Image analysis for radiomics: facts and challenges

This article sheds light on methods used for extracting high-dimensional data from clinical images – radiomics. The authors explain the standard steps of the radiomics process, discuss the origin of the term, and summarise the major issues in regard to this process....

DTI in children is influenced by pre-processing and tract selection

This study compares fractional anisotropy (FA) values derived with DTI analysis approaches, as well as FA values and number of tracts with the clinical motor outcome quantified by the functional independence measure for children (WeeFIM). The authors concluded that...

MRI automated assessment of muscle fat composition

Proton-density fat fraction of the paraspinal muscles has become an important surrogate biomarker in patients with intervertebral disc disease, osteoporosis, sarcopenia, and neuromuscular disorders. The authors developed an automatic segmentation algorithm of the...

Emphysema quantification by CT

To determine the influence of dose reduction and reconstruction methods on CT emphysema quantification, CT scans were performed on humans at routine radiation dose and at 45%, 60%, and 75% reduced radiation dose. Images were then reconstructed using filtered back...

Generative models: an innovation in musculoskeletal radiology?

New innovative technology, deep learning, is revolutionising many research and industrial fields. Growing interest in generative models inspired the authors to explore their potential in the field of magnetic resonance imaging of the spine. Tests showed that the...

Artificial intelligence (AI) in medical imaging: threat or opportunity?

This article deals with the potentials and consequences of integrating AI into radiology. The authors define basic terms such as “machine/deep learning” and analyse the use of AI in recent publications. Article: Artificial intelligence in medical imaging: threat or...