Adult Spinal Deformity (ASD) can lead to motor and sensory impairments, resulting in reduced Health-Related Quality of Life (HRQOL). Surgical decisions are typically based on spinal alignment parameters extracted from spinal X-ray images. However, the correlation between these parameters and HRQOL is often limited, and many patients experience insufficient improvement in HRQOL after surgery. Since HRQOL is a subjective, questionnaire-based measure, it is difficult to manually identify physical characteristics that are closely related to it. In this study, we constructed deep learning models to estimate HRQOL from spinal X-ray images and analyzed the trained models to identify anatomical regions potentially associated with HRQOL. We built multiple models using EfficientNetB0 and YOLO11, and evaluated them using data collected from multiple medical institutions. This work represents an initial step toward discovering physical features related to HRQOL, which may contribute to the development of more effective surgical decision-making criteria through future clinical investigations.