Systematic Biases from ARM-Based Sensitivity and Imaging in Compton Cameras

The angular resolution measure (ARM) is widely used in Compton camera analyses to characterize angular uncertainty and to define event selection criteria. However, the ARM distribution is fundamentally different from the point spread function (PSF), which represents the true spatial imaging response of the system on the celestial sphere. In this study, motivated by applications in gamma-ray astrophysics, we investigated the consequences of using ARM-based definitions in sensitivity estimation and image reconstruction. Through Monte Carlo simulations of 662 keV gamma-rays, we demonstrated that the background rejection power of ARM-based region-of-interest selections with the ARM resolution of 7.0$^{\circ}$ is equivalent to that of a PSF resolution of 33$^{\circ}$. As a result, treating the ARM resolution as a surrogate for the PSF resolution leads to significant overestimation of the signal-to-noise ratio, particularly in analyses assuming isotropic backgrounds. We further show that image reconstructions at pixel scales finer than the PSF resolution introduce substantial statistical uncertainties, reducing the effective number of independent spatial degrees of freedom. These findings underscore the necessity of PSF-based analysis for accurate sensitivity evaluation and statistically robust imaging in Compton cameras, particularly in applications where the background discrimination is critical.