(Sunday, 7/29/2018) 5:05 PM - 6:00 PM
Room: Room 209
Purpose: Patient-specific pre-treatment quality assurance (QA) is performed for most intensity-modulated plans to verify their clinical deliverability. However, a limited analysis on a 3%3mm gamma-analysis criteria is the most frequently used ground truth for QA irrespective of treatment site and clinical appropriateness. Furthermore, plan deliverability is usually determined at the last stage of plan-checking leaving little scope for adjustment. We introduce SMART (Statistical Measures and Analysis of quality assurance measurements for optimal Radiation Treatment) which for each treatment plans will provide an estimated gamma at the time of plan development based on accumulated data and machine learning.
Methods: This study consists of two steps. The first-step is aggregate planning and QA data to extract plan properties from ARIA database and corresponding QA results with a range of gamma-analysis (4%2mm,3%2mm,2%3mm and 2%2mm). The second-step is data analysis which summarises the aggregate data to find the correlation between gamma criterion and radiotherapy plan parameters, and predict the output of gamma-analysis for each plan prior to its physical QA. The predicted gamma output is then used as a decision-making support tool (SMART QA) for whether the current plan is acceptable or a replan is required.
Results: We aggregated 2390 QA datasets (1605 IMRT and 785 VMAT) together with ARIA data. With our current clinical criteria of 3%3mm 3D gamma-analysis, we found an average of 97.9% for IMRT and 98.5% for VMAT. Across five linear accelerators, the mean of gamma-passing rates was measured at 99.8%(175),97.3%(538),98.2%(379),97.5%(403), and 99.3%(110) for IMRT and 98.2%(65),97.8%(225),98.3%(122),98.4%(103) and 99.1%(265) for VMAT. We are presently undertaking additional gamma criteria analysis on the QA data.
Conclusion: We introduce SMART that automatically extracts ARIA data to correlate QA results with plan-specific information and can predict appropriate gamma analysis for plan-quality assessment at the time of treatment plan development that reflects clinical relevance.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
