Application And Reproducibility Of 3D Ultrasound Measurements In Assessing For Developmental Dysplasia Of The Hip

BACKGROUNDDevelopmental dysplasia of the hip (DDH) encompasses a spectrum of hip abnormalities that are either present at birth or develop during infancy. It is detected among 1.6-2.5%o of newborns[1,2]. Among them 1 %o children at birth is the dislocation hip. Radiographic imaging is very necessary, because in the neonatal period and the period of mild dysplasia, clinical examination, such as Ortolani and Barlow test lack of sensitivity [3]. For inexperienced examiners, Ortolani and Barlow test results are often unreliable, which lead to high false positives and false negatives. The proximal femur and acetabulum of the baby is composed of cartilage. The structure form of hip joint is not fully displayed on X ray. As X-ray has radioactivity injury, it is not suitable for baby hip checking. Ultrasound can clearly show the internal structure of baby hip, and the hip joint can be objectively measured. Compared with the clinical physical examination and X-ray, ultrasound has no radiation, the advantages of good visibility and early diagnosis.Two dimensional ultrasound (2D US) of the hip had been applied to the diagnosis of DDH in 1980s. Because of its good resolution of soft tissue and cartilage and noninvasive, no radiation, it have been widely used in the screening of DDH in foreign countries.Currently 2D US of DDH is also widely used, but the application of three-dimensional ultrasound (3D US)of DDH is rarely reported. The professor of Zhaoliang proposed that 2D US examination should be performed by experienced examiners. If ultrasound probe was not correctly placed on the baby, the Graf standard image could not be obtained. So that Measurement will appear deviation. It is necessary to correctly identify the baseline, bony roof, cartilage roof which play an important role on the measurements of a and β angles.2D ultrasound can only obtain one plane of the hip, the whole hip can’t completely be viewed. In practice, if the examiner was not fully skilled on hip ultrasound, it is inevitable to make the wrong diagnosis which Lead to excessive treatment or missed diagnosis on DDH.In recent years, the study of 3D US diagnosis of DDH has been rarely reported, this study aims to explore the feasibility and repeatability of US 3D application in the diagnosis of DDHChapter OneThe application of 3D ultrasoundmeasurements in assessing fordevelopmental dysplasia of the hipOBJECTIVE To explore the application of 3D ultrasound in assessing for infantile developmental dysplasia of the hip.METHODS1 clinical materialThis prospective study was approved by the Ethics Committee of the Hospital. We received informed consent from each parent/guardian of the infants used in our study. We selected 31 patients (26 females,5 males; aging from 9 days to 210 days) from March 2014 to October 2014. A total of 61 hips were evaluated with both 2D and 3D US.Inclusion criteria:Patients with high risk factors of DDH or clinical suspect of DDH for clinical manifestation such as leg length discrepancy, were recommended to perform ultrasound.Exclusion criteria:Patients with Nerve muscular dysplasia and obvious dislocation of the hip2 Equipment and examination methodsEquipmentA Voluson E8 ultrasound machine (GE Medical Systems, Zipf, Austria), equipped with both a RIC 5-9-D volume probe and 9L-D linear probe, was used for all 2D and 3D US examinations.ExaminerHe has 8 years of experience in the method of Graf in assessing for DDH.Examination methods1) 2D US The infant was placed on its right or left side on the examination bed by its mother. US was performed by an experienced examiner. Pacifiers, infant formula, toys or delighted music were prepared to make the baby relaxed. The infant’s hip was flexed slightly and the 9L-D linear probe was placed on the greater trochanter to obtain a coronal plane of the hip. The probe was moved laterally until the standard plane was observed.The image features and the anatomical structure should be displayed as follows:hyperecho of femoral neck epiphyseal plate below the hip, femoral head located in the central of hip,Scattered in the internal echoes dot of medium or high echo ovoid keyed acoustic area, hyperechoic plica, joint capsule, spoon lips and hypoechoic cartilaginous acetabulum. The lower limb of ilium, Straight iliac bone, labrum should be displayed on the standard plane.2) 3D US The RIC 5-9-D volume probe was placed on the greater trochanter of the infant and the view of the standard plane was obtained on the central screen. Required conditions were chosen on the machine and then the 3D ultrasound data was obtained.The femoral head was located at the center of the ultrasound image, the iliac bone was showed as long as possible. The 3D datas were stored for later review, analysis, and additional reconstruction3) 2D US images were analysed. According to Graf method, the iliac bone must be straight and both the acetabular labrum and the lower limb of the os ilium must be visible. Thenthe angle of acetabular inclination (a angle) and the acetabular roof