In general studies of RA patients involving ultrasound

In general, studies of RA patients involving ultrasound for joints other than the wrists and small joints of the hands and feet have described ultrasound findings commonly obtained for RA patients (Lehtinen et al. 1996; Riente et al. 2010; Soini et al. 2003).
To our knowledge, only Scheel et al. (2005) established a synovial hypertrophy quantitative ultrasound cutoff value differentiating between healthy persons and RA patients, as in our study. These authors studied only the palmar joint recesses of the MCP and PIP and included only 10 HVs for differentiation between the groups. Values greater than 0.6 mm, for both the PIP and MCP (approximately 90% specificity), were capable of differentiating the two groups. The values established by these authors are much lower than the values found in our study, in which there was greater concern over studying both the dorsal and palmar recesses, as well as a larger healthy sample.
Terslev et al. (2008) and Carotti et al. (2012) also established quantitative cutoff values in the recesses of the wrist and small hand joints, using color Doppler ultrasound technique. By contrast, in the present study, only semiquantitative PD measurements were made, preventing comparison with other investigations. Our univariate logistic regression revealed that for the analysis of synovial blood flow, the radiocarpal recess was the largest contributor in differentiating the two groups, because the guanylyl cyclase of scores from 0 to 3 increased the probability of RA detection by 66-fold. For other recesses, the chances were increased by up to 19-fold.
Millot et al. (2011) conducted an important study, with significant sampling, that compared healthy persons with patients with early RA. Similar to some of the recesses that were analyzed by Millot et al. (2011) (second and third MCP and fifth MTP joints), we observed that the probability of abnormalities also increased starting from a synovial hypertrophy score of 2. By contrast, for the semiquantitative erosion measurements, the probability of abnormalities greatly increased in our study starting from a score of 2 (true erosion), especially in the dorsal recesses of the second MCP (OR = 93), radial recess of the second MCP (OR = 92) and lateral recess of the fifth MTP (OR = 100). For the palmar recesses of the second MCP, the dorsal and palmar portions of the third MCP and dorsal portion of the fifth MTP, an increased probability was observed starting from evolution of the erosion score from 0 to 1. However, this increase was less significant, most likely because of insufficient sampling for this analysis. For the fifth dorsal MTP, however, PD analysis assisted in the differentiation between the groups because a score of 1 increased the probability of the patient presenting with RA by fivefold.
To our knowledge, Filer et al. (2011) published the only study using ultrasound that included small, medium and large joints with the goal of predicting RA. Their results also indicate the importance of involvement of the wrist, MCP and MTP joints as positive predictors of RA in a prospective study of a population starting with very early synovitis. In addition, in a prospective study in a similar sample of patients, Salaffi et al. (2010) observed that the greater the number of joints with positive PD at baseline, the greater was the chance of detecting RA, though they analyzed only the wrist, MCP and PIP joints.
The semiquantitative scores of synovial hypertrophy used in various joint recesses, especially in large joints, were determined by our group because there is no standardization in the literature, until recently (Hartung et al. 2012). Nevertheless, the worst inter-observer reproducibility was observed in the radiocarpal recess (k = 0.341) and posterior recess of the glenohumeral joint (k = 0.395), with concordance rates of 60.5% and 86.8%, respectively. Nevertheless, for semiquantitative measurements of synovial hypertrophy and PD, the reproducibility was similar to that found in the literature for the reading of captured images (Cheung et al. 2010).