Analysis Of Laterality Of Sacroiliac Joint Abnormalities In Patients With Developmental Dysplasia Of The Hip And Construction Of Predictive Models

Research Objectives:Developmental dysplasia of the hip（DDH）is a common deformity of the lower extremity and an important cause of osteoarthritis of the hip in adulthood.In adult patients with untreated DDH,osteoarthritis will inevitably develop in the diseased hip as the disease progresses.Total hip arthroplasty is an effective treatment for patients with end-stage hip osteoarthritis secondary to DDH,but these patients often experience pelvic tilt,which further leads to changes in the sacroiliac joint due to the involvement of the sacroiliac joint in compensation.In addition,sacroiliac joint abnormalities can lead to pain in the proximal thigh and hip area,etc.These symptoms may be considered to be caused by hip pathology on either the more severe or less severe side of DDH,and it is often to overlook the effects caused by the sacroiliac joint.There is still no"gold standard"for the diagnosis of sacroiliac joint pain.This study investigated the laterality of sacroiliac joint abnormalities in patients with end-stage hip osteoarthritis secondary to DDH using the widely used and economical anteroposterior pelvic radiographs,and developed nomogram prediction models to assist clinicians in noting the possible impact of the sacroiliac joint on symptoms while focusing on hip lesions,and the prediction results,combined with the location of positive clinical provocation tests,may increase the reliability of the diagnosis of sacroiliac joint lesions and sacroiliac pain in an economical,rapid and non-invasive manner.In patients who do not currently have sacroiliac joint pathology,the predicted results can alert the patient to abnormal changes in the sacroiliac joint and sacroiliac joint pain on that side in the future,in order to help with prevention.Methods:This retrospective study included 137 patients with end-stage hip osteoarthritis secondary to DDH between January 2017 and October 2022,and measured the preoperative anteroposterior pelvic radiographic parameters iliac obliquity（IO）and sacral obliquity（SO）,while using the gray level co-occurrence matrix（GLCM）method of image texture analysis to examine our grading of the images.After analyzing the influence of the interrelationship between IO and SO on the laterality,the data of 110 patients with the same direction of IO and SO and different grading on both sides,including the general clinical data,laboratory index data and radiographic parameters before surgery,were analyzed by univariate analysis and multivariate logistic regression analysis to identify the independent influencing factors affecting the lateralization of sacroiliac joint abnormalities.Combining the receiver operating characteristic（ROC）curves,nomogram models were constructed to predict the laterality of sacroiliac joint abnormalities in DDH patients,and calibration curves were plotted to validate the models.Results:The results of the univariate analysis showed significant differences（P<0.05）in SO,the relationship between the mutual magnitude of IO and SO,comparison of the number of highlighted pixels of the sacroiliac joint on both sides measured by Image J software,body mass index（BMI）and total carbon dioxide（TCO₂）（both with preoperative data and with the same direction of IO and SO）in the two groups with more severe ipsilateral and more severe contralateral sacroiliac joint abnormalities in the direction of pelvic tilt.Multivariate logistic regression analyses showed that the relationship between IO and SO,comparison of the number of highlighted pixels of the sacroiliac joint on both sides measured by Image J software,BMI,and TCO₂（all with preoperative data and the same direction of IO and SO）were independent influences on the laterality of SIJ abnormalities（P<0.05）.The ROC curve analysis showed that the area under the curve of the federated model for each independent factor and the federated model with the number of pixels measured by Image J removed was higher than that of the independent factors,0.924 and 0.909,respectively.The slopes of the calibration curves are all close to 1.The two federated models had better predictive ability than each independent factor.The nomogram model II constructed by the federated model with the number of pixels measured by Image J removed was well differentiated and had higher diagnostic value and better predictive ability,but compared with the nomogram model I constructed by the federated model with each independent factor,the nomogram model II had slightly lower predictive ability,diagnostic efficacy and accuracy.Conclusion:In patients with preoperative end-stage hip osteoarthritis secondary to DDH,the relationship between the mutual magnitude of IO and SO,comparison of the number of highlighted pixels of the sacroiliac joint on both sides measured by Image J software,BMI and TCO₂were independent influencing factors of the laterality of sacroiliac joint abnormal alterations in the same direction of IO and SO.In this study,two nomogram models were developed to predict which side of the sacroiliac joint has more abnormal alterations,with good performance of the model without the software and better performance of the model with the condition of using Image J software.