Increased diffusion and decreased perfusion in epiphyses of femoral heads detected via intravoxel incoherent motion after closed reduction in children with developmental dysplasia of the hip

Abstract

Background

Patients with developmental dysplasia of the hip (DDH) may have decreased blood supply to the femoral heads. Finding a non-invasive method to evaluate whether the femoral heads in patients with DDH are ischemic is paramount for orthopedic surgeons.

Purpose

To identify whether parameters reflecting perfusion and diffusion in intravoxel incoherent motion (IVIM) sequences can be used to assess ischemia in femoral heads in patients with DDH after closed reduction.

Material and Methods

Twenty-eight patients with DDH who had undergone closed reduction were enrolled. IVIM data were acquired using a 3-T magnetic resonance scanner, regions of interest were placed on the epiphyses; ADC_slow, ADC_fast, f, and ADC_fast×f were measured. A Mann–Whitney U test was performed to compare ADC_slow, ADC_fast, f, and ADC_fast×f between the lesion and control sides. Receiver operating characteristic curves were generated with respective cut-off values. The lesion sides were classified based on the International Hip Dysplasia Institute (IHDI) classification. ADC_slow, ADC_fast, f, and ADC_fast×f were compared among the groups.

Results

ADC_slow was higher and ADC_fast, f, and ADC_fast×f were lower on the lesion sides (P = 0.000–0.002). The optimal cut-off value for ADC_fast×f, ADC_fast, ADC_slow, and f were 0.030, 0.626, 0.000251, and 0.636, respectively. Higher IHDI classification scores on the lesion side were associated with lower ADC_fast, f, and ADC_fast×f, and higher ADC_slow values.

Conclusion

IVIM is a promising method to investigate the perfusion and diffusion of epiphyses of femoral heads.

Keywords

Closed reduction developmental dysplasia of the hip diffusion intravoxel incoherent motion perfusion

Introduction

Developmental dysplasia of the hip (DDH) is a common disease of hip joints. The epiphyses of femoral heads may have delayed development when DDH is present for a long time (1). Children with DDH who do not respond to treatment using the Pavlik harness require closed reduction followed by immobilization in spica casting (2). Salter et al. (3) concluded that excessive hip abduction could lead to avascular necrosis (AVN). Jaramillo et al. (4) suggested that AVN could be avoided if ischemia in the femoral heads is corrected within 6 h. Therefore, detection of ischemia of femoral heads as early as possible can help clinicians prevent AVN.

Finding a reliable and non-invasive method to assess the blood supply to epiphyses of femoral heads is paramount. Imaging modalities that are widely used to evaluate DDH cannot be used to assess the blood supply of femoral heads. Ultrasound (5) and X-rays hardly penetrate the spica casting; the radiation dose of computed tomography (CT) is too high for children and is difficult to identify the vasculature of femoral heads on contrast-enhanced CT images. Morphological magnetic resonance imaging (MRI) does not provide information about the blood supply of the femoral heads (6). To address these limitations, some researchers utilize contrast-enhanced MRI to study the blood supply to the femoral heads (7,8). However, this method requires use of contrast agents, which can lead to adverse effects, especially in infants and children as systemic sclerosis, allergies, nausea and vomiting, or other reactions to contrast agents. Therefore, development of a method to evaluate the perfusion of epiphyses of femoral heads accurately and safely is urgently required.

