Table 1.

Techniques for automatic vertebral fracture assessment and for vertebral segmentation

	Refs	Description
AI based techniques
Heat Map	[6]	Heat maps are color-coded probability maps overlaid on the input image to visualize the outcome such as vertebral fracture identification and localization of a neural network. Such an image may assist the radiologist in fracture assessment because the heat map typically does not contain information on the degree of fracture.
Localization and labeling of vertebra	[7–15]	In addition to the fracture identification and localization the algorithm also assigns the vertebral label such as T8 or L4 to each fractured or all vertebrae visible in the images. Labeling is typically an initial step of fully automated fracture detection and classification algorithms. Techniques like a support vector machine (SVM), random forest classification or CNNs are used for this task.
Segmentation	[7, 12–19]	AI-based automated vertebral fracture detection algorithms often require an automatic segmentation of the vertebrae or the vertebral bodies. Resulting segmentation masks facilitate the quantification of fracture. The masks can further be used to define the volume of interest for the measurement of BMD.
Vertebral morphometry and shape analysis	[20–22]	Morphological properties and shape models facilitate the determination of fracture grade and potentially the differentiation of osteoporotic from traumatic fractures and from degenerative deformities. Compared to 6-point X-ray-based morphometry a 3D shape analysis provides additional information, but it is not clear whether this information can be used to improve fracture risk prediction.
Grading of fracture severity	[19, 20, 23]	Severe vertebral fractures are associated with a higher risk of subsequent fractures than mild vertebral fractures. Thus, the determination of fracture grade is an important aim of vertebral fracture assessment.
Combinations	[15, 19, 24, 25]	Fully automated pipelines are increasingly being developed. They efficiently combine different algorithms for performing the tasks like vertebrae localization, labeling, segmentation, and classification to enable automatic fracture detection.
Classical techniques
Segmentation	[26–31]	Classical algorithms developed to automatically segment the vertebral body or the complete vertebra. The resulting masks can be used for the same purposes as those generated with AI-based algorithms
Vertebral morphometry and shape analysis	[32, 33]	Classical algorithms to automatically determine vertebral morphometry and shape.
Grading of fracture severity	[34–36]	Morphological measures to classify vertebra fractures.