Room: AAPM ePoster Library
Soft-tissue targets in the abdomen undergo significant inter-fractional changes requiring plan adaptation. We evaluate benefits of MR guided Adaptive RT (MRgART) for pancreas patients.
The adaptive workflow begins with a baseline plan on sim-MR. For each treatment, a new MR is acquired and base-plan recomputed. If OARs exceed dose threshold, plan is re-optimized, validated and delivered (adaptive RT); otherwise, treatment proceeds unchanged (non-adaptive). Using retrospective treatment data from 50 fractions (prescription dose 40Gy/5 fractions), we evaluated inter-fractional position changes in OARs via center-of-mass (COM) shifts and overlap-indices (OI). Next, three treatment approaches were evaluated on a planning-system capable of adaptive RT using five sets of CT data-sets, P1: baseline plan assuming "static" sim-anatomy for the entire treatment course, P2: baseline plan recomputed on daily imaging but not adapted; and P3: plan adapted on daily imaging for each treatment fraction. Average OAR doses (Dmean, Dmax) were compared for stomach, duodenum, and both bowels (SB & LB) via student t-test.
Significant inter-fractional OAR changes may occur near pancreas due to respiration, cardiac motion or peristalsis. On daily imaging, average COM shifts for stomach, duodenum, SB & LB were [mean (s.d.)] = 14(9), 5(3), 21(23) & 37(29) mm respectively. The corresponding OI were 0.83(0.09), 0.75(0.03), 0.55(0.21), & 0.67(0.25) respectively. Without plan adaptation, these shifts (P2 vs. P1) would result in unacceptable plans with large increases in OAR doses (average % increase in Dmean, Dmax): stomach (24.8%, 13.2%), duodenum (12%, 17.9%), SB (37.4%, 7.7%), and LB (39.1%, 8.3%) (p < 0.01). With adaptive RT (P3), all OAR doses were below threshold. In our patient cohort, adapted treatment fractions were characterized by large inter-fractional changes in OARs with both bowels showing the greatest variation.
MRgART is effective in assessing and mitigating dose impact of daily anatomical changes.