CONTEXT:

Distinguishing low-grade and high-grade noninvasive papillary urothelial carcinoma based on morphologic criteria can be challenging and adjunct markers are highly desirable. Survivin, presumably an antiapoptotic protein, was previously proposed as a prognostic marker for urothelial carcinoma.

OBJECTIVE:

To assess interobserver variability by 2004 World Health Organization classification and the value of survivin and Ki-67 as potential markers for grading noninvasive papillary urothelial carcinoma.

DESIGN:

Fifty-one bladder biopsies were graded blindly by 5 experienced general surgical pathologists. The protein and messenger RNA expression of survivin and Ki-67 was evaluated by immunohistochemistry and quantitative reverse transcription-polymerase chain reaction using paraffin-embedded tissue. The immunohistochemistry result was quantitatively analyzed using a computer-based color deconvolution module.

RESULTS:

The diagnostic agreement among 5 pathologists was fair to poor, with 32% of the cases graded differently by at least 2 raters. All cases were divided into 3 groups: consensus low-grade, consensus high-grade, and indeterminate. The percentage of urothelial cells with positive survivin nuclear staining (survivin score) was significantly higher in the high-grade than in the low-grade group (P < .001). Survivin score outperformed Ki-67 in separating the high-grade group from the low-grade group and showed a significantly higher predictive accuracy for high-grade recurrence than the histologic grade. The disagreement of grading for the indeterminate group could be resolved by their survivin scores in most cases. Survivin messenger RNA level correlated well with survivin score by immunohistochemistry but was not a more discriminating marker.

CONCLUSIONS:

Significant interobserver variability exists in grading low-grade versus high-grade papillary urothelial carcinoma. Survivin immunohistochemical staining can be a useful adjunct tool for the grading of challenging cases.