Automatization of CT Annotation: Combining AI Efficiency with Expert Precision

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Medical imaging diagnostics are being transformed by artificial intelligence (AI), machine learning (ML), and deep learning (DL).
• Accurate segmentation of anatomical structures in CT scans is crucial for diagnosis and research.

Purpose of the Study:

• To present a novel methodology for semantic segmentation of the vertebral column in CT scans.
• To develop a semi-automated annotation system for efficient and accurate labeling of vertebrae.

Main Methods:

• Utilized a dataset of 250 patients from Riga East Clinical University Hospital.
• Developed a semi-automated annotation system for initial data labeling by non-experts, followed by manual validation.
• Streamlined the segmentation process by eliminating the need for volume stitching.

Main Results:

• Achieved accurate identification and labeling of individual vertebrae from C1 to the sacrum-coccyx complex.
• Significantly reduced the time and resources required for vertebral column segmentation.
• Demonstrated enhanced external validity through meticulous patient selection and demographic balancing.

Conclusions:

• The proposed AI-driven methodology represents a substantial advancement in medical data semantic segmentation.
• This approach has the potential to revolutionize clinical and research practices in radiology by improving efficiency and accuracy.
• Challenges include the need for manual validation by skilled personnel and reliance on specialized hardware.