Program Information
Deriving Site-Specific Recommendations for Proton Therapy Range Margins
J Schuemann*, H Paganetti , (1) Massachusetts General Hospital & Harvard Medical School, Boston, MA
Presentations
TH-AB-605-3 (Thursday, August 3, 2017) 7:30 AM - 9:30 AM Room: 605
Purpose: To define new recommendations for range margins in proton therapy based on dose calculation uncertainties for various treatment sites, sub-divided into treatment field groups.
Methods: We analyze between 500 and 1000 proton treatment fields per treatment site for 7 sites. Each site was divided grouping beams with similar arrangement and patient heterogeneity. Differences in 2 dimensional distal dose profiles between prescribed (treatment planning system) and delivered (Monte Carlo (MC) simulation) were analyzed. The root-mean-square differences were fit with a first order polynomial to determine the dependence on range. Differences were separated into under- and over-shoot, the former to ensure target coverage, the latter to estimate effects on organs at risk. The slope of the fit plus 1.5 standard deviations were combined with published uncertainties not related to dose calculations.The analysis was repeated with normalized dose distributions to estimate potential range margin reductions if the prescribed dose were scaled based on MC simulations.
Results: We found treatment sites had to be split into 6 (medulloblastoma spine and brain), 5 (head and neck), 4 (breast, lung), 2 (liver) or 1 (prostate) groups based on the anatomy. The results show that, even without MC, currently accepted range margins in proton therapy (3.5%+1mm) can be safely reduced for some sites (such as prostate) to 2.7%+1.2mm. Other sites (lung, head and neck) have greatly varying range uncertainties depending on the field angles and position that require varying margins in contrast to the generic margin used today. We also found that MC-based treatment plans would allow site-independent margins of ~2.7%+1.2mm.
Conclusion: The concept of a constant range margin for proton therapy independent of treatment site and patient geometry does not reflect our current understanding of treatment planning uncertainties. We recommend new, site-specific uncertainties. Alternatively MC-based treatment plans would allow site-independent and reduced range margins.
Funding Support, Disclosures, and Conflict of Interest: NIH grant U19 CA021239. PI: Delaney
Contact Email: