In-vitro comparison of time-domain, frequency-domain and wavelet ultrasound parameters in diagnostics of cartilage degeneration.

Quantitative ultrasound imaging (QUI) is a promising preclinical method for detecting early osteoarthrotic (OA) changes in articular cartilage. The aim of this study was to compare time-domain, frequency-domain and wavelet transform (WT) QUI parameters in terms of their performance in revealing degenerative changes in cartilage in vitro. Mankin score and Cartilage Quality Index (CQI) were used as a reference for quantifying cartilage degeneration. Intact (n = 11, Mankin score = 0) and spontaneously degenerated (n = 21, Mankin score = 1-10, mean = 4) osteochondral samples (diameter 19 mm) from bovine patellae, prepared and scanned with an ultrasound instrument in our earlier study, were further analyzed. Ultrasound reflection coefficient (R), integrated reflection coefficient (IRC) and ultrasound roughness index (URI) for cartilage surfaces were obtained from our earlier study. In the present study, maximum magnitude (MM) and echo duration (ED) for the cartilage surface were determined from the WT analysis. All ultrasound (US) parameters were capable of distinguishing intact and degenerated cartilage groups (p < 0.01, Mann-Whitney U test). Significant correlations were established between all QUI parameters and CQI or Mankin score (p < 0.01, Spearman's correlation test). The receiver operating characteristic (ROC) analysis indicated that the simple time-domain parameters (R and URI) were diagnostically as sensitive and specific as the more complex frequency-domain (IRC) or WT (MM, ED) parameters. Although QUI shows significant potential for OA diagnostics, complex signal processing techniques may provide only limited additional benefits for diagnostic performance compared with simple time-domain methods. However, certain technical challenges must be met before any of these methods can be used clinically.