D90S Predicting Biochemical Tumor Control after 125I Prostate Brachytherapy: Effects of Day 1 and Day 30 Dosimetry and a Prostate Margin

To investigate cutoff levels of postimplant D90s of prostate and prostate with a margin on day 1 and day 30 predicting biochemical tumor control for patients with localized prostate cancer treated with 125I brachytherapy. Patients with low- to intermediate-risk prostate cancer treated with 125I brachytherapy were analyzed. Between 2003 and 2009, a total of 682 patients was observed from 24 to 108 months (median: 60 months). According to National Comprehensive Cancer Network risk classification, 457 patients with low-risk, 224 patients with intermediate-risk and one patient with high-risk disease were included. A dose delivered to 90% of the prostate (D90-P) on day 1 and day 30 postimplant CT-based dose-volume histogram were calculated. And D90 of prostate plus a 3-mm margin (D90-PM: three-dimensional expansion of the prostate with 3 mm in each direction except the posterior and cranial direction) on day 30 was assessed as well. All cases were contoured with VariSeed version 8 by one radiation oncologist (AY). A prostate-specific antigen (PSA) relapse was defined according to the ASTRO definition. Freedom from biochemical failure (FFBF) rates were calculated with the methods of Kaplan and Meier. Differences in rates were tested with a log-rank test. Neoadjuvant androgen deprivation therapy was given for 56 % of patients. The average D90-P was 156.5 Gy (median, 158.4; range, 91.8-196.3) on day 1 and 183.2 Gy (median, 184.7; range, 126.4-233.7) on day 30. The average D90-PM was 160.2 Gy (median, 161.1; range, 99.7-197.6) on day 30. The average prostate volume was 24.8 cc (100%) with TRUS before treatment, 32.3 cc (130 %) on day 1 and 28.1 cc (113%) on day 30. The average volume of prostate with a 3-mm margin was 38.4 cc (155%) on day 30. The 5-year FFBF rates for patients receiving D90-P of 130 Gy or higher on day 1 was 96.9% compared with 88.6% for those receiving a D90-P lower than 130 Gy on day 1 (p=0.022). According to a day 30 postplan, the 5-year FFBF for patients with D90-P <160 Gy, 160-180 Gy, and ≥180 Gy were 89.0%, 95.2%, and 98.7%, respectively (p<0.001). A level of 160 Gy was a good cutpoint of D90-P on day 30 (p<0.001), and so was even a level of 180 Gy (p=0.002). The 5-year FFBF for patients with D90-PM <140 Gy, 140-160 Gy, ≥160 Gy on day 30 were 87.9%, 96.3%, and 97.8%, respectively (p=0.002). A level of 140 Gy was a cutpoint of D90-PM on day 30 (p=0.001). A dose response was marginally observed at a cutoff level of 130 Gy of D90-P on day 1, and significantly observed at a level of 160 Gy of D90-P on day 30 and at a level of 140 Gy of D90-PM on day 30. Improvements in FFBF were more pronounced with increasing D90 levels on day 30 compared with day 1. The D90-P levels on day 30 were robust in predicting biochemical control. These dose response data could explain previous reports of various cutpoints for D90s predicting biochemical tumor control.