Standardized maps – an emerging approach to leverage quantitative information in knee imaging

Abstract

Objective

Property maps, which capture spatial variations across the entire joint, are emerging as a powerful means for extracting and analyzing quantitative information from knee 3D imaging datasets, particularly from CT and MRI data. This perspective paper aims to discuss the processing pipelines used so far, as well as the results they have enabled with respect to osteoarthritis.

Design

The key methodological steps for obtaining property maps, including segmentation, property calculation, and standardization are presented and analysis methods are discussed. Representative studies are also examined to illustrate the state-of-the-art in this field.

Results

Three main processing pipelines have been used, with the segmentation, property calculation, and standardization steps occurring in different orders. Many methods have been successfully considered for ordering these steps, without any looking generally preferable to the others. Thanks to recent advances in segmentation and standardization techniques, routine processing of property maps appears conceivable in the near future. Maps have been analyzed for multiple purposes, including group comparisons, pattern recognition, and cross-property modelling. Mostly maps of cartilage thickness and composition, as well as maps of bone shape and mineral density have been reported. They revealed distinct patterns associated with osteoarthritis severity, achieved high diagnostic accuracy, and identified relationships among tissue properties.

Conclusions

Property maps represent a promising approach for leveraging the extensive information in imaging data. They are particularly interesting for standardizing complex spatial variations in tissue properties, enabling global analysis and modelling. Once challenging to obtain and interpret, current mapping methods are being improved to the point that property maps may well be in routine use in the near future.