Automated Image Analysis of Microvascular Density in Clear Cell Renal Cell Carcinoma and Its Prognostic Utility
W Dubinski, V Iakovlev, M Gabril, Y Youssef, K Kovacs, S Metias, M Mankaruous, GM Yousef. St. Michael's Hospital, Toronto, ON, Canada; London Health Sciences Center, London, ON, Canada
Background: Tumor microvascular density (MVD) has been shown to correlate with the aggressiveness of several cancers and adjuvant anti-angiogenic therapy is currently being used. The significance of tumor vascularity in clear cell renal cell carcinoma (ccRCC) has been debated, with various publications showing contradictory results. Previous publications have been limited by manual quantification of MVD within only a small area of tumor. To overcome this, we employed a standardized image analysis approach, and normalized the data by assessing the cellular density of the tumor; we then assessed the vascularity of ccRCC and compared it with clinical outcome data.
Design: Clinico-pathologic data was collected for 50 cases of ccRCC. Each tumor and adjacent normal kidney underwent immunostaining for CD34 and whole slide scanning using an Aperio CS Scanner (Vista, CA) at 20x magnification. Then, computer image analysis was used to assess MVD and nuclear density within tumoral, and normal cortical and medullary tissue. We generated sets of normalized and raw non-normalized vascular data and tested these for correlations with clinical prognostic parameters and survival data.
Results: Tumors with increased MVD showed a statistically significant association with higher tumor stage compared to tumors with lower vascularity (p=0.017). There was a trend toward increased MVD in higher versus lower grade tumors. Increased MVD showed a statistically significant correlation with poor prognosis and a decrease in progression free survival versus those tumors with low MVD (p=0.016). In the univariate analysis, using MVD as a categorical variable, there was a hazard ratio of 9.14 (95% CI, 1.02-82.00). This was not statistically significant when used as a continuous variable. In the multivariate analysis, patients with highly vascular tumors had an increased hazard ratio of 4.82 compared to those with lower CD34 expression (95% CI, 0.51-45.6; p=0.17). There was a positive correlation between MVD and tumor size (rs=0.456; p=0.001).
Conclusions: The latest developments in digitized histology can be used for objective standardized assessment of prognostic markers. Our standardized assessment revealed that increased MVD in ccRCC was associated with higher tumor stage and size and a decreased disease free survival. These results may prove useful in the development of new prognostic tests and anti-cancer strategies.
Category: Genitourinary (including renal tumors)
Wednesday, March 24, 2010 9:30 AM
Poster Session V # 91, Wednesday Morning