Multimodality Image Fusion of the Liver Using Structure-Guided Deformable Image Registration in Velocity AI—What Is the Preferred Approach?

Emerging artificial intelligence applications in liver magnetic resonance imaging

Chronic liver diseases (CLDs) are becoming increasingly more prevalent in modern society. The use of imaging techniques for early detection, such as magnetic resonance imaging (MRI), is crucial in reducing the impact of these diseases on healthcare systems. Artificial intelligence (AI) algorithms have been shown over the past decade to excel at image-based analysis tasks such as detection and segmentation. When applied to liver MRI, they have the potential to improve clinical decision making, and increase throughput by automating analyses. With Liver diseases becoming more prevalent in society, the need to implement these techniques to utilize liver MRI to its full potential, is paramount. In this review, we report on the current methods and applications of AI methods in liver MRI, with a focus on machine learning and deep learning methods. We assess four main themes of segmentation, classification, image synthesis and artefact detection, and their respective potential in liver MRI and the wider clinic. We provide a brief explanation of some of the algorithms used and explore the current challenges affecting the field. Though there are many hurdles to overcome in implementing AI methods in the clinic, we conclude that AI methods have the potential to positively aid healthcare professionals for years to come.

Direct visualization and correlation of liver stereotactic body radiation therapy treatment delivery accuracy with interfractional motion

This study used the visualization of hypo‐intense regions on liver‐specific MRI to directly quantify stereotactic body radiation therapy (SBRT) spatial delivery accuracy. Additionally, the interfractional motion of the liver region was determined and compared with the MRI‐based evaluation of liver SBRT spatial treatment delivery accuracy. Primovist^®‐enhanced MRI scans were acquired from 17 patients, 8–12 weeks following the completion of liver SBRT treatment. Direct visualization of radiation‐induced focal liver reaction in the form of hypo‐intensity was determined. The auto‐delineation approach was used to localize these regions, and center‐of‐mass (COM) discrepancy was quantified between the MRI hypo‐intensity and the CT‐based treatment plan. To assess the interfractional motion of the liver region, a planning CT was registered to a Cone Beam CT obtained before each treatment fraction. The interfractional motion assessed from this approach was then compared against the localized hypo‐intense MRI regions. The mean ± SD COM discrepancy was 1.4 ± 1.3 mm in the left‐right direction, 2.6 ± 1.8 mm in an anteroposterior direction, and 1.9 ± 2.6 mm in the craniocaudal direction. A high correlation was observed between interfractional motion of visualized hypo‐intensity and interfractional motion of planning treatment volume (PTV); the quantified Pearson correlation coefficient was 0.96. The lack of correlation was observed between Primovist^® MRI‐based spatial accuracy and interfractional motion of the liver, where Pearson correlation coefficients ranged from −0.01 to −0.26. The highest random and systematic errors quantified from interfractional motion were in the craniocaudal direction. This work demonstrates a novel framework for the direct evaluation of liver SBRT spatial delivery accuracy.