Value of nuclear morphometry for differentiating localized from metastatic renal cell carcinoma.

Abstract

To evaluate the feasibility of using nuclear morphometry of primary renal cell carcinoma to differentiate cases with localized tumors from those with metastases. Using an interactive image analysis system, nuclear morphometry was performed on hematoxylin- and eosin-stained slides of 40 primary renal tumors. There were 18 cases with organ-confined disease and 22 with advanced stage. The most significant shape descriptor to differentiate between the two groups was mean nuclear regularity factor (889.5 +/- 24.4 vs. 701.1 +/- 157.3, p = 0.00015). The combination of mean nuclear area and mean nuclear regularity factor enabled accurate prediction of metastatic potential in 36 (85%) cases. Nuclear morphometry measured by image analysis provides prognostic information on patients with renal cell carcinoma.