In 1986, Le Bihanet et al. (9) reported that intravoxel incoherent motion (IVIM) diffusion-weighted imaging (DWI)-MRI sequences separately measured the pure diffusion coefficient (ADC_slow) and perfusion-related incoherent microcirculation (ADC_fast) in combination with the microvascular volume fraction (f) using multi-b-value DWI according to a biexponential curve fit algorithm. Many studies utilize IVIM in liver, pancreas, bowel, rectum, prostate, heart, breast, brain, etc., aiming to detect and stage tumors, evaluate therapeutic effects, and predict prognosis (10 –23). Gaeta et al. (24) utilized IVIM to evaluate bone metastases before and after radiotherapy and found significant changes in perfusion of bone metastases at different time points. Marchand et al. (25) suggested that the IVIM allows for measurement of diffusion and perfusion fraction in vertebral bone marrow in healthy volunteers. Zhao et al. (26) found that ADC_slow and f have significant differences in patients with ankylosing spondylitis in different groups. IVIM can also be used to measure muscle perfusion (27). These studies suggest that IVIM sequences are accurate and promising means to evaluate perfusion and diffusion in target tissues. Nevertheless, there are few articles discussing perfusion and diffusion of femoral heads. The objective of our study was to identify whether parameters in IVIM DWI-MRI sequences can be used to assess ischemia in epiphyses of femoral heads in patients with DDH after closed reduction.

Material and Methods

The study was approved by the Ethics Committee of the institution of our hospital (approval number 20160117) and conforms to the provisions of the Declaration of Helsinki. All patients and their guardians provided written informed consent.

Study cohort, and inclusion and exclusion criteria

From April 2016 to May 2017, children with DDH aged 6–24 months who had undergone closed reduction, spica casting, and MRI examinations were enrolled. The inclusion criteria for closed reduction in children with DDH were: (i) failed treatment with Pavlik harness; (ii) intractable dislocation of the hip; and (iii) delayed diagnosis of DDH (diagnosis after more than 3 months since the onset of DDH). The severity of DDH in each hip joint was classified based on the International Hip Dysplasia Institute (IHDI) classification system, which divides DDH into four grades (Fig. 1) (28). Because grade I DDH does not require closed reduction in our hospital, we did not include patients with grade I DDH. Patients who had previous surgery of the hips, infection, inflammatory diseases, tumor, or trauma around the hips, and those who were unable to cooperate to complete the MR examinations were excluded.

Fig. 1.

IHDI classification for developmental dysplasia of the hip for anteroposterior plain films of the hip joints. The horizontal line is the H-line or Hilgenreiner’s line, which goes through the bilateral tops of the tri-radiate cartilage. The line perpendicular to the H-line is the P-line or Perkin’s line, which is at the superolateral margin of the acetabulum. The diagonal line is the D-line, which is located at a 45° angle to the junction of the H-line and the P-line. The midpoint of the superior margin of the ossified metaphysis is called the H-point. Grade I: the H-point is medial to or at the P-line; grade II: the H-point is lateral to the P-line and medial to or at the D-line; grade III: the H-point is lateral to the D-line and inferior to or at the H-line; grade IV: the H-point is superior to the H-line.

MR protocols

MRI examinations of hip joints were performed using a 3-T MRI scanner (MR750; GE Healthcare; Milwaukee, WI, USA) with an eight-channel cardiac coil within 6 h after closed reduction. If the patients were awake, diluted chloral hydrate, prepared by dissolving 1 g of chloral hydrate in 10 mL of normal saline, was injected into the anus, 30 min before the MR examination. The injection volume of the diluted chloral hydrate was 0.5 mL per kg body weight. Routine MR protocols and the parameters of the imaging sequences are described in Table 1.

Table 1.

Routine sequences and parameters of MR examinations in the study.

Sequence	TR (ms)	TE (ms)	Bandwidth (kHz)	FOV (cm)	Slice thickness (mm)	Slice space (mm)	Matrix	NEX	TA (s)
Coronal FS PDWI	2500	39.5	83.3	22 × 22	3.3	0.5	320 × 224	3	1:15
Coronal T2WI	2590	125.5	62.5	22 × 22	3.3	0.5	320 × 256	2	1:08
Transverse FS PDWI	2864	38.6	83.3	24 × 14.4	3.3	0.5	320 × 224	6	1:26

MR, magnetic resonance; TR, repetition time; TE, echo time; FOV, field of view; NEX, number of excitations; TA, acquisition time; FS, fat suppression; PDWI, proton density-weighted imaging.

