Inverse planning anatomy-based dose optimization for HDR-brachytherapy of the prostate using fast simulated annealing algorithm and dedicated objective function.

摘要: ~Received 7 August 2000; accepted for publication 26 February 2001!An anatomy-based dose optimization algorithm is developed to automatically and rapidly producea highly conformal coverage of the target volume while minimizing urethra, bladder, andrectal doses in delivery an high dose-rate~HDR! brachytherapy boost treatment ofprostate cancer. The dwell times are optimized using inverse planning simulated annealingalgorithm ~IPSA! governed entirely from anatomy extracted a CT by dedicatedobjective function ~cost function! reflecting clinical prescription constraints. With this inverseplanning approach, focus on physician’s constraint instead thetechnical limitations. Consequently, control improved. Thecapacity represent presented pros-tate case. computation time ~CPU! IPSA less than 1 min ~41 s 142915iterations! typical case, allowing fast practical optimization. achieve-ment opens possibility deliver ahigher prostate without inducing overdosage urethra normal tissues surroundingthe prostate. Moreover, same concept, it will be possible adelimited tumor within Finally, method can easily extended otheranatomical sites. © 2001 American Association Physicists Medicine.@DOI: 10.1118/1.1368127#Key words: planning, brachytherapy, HDR, prostate, annealingI. INTRODUCTIONThe last decade has seen major changes way radiationtreatments delivered. century-old objective radia-tion therapy, i.e., curative whilepreserving tissues, now aimed at with highdegree sophistication. However, spite im-provements achieved 3D imaging modalities that allowthe properly defined, treat-ments have not yet fully benefit these important newpieces information. positioning seeds per-manent implant or catheters after-loading isnow image guided, greatly improving localization theapplicators. At stage, however, isusually only indirectly taken into account. For placed vol-ume assumed if distribution covers thecatheters, should also cover anatomy. Imaging com-monly used set margins, but dosedistributions based considerations such as catheter po-sitions desired few defined points. This neces-sarily results approximation shape theanatomy. case volumeoptimization essentiallycylindrical shape. may ‘‘average’’ pros-tate, each very specific Any ap-proximation its geometrical representation willresult surrounding theprostate. organs risk vicinity prostatemay included unnecessarily overdosed region.Recent studies