2827 Clincal IMRT Planning Testing for the Multi-Criteria Optimization (MCO) Methodology

C. Yang, L. Peng, S. Sim, M. Weiss

Monmouth Medical Center, Long Branch, NJ

Purpose/Objective(s): To apply a biological model based algorithm for acquiring optimized IMRT planning solutions. Thisinteractive planning tool will help users to select the best available plans in the IMRT solution space.

Materials/Methods: IMRT often is a time consuming iterative optimization process between evaluation of the dose distributionand redefinition of the object function. An IMRT planning optimization tool (Multi-Criteria Optimization, MCO™) has beenintroduced for non-clinical evaluation to acquire the best available solutions. Based on a Pareto’s solution concept, this toolcould search the solution space and offer users a limited set of deliverable IMRT plans. With this interactive process, users canset the target and critical structures dose constraints with the biological model (EUD) to obtain the best solution. We usedPinnacle system as the benchmark to compare the dosimetric gain from the MCO algorithm, DVH indicated excellent sparingwith better PTV coverage is achievable from the MCO process in KonRad system.

Results: Dosimetric findings are summarized as 1) MCO optimization testing shows that much better dose distribution can beachieved compared to the current planning results. Due to the confined solution space, the optimal results are easily achievable.2) MCO with Pareto’s approach is durable in the solution searching process. It is interactive with the graphical interface whichthe dose distribution along with the DVH can be compared simultaneously (Fig. 1). 3) IMRT dose optimization and summarybased on the MCO methodology are very conceivable. With pre-calculated IMRT solutions, final results help users to select thebest available plan from the solution domain in real time.

Conclusions: From this interactive MCO planning tool, we can calculate the best IMRT results in a very reasonable time frame.Human factors for determining an acceptable plan can be dramatically reduced.

Figure 1. An interactive tool for isodose and DVH evaluations

Author Disclosure: C. Yang, None; L. Peng, None; S. Sim, None; M. Weiss, None.

2828 Quantitative Delineation of PET Standard Uptake Values for Radiotherapy Treatment Planning:Validation and Application To Head and Neck Cancer

J. E. Bayouth, F. Qing, M. M. Graham, M. Yao

University of Iowa Hospitals & Clinics, Iowa City, IA

Purpose/Objective(s): Develop and validate a methodology to quantitatively delineate regions of abnormal FDG uptake forhead and neck (H&N) tumors based on standard uptake values (SUV) computed from raw pixel values within the treatmentplanning system, and apply this technique to evaluate regions of abnormal FDG uptake in relationship to anatomically definedclinical target volumes (CTV).

Materials/Methods: PET/CT images from 18 patients acquired on a Siemens Biograph with biopsy proven H&N cancer (10tonsil, 5 base of tongue, 1 pyriform sinus, 1 oral cavity, and 1 supraglottis) were analyzed. Maximum SUV was identified in5 separate anatomical sites (typically the gross tumor, abnormal lymph nodes, stomach, left and right kidney) for each patienton the Nuclear Medicine (NM) workstation (MS Viewer). Corresponding regions of interest were drawn on PET/CT imagestransferred into the Radiotherapy treatment (RT) planning system (Philips Pinnacle) and the maximum pixel values on PETimages were recorded. Information on patient weight, injected activity, and time from injection to initial scanning wasdetermined from the DICOM header of the PET images and used to compute the image set unique relationship between pixelvalues and SUV for each patient. SUV values between the NM workstation and the RT planning system were compared. Onceverified, the SUV/pixel value relationship was utilized to perform objective, quantitative tumor delineation on the RT planningsystem to contour voxels of FDG uptake with an SUV � 2.5. The volumes of SUV� 2.5 outside of the anatomically definedprimary tumor and grossly involved nodes (CTV1), high risk nodes and tissue (CTV2), and lower risk nodes and tissues (CTV3)regions were computed. The mean SUV outside the CTV(s), and the minimum expansion of the CTV(s) required to cover theSUV� 2.5 volume were also determined.

