SHAISTA PARVEEN ,Dr Raafiya Gulmeher
DOI: https://doi.org/Phase-contrast magnetic resonance imaging (PC-MRI) is a fundamental technique for non-invasive cardiovascular flow assessment; nevertheless, its precision is consistently compromised by through-plane cardiac and respiratory motion, which misaligns a static imaging slice with dynamic blood and tissue. These misregistrations lead to skewed assessments of stroke volume, regurgitant percentage, and shunt flow, which impacts clinical decision-making. Although 4D Flow MRI and prospective slice-following methods help reduce these inaccuracies, they are still costly and not commonly used. This survey brings together the many different ways that retroactive through-plane motion correction has been used with regular 2D PC-MRI. These include valve tracking, displacement-aware resampling, velocity-component subtraction, feature tracking, and deep learning. We categorize methods based on their assumptions, computing demands, and common failure mechanisms, and we integrate validation practices into a unified evaluation framework that includes reproducibility measures and benchmark datasets. This review statistically summarizes the reported enhancements and constraints pertaining to valves, pathologies, and acquisition techniques, thereby emphasizing the potential of advanced feature extraction and machine learning in enhancing dependability, alongside the significant deficiencies that persist. In the end, we suggest a plan for making motion-corrected PC-MRI a regular part of cardiovascular care, which will lead to more accurate and consistent flow measures.
