Partin nomogram effectively discriminates advanced prostate cancer
MedWire News: The Partin nomogram for prostate cancer staging discriminates excellently in the prediction of seminal vesicle (SV) invasion and positive lymph nodes (PLN), and performs best in men below 61 years of age, study findings show.
However, the prediction of extraprostatic extension (EPE) and organ confined (OC) disease with the nomogram is more limited, notes the research team.
“The pathologic stage of prostate cancer is critical to the success of curative local therapy,” explain James Yu, from the Yale School of Medicine in New Haven, Connecticut, USA, and colleagues.
“The Partin tables are a useful, widely applied clinical tool used by physicians to counsel patients about the risk of extraprostatic disease,” they add in The Journal of Urology.
The researchers identified 11,185 men who were treated with radical prostatectomy for prostate cancer and whose data were included in the National Cancer Institute Surveillance, Epidemiology, and End Results database.
The team evaluated the accuracy of the 2007 Partin tables for prediction of PLN, SV invasion, EPE, and OC disease in the cohort, using the clinical disease stage, preoperative prostate-specific antigen levels, and Gleason scores of each patient.
To illuminate the discriminative ability of the Partin tables, the researchers constructed receiver operating characteristic (ROC) curves, graded using an area-under-the-curve (AUC) score between 0.5 and 1, with 1 indicating perfect discrimination.
The AUC scores for PLN, SV invasion, EPE, and OC, were 0.77, 0.74, 0.62, and 0.68, respectively, demonstrating significant discrimination for PLN and SV invasion.
The accuracy of these results was then assessed using the Brier score, which compares the predicted probability of outcome and the actual outcome. With 0 representing perfect discriminative accuracy, the Brier scores for positive LN and SV invasion were 0.026 and 0.042, respectively, indicating excellent discriminative accuracy.
Additionally, the researchers considered predictive accuracy by age, choosing a boundary of 61 years, the median age among the cohort. The Partin table discriminated PLN and SV invasion better in younger than older men, with AUC curves of 0.80 versus 0.74, and 0.76 versus 0.71, respectively. EPE and OC disease discrimination was similar in each age group.
The researchers observe that this finding “may reflect the relative youth” of the cohort overall.
The team concludes: “The usefulness of the nomogram is its simplicity and inclusiveness with accurate prediction based on the most commonly available preoperative data.”
MedWire (www.medwire-news.md) is an independent clinical news service provided by Current Medicine Group, a trading division of Springer Healthcare Limited. © Springer Healthcare Ltd; 2009
By Sarah Guy