Work

802: Modulation of Ceramide Metabolism Enhances Tumor-Selective Viral Protein Apoptin’s Cytotoxicity in Prostate Cancer

Year: 2006

Type: article

Abstract: value could be identified to define what should constitute the extent of an 'extended PLND'.METHODS: We used a cohort of 1863 intermediate to high risk patients, selected for a PLND based on validated LNI risk stratification (Conrad S JU 2002).Between 1 and 29 nodes (mean 6.9, median 6) were removed.PSA values were 0.12 to 48.5 (mean 11.3, median 8.5).Biopsy Gleason sums were 6 or less in 519 (51.3%), 7 in 430 (42.5%) and 8-10 in 63 (6.2%) men.PSA, stage and Gleason were used as LNI predictors.To reduce overfit bias, 200 bootstrap resamples were used.Moreover, the multivariate nomogram method was used to graphically explore the increase in risk according to the extent of PLND after controlling for other covariates.RESULTS: LNI was diagnosed in 149 (14.7%).In univariate analyses, all variables statistically significantly predicted LNI (all p<0.001 ).The number of nodes was most informative (0.791 ), followed by Gleason (0.717), PSA (0673) and stage (0.607).The accuracy of a multivariate PSA, stage and Gleason sum-based model was 0.784 Vs 0.860, when total nodes were added (7.6% gain).The nomogram figure demonstrates that increasing number of total nodes (Totai#Nodes) directly increases the probability of LNI (in figure: PN+).For example, removal of 26 nodes instead of 5 results in 70 additional 'risks points', which increase from 10 to 90% the probability of LNI.CONCLUSIONS: The number of nodes represents the foremost determinant of LNI.A cut-off is difficult to determine, as LNI risk increases in a continuous fashion according to the number of nodes removed.However, our model accounts for this limitation and is capable of highly accurately predicting the probability of LNI, according to the number of removed nodes.

Source: The Journal of Urology

DOI: 10.1016/s0022-5347(18)33038-6