Brush cytology for the diagnosis of colorectal cancer.

Purpose: The purpose of this study was to analyze the results of brush cytology for the diagnosis of colorectal cancer and compare them with the results of endoscopic biopsy and histologic evaluation of the resected specimen.

Methods: Nine hundred eighteen patients who had brush cytology, endoscopic biopsy, and a definitive resection of a colorectal lesion between 1990 to 2006 were identified from our pathology database.

Results: Cytology alone had a sensitivity of 88.2 percent, a specificity of 94.1 percent, a positive predictive value of 98.6 percent, and a negative predictive value of 61.9 percent for the diagnosis of colorectal cancer. Brush cytology always recognized malignant cells, with a positive predictive value of 100 percent. There was no significant difference between brush cytology and biopsy, which had a sensitivity of 86.9 percent, specificity of 98.1 percent, positive predictive value of 99.5 percent, and a negative predictive value of 60.3 percent. Combining the results of brush cytology and biopsy resulted in a statistically significant increase in sensitivity to 97.4 percent (P < 0.001), a significant increase in the negative predictive value to 88.4 percent (P < 0.001), and a significant reduction in the false-negative rate to 0.03 percent (P < 0.001) for the diagnosis of colorectal cancer.

Conclusions: Brush cytology is as accurate as endoscopic biopsy for the diagnosis of colorectal cancer, and combining these two diagnostic modalities resulted in a significant improvement in the definitive diagnosis of cancer, which might reduce the need for further biopsy.