3049 - Adaptive MR-Guided Stereotactic Ablative Body Radiotherapy (MRgSABR) for Pancreas Cancer: Dosimetric Evaluation from EMERALD Trial

E. Moreno-Olmedo^1,2, B. George¹, S. Teoh², R. Owens^1,2, L. Swan³, J. Good¹, T. S. Maughan^3,4, and S. Mukherjee^1,2; ¹GenesisCare, Oxford, United Kingdom, ²Oxford University Hospitals NHS Foundation Trust, Oxford, United Kingdom, ³University of Oxford, Oxford, United Kingdom, ⁴University of Liverpool, Liverpool, United Kingdom

Purpose/Objective(s): The EMERALD trial evaluated the feasibility and safety of ultra-hypofractionation with MRgSABR in pancreatic cancer. We present the dosimetric impact of adaptive MRgSABR. Materials/

Methods: A single-center three-arm phase 1 non-randomized safety study (ISRCTN10557832). Patients with localized or locally recurrent pancreatic cancer deemed suitable for MRgSABR after chemotherapy received either 50 Gy/5 (Level1; BED₁₀ 100 Gy), 39 Gy/3 (Level2; BED₁₀ 90 Gy) or 25 Gy/1 fraction (Level3; BED₁₀ 87.5 Gy). Primary endpoint was the dose limiting toxicity. Each arm was assessed as independent cohorts. Targets were: CTV = GTV + 2 mm cropped to OARs; PTV = CTV + 3 mm and PTVHigh = PTV - (OARs + 5-8 mm margin). Objective was PTVHigh V(95%) = 98%. The minimum accepted PTV V (100%) was 60% - if this criterion was not met, planning at the next lower fractionation level was attempted. The max dose PTV 1 cc was =125% - = 140%. Prior to each fraction, contours were adjusted and a reoptimized plan calculated either to meet constraints or improve coverage. Treatment was delivered in breath-holds with respiratory-gating under MR-guidance. To quantify if there was any dosimetric benefit of daily adaption, changes in PTV V(100%) and PTVHigh V (95%) between predicted and reoptimized MRgSABR plans were compared.

Results: Between August 2022 - October 2023, 20 patients underwent MRgSABR (4 Level1, 8 Level2, 8 Level3). 1 patient in Level1 did not complete the treatment course due to development of obstructive jaundice requiring stenting. 100% fractions were adapted. Of the 37 of 49 fractions where dosimetric data was available, 28/37(76%) of predicted plans failed OAR constraints. Daily adaptation allowed all mandatory constraints to be met at the expense of PTV coverage. Adaptation resulted in increase in mean PTV High V (95%) for all levels compared to predicted plan. Despite larger tumor size in Level 3, PTVhigh coverage remained >95%, although total PTV was reduced by 10% (Table).

Conclusion: In the EMERALD trial, daily adaptive MRgSABR in pancreatic cancer allowed OAR constraints preservation and optimal PTVHigh coverage, even in Level3 despite larger average tumor size.

Abstract 3049 – Table 1

Level	N	PTV V (100%) % m (range)			PTV High V (95%) % m (range)			GTV (cc) m (range)		PTV (cc) m (range)
		B	P	R	B	P	R	B	A	B	A
1	4	88.32 (80-93.54)	85.05 (73.2-93.5)	84.24 (77.5-95)	99.78 (99.5-99.6)	96.75 (83.7-99)	99.67 (97.5-99.9)	31.91 (16-54.1)	35.64 (18.5-64.5)	70.15 (41.7-117.5)	76.33 (45.2-134.2)
2	8	82.9 (66.3-89.5)	82.44 (66,3-87,5)	83.57 (60-92.3)	99.37 (96.6-99.9)	97.62 (94.8-99.3)	98.76 (97.8-99.7)	30.67 (14.6-115.3)	34.12 (18.2-120.3)	73.08 (39.7-210.5)	80.21 (70.3-205.6)
3	8	77.45 (66.8-85.2)	67.91 (61.6-86.8)	72.72 (64.9-89)	99.3 (96.8-99.9)	94.78 (80.4-99.3)	96.98 (94.1-99.7)	42.19 (8.63-79.1)	48.70 (10.7-93.6)	89.16 (21.9-154.7)	83.70 (11.5-170.8)
B= Baseline; P= Predicted; R= Reoptimised; A= Adapted; m=mean