Nuclear long diameter as a new criterion to distinguish high-grade urothelial carcinoma cells from benign reactive cells.

OBJECTIVE: Recently, the nuclear area has attracted attention as a morphological parameter to differentiate high-grade urothelial carcinoma (HGUC) cells from benign reactive cells. The nuclear long diameter (NLD) strongly correlates with the nuclear area and is easy to subjectively estimate. Therefore, this study examined the usefulness of the NLD-to-neutrophil diameter ratio for detecting HGUC cells in urine cytology. METHODS: This study included 29, 26 and 18 patients with HGUC, glomerular disease and urolithiasis respectively. An image analysis system was used to measure the NLD of HGUC and benign reactive cells (reactive renal tubular cells and reactive urothelial cells) and the neutrophil diameter that appeared in the voided urine in these cases. The NLD index was calculated using the NLD-to-neutrophil diameter ratio. We subsequently compared HGUC and benign reactive cells with respect to NLD and NLD indices. In addition, the HGUC cell group and benign reactive cell group were compared by selecting the five cells with the largest NLD and NLD index on each slide. RESULTS: The NLD and NLD indices of HGUC cells were significantly higher than those of benign reactive cells in all cells and in the five cells with the largest NLD and NLD indices. The cut-off value of the NLD index for detecting HGUC cells was 1.25 in all cells and 1.80 in the five cells with the largest NLD index. CONCLUSIONS: The NLD index is a useful parameter that can be introduced into routine microscopic examinations to differentiate HGUC cells from benign reactive cells.

The usefulness of nuclear area in the diagnosis of high-grade urothelial carcinoma cells in voided urine cytology.

OBJECTIVE: The Paris System for Reporting Urinary Cytology considered the nuclear-to-cytoplasmic (N:C) ratio as the most important cytomorphological feature for detecting high-grade urothelial carcinoma (HGUC) cells. Few quantitative studies have been conducted on other features although quantitative studies on the N:C ratio have been reported. Therefore, this study quantitatively analysed important cytomorphological features in distinguishing benign reactive cells from HGUC cells. METHODS: We analysed 2866 cells from the urine of 52 patients. A digital image analyser was used to quantitatively measure the nuclear area, cell area, N:C ratio, and nuclear roundness for HGUC cells and benign reactive cells. Additionally, the diagnostic value of quantitative cytomorphological criteria in HGUC cells was evaluated by the receiver operating characteristic curve. RESULTS: The area under the curve for the prediction of HGUC cells for all cells and the top five cells was in the following order: nuclear area (0.920 and 0.992, respectively), N:C ratio (0.849 and 0.977), cell area (0.781 and 0.920), and nuclear roundness (0.624 and 0.605). The best cutoff value of the N:C ratio to differentiate HGUC cells from benign reactive cells was 0.438, and using the N:C ratio of 0.702, the positive predictive value obtained was 100%. CONCLUSIONS: Our study indicated that nuclear area is a more important cytomorphological criterion than the N:C ratio for HGUC cell detection. Moreover, extracted data of the top five cells were more valuable than the data of all cells, which can be helpful in the routine practice and future criteria definition in urine cytology.