TWO ORIGINAL ARTICLES PUBLISHED ON AUTOMATED ANAYLSIS OF CARTILAGE T2 (OA-BIO)

We are proud to announce publication of our two first full original papers generated from OA-BIO, a €4.2 Million Eurostars grant awarded to a consortium of 4 partners: 4Moving Biotech, AO Research Institute Davos (ARI), Utrecht University, and Chondrometrics GmbH. The four-year project aims at developing life-changing therapy for osteoarthritis, using a biomarker-driven approach

The two papers published have just appeared in Skeletal Radiology and in Quantitative Imaging in Medicine and Surgery, two well-established journals in the field of Musculoskeletal Imaging:

1. Evaluation of an automated laminar cartilage T2 relaxation time analysis method in an early osteoarthritis model by Wolfgang Wirth et al.
   Skeletal Radiol 2024 Sep 4. Online ahead of print.
2. Clinical validation of fully automated laminar knee cartilage transverse relaxation time (T2) analysis in anterior cruciate ligament (ACL)-injured knees, by Felix Eckstein et al.
   Quantitative Imaging in Medicine and Surgery (QIMS) 2024: 14: 4319-4332.

These make use of two (human) models of early osteoarthritis that were already outlined in our 2020 “Year in review” (imaging) article in Osteoarthritis & Cartilage

The first publication relies on target knees withOUT radiographic osteoarthritis, but with contra-lateral knee radiographic joint space narrowing (JSN); the second one on patients with recent (anterior) cruciate ligament (ACL) injury. The overarching purpose was to technically and clinically validate an automated approach of cartilage MRI transverserelaxation time (T2) analysis.

In the two above papers we demonstrate that:

1. Cartilage segmentation by deep-learning (CNN- and U net-) based methodology applied to multiple echo spin echo (MESE) MRIs shows high agreement (DSCs) with expert reader segmentation
2. Automated analysis of T2 in the deep and superficialcartilage layer is accurate when compared with that derived from manual segmentations
3. Sensitivity to clinically important T2 differences or longitudinal T2 changes is similar between the automated vs. the traditional (expert manual segmentation) approach.

The automated analysis method is currently extended to its application in a quantitative dual echo steady state (qDESS sequence), a more efficient, high-resolution and rapid-acquisition 3D-MRI method for evaluating cartilage T2 (Paper). This ongoing work was recently presented at the International Workshop of Osteoarthritis Imaging (IWOAI) in Marrakech, Morocco, and it is currently submitted for full publication as an original paper.

We thank our co-authors: Susanne Maschek, Anna Wisser, Jana Eder, Christian Baumgartner, Akshay Chaudhari, Francis Berenbaum, Nicholas Brisson, and Georg Duda, and we thank the Chondrometrics expert readers for diligent cartilage segmentation: Jana Daimer, Gudrun Goldmann, Sabine Mühlsimer, Annette Thebis, and Barbara Wehr.

The MRI image analysis in the above studies was funded through OA BIO (E! 114932; Eurostars 2/ Horizon 2020) that supports early clinical development of 4P004, a putative first-in-class disease modifying osteoarthritis drug (DMOAD), as well as validation of osteoarthritis biomarkers required in its clinical development.