Room: Karl Dean Ballroom B1
Purpose: To quantify the sensitivity of beam gating performance in MRI-guided radiotherapy to the effects of temporal resolution and latency in terms of confining the treatment target within the PTV.
Methods: Planar, intra-fractional MR images acquired during patient treatments with associated tracked target contours were collected for 20 liver and 20 pancreas patients treated on an MRI-guided RT system. Gating signals were retrospectively modified to adhere to a standardized criterion. GTVs were mapped onto the imaging plane for all fractions and conformal in-plane margins between 0-20 mm were applied for defining combinations of gating boundaries and in-plane PTVs. For every margin pair, worst-case target excursions beyond the PTV were estimated given the effects of finite temporal resolution and latency. For each treatment fraction, beam changes-of-state from beam on to beam off were identified and excursions were evaluated on images acquired within intervals immediately following these instances.
Results: When applying our most commonly used clinical conditions of 5 mm PTV with 3 mm gating boundary, for 80% of liver (pancreas) patients a finite temporal resolution causes no more than 10% (10%) of the in-plane GTV to move beyond the PTV at any instant. With 500 ms latency, the maximum target area beyond the PTV increases to greater than 40% of the in-plane GTV for liver and pancreas patients. Relationships between time spent beyond the PTV – expressed as a fraction of beam-on time – and treatment efficiency (a function of gating boundary) were derived.
Conclusion: Target motion beyond the PTV due to the effects of latency and temporal resolution was evaluated using clinical data. Heuristic metrics, such as maximum target area beyond the PTV, highlight the importance of carefully choosing gating and PTV margins. Clinically-implemented margin combinations were demonstrated to be reasonable even in the presence of worst-case estimates for uncertainties.
Not Applicable / None Entered.
Not Applicable / None Entered.