RESUMEN
<p><b>OBJECTIVE</b>To explore the method of predicting the stage of prostate cancer with serum prostate-specific antigen (PSA) and pathological grade.</p><p><b>METHODS</b>One hundred and eighty-seven patients were studied retrospectively with prostate cancer diagnosed by systemic biopsy in our hospital. The rank correlation analysis, rank sum test and stepwise discriminant multivariate analysis were used to assess the correlation of serum PSA level, ratio of free PSA to total PSA (FPSA/TPSA ratio) with Gleason score (GS) and stage.</p><p><b>RESULTS</b>Serum PSA level increased with GS for prostate cancer patients (r = 0.369, P < 0.001). With increasing stage, serum PSA level and GS increased (r = 0.398, 0.530, P < 0.001). Overall, FPSA/TPSA ratio was not correlated with stage (P > 0.70), but a significant negative correlation was demonstrated between them when serum PSA < or = 10 microg/L (r = -0.600, P < 0.05). When serum PSA > 20 microg/L, 67% - 87% patients with prostate cancer may be stage C or D. The equation using serum PSA and GS to predict the stage of patients with prostate cancer was: x = -3.488 + 0.041 x PSA + 0.428 x GS.</p><p><b>CONCLUSIONS</b>Serum PSA level is positively correlated with GS for prostate cancer patients. Serum PSA level and GS are positively correlated with stage. A negative correlation between FPSA/TPSA ratio and stage is demonstrated when serum PSA < or = 10 microg/L. The combination of serum PSA and GS may predict the stage of patients with prostate cancer.</p>