Quantitative Analysis of Deformity in Digital Model of Congenital Radioulnar Synostosis

Abstract

Objective

The deformity of congenital radioulnar synostosis is quite complicated and difficult. This study aims to find out the related factors of the “forearm rotation angle” (FR) which relate to the severity of congenital radioulnar synostosis (CRUS), and try to quantify the internal relations of each deformity and help to understand the reconstruction method in surgery treatment of this disease.

Methods

This study is case series research. We established 48 digital three‐dimensional forearm bone models of 48 patients with congenital radioulnar synostosis classified as Cleary and Omer type 3. All the patients were treated at our institution from January 2010 to June 2016. In total, 10 independent deformities (the rotation angle of forearm; the internal rotation, radial, and dorsal angulation of radius and ulna; the relative length of osseous fusion at PRUJ; the relative dislocation distance of distal radioulnar joint; the relative area of proximal radial epiphysis) involved in the CRUS complex deformity were measured. Pearson correlation analysis for each deformity which was mentioned above was performed, and multivariate linear regression analysis was also performed with FR as the dependent variable and the other deformities as the influential factors.

Results

The “dorsal angle of radius” (DAR, 21.69° ± 21.55°) had the strongest correlation with the FR (79.72° ± 40.39°), the Pearson correlation coefficient was 0.601 (p < 0.01), the internal rotation angle of the radius (IRAR, 82.69° ± 54.98°) had a moderate correlation with FR, the Pearson correlation coefficient was 0.552 (p < 0.01). A forearm deformity equation was established: FR = 35.896 + 0.271 DAR + 0.989 IRAR.

Conclusion

The dorsal angulation deformity of radius may be the most important deformity that effects the severity of CRUS and should be correct in the first place during reconstruction operation.

The digital evaluation of congenital radioulnar synostosis, what matters the most?

Introduction

Congenital radioulnar synostosis (CRUS) is a rare congenital deformity of the upper extremity, which usually presents as a pronation deformity of the forearm, and can be bilateral or unilateral. ¹ , ² , ³ This disease may cause difficult experiences for affected children in participating in activities like perineum sanitation and using utensils, thus leading them to develop a variety of psychological problems as they grow up. In severe cases, the patient cannot complete movements, such as carrying a bowl and writing, which affect the patient's daily life, work, social activities, and hobbies. ⁴

Cleary and Omer's research ⁵ classified the disease into four types according to the radial head dislocation and the osseous synostosis condition at the proximal radioulnar joint (PRUJ) on plain radiographs. The most common type is type 3, ⁶ , ⁷ which is characterized by the posterior dislocation of the radial head. Kanaya et al. treated the patients by reconstructing the shape of PRUJ and proximal radius, and overcame the problem of the high rate of postoperative re‐ankylosis with the traditional method. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³

Although the surgical treatment seems to be the only choice to correct the abnormal structure, the results were limited by the remaining deformity postoperatively. However, the outcome is not always satisfying. According to Oka and Sakamoto's study, the postoperative radial curvature is rather difficult to be shaped properly, which leads to a limited postoperative forearm rotation function. ¹⁴ , ¹⁵ , ¹⁶ As the bone structure of CRUS forearm is not as simple as we imagine (Figure 1), ¹⁷ , ¹⁸ , ¹⁹ it is hard to correct every abnormal structure by performing this procedure. And the postoperative bony deformities might be important influential factors. ¹⁴ , ¹⁵ , ¹⁶

FIGURE 1.

Abnormal osseous structures of CPRUS. (A–C) radial and ulnar lateral angulation deformity; (D–F) deformity at proximal part of radius and ulna; (G, H) internal rotation deformity of radius and ulna; (I–K) deformity at distal part of radius and ulna; (L) osseous synostosis of the PRUJ

Previous studies ¹⁷ , ¹⁸ , ¹⁹ have shown that the deformity of the CRUS forearm is complex, which consists of several independent deformities, including radial deformities, ulnar deformities, deformities at the proximal part of radius and ulna, and the deformities at the distal part of radius and ulna, among others. These independent deformities are associated with the severity of fixed forearm rotation deformity. As previous researchers always felt it to be troublesome to correct the shape of radius, we believe that as it is such a challenge to reconstruct the deformity, there should be something we have not recognized in it.

