On the determination of an effective planning volume for permanent prostate implants
Presented at the 41st Annual Meeting of ASTRO, San Antonio, Texas, Oct. 31–Nov. 3, 1999.
To view the full text, please login as a subscribed user or purchase a subscription. Click here to view the full text on ScienceDirect.
Fig. 1
A plot of coverage and conformity indices for the six single-period plans over a period of 30 days. EPV[t] = the plan generated by the single-period model using volumetric and dosimetric information based on time t, where t = 0, 6, 12, 18, 24, 30 days, respectively. The lines with values above 1.0 on the vertical axis correspond to the conformity indices for the six plans. Similarly, the lines with values below 1.0 on the vertical axis correspond to the coverage indices.
Fig. 2
A plot of coverage and conformity indices for five 2-period models over a period of 30 days. EPV[0,t] = the plan generated by the 2-period model using volumetric and dosimetric information based on time 0 and time t, where t = 6, 12, 18, 24, 30 days, respectively. For comparison, the single-period EPV[0] plan’s 30-day coverage and conformity plot is also shown. The lines with values above 1.0 on the vertical axis correspond to the conformity indices for the six plans. Similarly, the lines with values below 1.0 on the vertical axis correspond to the coverage indices.
Fig. 3
A plot of coverage and conformity scores over the 30-day horizon for several multi-period plans. For comparison, the single-period EPV[0] plan’s 30-day coverage and conformity plot is also shown. Note that whereas initial coverage is somewhat better for EPV[0] than for the multi-period plans, overall conformity for EPV[0] is much worse. The lines with values above 1.0 on the vertical axis correspond to the conformity indices for the five plans. Similarly, the lines with values below 1.0 on the vertical axis correspond to the coverage indices.
Abstract
Purpose: In current practice, planning for prostate brachytherapy is based on the state of the prostate at a particular instant in time. Because treatment occurs over an extended period, changes in the prostate volume (gland shrinkage) and seed displacement lead to disagreement between planned dosimetry to the prostate and the dose actually received by the prostate. Discrepancies between planned and actual dose to the rectum and urethra also occur. The purpose of this study is to investigate the possibility of defining an “effective planning volume” that compensates for changes in prostate volume and seed displacement.
Methods and Materials: Waterman’s formula is used to estimate prostate shrinkage and seed displacement. The prostate volume and potential seed positions at days 0, 6, 12, 18, 24, and 30 are used in formulating time-dependent dosimetric treatment planning models. Both single-period and multi-period models are proposed and analyzed. A state-of-the-art computational engine generates unbiased, high-quality treatment plans in a matter of minutes. Plans are evaluated using coverage and conformity indices computed at specific times over a period of 30 days. The models allow dose to urethra and rectum to be strictly controlled at specific instants in time, or throughout the 30-day horizon.
Results: For plans generated from the single-period models—based on projected prostate volumes and potential seed positions on days t = 0, 6, 12, 18, 24, 30, respectively—as t increases, the conformity index improves while the coverage worsens. In particular, the best coverage and worst conformity are achieved for the plan generated using t = 0 (day 0) information. This plan provides over 99% coverage over the entire 30-day period, and while it has initial conformity index 1.24, the conformity index climbs to 1.58 by day 30. Conversely, the worst coverage and best conformity are achieved when the plan is generated using projected information from t = 30 (day 30). Plans based on projected data at day 30 yield an initial coverage of only 84%, with conformity scores less than 1.34 over the entire 30-day period. Among the multi-period plans, with the exception of the two-period plan obtained using day 0 and projected day 6 data, the average coverage is 98% while conformity indices below 1.46 are maintained throughout the 30-day horizon. Excessive dose to the urethra and rectum is observed when only day 0 dosimetric and volumetric data are imposed in the planning procedure. In this case, by day 30, 89% of urethra volume receives dose in excess of 120% of the remaining prescription dose. Similarly, 40% of rectum volume receives dose in excess of the prescribed upper dose bound of 78% of the remaining prescription dose. When multi-period dosimetric constraints for urethra and rectum are imposed, dose to these structures is controlled throughout the 30-day period.
Conclusions: A planning method that takes into account prostate shrinkage and seed displacement over time can be used to adjust the balance between coverage and conformity. Incorporating projected future volumetric information is useful in providing more conformal plans, in some cases improving conformity by as much as 21% while sacrificing roughly 7% of initial coverage. Evidence of possible morbidity reduction to urethra and rectum via the use of multi-period dosimetric constraints on these structures is demonstrated. Among all plans considered, the plan obtained via the six-period model provides the best coverage and conformity over the 30-day horizon.
Keywords:
Prostate implants, Effective planning volume, Edema, Urethra complications, Automated treatment planningTo access this article, please choose from the options below
Society Member Login
Society Members, full access to the journal is a member benefit. Use your society credentials to access all journal content and features
Non-Member Login
Register
Create a new accountPurchase access to this article
Claim Access
If you are a current subscriber with Society Membership or an Account Number, claim your access now.
Subscribe to this title
Purchase a subscription to gain access to this and all other articles in this journal.
Institutional Access
Visit ScienceDirect to see if you have access via your institution.
☆This research is partially supported by the Whitaker Foundation and by CPLEX, a division of ILOG Inc.
