Differentiation of Cartilage Repair Techniques Using Texture Analysis from T2 Maps.

Vladimir Juras,Veronika Janáčová,Peter Angele,Alexandra Kirner,Siegfried Trattnig,Pavol Szomolanyi

doi:10.1177/19476035211029698

Abstract

ObjectiveThe aim of this study was to investigate texture features from T2 maps as a marker for distinguishing the maturation of repair tissue after 2 different cartilage repair procedures.DesignSeventy-nine patients, after either microfracture (MFX) or matrix-associated chondrocyte transplantation (MACT), were examined on a 3-T magnetic resonance (MR) scanner with morphological and quantitative (T2 mapping) MR sequences 2 years after surgery. Twenty-one texture features from a gray-level co-occurrence matrix (GLCM) were extracted. The texture feature difference between 2 repair types was assessed individually for the femoral condyle and trochlea/anterior condyle using linear regression models. The stability and reproducibility of texture features for focal cartilage were calculated using intra-observer variability and area under curve from receiver operating characteristics.ResultsThere was no statistical significance found between MFX and MACT for T2 values (P = 0.96). There was, however, found a statistical significance between MFX and MACT in femoral condyle in GLCM features autocorrelation (P < 0.001), sum of squares (P = 0.023), sum average (P = 0.005), sum variance (P = 0.0048), and sum entropy (P = 0.05); and in anterior condyle/trochlea homogeneity (P = 0.02) and dissimilarity (P < 0.001).ConclusionTexture analysis using GLCM provides a useful extension to T2 mapping for the characterization of cartilage repair tissue by increasing its sensitivity to tissue structure. Some texture features were able to distinguish between repair tissue after different cartilage repair procedures, as repair tissue texture (and hence, probably collagen organization) 24 months after MACT more closely resembled healthy cartilage than did MFX repair tissue.

Highlights

To circumvent the invasive biopsy and limitations of clinical evaluation used to monitor patients after cartilage repair surgery, a number of magnetic resonance imaging (MRI) methods have been developed and validated as sensitive approaches with which to assess cartilage repair tissue maturation
Cartilage repair tissue is characterized by a mean T2 value that can be misleading in some cases as T2 mapping is limited in ability to detect the subtle details of the repair cartilage architecture and composition.[13]
The results presented here are derived from the subgroup of patients who participated in the MRI substudy of a prospective, multicenter, randomized, controlled, open-label, phase III study comparing the efficacy and safety of matrix-associated chondrocyte transplantation (MACT) using NOVOCART 3D plus (TETEC AG, Reutlingen, Germany) versus MFX in patients with cartilage defects of the knee

Summary

Introduction

To circumvent the invasive biopsy and limitations of clinical evaluation used to monitor patients after cartilage repair surgery, a number of magnetic resonance imaging (MRI) methods have been developed and validated as sensitive approaches with which to assess cartilage repair tissue maturation. MRI methods comprise morphological evaluation, typically using semiquantitative scoring systems, such as Magnetic Resonance Observation of Cartilage Repair Tissue (MOCART)[1,2,3] and glycosaminoglycan- and collagen-specific quantitative MRI techniques.[4,5,6,7] T2 mapping, in particular, is the most often used technique for cartilage repair tissue assessment as it reflects the collagen fiber network organization in cartilage, which is represented by the formation of different zones, which is a positive sign of successful tissue maturation after cartilage repair surgery.[7,8,9] Unlike the modern glycosaminoglycan-specific techniques, such as sodium MRI and glycosaminoglycan chemical exchange saturation transfer (gagCEST), it does not require an ultra-high field MR scanner to be clinically feasible.[6,10] As T2 is sensitive to the loading applied to the cartilage, it has been successfully used for functional cartilage repair evaluation.[11,12] In a clinical setup, cartilage repair tissue is characterized by a mean T2 value that can be misleading in some cases as T2 mapping is limited in ability to detect the subtle details of the repair cartilage architecture and composition.[13]

Methods

Results

Discussion

Conclusion