IVIM data were obtained using a multi-b single shot spin echo-echo planar imaging pulse sequence in three orthogonal directions and traverse plane images were acquired. For each individual, the following nine b-values sets at (number of excitation) were acquired: 0 (n = 2), 10 (n = 4), 30 (n = 4), 50 (n = 4), 80 (n = 4), 120 (n = 4), 200 (n = 4), 500 (n = 6), and 800 (n = 6) s/mm². The acquisition parameters were as follows: TR/TE = 2000/73.5 ms; field of view (FOV) = 300 × 300 mm²; slice thickness = 2.5 mm; slice space = 1.0 mm; matrix = 160 × 192; acceleration factor = 2; bandwidth = 250 kHz, and duration of examination = 3 min 20 s.

Image post-processing and analysis

IVIM sequence data were transferred to an AW4.6 workstation (GE Healthcare). Based on IVIM theory, the bi-exponential model was described by the following equation (9):

\begin{matrix} S_{b} / S_{0} = (1 - f) * exp (- b * {ADC}_{slow}) \\ + f * exp (- b * ({ADC}_{fast} + {ADC}_{slow})) \end{matrix}

where S_b is the signal intensity of the pixel with diffusion gradient b, S₀ is the signal intensity of the pixel without a diffusion gradient (b = 0 s/mm²), f is the fraction of perfusion related to the microcirculation, ADC_slow is the true diffusion coefficient as reflected by pure water molecular diffusion, and ADC_fast is the pseudo-diffusion coefficient representing perfusion.

Measurements on MRI

The angle of hip abduction, which is formed by the long axis of the femoral shaft and the center line of the body, was measured on coronal T2-weighted (T2W) images with the maximal epiphysis (Fig. 2). The elliptical regions of interest (ROIs) were placed on two or three consecutive slices of epiphyses of femoral heads. Based on the volumes of the epiphyses, the maximal epiphysis image was set as the central slice on the b = 0 DWI image. The ROIs were carefully placed to avoid the cartilage of the epiphyses, the physis, and synovial fluid. The areas of the ROIs were 25–57 mm². Two or three ROIs in each epiphysis, resulting in a total of four or six ROIs in both hip joints, were obtained for each patient, and mean values and standard variations were calculated for each epiphysis. ADC_slow, ADC_fast, f, and ADC_fast×f were measured for each ROI (Fig. 2).

Fig. 2.

A 12-month-old girl with developmental dysplasia of the left hip who had undergone closed reduction and spica casting. (a, b) Coronal least-squares estimation (IDEAL) proton density-weighted (PDW) images and T2W images of the hips. (c) Traverse IDEAL PDWI of the hips, which shows that both hips were reduced and abducted, the joint space of the left hip was broadened relative to the control side, and the epiphysis of the left hip was smaller than that on the control side. (d) Transverse DWI with b = 0 s/mm². The ROIs were placed on the epiphyses of both hips. (e–h) Parameter diagrams showing the ADC_slow, ADC_fast, f, and ADC_fast×f values of the epiphyses of both hips. The ADC_fast, ADC_fast×f, and f values of the epiphysis of the left hip were lower than those of the right hip were, and the ADC_slow value of the epiphysis of the left hip was higher than that of the right hip.

Statistical analysis

A Mann–Whitney U test was used to compare the angle of hip abduction and the ADC_slow, ADC_fast, f, and ADC_fast×f values of the ROIs between the lesion and control sides of the hip. Intraclass correlation coefficient (ICC) estimates and their 95% confidence intervals (CIs) were calculated based on a mean-rating (k = 3), absolute-agreement, two-way random-effects model for the inter-observers, and an absolute-agreement, two-way mixed-effects model for the intra-observer. ICCs were calculated for both sides. ICC values < 0.5 are indicative of poor reliability, values in the range of 0.5–0.75 indicate moderate reliability, values of 0.75–0.9 indicate good reliability, and values >0.90 indicate excellent reliability (29). Three radiologists (MXH, ZXN, and WZ, with six, 30, and 28 years of experience in musculoskeletal radiology, respectively) measured the parameters for all patients. One radiologist (MXH) repeated the measurements three times on three consecutive days. Receiver operating characteristic (ROC) curves were generated with respective cut-off values based on the Youden index. ADC_slow, ADC_fast, f, and ADC_fast×f were compared among lesions with different IHDI classification grades using analysis of variance and least significant difference (LSD) tests. Correlations of ADC_slow, ADC_fast, f, and ADC_fast×f with IHDI grades were assessed using Spearman rank correlation. SPSS (version 23.0; SPSS Inc.; Chicago, IL, USA) was used for statistical analysis. Two-tailed P values < 0.05 were considered significant.

