Abstract
OBJECTIVES To determine the relationship between dosimetric parameters obtained on postimplantation Day 0 and biochemical relapse-free survival (bRFS) in patients treated with (125)I transperineal interstitial permanent prostate brachytherapy (TIPPB). METHODS Two-hundred twenty men with low-risk (n=155, 70.4%), low-volume intermediate-risk (n=63, 28.7%), or high-risk (n=2, 0.9%) prostate cancer were treated with TIPPB between December 2000 and June 2006. Seventy-four (33.6%) patients received short-term (3-6 months) androgen suppression therapy before TIPPB. The median followup for patients free of biochemical failure was of 37.9 months (range, 24.0-84.5 months). RESULTS The receiver operating characteristic (ROC) analysis established a best-fit cutoff value for the quantifiers D(90) and V(100) of 147Gy and 92%, respectively. The Kaplan-Meier analysis of bRFS at the cutoff value of D(90)=147Gy using the ASTRO, nadir+2, and combined (ASTRO and nadir+2) definitions showed a trend toward statistical significance for the ASTRO (p=0.076) and nadir+2 (p=0.064) definitions and a statistically significant correlation for the combined definition (p=0.033). The corresponding 7-year bRFS for the D(90) >147Gy and D(90) ≤147Gy subsets using the ASTRO, nadir+2, and combined definitions were 96.5% vs. 89.7% (ASTRO, p=0.076); 93.7% vs. 70.5% (nadir+2, p=0.064); and 94.4 vs. 75.5% (combined, p=0.033). The V(100) (%) cutoff value of 92% predicted by the ROC analysis was not significant. Among other cutoff values, only D(90)=140Gy (p=0.050) and D(90)=160Gy (p=0.098) showed a trend toward statistical significance when the nadir+2 and the ASTRO definitions were used. The rest of dosimetric, tumor, and patient parameters did not show statistical correlation with bRFS in the Kaplan-Meier analysis. CONCLUSIONS The cutoff value of D(90)=147Gy obtained on postimplantation Day 0 showed a trend toward significant correlation with bRFS when the standard ASTRO and nadir+2 definitions were used and a weak but statistically significant correlation with bRFS as per the nonstandard combined definition in a series of patients with predominantly low-risk disease (70.4%) treated at high radiation doses (median D(90)=152.9Gy, median V(100)=92.5%).