S675Proceedings of the 48th Annual ASTRO Meeting

Results: Maximum SUV computed from RT pixel values for the anatomical sites considered (mean 8.2, range 1.7–27.6)corresponded well with those measured on the NM workstation (r2�0.9985, n�90). Kidney maximum SUV only correlatedwell when the maximum value for the entire organ was considered, due to the dramatic SUV gradient and differences in slicethickness displayed on the two workstations (1 mm - NM, 3.5 mm RT). SUV threshold delineation of CTV was accomplishedon 12 of 18 patients whose PET study was within 16 days of the planning CT. Regions were modified to exclude knownfalse-positive areas of uptake (e.g., contralateral tonsil). The volume of SUV� 2.5 around CTV1 had a range of 12.7–132.2 cc(mean 59.6). The SUV volume delineated outside of the CTV1 ranged from 0.3–26.8 cc (mean 17. 4cc or 24.4%); the volumeof SUV � 2.5 outside all CTV regions ranged from 0–12.6 cc (mean 4.2 cc or 7.7%). The mean SUV in these regions identifiedoutside of CTV1 and all CTVs were 3.1 and 3.4, respectively. The mean minimum expansion of CTV1 and all CTVs requiredto cover the SUV� 2.5 volume were 16.5 and 6.4 mm, respectively.

Conclusions: An objective and quantitative tumor delineation approach was developed, using pixel value thresholds forcontouring voxels of specified FDG uptake. This methodology was validated for a wide range of SUVs for H&N patients, andwhen applied to delineate regions around tumors with an SUV� 2.5, clinically significant regions outside of both the primaryCTV, as well as all CTVs were identified.

Author Disclosure: J.E. Bayouth, None; F. Qing, None; M.M. Graham, None; M. Yao, None.

2829 Jaws-Only IMRT Using Direct Aperture Optimization

M. K. N. Afghan, M. A. Earl, C. X. Yu, Z. Jiang, D. M. Shepard

University of Maryland School of Medicine, Baltimore, MD

Purpose/Objective(s): Clinical implementation of intensity modulated radiation therapy (IMRT) is significantly more com-plicated for facilities without a multileaf collimator (MLC). In this study, we have examined the use of Direct ApertureOptimization (DAO) to produce efficient IMRT treatment plans that can be delivered without an MLC.

Materials/Methods: This “jaws-only” approach to IMRT uses a series of rectangular field shapes to achieve a modulatedintensity pattern from each beam direction. The user specifies the number of allowable jaws-only segments, and theoptimizer directly optimizes the jaw positions and the relative weights assigned to each aperture. Because all of theconstraints imposed by the jaws are incorporated into the optimization, the need for leaf sequencing is eliminated. Theoptimization is performed using a simulated annealing algorithm. For fives clinical cases (2 prostates, 1 pancreas, 1abdomen, and 1 breast), treatment plans were created using 5, 10, 15, 20, and 25 apertures per beam direction. The planswere compared with the benchmark MLC-based DAO plan created for each case using the same treatment objectives. Eachplan was delivered to a phantom where absolute dose was measured with an ionization chamber and relative dose wasmeasured with Kodak EDR film.

Results: For the jaws-only treatment plans, only minor improvements were observed as the number of segments per beamdirection was increased beyond 20. Figure 1 plots a DVH comparison between an MLC plan (5 apertures per direction) versusa jaws-only plan (20 apertures per direction) for a prostate patient. DVH and isodose comparisons for the cases included in thisstudy reveal that jaws-only IMRT is generally able to approach the plan quality provided by MLC-based IMRT using 15 to 20jaw settings per beam direction. For the most complex cases involving multiple target shapes or multiple prescription levels,jaws-only IMRT is unable to achieve comparable critical structure sparing while maintaining an acceptable number of segmentsper beam direction. The 5 plans in this study were all delivered in 15 minutes or less and passed our IMRT verification criteria.

Conclusions: The ability of DAO to generate efficient jaws-only IMRT plans has been explored. Our results demonstrate thatthe “jaws-only” approach used in conjunction with DAO provides a feasible IMRT delivery option for clinics without amultileaf collimator.

Author Disclosure: M.K.N. Afghan, None; M.A. Earl, None; C.X. Yu, None; Z. Jiang, None; D.M. Shepard, None.

S676 I. J. Radiation Oncology ● Biology ● Physics Volume 66, Number 3, Supplement, 2006