Results

Patient demographics

Thirty-three patients with DDH who had undergone closed reduction underwent MR examinations of the hip joints. Five of these patients were excluded due to movement artifacts during the MR scan. Therefore, 28 patients (3 boys, 25 girls) were enrolled in the study. The patients’ ages were in the range of 6–24 months (average age = 16.14 ± 7.62 months). Fourteen patients had DDH of the left hip, seven had DDH of the right hip, and seven had DDH of both hips. The angle of hip abduction on the lesion side was in the range of 52.1–61.3° and the average angle was 57.89 ± 2.72°; on the control side, the range of angle of hip abduction was 52.3–61.6° and the average angle was 57.44 ± 3.06°. There was no significant difference in abduction angle between the two sides (Z = –0.302, P = 0.762).

Reliability of ROC analyses

The intra-observer ICC for ADC_slow, ADC_fast, f, and ADC_fast×f on the lesion side was 0.989 (95% CI = 0.882–0.999, F = 11.809, P = 0.004), 0.995 (95% CI = 0.946–0.999, F = 10.547, P = 0.006), 0.992 (95% CI = 0.931–0.999, F = 8.820, P = 0.009), and 0.989 (95% CI = 0.889–0.999, F = 11.046, P = 0.005), respectively. The intra-observer ICC for ADC_slow, ADC_fast, f, and ADC_fast×f on the control sides was 0.994 (95% CI = 0.969–0.999, F = 3.272, P = 0.092), 0.995 (95% CI = 0.980–0.999, F = 0.637, P = 0.554), 0.994 (95% CI = 0.973–0.999, F = 0.913, P = 0.439), and 0.994 (95% CI = 0.975–0.999, F = 0.837, P = 0.468), respectively. The inter-observer ICC for ADC_slow, ADC_fast, f, and ADC_fast×f on the lesion side was 0.979 (95% CI = 0.931–0.994, F = 4.158, P = 0.033), 0.973 (95% CI = 0.860–0.994, F = 11.948, P = 0.000), 0.987 (95% CI = 0.806–0.998, F = 57.164, P = 0.000), and 0.973 (95% CI = 0.806–0.994, F = 18.969, P = 0.000), respectively. The inter-observer ICC for ADC_slow, ADC_fast, f, and ADC_fast×f on the control side was 0.987 (95% CI = 0.801–0.999, F = 24.212, P = 0.000), 0.992 (95% CI = 0.815–0.999, F = 75.463, P = 0.000), 0.989 (95% CI = 0.951–0.999, F = 1.568, P = 0.266), and 0.993 (95% CI = 0.911–0.989, F = 15.808, P = 0.002), respectively.

IVIM parameter differences between the lesion and control sides

Mean ADC_slow was 0.000515 ± 0.00015 mm²/s on the lesion side, higher than on the control side (0.000300 ± 0.00009, Z = –4.866, P = 0.000). ADC_fast, f, and ADC_fast×f were lower on the lesion side than on the control side (ADC_fast: 0.0283 ± 0.013 mm²/s vs. 0.0536 ± 0.076 mm²/s, Z = –5.475, P = 0.000; f: 0.438 ± 0.090 vs. 0.521 ± 0.090, Z = –3.174, P = 0.002; ADC_fast×f: 0.013 ± 0.007 mm²/s vs. 0.028 ± 0.007 mm²/s, Z = –5.340, P = 0.000).

