Body composition analysis by radiological imaging – methods, applications, and prospectsRadiologische Bestimmung der Gewebezusammensetzung im menschlichen Körper (Body Composition) – Methoden, Anwendungen und Aussichten

Nicolas Linder,Timm Denecke,Harald Busse

DOI: https://doi.org/10.1055/a-2263-1501

2024-04-04

RöFo - Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren

Abstract:Background This review discusses the quantitative assessment of tissue composition in the human body (body composition, BC) using radiological methods. Such analyses are gaining importance, in particular, for oncological and metabolic problems. The aim is to present the different methods and definitions in this field to a radiological readership in order to facilitate application and dissemination of BC methods. The main focus is on radiological cross-sectional imaging. Methods The review is based on a recent literature search in the US National Library of Medicine catalog (pubmed.gov) using appropriate search terms (body composition, obesity, sarcopenia, osteopenia in conjunction with imaging and radiology, respectively), as well as our own work and experience, particularly with MRI- and CT-based analyses of abdominal fat compartments and muscle groups. Results and Conclusion Key post-processing methods such as segmentation of tomographic datasets are now well established and used in numerous clinical disciplines, including bariatric surgery. Validated reference values are required for a reliable assessment of radiological measures, such as fatty liver or muscle. Artificial intelligence approaches (deep learning) already enable the automated segmentation of different tissues and compartments so that the extensive datasets can be processed in a time-efficient manner – in the case of so-called opportunistic screening, even retrospectively from diagnostic examinations. The availability of analysis tools and suitable datasets for AI training is considered a limitation. Key Points Citation Format Hintergrund Die vorliegende Arbeit stellt die quantitative Erfassung der Gewebezusammensetzung im menschlichen Körper (Body Composition) mit den Mitteln der Radiologie vor. Derartige Analysen gewinnen vor allem bei onkologischen und metabolischen Fragestellungen an Bedeutung. Zielsetzung ist es, einer radiologischen Leserschaft die unterschiedlichen Methoden und Definitionen auf diesem Gebiet vorzustellen, um deren Anwendung und Verbreitung zu erleichtern. Das Hauptaugenmerk gilt dabei der radiologischen Schnittbildgebung. Methoden Die Übersicht stützt sich auf eine aktuelle Literaturrecherche im Katalog der US-amerikanischen National Library of Medicine (pubmed.gov) mit entsprechenden Suchbegriffen (body composition, obesity, sarcopenia, osteopenia in Verbindung mit imaging bzw. radiology), sowie auf eigene Arbeiten und Erfahrungen, insbesondere mit der MRT- und CT-gestützten Analyse abdomineller Fettkompartimente und Muskelgruppen. Ergebnisse und Schlussfolgerung Zentrale Nachverarbeitungsmethoden wie die Segmentierung von tomografischen Datensätzen sind inzwischen gut etabliert und finden in zahlreichen klinischen Studien Anwendung, u. a. in der Adipositas-Chirurgie. Für die verlässliche Beurteilung der radiologischen Messgrößen, z. B. einer Verfettung von Leber oder Muskulatur, sind validierte Referenzwerte erforderlich. Ansätze der Künstlichen Intelligenz (Deep Learning) ermöglichen bereits heute die automatisierte Segmentierung unterschiedlicher Gewebe und Kompartimente, damit die umfänglichen Datensätze zeiteffizient bearbeitet werden können – beim sogenannten opportunistischen Screening sogar retrospektiv aus diagnostischen Untersuchungen. Als Limitation gilt die Verfügbarkeit von Analyse-Werkzeugen sowie geeigneter Datensätze für das KI-Training. Kernaussagen Zitierweise

radiology, nuclear medicine & medical imaging

What problem does this paper attempt to address?

The paper attempts to address the issue of quantitatively assessing body composition (BC) through radiological imaging methods, particularly in the context of oncology and metabolic diseases. Specifically, the paper aims to introduce radiology readers to different BC analysis methods and definitions to promote the application and dissemination of these methods. The main focus is on tomographic imaging techniques such as CT and MRI, which can provide spatial distribution information of tissues and are widely used in multiple clinical disciplines.

### Main Issues:
1. **Quantitative Assessment of Body Composition**: How to use radiological imaging methods (such as CT and MRI) to quantify tissue components like fat, muscle, and bone in the human body.
2. **Clinical Applications**: The specific applications of these methods in fields such as oncology and metabolic diseases (e.g., obesity, sarcopenia, osteopenia, and osteoporosis).
3. **Technological Advances**: How to use artificial intelligence (especially deep learning) to automate the processing of large amounts of imaging data and improve analysis efficiency.
4. **Standardization and Reference Values**: How to establish reliable reference values to more accurately assess radiological measurement results, such as fatty liver or muscle mass.
5. **Future Prospects**: Exploring future research directions and application prospects, including applications in large-scale cohort studies and promotion in clinical routine.

### Background and Methods:
- **Background**: Quantitative assessment of body composition is becoming increasingly important in oncology and metabolic diseases. Imaging techniques such as CT and MRI provide good tissue visualization and spatial distribution information.
- **Methods**: The paper is based on a literature search on PubMed over the past 5 years, using keywords related to "body composition," "obesity," "sarcopenia," "osteopenia," etc., combined with the authors' own work and experience, particularly regarding MRI and CT analysis of abdominal fat and muscle groups.

### Results and Conclusions:
- **Key Post-Processing Methods**: Segmentation of tomographic datasets has been widely applied in multiple clinical disciplines, including bariatric surgery. Validated reference values are needed for reliable assessment of radiological measurement results.
- **Application of Artificial Intelligence**: Deep learning methods have been able to achieve automated segmentation of different tissues and compartments, allowing large datasets to be processed efficiently and even enabling retrospective analysis in diagnostic examinations.
- **Limitations**: The availability of analysis tools and datasets suitable for AI training is a limiting factor.

### Key Points:
- Radiological imaging methods are increasingly used to determine body composition (BC).
- BC parameters are typically quantitative and reproducible.
- CT image data from routine clinical examinations can be used for retrospective BC analysis.
- In the future, MRI examinations can be used to determine organ-specific BC parameters.
- Automated and in-depth analysis methods (such as deep learning or radiomics) will become increasingly important in the future.