PROJECT SUMMARY Even though cone beam computed tomography (CBCT) is the most commonly used volumetric image guidance modality, its role has been severely limited in the context of treatment monitoring and patient-specific treatment modifications in radiation therapy. Due to CBCT’s poor image quality, clinicians cannot clearly visualize soft tissues to assess anatomical changes, thus affecting their clinical decision-making. Moreover, tools for treatment monitoring, such as deformable registration and dose calculation, do not function robustly with today’s CBCT images due to the lack of CT number accuracy. Scattered radiation remains to be the fundamental problem in improving CBCT image quality. Thus, in this project, we propose the two-dimensional antiscatter grid (2D Grid) as a novel device to address the scatter problem and achieve high-quality CBCT images that are suitable for treatment monitoring. Our device has fundamentally different architecture and fabrication than existing antiscatter grids for CBCT. Due to its optimized grid structure, our 2D Grid provides both higher primary transmission and better scatter rejection performance than today’s state-of-the-art antiscatter grids. Due to its favorable primary transmission and scatter rejection performance, our 2D Grid improves the contrast-to-noise ratio and CT number accuracy to levels not achievable with existing antiscatter grids. We hypothesize that our 2D Grid will provide significantly better soft tissue visualization and CT number accuracy, and deformable registration algorithms are expected to perform significantly better. To test our hypotheses, we will develop and optimize data processing methods for 2D Grid implementation in CBCT (Aim 1). Subsequently, we will fabricate 2D Grid prototypes and evaluate their performance in clinical CBCT systems for photon and proton therapy (Aim 2). Following phantom based evaluations, we will conduct a prospective clinical trial to evaluate the clinical utility of improved image quality (Aim 3). We will perform observer studies to quantify the improvement in soft tissue visualization with respect to existing clinical CBCT and gold-standard Helical CT, assess the improvement in accuracy of deformable image registration algorithms, and evaluate the improvement in consistency of image intensity and texture features. While our application is focused on radiation therapy, the 2D Grid can play a key role in other CBCT applications, such as interventional radiology, extremity imaging, and intraoperative imaging. Due to its improved low-contrast visualization performance, our 2D Grid may also allow reduction of the imaging dose in CBCT.