Publications
Below are peer-reviewed publications detailing the scientific underpinnings of SliceVault’s AI-based innovations. Each article highlights key features such as quality control, image integrity, and organ segmentation in clinical trial imaging.
Published Papers
1. Fingerprint: An AI‑based method for detection of mislabeled CT studies in clinical trials
Edenbrandt L et al., Journal of Radiology and Clinical Imaging 7 (2023): 12–15.
Introduces a novel algorithm that uses anatomical “fingerprinting” of bones (hips and scapulae) to detect mismatches between paired CT scans, achieving 100 % accuracy in test datasets.
Download the paper (PDF)
2. AI‑Based Image Quality Assessment in CT
Edenbrandt L et al., Archives of Clinical and Biomedical Research 6 (2022): 869–872.
Describes an AI model trained to assess CT image features—such as anatomical coverage (head, chest, abdomen, pelvis), presence of hip prostheses, and indicator of IV/oral contrast—with high accuracy (98.4–100 % for body region detection; 89.6 % IV contrast; 82.4 % oral contrast).
Download the paper (PDF)
3. Organ Finder – A New AI‑based Organ Segmentation Tool for CT
Edenbrandt L et al., Journal of Radiology and Clinical Imaging 5 (2022): 65–70.
Presents Organ Finder 2.0, a deep learning tool trained on a diverse CT dataset (contrast and non-contrast) and manually annotated organs. It achieved a high average Dice coefficient of 0.93 across 22 organs, demonstrating reliable segmentation performance.
Download the paper (PDF)
Why It Matters
Quality & Integrity: These publications underscore SliceVault’s AI capabilities—ranging from labeling validation to image quality control and accurate anatomical segmentation. Clinical Compliance: Designed with trial workflows in mind, these tools support data reliability, regulatory standards, and streamlined operations. Accessible Evidence: Full-text PDFs are provided for internal review, external validation, or learning purposes.