This review explores modern image registration techniques in the context of managing respiratory-induced motion in the thoracic region. The respiratory cycle introduces anatomic variability that poses challenges for radiation therapy, breathing dynamics modeling, and diverse research applications. Conventional and advanced motion management strategies are presented, emphasizing the critical role of robust image registration. Emerging approaches such as multimodal and interpatient registration are discussed, alongside metrics for assessing registration quality. This review aims to enhance understanding of recent developments in image registration and thoracic motion management. Continued progress in modeling respiratory dynamics is essential to support advanced research applications and clinical innovation. Keywords: Thorax, Deep Learning, Machine Learning, Radiation Therapy Supplemental material is available for this article. © RSNA, 2025.
Keywords: Deep Learning; Machine Learning; Radiation Therapy; Thorax.