Although recurrent rotator cuff tears are not uncommon, and imaging evaluation of a postoperative rotator cuff plays a critical role, as noted in the introduction, the temporal changes in the postoperative tendon on sonography have not been well investigated. This study aimed to address the uncertainties regarding the postoperative rotator cuff on serial follow-up sonographic examinations. In our study, serial sonographic evaluations of the repaired rotator cuff revealed mild thinning of the tendon over time. Recurrent tears of the repaired tendon were not frequent (4 of 65 [6%]), but if they happened, they always occurred within the first 3 months of surgical repair, which was concordant with results from a previous study.5 The morphologic appearance and peritendinous vascularity of the tendon were gradually normalized, although mild bursal thickening remained 6 months after surgery. Crim et al19 described the temporal evolution of MRI findings after arthroscopic rotator cuff repair, with serial MRI examinations at 6 weeks, 3 months, and 12 months after surgery. The tendons were the most disorganized compared to the native tendons 3 months after surgery, and they generally improved between 3 and 12 months after surgery. Fealy et al14 also reported that it was not uncommon to detect a full-thickness defect on sonography in the early postoperative period; however, they hypothesized that this defect was a reparative scar rather than a true retear because it gradually improved over time. Previous histologic studies supported these results. Four to 6 weeks after surgery, there was disorganized collagen at the bonetendon interface as well as an irregular zone of edema between the collagen bundles with neovascularization immediately proximal to the bone-tendon interface.20,21 The interface tissue became progressively more organized with time, and the tendon fibroblasts were increasingly oriented along the tendon.21 Similarly, in this study, there was also a disorganized appearance, including a decreased echo texture, absence of a fibrillar pattern, and surface irregularity, in the repaired tendon at 5 weeks and 3 months; however, this disorganization normalized through remodeling by 6 months after surgery. Early postoperative tendons frequently had a hypo - echoic echo texture and the absence of a fibrillar pattern, which might be misinterpreted as recurrent tears; however, these features often normalized into tendons with an increased echo texture and the reappearance of a fibrillar pattern at 6 months (Figure 7). Based on these sequential findings, the sonographic findings within 3 months after surgery should be interpreted with caution to accurately understand and monitor the repaired tendon status. The defect described by Fealy et al14 might be similar to the finding mentioned above, although it could not be confirmed because the authors did not provide an image illustrating the defect in their article. In terms of the tendon thickness, Tham et al15 demonstrated that the repaired supraspinatus tendon thickness remained unchanged throughout 6 months of sonographic analysis, whereas Lasbleiz et al9 reported an inverse correlation between the tendon thickness and the time between measurements. In our study, the tendon thickness decreased over time after surgery, resulting in a 5% to 10% difference in the thickness. Some patients had marked changes exceeding 30% to 50% of the tendon thickness 5 weeks after surgery (Figure 6). After careful review of the arthroscopic findings and intense discussion with the surgeon, this change was thought to be a postoperative deformity, a “dog ear” deformity, at the repaired tendon with spontaneous remodeling over time.22 Subacromial-subdeltoid bursitis decreased significantly over time; however, mild bursal thickening was frequently observed at 6 months. This finding was consistent with a study performed by Tham et al,15 which demonstrated a significant decrease in the bursal thickness, capsular thickness, shoulder stiffness, and level of pain over time. Another interesting issue in the healing of a rotator cuff is the vascularity of the rotator cuff after surgical repair because blood at the site of a repaired tendon encourages the reestablishment of the bone-tendon attachment.13,14 Several studies have shown the vascular pattern in the repaired tendon on Doppler studies with or without a contrast agent.13–17These studies reported initial high vascular flow at the peritendinous region that decreased with time, whereas the repaired tendon showed either sparse or no blood flow. However, it was uncertain whether the vascularity at the bone anchor site was increased. The bone anchor site had the lowest blood flow on conventional Doppler studies,14whereas marked enhancement in the suture anchor and the peribursal regions were observed on contrast-enhanced sonography.13,16,17 The authors explained that the discrepancy between the findings from the contrast-enhanced and conventional Doppler studies could be attributed to the increased sensitivity of microbubble contrast techniques. In our study, the peritendinous region had the greatest blood flow, which decreased on follow-up sonography, and the bone anchor site and tendon remained relatively avascular. This result was consistent with studies performed using conventional Doppler analysis without microbubble contrast. On the basis of the finding of the most robust, highest vascularity in the peribursal tissue, some authors have suggested that the peribursal tissue might be the greatest conduit of blood flow for promoting healing of the repaired tendon.13,14,17 However, no data have shown the relationship between the peribursal vascularity and the retear rate or clinical outcomes, and further investigation is warranted. Our study had several limitations. First, subjective criteria were used to assess the morphologic tendon characteristics. Although efforts were made to use reproducible criteria, it was difficult to apply consistent and reproducible criteria between patients and examinations because of the operator-dependent nature of sonography. Second, it was at times difficult to assess the repaired tendon in the early postoperative period in patients with severe pain and a limited range of motion. Third, the follow-up was rather short, and further investigations of the long-term outcomes or prognoses 1 and 2 years after surgery are needed. Nevertheless, sonographic assessment of the repaired tendon 3 months after surgical repair might be important because the clinical examination at this time might be limited by immobility and pain, and most recurrent tears occur within the first 3 months of surgical repair. Fourth, the sizes of the rotator cuff tears were not uniform in our study. Fifth, there was no surgical proof of retears. Finally, clinical findings, including stiffness and the level of pain, were not assessed in our study. Further study is warranted to determine whether the sonographic appearance of a healing tendon correlates with the level of pain. In conclusion, we have demonstrated that the occurrence of a retear developed within 3 months after surgery, and the tendon thickness and morphologic appearance of the repaired tendon improved over time and nearly normalized within 6 months after repair. In our study, the sonographic findings in the first 3 months after surgical repair should be interpreted with caution to accurately understand and monitor the repaired tendon status.