According to our experience, it is easier to meet the condition of perpendicularity of the upcoming beam just in the location of the intercondylar notch area than at the adjacent condyles due to their naturally curved shape. Maintaining a consistent angle of the US beam against overnight delivery the femur might be difficult with manual placement, and thus might affect the US intensity or slope of the cartilage-bone interface. One approach to overcome this problem could be mechanical scanning, as described by Ohashi et al. [29]. In femoral LAT compartment, the statistically insignificant comparisons were probably entailed by a limited acoustic window. In knee flexion, the patella is shifted over the LAT condyle causing acoustic shadow allowing the US beam to reach only small area of the LAT compartment.

Therefore, the possible damage visible in radiography or arthroscopy might not have been detected by US. The US images were obtained of a previous study protocol and then reviewed for this study what may be considered as a limitation of the present work. Hence, a rather small patient group was enrolled in the study which could explain the relatively large US intensity variation in K-L grade 0 and 1 which can be noticed from Figures 2(a) and 2(c). Additionally, US data in moderate and severe K-L grades were lacking, as well as in completely normal OA (grade 0) and early OA summed arthroscopic grades. Consequently, a comparison of FAS2 groups 0 and 1 was not possible to conduct due to statistically insufficient amount of data in group 0.

Therefore, we suggest that more healthy volunteers and symptomatic patients should be enrolled into next studies in which ethically convenient, reliable, quantitative, and noninvasive diagnostic method would be used as a reference (e.g., MRI) in order to verify and validate this method. New low dose, high resolution cone beam computed tomography [30] could be also used as a minimally invasive reference method to quantitative US imaging of subchondral bone and articular cartilage (contrast agent injection needed for visualization of the cartilage).In the US image analysis, some errors might be caused by the subjective segmentation of ROIs due to possible inclusion of cartilage tissue into the processed ROI. This could happen especially in MED and LAT condyle image segments where the entire bone-cartilage interface was not always totally perpendicular to the upcoming US beam. In future studies, the above-mentioned limitations should be taken into account during both, the preparation of image acquisition protocol as well as the image processing and analysis. For instance, the US operator-dependent parameters Drug_discovery should be always kept constant in order to compare absolute reflection values within the investigated population.