ROC curves and optimal cut-off values

ADC_fast was the parameter with the best predictive ability, with an area under the curve (AUC) of 0.940 (P = 0.000). It was followed by ADC_fast×f (AUC =0.929, P = 0.000), ADC_slow (AUC = 0.891, P = 0000) and f (AUC = 0.755, P = 0.002). The optimal cut-off values for ADC_fast, ADC_fast×f, ADC_slow, and f were 0.0626, 0.030, 0.000251, and 0.636, respectively (Fig. 3).

Fig. 3.

ROC curves for ADC_slow, ADC_fast, f, and ADC_fast×f. ADC_fast was the parameter with the best predictive ability, with an AUC of 0.940 (P = 0.000), followed by ADC_fast×f, which had an AUC of 0.929 (P = 0.000). ADC_slow and f had lower predictive ability, with AUCs of 0.891 (P = 0000) and 0.755 (P = 0.002), respectively. The optimal cut-off value for ADC_fast was 0.0626, which resulted in a sensitivity of 0.97 and a specificity of 0.91. The cut-off value for ADC_fast×f was 0.030, which resulted in a sensitivity of 0.97 and a specificity of 0.62. The cut-off value for ADC_slow was 0.000251, which led to a sensitivity of 0.97 and a specificity of 0.67. The cut-off value for f was 0.636, resulting in a sensitivity of 0.97 and a specificity of 0.95.

Comparisons among lesions with different IHDI classifications

ADC_slow, ADC_fast, f, and ADC_fast×f were significantly different among grade II, III, and IV hips (Table 2). LSD tests revealed that ADC_slow was not different between grade III and IV (P = 0.377), and that it was higher in grade II and IV than in grade II (P = 0.000). ADC_fast, f, and ADC_fast×f showed significant differences between grades II and III, grades II and IV, and grades III and IV (P = 0.000–0.017). Higher IHDI grades were associated with decreased ADC_fast, f, and ADC_fast×f, and increased ADC_slow (Fig. 4).

Table 2.

Comparisons of IVIM parameters on the lesion side among patients with DDH lesions with IHDI classification grades II–IV using ANOVA.

	ADC_slow*	ADC_fast*	f*	ADC_fast×f*
Grade II (n = 12)	0.00037 ± 0.00008^†	0.042 ± 0.010^†	0.51 ± 0.07^†	0.021 ± 0.005^†
Grade III (n = 10)	0.00057 ± 0.00009^‡	0.028 ± 0.004^‡	0.45 ± 0.07^‡	0.012 ± 0.003^‡
Grade IV (n = 13)	0.00061 ± 0.00012^‡	0.017 ± 0.003^§	0.36 ± 0.05^§	0.006 ± 0.002^§
F values	20.800	45.340	17.791	62.130
P values	0.000	0.000	0.000	0.000

P < 0.05, indicating significant statistical difference among the three groups.

†

used to annotate the values in grade II, if the values in grade III or grade IV had no statistical differences from that in grade II, they also were annotated by this symbol.

‡

used to annotate the values in group III or group IV which had statistical differences from that in group II, and if the values in group III and group IV had no statistical differences, the values in group IV were annotated by this symbol too.

used to annotate the values in group IV which had statistical differences from that in group II and group III.

IVIM, intravoxel incoherent motion; IHDI, International Hip Dysplasia Institute, DDH, developmental dysplasia of the hip; ANOVA, analysis of variance; ADC, apparent diffusion coefficient.

Fig. 4.

Correlations of ADC_slow (a)_, ADC_fast (b), f (c), and ADC_fast×f (d) with IHDI classification grades for developmental dysplasia of the hip. Higher IHDI grades were associated with decreased ADC_fast, f, and ADC_fast×f, and increased ADC_slow.

