Do the Tables and Nomograms Commonly Used by Clinicians To Predict Pathological Stage in Men with Biopsy Gleason Score (GS) 8-10 Adenocarcinoma of the Prostate Match What Is Seen at Radical Prostatectomy (RP)?
Brian Robinson, Nilda Diana Gonzalez Roibon, Bruce Trock, Jonathan Epstein. The Johns Hopkins Hospital, Baltimore
Background: The Partin Tables and the Kattan Nomogram are the two most commonly used tools to predict the pathological stage at RP based on preoperative clinical findings and biopsy pathology. These tools use mathematical equations (logistic regression) which are primarily based on the majority of men with prostate cancer (ie. biopsy GS ≤7) to predict the outcome for all men, including those with higher grade cancer on biopsy. It is unknown how accurate these predictions are which can have significant implications for deciding therapy.
Design: 411 men with GS 8-10 prostate cancer on biopsy (2004-2010) who under RP at our institution were studied. All RPs were serially sectioned and submitted in their entirety. Clinical stage, biopsy GS, and serum PSA for each patient was inputted into both the Partin Tables and the Kattan Nomograms to generate a predicted probabilities of each stage which was then compared to the actual findings at RP.
Results: For the Partin Table there were no significant differences between observed and predicted organ confined (n=175 and 178.5, respectively, p=0.898), seminal vesicle invasion (SVI) (n=55 and 56.2, respectively, p=0.224), or lymph node metastases (LN met.) (n=32 and 26.3, respectively, p=0.142). Using the Kattan Nomograms, it predicted 215 organ confined cases compared to 175 RPs which had organ confined disease (p=0.022). The Kattan nomograms also predicted 75 cases with SVI, whereas there were 80.5 RPs with SVI (p=0.0002). There were no difference between observed and predicted LN met. using the Kattan nomogram (n=32 and 31.0, respectively, p=0.395).
Conclusions: Utilizing a large number of cases with high grade cancer at biopsy and matched entirely submitted RP specimens, the current study was able to assess the accuracy of the two most commonly predictive tools used by clinicians to predict pathological stage. Whereas the Partin Tables accurately estimated pathological stage, the Kattan Nomogram underestimated stage at RP. Using the Kattan Nomograms, men with GS 8-10 disease on biopsy may be counseled that they have more curable disease than they in fact do. Some men may choose RP based on these overly optimistic predictions as opposed to other therapies, such as radiation therapy. Pathologists can play a key role in the treatment and prognostication of prostate cancer by testing the validity of commonly used clinical tools using RP data.
Category: Genitourinary (including renal tumors)
Tuesday, March 1, 2011 11:00 AM
Platform Session: Section A, Tuesday Morning