Abstract
INTRODUCTION: Urinary cytology (UrCy) is highly sensitive to diagnosing high-grade urothelial carcinoma (HGUC) but cannot predict muscularis propria invasion. Further, the atypical urothelial cell category (AUC) may have variable outcomes. Image morphometry (IM) may be a valuable adjunct technique in this setting. Hence, we evaluated IM in the AUC and HGUC categories to improve the diagnostic performance. MATERIALS AND METHODS: The following six nuclear parameters were evaluated by IM on 3150 cells: nucleo-cytoplasmic (N:C) ratio, nuclear area, diameter, perimeter, standard deviation of the nuclear area (SDNA; pleomorphism) and integrated density (ID; nuclear chromasia), using the ImageJ software, in three cohorts based on the histopathology outcome: 20 cases of AUC - benign non-neoplastic outcome (AUC-B); 22 cases of HGUC Muscle invasive (HGUC-MI) and 21 cases of HGUC non-muscle invasive (HGUC-MF). RESULTS: A retrospective analysis of urine cytology. The patient's ages ranged from 36 to 85 years, with a mean age of 60.6. The male-to-female ratio was 5.4:1. A total of 20 cases of AUC-B and 43 cases of HGUC were selected for IM analysis. HGUC cases had higher nuclear parameters than AUC-B, and HGUC-MI had higher SDNA, ID, diameter, and area than HGUC-MF. SDNA and ID predict muscularis propria invasion in HGUC. CONCLUSIONS: Image morphometry successfully differentiates HGUC cases from benign non-neoplastic ones and might help to identify muscularis propria invasion in HGUC using a combination of nuclear parameters.