Impact of postimplant edema on V100 and D90 in prostate brachytherapy: can implant quality be predicted on day 0?

Purpose: To determine the effect of edema on the dosimetric parameters V100 (percentage of prostate volume that received a dose equal to or greater than the prescribed dose) and D90 (minimal dose delivered to 90% of prostate volume) in 125I prostate brachytherapy and to determine whether the edema can be used to predict implant quality on the day of the implant (Day 0). Methods and Materials: Fifty consecutive patients treated with 125I implants who had two postimplant CT scans were selected for this study. The mean interval between the studies was 46 ± 23 days. The implants were preplanned to deliver 150 Gy to the prostate plus a 3-5-mm symmetric dose margin using peripherally loaded 0.4-0.6-mCi (NIST-99) 125I seeds. A dose-volume histogram was compiled for each postimplant CT scan. The V100 and D90 from the first and second CT scans were compared to determine the effect of edema on these parameters. A multivariate regression analysis was performed to define the linear relationships for predicting the V100 or D90 at 30-60 days after implant from the magnitude of the edema and the values of V100 and D90 on Day 0. Results: V100 and D90 increased by 5% ± 6% and 15% ± 17%, respectively, during the interval between the first and second postimplant CT scans. The mean edema was 1.53 ± 0.20. The increases in V100 and D90 were found to be proportional to the edema and the values of V100 and D90 on Day 0. The increase in V100 was also found to depend on the width of the preplan dose margin. Linear relationships were derived that predict the V100 and D90 at 30-60 days after implant with a standard error of ±4% and ±24 Gy, respectively. Conclusion: V100 and D90 increased by 5% ± 6% and 15% ± 17%, respectively, during the first 30-60 days after implant. The results of a multivariate linear regression analysis showed that the increases in V100 and D90 were proportional to both the magnitude of the edema and the values of these parameters on Day 0. The relationships derived by linear regression analysis predict V100 and D90 at 30-60 days after implant to within ±4% and ±24 Gy, respectively. However, predicting the 30-60-day V100 and D90 on Day 0 is a poor substitute for obtaining a 30-60-day CT scan, because the uncertainty in the predicted values is greater by a factor of ≥2. Nevertheless, on average, the predicted values should provide a more reliable estimate of the actual V100 and D90 than the Day 0 values that ignore the effect of edema altogether. The increase in V100 was also found to depend on the width of the preplan dose margin; therefore, our results for V100 are only valid for implants planned with a 3-5-mm margin