Detection and evaluation of initial cartilage pathology in man: A comparison between MRT, arthroscopy and near-infrared spectroscopy (NIR) in their relation to initial knee pain.

BACKGROUND AND AIMS: MRI and arthroscopy are important methods in the evaluation of cartilage pathology. But frequently initial changes of cartilage in combination with chronic knee pain cannot be detected by employing these two methods. Better diagnostic tools for the detection of the early stages of osteoarthritis (OA) are required. The objective of this study was to show that near-infrared spectroscopy (NIRS) can be incorporated into routine arthroscopy to improve detection and assessment of the initial cartilage pathology. Furthermore correlations between findings in MRI, arthroscopy and NIRS in patients with initial symptoms of OA have studied.
METHODS: Patients (n=21, 12 women, 9 men, age: 15-59 years, mean 34.19 years) with knee pain lasting for at least half a year without any trauma of the knee in their history were interviewed (body weight, smoking behaviour) and clinically evaluated using the Knee Injury and Osteoarthritis Outcome Score (KOOS). Also serum parameters (cholesterol, lipids) were analysed, conventional X-rays in three directions (evaluated according to Kellgren and Lawrence) and MRI (evaluation of cartilage damage according to the ICRS-score) were performed preoperatively in all patients. During subsequent arthroscopy cartilage damage was evaluated according to the ICRS-score. In addition the spectral reflection of cartilage was investigated in all knees using a special micro-glass-fiber probe in the near-infrared light region (spectral range between 1150 and 1475nm). To characterize relations between the investigated parameters the Spearman's rank correlation coefficient was used. Inter-observer variance was calculated employing the Cohens-Kappa-test.
RESULTS: MRI demonstrated a strong inter-observer variance with no significant correlations to other parameters. The same was observed for arthroscopic findings. Only NIRS showed significant correlations with three out of five KOOS subscores. Within the general parameters only smoking behaviour showed a significant correlation with two of the KOOS-scores. NIRS therefore seemed to be a sensitive diagnostic tool in detection of initial pathology in human cartilage. The additional necessary time for the spectroscopic investigation as part of the routine arthroscopy ranged between 3 and 7min (mean: 4min 18s).
CONCLUSION: Particularly for early-stage cartilage lesions (ICRS 0/I) MRI and arthroscopy have rather low predictive value. The inter-observer variance is very high (Cohens-Kappa<0.4). Correlations found between NIRS and KOOS suggest that NIRS potentially can be used for detection of initial cartilage pathology and may be helpful in the evaluation of the benefit of different medical or surgical interventions at early-stage of articular cartilage damage.