angle (β angle) was measured to permit description of acetabular dysplasia and classification of the infant hip into four types. The a angle between the bony roof line and baseline quantifies the bony socket. The P angle between the baseline and the cartilage roof line quantifies the cartilaginous acetabular roof.With the Graf method, a hip with an alpha angle greater than 60°s classified as type I, a hip with an alpha angle from 50° to 59°is classified as type Ila/IIb, a hip with an alpha angle from 43°to 49°, a beta angle less than 77°is classified as type Ⅱc, and a hip with an alpha angle of less than 43°is classified as type Ⅲ, type Ⅳ.4) 3D US images were analysed. Until the children leaved the examination room,3D US data was extracted from the machine. Due to 3D US is a volume data, the Graf standard plane need to be obtained through rotating the X, Y, Z axis. The specific method as follows:A, B, C represented the coronal, axial, and longitudinal planes on the 3D image. First, A-plane was activated to locate the peak of the bony roof, which could be calibrated on the C-plane. Next, the Y-axis was rotated to show the straight plane of the ilium, and the X-axis was rotated to locate the lower limb of the os ilium. The labrum was displayed naturally. The Z-axis was perpendicular to the X, and Y-axes and the rotation was not required. Then, the 3D view, including the full lower limb of the os ilium, the straight ilium and the labrum was displayed. Finally, the bony roof, cartilaginous roof, joint capsule, synovial fold, femoral head, and femoral neck were also seen on the standard plane. The Graf classification was made according to the a and β angles on 3D US. The 3D surface imaging was also rendered to diagnose DDH.5)The examination time from the probe being placed on the hip to 2D images and 3D volume sweeps being obtained was recorded, respectively.6) Statistical Analysis For analyses of a and β angles, we averaged the values of the right and left hips. Consistency of measurement of a and β angles was assessed using intraclass correlation coefficients and The Bland-Altamn plots on 2D and 3D US. ICC values ranged from 0 to 1, with higher values reflecting more reliable measurements. ICC value more than 0.75 was taken to indicate satisfactory measurement reliability.The consistency of Graf Classifications was assessed by using Cohen’s kappa coefficient on 2D and 3D US. Levels of agreement for k were determined, as proposed by Landis et al., with k values 0.00-0.20 considered slight agreement; 0.20-0.40, fair agreement; 0.40-0.60, moderate agreement; 0.60-0.80, substantial agreement, and 0.80-1.00, almost perfect agreement. Paired t-tests was used to show the variation of the examination time on 2D US and 3D US. (P <0.05 was considered significant)ResultsGeneral informationOf the 31 infants, 1hip with obvious dislocation of the hip was excluded. A total of 61 hips were included. Of the 61 hip,there 53 hips with type I,3hips with type Ⅱa, 3 hips with IIb,2 hips with type Ⅱc.Consistency of 2D US and 3D US on a angle and β angle.The 3D US has substantial consistency with 2D US on the measurement of a angle. And the consistency of 2D US and 3D US on a angle is better than that on β angle (ICCa=0.91;ICCβ=0.74)2 The Graf classification consistency between 2D US and 3D US on DDHAgreement on the Graf classification between two methods was substantial with a kappa coefficient of 0.87 (k=0.87).Analysis of the surface imaging of bony acetabular roof according to 3D US.The morphology of bony acetabular roof was classified sharp, round, and flat. There were 54 hips with sharp,4 hips with round,3 hips with flat.Comparison between 2D and 3D ultrasound about operation time (Paired t-test)The average time from we placing the probe on the greater trochanter to capture storing data image was 12.2s±7.0s (2D),10.0s±7.4s (3D).The average time of 3D US is shorter compared to that of 2D ultrasound. The difference between 2D and 3D US about operation time was statistically significant(P<0.01).CONCLUSIONThe consistency of 2D US and 3D US both on a angle and Graf classilication were good,according to experienced examiner.3D US of the infant hip is feasible and images can be analyzed according to Graf method.The morphology of bony acetabular roof can be showed on 3D US. The development of acetabulum was assessed intuitively through the 3D US of the hip which can make more reliable diagnosis of DDH.3D US is a promising technology for the baby’s hip joint. If combined with the morphology of the acetabulum, it will play an important role on the diagnosis and treatment of DDH.The time of 3D US when we place the probe on the greater trochanter to capture figures is shorter compared to that of 2D US.3D US reduced the children and their families’anxiety, which is suitable for its intended use.Chapter TwoReproducibility of 3D ultrasound measurements in assessing for infantile developmental dysplasia of the hipMethods1 Study populationThis prospective study was approved by the Ethics Committee of Hospital. We received informed consent from each parent/guardian of the infants used in our study. We selected 