Discussion

The ICC values for the different IVIM parameters revealed good to excellent intra- and inter-observer consistency. ADC_slow was higher, and ADC_fast, f, and ADC_fast×f were lower in epiphyses of femoral heads on the lesion sides. ADC_fast×f had the best ability to detect ischemia in epiphyses of femoral heads. Higher IHDI grades were associated with decreased ADC_fast, f, and ADC_fast×f, and increased ADC_slow.

Ischemia of femoral heads remains a severe sequela in patients with longstanding DDH (1), and is a serious complication of treatment with closed reduction (3). AVN is more likely to occur in patients with insults to the blood supply of the femoral heads after closed reduction (30,31). Necrosis of femoral heads results from occlusion of the femoral circumflex arteries and small vessels supplying the femoral heads (3,32). The degree of hip abduction is an important factor for future AVN of femoral heads. Hip abduction angles of 55–60° are generally considered to be the “safe area.” Excessive abduction may lead to ischemia to the femoral heads and AVN of the epiphyses (3,4,33). In our study, the hip abduction angles were not different between the lesion and control sides, and both within normal limits. Therefore, excessive abduction of the hip in patients undergoing closed reduction may not explain the ischemia of the epiphyses in our study.

Evaluation of the blood supply of the epiphyses of femoral heads to prevent AVN is a major issue for clinicians. IVIM sequences can be used to distinguish intravascular perfusion from pure water diffusion, and simultaneously acquire perfusion and diffusion data. Contrast-enhanced MRI is currently widely used to evaluate perfusion in femoral heads (2). In patients with impeded blood supply, a global non-enhancement pattern in epiphyses of femoral heads has been associated with a higher incidence of AVN (33). Sebag et al. (8) referred to ischemia as a widespread absence of enhancement, using dynamic gadolinium-enhanced subtraction MRI. These findings are concordant with our results. In our study, the parameters for perfusion were significantly decreased in the femoral heads in DDH lesions. The severity of the lesions was higher in patients with higher IHDI classifications. This suggests that ischemia occurs in the femoral heads in DDH, and that patients with higher IHDI grades had more severe ischemia.

Some researchers have evaluated diffusion in femoral heads in patients with Legg–Calve–Perthes (LCP) disease, which also occurs due to occlusion of blood flow to femoral heads in children, using DWI sequences. Yoo et al. (34) reported that epiphyseal diffusion increased early and remained elevated through the healing stage in LCP lesions. Ozel et al. (35) and Merlini et al. (36) also found increased mean ADC values in femoral heads in patients with LCP disease. In our study, ADC_slow, which reflects the diffusion of water molecules, was higher on the lesion sides than the control sides. ADC_slow values were higher in patients with high IHDI grades. However, the increase in ADC was due to AVN in femoral heads in patients with LCP disease, while ADC_slow values in our study increased in the absence of obvious necrosis of femoral heads, which requires further investigation.

Our study has some limitations. First, the patients we enrolled had had hip dislocation for a long time and we obtained the IVIM sequence values after closed reduction. Therefore, we are unable to determine whether the higher ADC_slow value and lower ADC_fast×f, ADC_fast, and f values in lesion sides were due to the longstanding state of the dislocation or to the closed reduction operation. In the future, we will compare the IVIM parameters before and after close reduction in patients with DDH. We will then compare these parameters among normal subjects, patients with ischemic femoral epiphyses, and the patients with AVN in the femoral epiphyses. Second, we treated the normal hip joint in patients with DDH as the control side instead of using healthy children as controls. This may have led to bias in the measurements. However, we were unable to find healthy children whose parents were willing to allow them to wear spica castings and undergo chloral hydrate injections. Third, we did not analyze correlations between IVIM findings and post-enhanced images in patients with DDH after closed reduction. We will perform such studies in the future.

In conclusion, lower ADC_fast, f, and ADC_fast×f, and higher ADC_slow values suggest ischemia of epiphyses of femoral heads in patients with DDH after closed reduction. IVIM is a promising method to investigate perfusion and diffusion in epiphyses of femoral heads.

