3068 - Collimator and Detector Optimization for Compton Scatter Detection: A First Step Towards Scatter Imaging for Real-Time Motion Management

Purpose/Objective(s):

During lung SBRT, tumor motion is managed using 4DCT-based ITV expansion, respiratory or external gating, or breath-hold techniques, each with limitations. Existing real-time tracking methods require implanted fiducials or additional imaging dose. A non-invasive, dose-free method for real-time tumor tracking is warranted. Passive detection of Compton-scattered, therapeutic photons offers a novel solution. A comprehensive Monte Carlo (MC) framework was utilized to optimize the collimator and detector towards motion management with Compton scatter detection.

Materials/Methods:

The validated GATE MC toolkit, which allows accurate scintillation system modelling was utilized. Phase-space analyses first determined the optimal angular acceptance range (AAR) of scattered photons that preserved geometry of a layered cuboid lung tumor phantom, selected for its discernability in imaging development, for a clinical 6 MV FFF beam. The detector position, scintillation composition, and collimator properties: septal height (H), thickness (T), interspace (I), and material of parallel-focused 2D anti-scatter grids (ASGs) were varied to accept scattered photons within the determined optimal AAR. The angular detection significance metric (ADSM), defined as the number of detected scattered photons within the optimal AAR divided by the square root of the total detected photons, was the performance metric. The highest-ADSM configuration was used to simulate imaging at clinically realistic frame rates and doses for contrast-to-noise (CNR) assessment.

Results:

The phantom image was geometrically accurate when the scattered photons AAR was restricted to ±0.5° about the detector position (e.g., 44.5–45.5° for a detector at 45°). The highest ADSM value (54.1, Table 1), corresponding to the optimal design, was achieved using a tungsten ASG with H, T, and I of 75, 0.15, and 1.0 mm, respectively, positioned at 45° relative to the treatment head (0°) coupled to a 3.5 mm thick lead tungstate (PbWO4) scintillation crystal. Preliminary image formation demonstrated promising CNR for potential tumor tracking.

Conclusion:

MC modelling demonstrated that optimizing the detector and ASG design can selectively accept scattered photons, preserve geometric fidelity, and produce images toward Compton scatter–based motion management for lung SBRT.

Table 1: A sample of the evaluated detector and ASG configurations, showing the best- and worst-performing cases based on ADSMs.

Detector pos.	Scin. crystal	Scin. crystal thickness [mm]	ASG material	ASG shape	H [mm]	T [mm]	I [mm]	ADSM
90°	Gd2O2S:Tb	1.0	Lead	2D Square	57	0.05	1.0	9.97
90°	Gd2O2S:Tb	1.0	Tung	2D Square	57	0.05	1.0	12.2
45°	Gd2O2S:Tb	1.0	Tung	2D Cylinder	57	0.1	1.0	33.1
45°	CsI:Tl	1.0	Tung	2D Hexagon	57	0.1	1.0	27.5
45°	PbWO4	5.0	Tung	2D Hexagon	57	0.1	1.0	49.8
45°	PbWO4	3.5	Tung	2D Hexagon	30	0.1	0.52	41.2
45°	PbWO4	3.5	Tung	2D Hexagon	75	0.15	1.0	54.1