Stereological estimation of mean nuclear volume as a prognostic factor in canine subcutaneous mast cell tumors

Classification schemes regarding canine subcutaneous mast cell tumors (csMCTs) remain elusive, lack consensus, and are prone to interobserver variability and bias. This observational study aimed to assess the reproducibility and the prognostic significance of volume-weighted mean nuclear volume ($\overline{v_v}$), a stereological estimation offering insights into nuclear size and its variability, in csMCTs. Thirty csMCTs were selected with information regarding outcome, and $\overline{v_v}$ was estimated using the "point-sampled intercept" method. Interobserver and intraobserver $\overline{v_v}$ reproducibility yielded concordance coefficients near or above 0.90. Regarding previously reported risk factors (pattern, mitotic count, and multinucleated cells), no statistically significant differences were identified between patterns and clinical outcome, nor between patterns and $\overline{v_v}$; however, the infiltrative pattern was represented more in the poorer outcome group and had higher $\overline{v_v}$ values. When comparing $\overline{v_v}$ and clinical outcome, a statistically significant difference emerged. Cases with poorer outcomes had higher $\overline{v_v}$ values ($\tilde{x}$ = 192.9) than cases with more favorable outcomes ($\tilde{x}$ = 120.5), and this association was statistically significant on both univariable and multivariable analyses. This study suggests that $\overline{v_v}$ is highly reproducible and is associated with clinical outcome in csMCTs.