Footnotes

Acknowledgments

The authors thank Dan Dan Zheng from General Electric for her contribution in the optimization of the sequences and language help. The authors also thank Editage () for English language editing.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Bracken

Tran

Ditchfield

. Developmental dysplasia of the hip: controversies and current concepts. J Paediatr Child Health 2012; 48: 963–973.

Rosenbaum

Servaes

Bogner

et al.

MR Imaging in postreduction assessment of developmental dysplasia of the hip:goals and obstacles. Radiographics 2016; 36: 840–854.

Salter

Kostuik

Dallas

. Avascular necrosis of the femoral head as a complication of treatment for congenital dislocation of the hip in young children: a clinical and experimental investigation. Can J Surg 1969; 12: 44–61.

Jaramillo

Villegas-Medina

Doty

et al.

Gadolinium-enhanced MR imaging demonstrates abduction-caused hip ischemia and its reversal in piglets. Am J Roentgenol 1996; 166: 879–887.

Kishore

Shah

et al.

Diagnosing developmental dysplasia of hip in newborns using clinical screen and ultrasound of hips-an Indian experience. J Trop Pediatr 2016; 62: 241–245.

Kim

HKW

Kaste

Dempsey

et al.

A comparison of non-contrast and contrast-enhanced MRI in the initial stage of Legg-Calvé-Perthes disease. Pediatr Radiol 2013; 43: 1166–1173.

Dempsey

et al.

MR perfusion index as a quantitative method of evaluating epiphyseal perfusion in Legg-Calve-Perthes disease and correlation with short-term radiographic outcome: a preliminary study. J Pediatr Orthop 2013; 33: 707–713.

Sebag

Ducou

LPH

Klein

et al.

Dynamic gadolinium-enhanced subtraction MR imaging–a simple technique for the early diagnosis of Legg-Calve-Perthes disease: preliminary results. Pediatr Radiol 1997; 27: 216–220.

Le Bihan

Breton

Lallemand

et al.

MR imaging of intravoxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology 1986; 161: 401–407.

10.

Zhang

et al.

Intravoxel incoherent motion MR imaging. Medicine 2015; 94: e1028–e1028.

11.

Wurnig

Donati

Ulbrich

et al.

Systematic analysis of the intravoxel incoherent motion threshold separating perfusion and diffusion effects: proposal of a standardized algorithm. Magn Reson Med 2015; 74: 1414–1422.

12.

Hwang

Lee

Yoon

et al.

Intravoxel incoherent motion diffusion-weighted imaging of pancreatic neuroendocrine tumors: prediction of the histologic grade using pure diffusion coefficient and tumor size. Invest Radiol 2014; 49: 396–402.

13.

Scott

Ferreira

Nielles-Vallespin

et al.

Optimal diffusion weighting for in vivo cardiac diffusion tensor imaging. Magn Reson Med 2015; 74: 420–430.

14.

Chung

Lee

Kim

et al.

Intravoxel incoherent motion MRI for liver fibrosis assessment: a pilot study. Acta Radiol 2015; 56: 1428–1436.

15.

Wang

Lin

Liu

et al.

Intravoxel incoherent motion diffusion-weighted MR imaging in differentiation of lung cancer from obstructive lung consolidation: comparison and correlation with pharmacokinetic analysis from dynamic contrast-enhanced MR imaging. Eur Radiol 2014; 24: 1914–1922.

16.

Lai

Lee

VHF

et al.

Nasopharyngeal carcinoma: comparison of diffusion and perfusion characteristics between different tumour stages using intravoxel incoherent motion MR imaging. Eur Radiol 2014; 24: 176–183.

17.

Sumi

Nakamura

. Head and neck tumours: combined MRI assessment based on IVIM and TIC analyses for the differentiation of tumors of different histological types. Eur Radiol 2014; 24: 223–231.

18.

Pang

Turkbey

Bernardo

et al.

Intravoxel incoherent motion MR imaging for prostate cancer: an evaluation of perfusion fraction and diffusion coefficient derived from different b-value combinations. Magn Reson Med 2013; 69: 553–562.

