Machine learning classification on texture analyzed T2 maps of osteoarthritic cartilage: oulu knee osteoarthritis study.

OBJECTIVE: To introduce local binary pattern (LBP) texture analysis to cartilage osteoarthritis (OA) research and compare the performance of different classification systems in discrimination of OA subjects from healthy controls using gray-level co-occurrence matrix (GLCM) and LBP texture data. Classification algorithms were used to reduce the dimensionality of texture data into a likelihood of subject belonging to the reference class. METHOD: T2 relaxation time mapping with multi-slice multi-echo spin echo sequence was performed for eighty symptomatic OA patients and 63 asymptomatic controls on a 3T clinical MRI scanner. Relaxation time maps were subjected to GLCM and LBP texture analysis, and classification algorithms were deployed with an in-house developed software. Implemented algorithms were K nearest neighbors, support vector machine, and neural network classifier. RESULTS: LBP and GLCM discerned OA patients from controls with a significant difference in all studied regions. Classification models comprising GLCM and LBP showed high accuracy in classing OA patients and controls. The best performance was obtained with a multilayer perceptron type classifier with an overall accuracy of 90.2 %. CONCLUSION: LBP texture analysis complements prior results with GLCM, and together LBP and GLCM serve as significant input data for classification algorithms trained for OA assessment. Presented algorithms are adaptable to versatile OA evaluations also for future gradational or predictive approaches.

Defective WNT signaling may protect from articular cartilage deterioration - a quantitative MRI study on subjects with a heterozygous WNT1 mutation.

OBJECTIVE: WNT signaling is of key importance in chondrogenesis and defective WNT signaling may contribute to the pathogenesis of osteoarthritis and other cartilage diseases. Biochemical composition of articular cartilage in patients with aberrant WNT signaling has not been studied. Our objective was to assess the knee articular cartilage in WNT1 mutation-positive individuals using a 3.0T MRI unit to measure cartilage thickness, relaxation times, and texture features. DESIGN: Cohort comprised mutation-positive (N = 13; age 17-76 years) and mutation-negative (N = 13; 16-77 years) subjects from two Finnish families with autosomal dominant WNT1 osteoporosis due to a heterozygous missense mutation c.652T>G (p.C218G) in WNT1. All subjects were imaged with a 3.0T MRI unit and assessed for cartilage thickness, T2 and T1ρ relaxation times, and T2 texture features contrast, dissimilarity and homogeneity of T2 relaxation time maps in six regions of interest (ROIs) in the tibiofemoral cartilage. RESULTS: All three texture features showed opposing trends with age between the groups in the medial tibiofemoral cartilage (P = 0.020-0.085 for the difference of the regression coefficients), the mutation-positive individuals showing signs of cartilage preservation. No significant differences were observed in the lateral tibiofemoral cartilage. Cartilage thickness and means of T2 relaxation time did not differ between groups. Means of T1ρ relaxation time were significantly different in one ROI but the regression analysis displayed no differences. CONCLUSIONS: Our results show less age-related cartilage deterioration in the WNT1 mutation-positive than the mutation-negative subjects. This suggests, that the WNT1 mutation may alter cartilage turnover and even have a potential cartilage-preserving effect.

Assessment of meniscus with adiabatic T1ρ and T2ρ relaxation time in asymptomatic subjects and patients with mild osteoarthritis: a feasibility study.

OBJECTIVE: To investigate the ability of magnetic resonance imaging (MRI) adiabatic relaxation times in the rotating frame (adiabatic T1ρ and T2ρ) to detect structural alterations in meniscus tissue of mild OA patients and asymptomatic volunteers. METHOD: MR images of 24 subjects (age range: 50-67 years, 12 male), including 12 patients with mild osteoarthritis (OA) (Kellgren-Lawrence (KL) = 1, 2) and 12 asymptomatic volunteers, were acquired using a 3 T clinical MRI system. Morphological assessment was performed using semiquantitative MRI OA Knee Score (MOAKS). Adiabatic T1ρ and T2ρ (AdT1ρ, AdT2ρ) relaxation time maps were calculated in regions of interest (ROIs) containing medial and lateral horns of menisci. The median relaxation time values of the ROIs were compared between subjects classified based on radiographic findings and MOAKS evaluations. RESULTS: MOAKS assessment of patients and volunteers indicated the presence of meniscal and cartilage lesions in both groups. For the combined cohort group, prolonged AdT1ρ was observed in the posterior horn of the medial meniscus (PHMED) in subjects with MOAKS meniscal tear (P < 0.05). AdT2ρ was statistically significantly longer in PHMED of subjects with MOAKS full-thickness cartilage loss (P < 0.05). After adjusting for multiple comparisons, differences in medians of observed AdT1ρ and AdT2ρ values between mild OA patients and asymptomatic volunteers did not reach statistical significance. CONCLUSION: AdT1ρ and AdT2ρ measurements have the potential to identify changes in structural composition of meniscus tissue associated with meniscal tear and cartilage loss in a cohort group of mild OA patients and asymptomatic volunteers.

Efficacy of progressive aquatic resistance training for tibiofemoral cartilage in postmenopausal women with mild knee osteoarthritis: a randomised controlled trial.

OBJECTIVE: To study the efficacy of aquatic resistance training on biochemical composition of tibiofemoral cartilage in postmenopausal women with mild knee osteoarthritis (OA). DESIGN: Eighty seven volunteer postmenopausal women, aged 60-68 years, with mild knee OA (Kellgren-Lawrence grades I/II and knee pain) were recruited and randomly assigned to an intervention (n = 43) and control (n = 44) group. The intervention group participated in 48 supervised aquatic resistance training sessions over 16 weeks while the control group maintained usual level of physical activity. The biochemical composition of the medial and lateral tibiofemoral cartilage was estimated using single-slice transverse relaxation time (T2) mapping and delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC index). Secondary outcomes were cardiorespiratory fitness, isometric knee extension and flexion force and knee injury and OA outcome (KOOS) questionnaire. RESULTS: After 4-months aquatic training, there was a significant decrease in both T2 -1.2 ms (95% confidence interval (CI): -2.3 to -0.1, P = 0.021) and dGEMRIC index -23 ms (-43 to -3, P = 0.016) in the training group compared to controls in the full thickness posterior region of interest (ROI) of the medial femoral cartilage. Cardiorespiratory fitness significantly improved in the intervention group by 9.8% (P = 0.010). CONCLUSIONS: Our results suggest that, in postmenopausal women with mild knee OA, the integrity of the collagen-interstitial water environment (T2) of the tibiofemoral cartilage may be responsive to low shear and compressive forces during aquatic resistance training. More research is required to understand the exact nature of acute responses in dGEMRIC index to this type of loading. Further, aquatic resistance training improves cardiorespiratory fitness. TRIAL REGISTRATION NUMBER: ISRCTN65346593.