94hips of 48 patients (37 females,11 males; aging from 40 days to 180 days) from December 2014 to October 2015. Two dislocated hips that had been diagnosed were excluded from the study. A total of94 hips were evaluated with both 2D and 3D US.Inclusion criteria:The same as Chapter One.Exclusion criteria:The same as Chapter One.2 Equipment and SettingsThe same as Chapter One.3 ExaminersThe experienced one had over 8 years’experience in sonography and Graf hip ultrasound. The inexperienced one had only 1 year experience in sonography and Graf hip ultrasound. To avoid bias, the acquisition of measurements was completely independent, with only one examiner in the examination room during the evaluation. To avoid memory effect on the analysis during the scan and interpretation, we masked the numeric results and only the images were recorded on the hard drive for later assessment.4 Preparation and examination for 2D and 3D US on DDHThe infant was placed on its right or left side on the examination bed by its mother. US was performed by two different examiners respectively. Pacifiers, infant formula, toys or delighted music were prepared to make the baby relaxed. The infant’s hip was flexed slightly and the probe was placed on the greater trochanter. The US examination was then started and the procedure was as followed:1) 2D US technique:The same as Chapter One.2) 3D US technique:The same as Chapter One.3) The infant was rolled on its right side by its mother and the 2D and 3D US examination above was repeated.4) The examination time from the probe being placed on the hip to 2D images and 3D volume sweeps being obtained was recorded, respectively.5) 2D and 3D US images were analysed respectively:The same as Chapter One.Statistical AnalysisAll statistics were analyzed with SPSS version 19.0 software for Windows. Descriptive statistics were recorded as the mean ± standard deviation (x±SD). For analyses of α and β angles, we averaged the values of the right and left hips. Reproducibility of measurement of a and β angles was assessed using intraclass correlation coefficients and the Bland-Altamn plots. ICC values ranged from 0 to 1, with higher values reflecting more reliable measurements. ICC value more than 0.75 was taken to indicate satisfactory measurement reliability.The reproducibility of Graf Classifications was assessed by using Cohen’s kappa coefficient. Levels of agreement for k were determined, as proposed by Landis et al.with k values 0.00-0.20 considered slight agreement; 0.21-0.40, fair agreement; 0.41-0.60, moderate agreement; 0.61-0.80, substantial agreement, and 0.81-1.00, almost perfect agreement. Paired t-tests was used to show the variation of the examination time on 2D US versus 3D US. (P<0.05 was considered significant)Results1 General informationAmong 48 infants, two hips showing severe pathological conditions were looked at separately for the analysis of angle measurements. In this study, of 94 hips, there were 84 hips with type I,7 hips with type Ⅱa,3hips with type Ⅱb,2 hips with type Ⅱc. The average age of the infants was 105days ± 35days. There were more girls (77%) in the study than boys (23%). There were 33 infants with high risk factor of DDH,15 infants with clinical manifestations of suspected DDH2 The intra-examiner reproducibility of 2D US and 3D US in the measurement of a and β angles according to two examiners.For the experienced examiner, the intra-examiner reproducibility of 2D and 3D US in the measurements of a and p angles were excellent (ICCa=0.91, ICCp=0.81). For the inexperienced one, the intra-examiner reproducibility of 2D and 3D US in the measurements of a and β angles was poor (ICCa=0.52, ICCp=0.45)3 Inter-examiner reproducibility of alpha and beta angles measured by different examiners on 2D and 3D US.The inter-examiner reproducibility of 3D US in the measurement a angle was excellent (ICCa=0.95), The inter-examiner reproducibility of 3D US in the measurement of β angle was poor(ICCp=0.54), whereas the inter-examiner reproducibility of 2D US was relatively poor (ICCa=0.47, ICCp=0.36).4 The reproducibility of Graf classification both on 2D and 3D US between examiners respectively.The mean kappa-coefficients for Graf classification demonstrated high inter-observer reproducibility both on 2D and 3D US (k2D=:0.92, k3D=0.98)5 Comparison of examination time between 2D and 3D US according to two examinersThe average examination time of 3D US is shorter than that of 2D US, unincluding preparation and images analysis. The difference between 2D and 3D US about examination time was statistically significant according to two examiners (P<0.01).CONCLUSION2D US was affected by the operator experience, the results of the diagnosis of different experience of the examiner may have a greater difference.The repeatability of 3D US is good, which reduces the experience and operator dependency in a certain extent. The operation time of 3D US is shorter compared to that of 2D US, which reduces the children and their families’anxiety.3D ultrasound is a simple method for diagnosis of DDH, especially for inexperienced examiner, and is worthy of clinical promotion.