19.

Federau

Hagmann

Maeder

et al.

Dependence of brain intravoxel incoherent motion perfusion parameters on the cardiac cycle. PLoS One 2013; 8: e72856–e72856.

20.

Chow

Gao

Fan

et al.

Liver fibrosis: An intravoxel incoherent motion (IVIM) study. J Magn Reson Imaging 2012; 36: 159–167.

21.

Lee

Liau

Murphy

et al.

Cross-sectional investigation of correlation between hepatic steatosis and IVIM perfusion on MR imaging. Magn Reson Imaging 2012; 30: 572–578.

22.

Federau

Maeder

O’Brien

et al.

Quantitative measurement of brain perfusion with intravoxel incoherent motion MR imaging. Radiology 2012; 265: 874–881.

23.

Müller

Prasad

Edelman

. Can the IVIM model be used for renal perfusion imaging? Eur J Radiol 1998; 26: 297–303.

24.

Gaeta

Benedetto

Minutoli

et al.

Use of diffusion-weighted, introvoxel incoherent motion, and dynamic contrast-enhanced MR Imaging in the assessment of response to fadiotherapy of lytic bone metastases from breast cancer. Acad Radiol 2014; 21: 1286–1293.

25.

Marchand

Hitti

Monge

et al.

MRI quantification of diffusion and perfusion in bone marrow by intravoxel incoherent motion (IVIM) and non-negative least square (NNLS) analysis. Magn Reson Imaging 2014; 32: 1091–1096.

26.

Zhao

Liu

et al.

Detection of active sacroiliitis with ankylosing spondylitis through intravoxel incoherent motion diffusion-weighted MR imaging. Eur Radiol 2015; 25: 2754–2763.

27.

Nguyen

Ledoux

Omoumi

et al.

Application of intravoxel incoherent motion perfusion imaging to shoulder muscles after a lift-off test of varying duration. NMR Biomed 2016; 29: 66–73.

28.

Narayanan

Mulpuri

Sankar

et al.

Reliability of a new radiographic classification for developmental dysplasia of the hip. J Pediatr Orthop 2015; 35: 478–484.

29.

Koo

. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med 2016; 15: 155–163.

30.

Gornitzky

Georgiadis

Seeley

et al.

Does perfusion MRI after closed reduction of developmental dysplasia of the hip reduce the incidence of avascular necrosis?

Clin Orthop Relat Res 2016; 474: 1153–1165.

31.

Tiderius

Jaramillo

Connolly

et al.

Post-closed reduction perfusion magnetic resonance imaging as a predictor of avascular necrosis in developmental hip dysplasia: a preliminary report. J Pediatr Orthop 2009; 29: 14–20.

32.

Kruczynski

. Avascular necrosis of the proximal femur in developmental dislocation of the hip: Incidence, risk factors, sequelae and MR imaging for diagnosis and prognosis. Acta Orthop Scand Suppl 1996; 268: 1–48.

33.

Jaramillo

Villegas-Medina

Laoret

et al.

Gadolinium-enhanced MR imaging of pediatric patients after reduction of dysplastic hips: assessment of femoral head position, factors impeding reduction, and femoral head ischemia. Am J Roentgenol 1998; 170: 1633–1637.

34.

Yoo

Kim

Menezes

et al.

Diffusion-weighted MRI reveals epiphyseal and metaphyseal abnormalities in Legg–Calve–Perthes disease: a pilot study. Clin Orthop Relat Res 2011; 469: 2881–2888.

35.

Ozel

Ozkan

et al.

Diffusion-weighted magnetic resonance imaging of femoral head osteonecrosis in two groups of patients: Legg–Perthes–Calve and avascular necrosis. Radiol Med 2016; 121: 206–213.

36.

Merlini

Combescure

De Rosa

et al.

Diffusion-weighted imaging findings in Perthes disease with dynamic gadolinium-enhanced subtracted (DGS) MR correlation: a preliminary study. Pediatr Radiol 2010; 40: 318–325.