Simultaneous Plan Optimization

Implementation of in-house software for simultaneous optimization of multiple treatment plans per patient. Objectives are defined on individual treatment plans, their dose distributions, and on dose distributions resulting from the summation of individual treatment plan doses. Simultaneous optimization of multiple linked treatment plans (SOLTP) was developed as a module to the GRATIS treatment planning system. All treatment planning CTs are linked by rigid transform matrices or by deformation fields. Dose objectives and constraints are defined on the individual treatment plans and on one or more accumulated dose distributions. During the optimization, the dose per treatment plan as well as the accumulated dose distribution is updated according to the changes in leaf positions and monitor units. The combined objective function value of each treatment plan and each accumulated dose distribution is used in the optimization. To demonstrate the possibilities of SOLTP, a cadaver was scanned 5 times on CT. Five CT scans with different bladder fillings (0, 60, 120, 180, and 240 mL) were obtained. On each CT, a treatment plan was created. The dose prescribed to 50% of the gross tumor volume (GTV) and clinical target volume was 62.5 and 40 Gy, respectively. To avoid late grade ≥2 rectal toxicity, the rectum and sigmoid volumes receiving 34.3 Gy and 34.4 Gy were constrained to 64% and 35%, respectively. Two sets of optimizations were performed separately. The first optimization was performed with objectives defined on individual treatment plans only. The second optimization had objectives defined on the individual treatment plans, as well as on the accumulated dose distribution from all 5 treatments plans on the 120-mL CT. Dose computations were performed with the collapsed cone convolution/superposition algorithm of a treatment planning system. The minimum dose to the GTV on the dose distribution accumulated on the 120-mL CT was higher in SOLTP (D98% : 58.74 Gy) than in the individually optimized treatment plans (IOTP) (D98% : 55.30 Gy). The V34.3Gy and V34.4Gy for the rectum with IOTP was respectively 3.20 and 3.12% while with SOLTP it was reduced to 1.65 and 1.56%. The V34.3Gy and V34.4Gy for the sigmoid with IOTP was respectively 0.279 and 0.278% while with SOLTP it was 0.279 and 0.269%. The D2% in IOTP for the femoral head left and right was respectively 29.70 and 31.29 Gy, while for SOLTP the D2% was respectively 29.03 and 29.61 Gy. The small intestine received a higher D2% using SOLTP (39.01 Gy) than using IOTP (36.78 Gy). A better target coverage was obtained with the optimization taking into account the objectives defined on the accumulated dose distribution. SOLTP reached all clinical constraints on the individual treatment plans and on the accumulated dose distribution.