Prospective evaluation of optimal number of biopsy specimens and brush cytology in the diagnosis of cancer of the colorectum.

We prospectively studied the diagnostic accuracy of endoscopic biopsy and cytology in the diagnosis of colorectal cancer and sought to ascertain the optimal number of biopsy specimens that should be taken. Seventy consecutive patients with colorectal masses or polypoid lesions that could not be removed by endoscopic polypectomy and which were referred for surgery were studied. Lesions were first brushed for cytology. A series of 10 biopsies were then obtained from each lesion; the first four were placed into the first container, and two were placed into each of three other containers. Cytology and biopsies were read blindly by one pathologist. Findings were compared with the resection specimen results. Ten lesions were benign; all had negative biopsies and cytology. Sixty lesions were malignant: cytology was positive in 77.2% and biopsies in 78.3%. The yield was increased to 91.7% when both were done (cytology plus six biopsies). Cytology detected seven cancers that were missed by biopsy. Biopsy detected eight cancers not diagnosed by cytology. Five cancers were missed by both. Of the 47 cancers detected by biopsy, the first four biopsies made the diagnosis in 41 cases. Taking two more biopsies (six total) identified six more cancers. No additional cancers were identified by taking more biopsies (eight or 10 total). When six different biopsy-cytology strategies were examined: 1) cytology alone, 2) four biopsies (Bx), 3) six Bx, 4) 10 Bx, 5) cytology + four Bx, and 6) cytology + six Bx, the yield was 77.2%, 68.3%, 78.3%, 78.3%, 90.0%, and 91.7%, respectively. We conclude that biopsy and cytology are complementary in the diagnosis of colorectal cancer. The combination of cytology and four to six biopsies is recommended. There appears to be little additional yield to taking more than six biopsies. Even combined biopsy and cytology will miss 8-10% of cancers.