Prediction Of Collapse With MRI Signs And Necrosis Volume Measurement For Non-traumatic Osteonecrosis Of Femoral Head

Objective: To evaluate the value of prediction of collapse in non-traumaticosteonecrosis of femoral head with MRI sign and necrosis volume ratio usingoctant-based lesion measurement method proposed by Malizos.Materials and Methods:56patients （76hips） diagnosed non-traumaticosteonecrosis of femoral head (pre-collapse) by X-ray film and MRI were enrolled inthe study. There were42men and14women with a mean age of49±13.41(range,24to76years) at the time of the diagnosis. The disease was attributed to several etiologicfactors, such as steroid-induced necrosis in15patients (17hips), alcohol necrosis15patients (22hips), idiopathic necrosis in24patients (35hips) and DevelopmentalDysplasia of Hip (DDH) in2patients (2hips). According to the Association ResearchCirculation Osseous international staging system，32hips were in stage I (normalfindings on radiographs but abnormal findings on magnetic resonance images) and44hips were in stage II (radiographs showing scleroses or cysts but no crescent sign) at thetime of diagnosis. All patients underwent initial examination of hip withanteroposterior(AP) and frog-position X-ray and MRI scans at the time of diagnosis. Toevaluate the collapse of the femoral head at each X-ray examination; observe the hipjoint effusion, bone marrow edema, necrosis signal features, location of necrotic lesionsand necrotic morphology at the first MRI images, and measure the volume of femoralhead necrosis using MRI post-processing workstation. Necrotic volume was measuredat the femoral head which was regarded as a standard sphere on a mathematical concepton the coronal TSE/T1WI. First，determine the center of the sphere，and then through this central point，orthogonal coronal, sagittal and horizontal plane divided it into eightparts equally，namely Anterior Superior Lateral (ASL)，Posterior Superior Lateral(PSL)，Anterior Superior Medial (ASM)，Posterior Superior Medial (PSM)，AnteriorInferior Lateral (AIL)，Posterior Inferior Lateral (PIL)，Anterior Inferior Medial (AIM)and Posterior Inferior Medial (PIM). Follow by a low signal band on the outer edge ofthe lesion on T1WI，measure the necrotic area manually，multiply by the thickness ofthe image and calculate the volume of necrotic lesions of each layer. Eventually，volumeof each layer would be added to calculate the total necrosis volume and necrosis volumepercentage of each quadrant. The above observations and measurementof the volumewere made by two experienced radiologists using a double-blind design.All of thosehips were followed until collapse developed or for at least15months.Meanwhile，everypatients reviewed anteroposterior(AP) and frog-position X-ray for observing whetherthe collapse of the femoral head or not. Then，divide these hips into two groups，collapse groups and non-collapse groups，and record the time.Demographic and clinical parameters, such as patient age and gender, etiology andextent of the necrotic lesion were compared between two groups: hips in which collapseoccurred versus hips in which collapse did not occur. Nominal data were analyzed withthe chi-square test and measurement data were analyzed with independent sample T-test.Logistic regression was used to analyze the risk factors. Receiver operatingcharacteristic curve (ROC) was used to analyze sensitivity and specificity. Statisticalanalysis was done with SPSS20.0software. The measurement data presented as mean±standard deviation, P value <0.05of each test method denominated statisticallysignificant.Result:1.Of these56patient（s76hips），collapse of the femoral head was found in31hips,collapse had not occurred for the others（45hips）. The average follow-up period was21.6±9.1months2.For general information on the patient, between collapse group and non-collapsegroup，sex (P=0.078), the age (P=0.631), etiology (P=0.604) and necrosis signal (P= 0.071) was not statistically significance; there was a statistically significant aboutARCO stage,joint effusion,edema of the bone marrow and necrosis morphology, P <0.0001.3.Analysis of independent sample T-test between the groups demonstrated that allnecrosis volume ratio values were statistically significant (P <0.05) except the PIMquadrant (P=0.143).4.Logistic regression analysis showed that ARCO stage Ⅱ, peripheral necrosis,bone marrow edema, the total necrosis volume ratio (%), ASL quadrant necrosisvolume ratio (%) and PSL quadrant necrosis volume ratio (%) were risk factors forcollapse of the femoral head, and the occurrence rate of collapse of femoral headnecrosis in the lateral and anterior quadrant was higher than the other position.5.ROC curve analysis showed the optimal diagnostic value of the total necrosisvolume ratio (%), ASL quadrant necrosis volume ratio (%) and PSL quadrant necrosisvolume ratio (%) for predicting collapse were19.51%,47.78%,21.00%,corresponding sensitivity and specificity were93.5%,83.9%,77.4%and71.1%,68.9%,86.7%,respectively.Conclusion:Based on MRI signs of hip and quantitative measurement of non-traumaticosteonecrosis of patients, conclusions were explore and summarize dynamicly asfollows:1.ARCO stage, necrotic morphology, bone marrow edema, the total necrosisvolume ratio (%), ASL quadrant necrosis volume ratio (%) and PSL quadrant necrosisvolume ratio (%) are risk factors for collapse of the femoral head, and the necrosislocated in the lateral and anterior quadrant of femoral head is greatly contributed tocollapse.2. The optimal diagnostic value of the total necrosis volume ratio (%), ASLquadrant necrosis volume ratio (%) and PSL quadrant necrosis volume ratio (%) forpredicting collapse has better clinical value.3. MRI examination and quantitative measurement of osteonecrosis has important clinical value for evaluation and prediction of the collapse of femoral head, but variousfactors should be taken into account for predicting collapse of the femoral head, thusimproving the accuracy of predictions.