Diagnostic approach of atypical cells in effusion cytology using computerized image analysis

Cytology is one of the important diagnostic tests done on effusion fluid. It can detect malignant cells in up to 60% of malignant cases. The most important benign cell present in these effusions is the mesothelial cell. Mesothelial atypia can be striking and may simulate metastatic carcinoma. Many clinical conditions may produce such a reactive atypical cells as in anemia, SLE, liver cirrhosis and many other conditions. Recently many studies showed the value of computerized image analysis in differentiating atypical cells from malignant adenocarcinoma cells in effusion smears. Other studies support the reliability of the quantitative analysis and morphometric features and proved that they are objective prognostic indices. Sixty three cases of pleural and peritoneal smears, previously reported as benign [19] cases, malignant [21] cases or atypical [23] cases, were retrieved from the files. In each of these smears; nuclear area, perimeter, and roundness coefficient of 80-100 cell were determined at x400 magnification by the use of image analysis system. Statistical analysis was performed using analysis of variance and Tukey's HSD test. The mean values of nuclear roundness, nuclear perimeter and nuclear area vary between the three groups [benign, atypical and malignant cells] by using analysis of variance [p > 0.01]. The value of nuclear roundness, perimeter and area did not differ significantly between benign and atypical cells [Tukey's test: p<0.01]. On the other hand, the value of nuclear roundness, perimeter and area showed a significant difference between malignant and atypical cells [Tukey's test: p> 0.01]. In conclusion, our data suggest that cytomorphometry performed on effusion smear cells may provide important information for the differentiation of atypical cells from malignant cells, in which the values of atypical cells are closer to those of benign cells during the examination of pleural and peritoneal smears by the use of image analysis system

Correlation between the conventional, routine histological grading of transitional cell carcinoma of the urinary bladder and morphometric analysis

Transitional cell carcinoma of the urinary bladder is one of the important malignancies in both sex groups. It is considered as a heterogenous neoplasm with different biological behavior, in which the majority are early non invasive with tendency for recurrence and some may progress to invasive tumor. Important clinicopathological features are, the tumor stage and histological grade which are used as prognostic parameters of the tumor and play an important role in therapy. Due to the subjectivity of the histological grading, the reproducibility was low. Many studies showed the value of quantitative analysis of the tumor as an important method in determining the recurrence of the tumor and muscular invasion, some other studies showed the value of nuclear measurement as a prognostic tool for bladder carcinoma. To evaluate the benefit of nuclear image analysis as an objective method for grading of transitional cell carcinoma of the urinary bladder and compare it with the subjective routine histological grading. Sixty two cases of transitional cell carcinoma of the urinary bladder, histologically diagnosed and graded according to WHO grading system, were selected. In each case 8-10 HPF were examined and 10-20 consecutive cells were studied, also we measured the largest 10 nuclei for each case separately by a computerized image analysis system at x400 magnification. Nuclear area and roundness were determined. Statistical analysis was performed using the analysis of variance and Tukey's test [HSD]. There was no statistical difference in the mean value of nuclear roundness between the three grades [P<0.05], while there was a statistical difference between grade I and III of their mean nuclear area [MNA] and MNA of the largest 10 nuclei [P<0.05]. No such difference was found between grade I and II or grade II and III [P<0.05]. Morphometric analysis should be based on the selection of special areas and not by random measurement as done on routine histological grading