1. Fig. 4. Percentage of passing rate between clinical and 544 plans.
Purpose/Objectives
Results
Discussions
The purpose of this study was to identify a set of intensity-modulated radiation therapy (IMRT) planning parameters that attempts to optimize the balance among quality assurance (QA) passing rate, plan quality, dose calculation accuracy, and delivery time for Spine Stereotactic Radiosurgery (SSRS) plans.
The average coverage of GTV and CTV was 82.5 ± 13% (clinical) vs ± 13% (544) and 92.3 ± 8% (clinical) vs ± 8% (544). Maximum point dose to cord was 11.4 ± 3.5 Gy (clinical) vs ±4 Gy (544). Max point dose to GTV increased slightly from an average of 115% in clinically approved plans to an average of 117% in the 544 trials. Total plan delivery time was decreased by an average of 11.3% for the 544 plans. The average QA passing rate for the clinical plan versus the 544 plan was 90.3% and 91.9%.
Even though IMRT plan quality depends on the degree of modulation the TPS provides, for the studied SSRS cases, 5 segments per beam and 4 cm2 minimum segment area are more efficient in generating clinically acceptable plans compared to 10 segments per beam and 2-3 cm2 minimum segment area. DVH analysis showed similar doses to target and critical structures. The largest challenge was the increase in max point dose from an average of 115% in clinically approved plans to an average of 117% in the 544 trials, which is due to the decrease in the degree of modulation. However, paired t-test of the maximum point dose to spinal cord, cauda equina and GTV Paddick’s conformity index2 proves no statistical difference between the 544 and its clinical counterpart, which confirms the comparability between 544 and clinically approved plans.
The complexity of SSRS plans and the risks associated with dose escalation near critical structures necessitates stringent QA procedures. On average, QA passing rates for 544 plans increased by 1.6% compared to the clinically approved plans. This translates to higher dose accuracy between the planned and delivered treatment due to the decrease in segment number and increase in segment area.
Not only was plan quality maintained with improved QA passing rate, actual beam delivery time decreased as well. Treatment time plays a crucial role in the proper delivery of the SSRS, due to the implications involved with administering a high dose in a single fraction. A reduction in delivery time of 11.3% for the 544 plans compared to the clinical plans was noted. Additionally, half of the 544 plans showed a decrease of delivery times greater than 17%, of which all four types of vertebra were represented in this group. This implies that vertebral body location may not be a significant factor in adversely affecting decreased treatment time. Conversely, tumor complexity may be the driving force negatively impacting delivery time. If so, SSRS cases with high complexity may not see a significant benefit in delivery time with the adjusted IMRT parameters.
Introduction
Table 1. Table of the prescriptions for each disease sites and planned MU
Spine stereotactic radiosurgery (SSRS) is a noninvasive treatment for metastatic spine lesions. Some advantages that SSRS offers over conventional external beam radiation therapy (EBRT) are an incremental benefit of palliation, better local disease control, shorter treatment course, and minimal toxicity to organs at risk (OAR).1 The complexity of SSRS plans and the risk of severe adverse effects necessitates stringent quality assurance (QA) procedures. MD Anderson Cancer Center has observed an SSRS QA failure rate that approaches 15%. The passing rate of SSRS plans can depend on the complexity of the plan, the intricacy of the target volume, and selected IMRT planning parameters. Our intent is to determine which set of IMRT optimization planning parameters will result in improved QA measurements while maintaining plan quality.
Fig. 2. Transversal views of SSRS plans centered on treatment isocenter of clinical (right) and 544 (left) plans for A-Cervical (C4-7), B-Thoracic (T-6), C-Lumbar (L2-3), and D- Sacral (S1-2) vertebral cases.
Methods and Materials
Fig. 3. DVH comparison of clinical (dashed) and 544 (solid) plans for A-Cervical (C4-7), B-Thoracic (T-6), C-Lumbar (L2-3), and D- Sacral (S1-2) vertebral cases. A C D B
In this study, 10 previously treated SSRS cases were selected with various spinal sites: cervical (1), thoracic (4), lumbar (4), and sacral (1). Using Phillips Pinnacle 9.10, we re-optimized and varied 2 IMRT planning parameters: maximum number of segments and minimum segment area, while keeping minimum MU per segment as 4. The original IMRT planning objectives and beam angles were conserved to minimize planner bias.
In the preliminary phase, we evaluated 4, 7, or 10 segments per beam along with 2, 4, or 6 cm2 minimum segment area for a total of 9 combinations on 4 cases. The created plans were compared based on target coverage, critical structure constraints, and 3D isodose distribution to find the best workable set of parameters. 5 segments per beam, 4 minimum segment area and 4 MU per segment (544) provided the best balance. We tested this set of IMRT planning parameters on 6 additional cases. All 544 plans were evaluated by a certified medical dosimetrist and medical physicists and deemed clinically acceptable in reference to the physician approved plans and planning directives. IMRT QA was performed on all clinical and 544 plans using an OCTAVIUS 4D phantom with a 2D detector array. The delivery time per beam  was manually recorded during the QA process.
For statistical analysis, two-tailed t-test for paired two sample for means was performed on the results using the Excel add-in Analysis ToolPak. The hypothesized mean difference was set to zero and alpha was selected at 0.05.
Conclusion
In conclusion, IMRT parameters of five segments per beam and four cm2 minimum segment area provided the better balance of plan quality, delivery efficiency, and plan dose calculation accuracy. We propose reducing the segments per beam and increasing minimum segment area from the conventional ten segments per beam and two cm2 minimum segment area to improve QA passing rate. A more robust study which focuses on tumor complexity and decreased segments per beam is needed in order to provide more comprehensive guidelines for IMRT planning parameter selection.
Fig. 4. Comparison of A) GTV and B) CTV coverage between clinical and 544 plans
References
P. C. Gerszten, E. Mendel, and Y. Yamada, Spine 34, S78 (2009).
I. Paddick, Journal of Neurosurgery 93, 219 (2000).
Fig. 5. Comparison of delivery time (min) between clinical and 544 plans
Fig. 4. Percentage of passing rate between clinical and 544 plans.
Contact
Gabriel Ayala
(832)