Performance and clinical applicability of machine learning in liver computed tomography imaging: a systematic review - Takeaways - MDSpire

Performance and clinical applicability of machine learning in liver computed tomography imaging: a systematic review

  • By

  • Keyur Radiya

  • Henrik Lykke Joakimsen

  • Karl Øyvind Mikalsen

  • Eirik Kjus Aahlin

  • Rolv-Ole Lindsetmo

  • Kim Erlend Mortensen

  • May 12, 2023

  • 0 min

Share

  • 1

    Machine learning (ML) is increasingly utilized in liver computed tomography (CT) imaging, showing high performance in various diagnostic tasks.

  • 2

    A systematic review identified 191 studies focusing on ML applications in liver imaging, highlighting the need for comprehensive evaluation of their clinical applicability.

  • 3

    Liver segmentation was the primary aim in 84 studies, demonstrating significant advancements in ML techniques, particularly deep learning.

  • 4

    The quality of studies improved with external validation methods, yet only 11 studies compared ML results directly with human experts.

  • 5

    Models showed strong performance in segmenting larger lesions, but challenges remain for accurately segmenting lesions smaller than 1 cm.

Original Source(s)

Performance and clinical applicability of machine learning in liver computed tomography imaging: a systematic review

  • by Keyur Radiya, Henrik Lykke Joakimsen, Karl Øyvind Mikalsen, Eirik Kjus Aahlin, Rolv-Ole Lindsetmo, Kim Erlend Mortensen

  • May 12, 2023

Related Content

Bridging radiology and pathology: domain-generalized cross-modal learning for clinical applications

Suspended Lead Suits Linked to Lower Radiation in LAAO

Study shows suspended lead suits reduce head-level radiation exposure during left atrial appendage occlusion procedures

Diagnostic accuracy of deep learning vs. human raters for detecting osteoporotic vertebral compression fractures in routine CT scans