In Vivo Quantitative Ultrasound Image Analysis of Femoral Subchondral Bone in Knee Osteoarthritis

A potential of quantitative noninvasive knee ultrasonography (US) for detecting changes in femoral subchondral bone related to knee osteoarthritis (OA) was investigated. Thirty-nine patients referred to a knee arthroscopy underwent dynamic noninvasive US examination of the knee joint. The subchondral bone was semiautomatically segmented from representative US images of femoral medial and lateral condyles and intercondylar notch area. Subsequently, the normalized mean gray-level intensity profile, starting from the cartilage-bone interface and extending to the subchondral bone depth of ~1.7 mm, was calculated. The obtained profile was divided into 5 depth levels and the mean of each level, as well as the slope of the profile within the first two levels, was calculated. The US quantitative data were compared with the arthroscopic Noyes' grading and radiographic Kellgren-Lawrence (K-L) grading. Qualitatively, an increase in relative subchondral bone US gray-level values was observed as OA progressed. Statistically significant correlations were observed between normalized US mean intensity or intensity slope especially in subchondral bone depth level 2 and K-L grading (r = 0.600, P < 0.001; r = 0.486, P = 0.006, resp.) or femoral arthroscopic scoring (r = 0.332, P = 0.039; r = 0.335, P = 0.037, resp.). This novel quantitative noninvasive US analysis technique is promising for detection of femoral subchondral bone changes in knee OA.