times and larger storage needs compared to those without Z‐stack.

This publication has 20 references indexed in Scilit: