SU‐FF‐T‐94: Clinical Evaluation of Direct Machine Parameter Optimization Algorithm for Head and Neck IMRT Treatment

Abstract

Purpose: Because many critical structures are in close proximity to target volumes, cancers of the head and neck (H&N) are often suited for treatment with IMRT. However, the time required to generate and deliver a clinically acceptable IMRT plan can be significantly longer than a conventional plan. This study evaluated a new inverse planning algorithm, DMPO (direct machine parameter optimization), with attention to parameter settings, plan quality and treatment efficiency for H&N cancers. Method and Materials: The Pinnacle treatment planning system version 7.4 was used. The DMPO allows users to limit the number of total MLC segments (N) for treatment. After a user‐defined number of iterations (n) for pencil beam optimization, the DMPO generates MLC segments for each field for dose calculations using a convolution algorithm. Both the MLC leaf positions and the weight of each segment are then optimized until cost tolerance or iteration number is reached. Treatment plans generated using DMPO were compared with H&N cases that were previously treated using an older version (6.2). The plan quality was compared using cost functions and DVHs of target volumes and critical structures. The total monitor units and MLC segments for treatment were compared for different combinations of n and N. Results: The DMPO provided plans of DVHs similar to clinical cases with significantly less planning time. More importantly, the total MU and MLC segments for treatment delivery were reduced by 40% to 50%. Cost functions changed only slightly on n and N and total MU increased as n increased, but was independent of N. Our preliminary data indicated a combination of n=10–15 with 10 segments per field appeared to be optimal for most H&N cases. Conclusion: The DMPO algorithm generated more efficient plans while providing equal or better quality than the previous plans for IMRT treatment.

Fluence field optimization for noise and dose objectives in CT

“…Such approaches are currently used in the field of IMRT, where optimization of fluence patterns for dose delivery is constrained by the shape and design of beam-shaping collimator leaves, via direct machine parameter optimization (DMPO). 49,50 A similar multileaf collimation approach for altering fluence patterns might also be applicable to FFMCT. Current MeV CT scanners for example, are already equipped with a multileaf collimation for dose delivery application.…”

Combining segment generation with direct step‐and‐shoot optimization in intensity‐modulated radiation therapy

“…[6]. In addition, a clinical study has found that DSS optimization simplifies plans without reducing plan quality, compared to plans generated with the two-step approach [26]. Our segment generation module is based on the column generation approach for IMRT presented in [22].…”

Comparison of direct machine parameter optimization versus fluence optimization with sequential sequencing in IMRT of hypopharyngeal carcinoma

“…Because of limitations of the MLC settings the resulting fluence is different from the optimization result and therefore no longer optimal [ 15 ]. Other systems incorporate the MLC sequencing in the optimization process [ 16 , 17 ], or optimize the machine parameters directly [ 18 , 19 ]. In both cases the MLC position is taken into account in the optimization process and the resulting optimal fluence can be delivered by the linac without further approximations [ 15 ].…”