25dkou498
10.3205/25dkou498
urn:nbn:de:0183-25dkou4987
Meeting Abstract
Diagnostic performance of hip MRI for the detection of labral and cartilage abnormalities: A comparative study of deep learning and standard techniques using arthroscopy as standard of reference
Twardy
Twardy
Vanessa
V
Klinik und Poliklinik für Orthopädie und Sportorthopädie des Klinikums rechts der Isar, München, Deutschland
author
Marka
Marka
Alexander
A
Institut für Diagnostische und Interventionelle Radiologie des Klinikums rechts der Isar, München, Deutschland
author
Meurer
Meurer
Felix
F
Institut für Diagnostische und Interventionelle Radiologie des Klinikums rechts der Isar, München, Deutschland
author
Graf
Graf
Markus
M
Institut für Diagnostische und Interventionelle Radiologie des Klinikums rechts der Isar, München, Deutschland
author
Weiss
Weiss
Kilian
K
Philips GmbH, Hamburg, Deutschland
author
Makowski
Makowski
Marcus
M
Institut für Diagnostische und Interventionelle Radiologie des Klinikums rechts der Isar, München, Deutschland
author
Karampinos
Karampinos
Dimitros
D
Institut für Diagnostische und Interventionelle Radiologie des Klinikums rechts der Isar, München, Deutschland
author
Neumann
Neumann
Jan
J
Institut für Diagnostische und Interventionelle Radiologie des Klinikums rechts der Isar, München, Deutschland
author
Wörtler
Wörtler
Klaus
K
Institut für Diagnostische und Interventionelle Radiologie des Klinikums rechts der Isar, München, Deutschland
author
Foreman
Foreman
Sarah
S
Institut für Diagnostische und Interventionelle Radiologie des Klinikums rechts der Isar, München, Deutschland
author
Banke
Banke
Ingo
I
Klinik und Poliklinik für Orthopädie und Sportorthopädie des Klinikums rechts der Isar, München, Deutschland
author
German Medical Science GMS Publishing House
Düsseldorf
610
20251031
engl
This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License.
Dieser Artikel ist ein Open-Access-Artikel und steht unter den Lizenzbedingungen der Creative Commons Attribution 4.0 License (Namensnennung).
M0634
498
Deutsche Gesellschaft für Orthopädie und Unfallchirurgie
Deutsche Gesellschaft für Orthopädie und Orthopädische Chirurgie
Deutsche Gesellschaft für Unfallchirurgie
Berufsverband für Orthopädie und Unfallchirurgie
Deutscher Kongress für Orthopädie und Unfallchirurgie (DKOU 2025)
Abstracts | Digitalisierung 4
Berlin
20251028
20251031
AB79-3237
TextObjectives and questions: Underestimating cartilage or labral damage in magnet resonance imaging (MRI) for femoroacetabular impingement syndrome (FAIS) can result in delayed surgical treatment, unnecessary osteoarthritis progression, and poor patient outcomes. This study aimed to assess the diagnostic performance of high-resolution artificial intelligence-driven deep learning-based MRI sequences versus standard-resolution compressed sense (CS) sequences in detecting labral and cartilage lesions, with hip arthroscopy serving as the reference gold standard.Material and methods: Thirty-two patients with FAIS underwent 3-Tesla hip MRI prior to hip arthroscopy. Coronal and sagittal intermediate-weighted TSE sequences with fat saturation were obtained using CS (0.6×0.8 mm) and high-resolution compressed senses artificial intelligence (CSAI) (0.3 × 0.4 mm) protocols, both with a 3 mm slice thickness and similar acquisition times (3:55–4:12 min). Three radiologists and a hip arthroscopy specialist assessed labral and cartilage abnormalities across five acetabular and femoral zones. Sensitivity, specificity, and accuracy were calculated, using high resolution hip arthroscopy as standard of reference. Results: Detection of labral abnormalities showed high sensitivity (97–100% for radiologists; 81% [CS]–91% [CSAI] for the surgeon) and perfect specificity (100% for all readers). Cartilage lesion detection had low sensitivity for CS (radiologists: 37–45%; surgeon: 27%) with only moderate improvement in high resolution CSAI sequences (radiologists: 45%; surgeon: 32%). Averaged for all readers, the assessment of superolateral acetabulum yielded the highest sensitivity (CS: 81%, CSAI: 88%), while the assessment of anteroinferior acetabulum showed the highest specificity (CS and CSAI: 100%). Averaged for all readers, the sensitivity was highest in assessment of the superolateral acetabulum (CS: 81%, CSAI: 88%). Sensitivity was lowest in evaluation of anteroinferior and posterior acetabular zones, and inferior and posterior femoral zones (CS and CSAI <6%).Discussion and conclusion: CS and CSAI hip MRI showed excellent performance for labral abnormalities, but detection of cartilage issues remained poor, even with next-generation high-resolution CSAI. Integrating CSAI may improve accuracy, but further technical advancements are needed for better patient treatment. In clinical routine, hip cartilage damage continues to be significantly underestimated in preoperative imaging.
0
0
0
0