Accuracy and Reproducibility of Nuclear/Cytoplasmic Ratio Assessments in Urinary Cytology Specimens

Abstract

Visual quantitation of N/C ratio showed only a fair correlation with actual N/C ratio with correlation decreasing with increasing N/C ratio. In the critical range, 0.5-0.7 N/C ratio both interobserver correlation and correlation with true N/C ratio may be insufficiently accurate for precise category assignment as used in the Paris System. Diagn. Cytopathol. 2017;45:107-112. © 2016 Wiley Periodicals, Inc.

Digital image analysis supports a nuclear‐to‐cytoplasmic ratio cutoff value of 0.5 for atypical urothelial cells

“…Previous studies have shown that morphologists tended to overestimate the N:C ratio of urothelial cells, especially at N:C ratios close to cutoff values in TPS. 8,19,20 The current study data suggest that overestimating the N:C ratio would result in a slightly increased sensitivity but a markedly decreased specificity for HGUC. These data support the possible integration of automated digital imaging and/or computerized measurement of N:C ratios in urinary tract specimens, because the added precision could have a significant impact.…”

“…This finding underscores the difficulty of determining such small changes in the N:C ratios when estimated by the human eye with light microscopy alone. Previous studies have shown that morphologists tended to overestimate the N:C ratio of urothelial cells, especially at N:C ratios close to cutoff values in TPS . The current study data suggest that overestimating the N:C ratio would result in a slightly increased sensitivity but a markedly decreased specificity for HGUC.…”

Digital image analysis supports a nuclear‐to‐cytoplasmic ratio cutoff value below 0.7 for positive for high‐grade urothelial carcinoma and suspicious for high‐grade urothelial carcinoma in urine cytology specimens

“…Hang et al used DIA to validate an N:C ratio cutoff of 0.5 for the AUC category, in keeping with TPS . To date, there have been no studies validating a N:C ratio cutoff value ≥0.7 for the HGUC and SHGUC categories Such a study is necessary, because Layfield and colleagues demonstrated that, at the critical N:C ratio (range, 0.5‐0.7), morphologists were less accurate and had moderate‐to‐poor interobserver concordance . Zheng et al also demonstrated that morphologists tended to overestimate N:C ratio in this range …”

Performance of an artificial intelligence algorithm for reporting urine cytopathology

“…Unfortunately, humans are imprecise and not reproducible when performing visual quantitation. Layfield et al demonstrated that in the critical range of 0.5 to 0.7 for the N/C ratio, interobserver correlation and accuracy were insufficient to render a correct diagnosis …”