External Beam Radiation Therapy (EBRT) remains a mainstay in cancer treatment. Accurate patient positioning
and beam alignment are essential for optimal therapeutic outcomes. However, ensuring the correct patient
positioning can be challenging during radiation treatment sessions. This study focuses on the development and
validation of a kV X-ray based rectilinear geometry digital tomosynthesis (DTS) system tailored for the LINAC
system within EBRT. Additionally, we emphasize the enhancement of the DTS image reconstruction speed via
GPU-based CUDA programming and its potential implications for real-time patient alignment and monitoring.
Through modeling the rectilinear geometry DTS system, we obtained the projection images crucial for
reconstruction. The acquired images were reconstructed using various algorithms. The impact of GPU-based
CUDA programming on speeding up the reconstruction process was also examined. As a result, acceleration
factors of up to 30 % times were achieved compared to a single thread CPU implementation. This study underscores
the potential of GPU-Accelerated image reconstruction using the distance-driven method in DTS with
rectilinear geometry for the LINAC System. By applying this technique to the EBRT system, it is feasible to
position patients using reconstructed tomographic images in real-time using kV tomography images, even in
the absence of an On-board imaging (OBI).