| ObjectiveAlthough the sensitivity of MRI in the diagnosis of supraspinatus(SSP)and infraspinatus(ISP)tendon tears was more than 90%,its sensitivity in subscapularis(SSC)tendon tears was unsatisfactory.Among them,type Ⅰ and type Ⅱ injuries(Yoo and Rhee classification)can account for 80%of the entire subscapularis tendon tear with fraying and partial-thickness tear.The characteristics of small partial tear will decrease the overall diagnostic sensitivity of MRI in subscapularis tendon tears.The omission of SSC tendon tear can lead to muscle atrophy,fatty infiltration,and increased tear accompanied by aggravated shoulder pain and loss of function.An effective noninvasive evaluation tool will be beneficial to early identification and intervention.The purpose of this study was to identify and screen sensitive predictors associated with SSC tendon tear and develop a web-based dynamic nomogram to assist clinicians in early identification and intervention.MethodsFrom July 2016 to August 2021,528 consecutive cases of patients who underwent shoulder arthroscopic surgery with completely magnetic resonance imaging and clinical data were retrospectively analyzed.According to the arthroscopic diagnosis of subscapularis tendon tear,the patients were divided into subscapularis tendon tear group and non-tear group.Among them,patients admitted from July 2016 to July 2019 were included in the training cohort,and patients admitted from August 2019 to August 2021 were included in the validation cohort.Univariate analysis,multivarite logictic regression analysis,least absolute shrinkage and selection operator(LASSO)method and cross-validation method were used to screen for reliable predictors highly associated with subscapularis tendon tear in training cohort.Developing and validating a nomogram prediction model of subscapularis tendon tear with these key predictors.The prediction performance of the nomogram was evaluated by concordance index(C index).Calibration with 1000 bootstrap samples was used to assess how far the model predictions were from actual risk outcomes.Decision-curve analysis(DCA)was used to describe and compare the clinical implications of using each risk model.Receiver operating characteristic curve(ROC)was used to evaluate the performance of the predictive model and traditional diagnosis method(MRI diagnosis based on directsigns combined with clinical test)in patients with subscapularis tendon tears.ResultsAccording to the inclusion and exclusion criteria,three hundred and sixty-two and one hundred and sixty-six patients entered the training set cohort and the validation set cohort,respectively.The outcomes of LASSO method and 10-fold cross validaiton indicated that six items including coracohumeral distance(oblique sagittal plane),coracoid angle,malposition of the long head tendon of the biceps(subluxation/dislocation),multiple posterosuperior rotator cuff tears,considering SSC tendon tear on MRI(based on direct signs),and positive clinical test(at least 2 methods)were determined as sensitive predictors of subscapularis tendon tears.In training cohort,the nomogram achieved a good C index of 0.879(95%CI,0.838-0.920)with a good agreement on the risk estimation of calibration plots(mean absolute error=0.009).Meanwhile,the results of validation cohort also showed a good C index of 0.884(95%CI,0.831-0.938)with a good agreement on the risk estimation of calibration plots(mean absolute error=0.025).The areas under the receiver operator characteristic curves(ROC)of the two methods showed that the dynamic nomogram model has achieved a better prediction performance in sentivity(76.74%vs 56.98%),specificity(84.78%vs 84.42%),positive predicition value(61.11%vs 53.26%),negative prediction value(92.13%vs 86.30%),positive likelihood ratio(5.04 vs 3.66),negative likelihood ratio(0.27 vs 0.51)than conventional diagnosis method(MRI diagnosis based on direct-signs combined with clinical test).The result suggested that when the threshold probability is between 3%and 93%,the prediction model can yield a good net benefit with higher clinical application value and better clinical practicability than conventional diagnosis method.Based on the nomogram model,we developed a webbased rating system(SSC-DynNom score).According to the area under the ROC curve of this prediction model,we found that the best diagnostic sensitivity(76.7%)and specificity(84.8%)were obtained when the total nomogram score was 110,corresponding to a risk prediction value of 28%.When the risk prediction value was>28%,we defined it as the high-risk tear group,and when the risk prediction value was≤28%.we defined it as the low-risk tear group.Conclusion Compared with traditional diagnostic methods,this model provided an individualized probability of risk prediction,which is convenient for clinicians to identify patients at high risk for SSC tendon tears with useful clinical suggestion. |
PDF Full Text Request