A vertebral shape-based classification for labeling the thoracolumbar junction on CT imaging.
retrospective_cohort · Level III
Where this comes from
- Record sourced from PubMed, PMID 41003722.
- Also identified by DOI 10.1007/s00586-025-09364-0.
- No licence information is recorded for this record.
- Because redistribution is not established, this page shows the abstract only. Follow the links below for the full text.
Abstract
The correct identification of the thoracolumbar junction is mandatory for labeling of the lumbar spine, particularly if visualized incompletely. The current state-of-the-art relies on a continuous counting-based (CCBC), a rib-length-based (RLBC), or rib-based classification (RBC) of the spine enumeration. However, the CCBC and RBC often lead to inconsistent labels depending on the visualized spine segments. In this retrospective study, three classifications for assessing the thoracolumbar junction were evaluated in CT scans from 2005 to 2018, covering the complete thoracolumbar spine: (1) CCBC, (2) RLBC/RBC, and (3) a vertebra-shape-based classification (VSBC). The classifications were compared regarding consistent labels depending on the visualized spine segments, agreement with nerve morphology, and inter-rater reliability. CT-imaging data from 1,242 subjects (mean age, 63 years ± 12; 771 women) were included. The VSBC led to more consistent labels than all other classifications. Only the VSBC perfectly complied with the nerve morphology (VSBC, 100%; RBC, 97.2%, RLBC, 96.3%; CCBC, 92.6%). The VSBC had slightly superior inter-rater reliability than the RBC (0.61 vs. 0.59). The VSBC performs more reliably and consistently than other classifications in labeling the thoracolumbar junction, even for anatomical variants or incompletely visualized spines. Compared with rib-based classifications, the VSBC demonstrated modest but clear advantages in consistency and agreement with nerve morphology.
Anatomy
- lumbar spine