Please select Print from the file menu or right mouse click and choose print from the menu to print this page.

[348] Impact of Urine Cytology Errors on Patient Care

SS Raab, DM Grzybicki, CM Vrbin, KR Geisinger. University of Pittsburgh, Pittsburgh, PA; Wake Forest University, Winston-Salem, NC

Background: Although urine cytology is a mainstay for the diagnosis of urothelial cancer, there has been little study of the frequency, causes, and outcomes of urine cytology error.
Design: We obtained histologic follow-up in 362 voided (6.2%), 125 lower tract instrumented (19.5%), and 25 upper tract (34.2%) urinary cytologic specimens from 1 institution over a 2-year timeframe to determine diagnostic accuracy. A total of 4102 patients had a urine cytology during this timeframe. We reviewed the slides from specimens in which there was a diagnostic disagreement between the cytologic and histologic diagnoses and adjudicated the cause of discrepancy as sampling or interpretation failure. Clinical follow-up was obtained from the medical records in patients with discrepant diagnoses and outcomes were classified as no harm, near miss, and harm, and harm was further subdivided into minimal, mild, moderate and severe.
Results: Cytologic-histologic discrepancies were observed in 209 (40.8%) of cases with histologic follow-up, and the cause of discrepancy was interpretation and sampling in 34.9% and 63.2%, respectively. Of all discrepancies, 101 (48.3%) resulted in minimal or mild harm, consisting mainly of repeat invasive or non-invasive testing and diagnostic delays. Severe harm never was observed, as no patient developed invasive cancer as a result of a false negative urine cytologic diagnosis. Causes of error included poor specimen quality, diagnostic undercalls, and variable use of malignant and non-definitive diagnostic categories (P < 0.001). Only one of 7 pathologists was able to diagnose low grade cancers using urine cytology. For voided specimens, more cancers were detected by an atypical diagnosis (31% of all cancers) than by a malignant diagnosis (19% of all cancers). Considering an atypical diagnosis as predictive of a benign or neoplastic lesion markedly shifted diagnostic sensitivity for voided specimens from 49.2% to 74.6% for high grade and 9.7% to 50.5% for low grade cancer.
Conclusions: These findings indicate that cancer screening protocols with urine cytology are exquisitely accurate in not missing cancer but that potentially reducible errors result in unnecessary testing and diagnostic delays. Improved diagnostic standardization across pathologists and improving specimen processing techniques could effectively low errors.
Category: Cytopathology

Tuesday, March 27, 2007

Poster # 29, Tuesday Afternoon