Validation of cytokeratin 5/6 as an effective substitute for keratin 903 in the differentiation of benign from malignant glands in prostate needle biopsies.

AIMS

Keratin 903 (also known as anti-cytokeratin antibody 34betaE12) is widely used to differentiate benign glands from malignant glands in prostate needle biopsies. However, it is subject to considerable staining heterogeneity. We sought to evaluate the use of cytokeratin 5/6 (CK5/6) as an effective alternative to K903 in the evaluation of prostate needle biopsies in clinical practice.

METHODS AND RESULTS

Thirty Hollandes-fixed prostate needle biopsies were randomly selected over a period of 2 months from the surgical specimens accessioned over that period of time. Twelve cases had diagnosed prostatic adenocarcinoma (Gleason scores 3 + 3, 3 + 4 and 4 + 4) and the remaining cases (n = 18) were negative for carcinoma. Four sequential sections were stained with H&E (x2), K903, and CK5/6. Care was taken to preserve tissue so that matching glands were evaluated on all four sections. All cases were run routinely over a period of 3 weeks on a daily basis with matching positive controls. All slides were evaluated in a blinded fashion independently by two pathologists using a semiquantitative analysis of staining: <25%, 25-50%, 50-75%, >75% and >95% of benign glands (verified on H&E). Cases that showed no staining were repeated to ensure no false negatives. Both observers agreed with respect to percentage of staining in 96% of the cases. Twenty-nine of 30 cases (97%) showed staining in >95% of benign glands with CK5/6. In contrast, K903 staining was seen in <50% of benign glands in five of 30 (17%), 50-75% in nine of 30 (30%), and >75% in 10 of 30 (33%), with only two cases (7%) showing >95% staining for K903. In four cases (13%) the K903 failed to stain any tissue even after repeat staining. K903 was conspicuously negative in atrophic glands in three of 30 cases (10%). Neither K903 nor CK5/6 stained malignant glands. Using a cut-off of >75% staining in benign glands the sensitivity of CK5/6 and K903 was 97% and 40%, respectively.

CONCLUSIONS

CK5/6 has superior sensitivity and reliability compared with that of K903 when evaluating routine prostate needle biopsies, including improved staining of atrophic prostatic glands. While K903 is traditionally used to differentiate benign glands from malignant glands, these results support the use of CK5/6 as an effective and reliable substitute for K903 in routine clinical practice.