Predicting pelvic lymph node involvement in current-era prostate cancer

The Roach formula [2/3 × prostate-specific antigen + (Gleason score--6) × 10], derived in 1993 during the early prostate specific antigen (PSA) screening era, has been used to predict the risk of pelvic lymph node involvement in patients with prostate cancer. In the current era of widespread PSA screening with a shift to earlier disease stages, there is evidence to suggest that the Roach score overestimates risk of nodal metastasis. This study retrospectively reviews the validity of this formula as a prediction tool.

We conducted a retrospective institutional review including men with clinical T1c-T3 prostate cancer, with baseline PSA levels and biopsy-obtained Gleason scores who underwent radical prostatectomy with pelvic node dissection from 2001 through 2009 (N = 1,022). The predicted risk of nodal involvement was calculated for each patient using the Roach formula and then compared with actual rates following surgery.

The study included 1,022 patients; 99.6% had clinical T1c/T2 disease, with a mean of 10.3 lymph nodes surgically evaluated. Overall, 42 patients (4.1%) had nodal metastasis. For every range of scores, the Roach formula overestimates the risk of nodal involvement. Observed nodal positivity was 1%, 6.3%, 10%, 15.2%, and 16.7% for Roach scores ≤ 10%, >10%-20%, >20%-30%, >30%-40%, and >40%, respectively. The Roach score overestimates the risk by approximately 4.5-fold in patients with scores ≤ 10%, by 2.5-fold for all scores between 10% and 40%, and by 4-fold for scores >40%.

The Roach formula overpredicts the risk of pelvic nodal involvement in current-era prostate cancer patients undergoing regular PSA screening and with mainly T1c/T2 disease. Contemporary patients are much less likely to have nodal involvement for a given PSA and Gleason score.