Objective: To relate the predictive value of serum prostate specific antigen (PSA) levels, histological grade and intraductal carcinoma (IDC-P, a possible marker of poor prognosis) to pathological stage and subsequent clinical outcome, and thus derive an improved predictive model to aid the decision to initiate potentially curative therapy in localized prostate cancer.
Materials and methods: Fifty-nine radical prostatectomy specimens were histologically graded, allocated a pathological stage and the tumour volume determined by image analysis. Pre-operative (needle biopsy) tumour grade, the presence or absence of IDC-P, and serum PSA levels were correlated with the pathological stage. This was used to define the sequence and values that would be incorporated into a predictive model for pathological stage and clinical outcome.
Results: There were close correlations between cancer volume and tumour grade (P = 0.004) and between cancer volume and serum PSA level (P = 0.003). However, in tumours with IDC-P, serum PSA level did not correlate with tumour volume of IDC-P (P > 0.9). IDC-P was an independent variable that significantly improved the prediction of pathological stage and tumour volume, and furthermore, was closely related to (r = 0.53, P = 0.001) and accurately predicted treatment failure.
Conclusion: The model which best predicted pathological stage and clinical outcome involved first identifying those cancers with IDC-P as having the poorest outcome. Cancers without IDC-P were then separated into low- and high-risk groups on the basis of serum PSA levels below and above 10 ng/mL, and those in the high-risk group further stratified using Gleason grading. Furthermore, the use of a sequential consideration of pre-operative variables including IDC-P allowed cases to be grouped which, after radical surgery, closely correlated with clinical outcome.