Knee osteoarthritis(KOA)is a chronic degenerative joint disease,which poses a major challenge particularly among the elderly population due to its high incidence and high disability.Imaging examination has been used commonly to diagnose KOA.However,it faces imitations in predicting disease progression due to the lack of prior information and constraints in manpower and time.With the rapid evolution of big data and computational technologies,artificial intelligence(AI)is progressively integrating into various healthcare domains.Therefore,the integration of artificial intelligence(AI)into healthcare holds promise for revolutionizing KOA diagnosis and treatment.AI-assisted diagnostic models have demonstrated the potential to automate diagnosis,classify disease severity,and predict disease progression with improved efficiency and accuracy.In addition,these models provide personalized diagnosis and treatment options,as well as accurate disease progression risk assessment.Despite these promising outcomes,challenges such as high costs associated with data annotation and limitations in model generalization capabilities persist.This paper reviews recent advancements in AI applications and summarizes the potential value of utilizing AI applications for KOA.To further enhance the utilization of AI in KOA management to overcome current limitations,future efforts should focus on standardizing clinical sample databases,optimizing AI algorithms,and enhancing external verification sets.

OBJECTIVETo accurately define the injury position of medial patellar retinaculum with acute injury under the guiding of high frequency ultrasonography, and to study therapeutic effects of suture operation on medial patellar retinaculum in the injuried position.

METHODSFrom June 2009 to March 2014, there were 17 patients with acute patellar dislocation, 6 males and 11 females with average age of (16.2±6.2) years old. The duration time of patellar dislocation was 2 weeks. Before operation, the medial patellar retinaculum of all patients were examined with the high frequency ultrasonography, and the skin with the non-continuous fiber was iudicated as the surface mark under the high frequency ultrasonography. The injury position of medial retinaculum was in the middle of 5 patients who were treated with suture operation of arthroscopic medial retinaculum. The injury position was in the patellar edge in 12 patients who were treated with fixing bone anchor on patella and arthroscopic suture operation of medial retinaculum. The CT examination and Kujala scores, patellar tilt angle on CT film, measured maximal angles of passive or active knee flexion and apprehension test were observed before treatment and postoperative 18 months.

RESULTSEighteen months after treatment, Kujala scores were 92.2±11.1 and patellar tilt angle were(11.5±4.2) °, and there was no statistical difference between post-operation and pre-operation. The difference between maximal angles of passive knee flexion(133.5±4.2) ° and normal had no statistically significance. Maximal angles of active knee flexion were(153.5±4.6) °. Ultrasonography showed the continuous fiber of medial retinaculum. A patient showed positive apprehension test and no patient had the recurrence patella instability after operation.

CONCLUSIONSThe injury position of medial patellar retinaculum was accurately shown by high frequency ultrasonography and treated with arthroscopic suture operation. Knee immobilization time after operation was shorten. Eighteen months after operation, knee joint function was good, and no patient had the recurrence patella instability.