Volume-weighted mean nuclear volume. Is this new prognosticator comparable in different institutions?

To determine whether volume-weighted mean nuclear volume (MNV) obtained at one institution is comparable to that from other institutions.

MNV calculated from histologic slides obtained at three hospitals--Shizuoka Prefectural Hospital (SPH), Shimada Municipal Hospital (SMH) and Shizuoka City Hospital (SCH)--were compared. Between December 1994 and June 1996, transurethral resection of bladder tumor or transurethral resection of the prostate was performed on 37 patients at SPH and 50 patients at SMH; histologic specimens from 40 cases of bladder tumors, 63 cases of normal bladder mucosa, 28 cases of benign prostatic hyperplasia and 1 case of prostate cancer were obtained. A portion of each specimen obtained at SPH or SMH was carried to SCH, and histologic slides were made at SCH using it. Using the remaining position of each specimen, histologic slides were then prepared at each hospital. Estimates of MNV were made from all histologic slides from each hospital, and the differences in MNV between the hospitals were analyzed. In addition, intraobserver and interobserver reproducibility were analyzed using 50 specimens obtained between December 1994 and August 1995.

On linear regression analysis, comparison of MNVs calculated from the histologic slides from SPH and SCH and those calculated from SMH and SCH revealed high correlation coefficients (R = .966 and .966, respectively), and the slope of the regression line did not differ significantly from unity. The paired t test also disclosed no significant difference between MNVs calculated at the two hospitals. Furthermore, the correlation coefficients for intraobserver and interobserver reproducibility of MNV estimates were also high (R = .918 and .949, respectively).

The results of this study indicate that estimates of MNV are comparable in multiple institutions, and we recommend that they be used to support subjective histologic grading.