Therefore, it is very important to determine the bony malformation relationship of the CRUS forearm. ¹⁹ However, there is no relevant research investigating which of these independent deformities is the most important factor affecting the fixed forearm rotation deformity. Also, the mechanisms on how the independent deformities related to the forearm pronation deformity and how they relate to each other have not been determined. ²⁰ Thus, we aimed to describe the deformities of CRUS more reliably by using digital three‐dimensional (3D) reconstruction technology, and to investigate the degree of correlation of each individual deformity perimeter on the forearm pronation deformity and the correlation among the individual deformity perimeters, so as to provide a theoretical basis for improving the surgical treatment strategy.

Methods

This case series study obtained approval from the ethical review board of our institution, and all of the included patients provided informed consent. The experimental protocol was established according to the ethical guidelines of the Helsinki Declaration. The inclusion criteria were as follows: (1) treated at our institution from January 2010 to June 2016, (2) classified as having type 3 deformities according to Cleary and Omer classification. The exclusion criteria were as follows: (1) patients who were unwilling to undergo the computed tomography (CT) examinations, (2) patients who were unable to complete computed tomography (CT) examinations.

Altogether 48 forearms of 48 CRUS patients (36 males and 12 females; age range: 1 to 28 years old) were included in the analysis. There were 15 and 33 patients with bilateral and unilateral involvement, respectively; of these, 22 and 11 patients had a deformity on the left and right sides, respectively. CRUS patients with a type 3 deformity on one side and another type on the other side were considered as having unilateral involvement.

Preoperative bilateral forearm CT examination was performed. During the examination, the patient was placed in the prone position, with both forearms raised, palms down, and elbows slightly flexed, and the forearms were scanned simultaneously. For bilateral cases, one side was randomly selected to be included in the study. Among the 15 bilateral cases, eight right and seven left forearms were included in the study. Digital model reconstruction software (Mimics Medical 20.0, materialize NV Technologielaan, Belgium) was used to conduct digital 3D modeling using the DICOM files of forearm CT. The independent osseous deformities of CRUS were as follows ¹⁷ , ¹⁸ , ¹⁹ : (1) the radial diaphysis is in an arch shape, and its deformity is characterized by internal rotation, radial, and dorsal angulations; (2) the ulnar diaphysis deformity is characterized by internal rotation, ulnar, and palmar angulations; (3) osseous fusion of the PRUJ; (4) posterior dislocation of the radial head with a dysplastic and abnormal shape; (5) abnormal alignment of the humeroradial joint; and (6) dorsal dislocation of the ulnar head at the distal radioulnar joint (DRUJ).

The ulna and radius were separated according to the physiological traces at the proximal osseous synostosis part of CRUS in mimics medical 20.0 software, and the relevant data of the independent deformities were measured independently by three hand surgeons specializing in treating CRUS, and the average of each index in each sample was used for statistical analysis.

Forearm Rotation Angulation Deformity

The angle between the axis of the humerus and the projection of the connecting line of the radial styloid process and the middle point of the ulnar head on the transverse plane was used as the forearm rotation (FR) angle (Figure 2).

FIGURE 2.

Measurement method of the forearm rotation angle. Angle A represented the forearm rotation (FR) angle

Radial Deformities

In unilateral cases, the digital 3D model of the bilateral radius was established, the model of the healthy side was reversed on the sagittal plane, and its distal articular surface and that of the affected side were superimposed using the interactive closest‐point algorithm. In the bilateral cases, the model of the ipsilateral healthy radius from people of similar age was established and rescaled, and its distal articular surface was superimposed with the affected radius using the interactive closest‐point algorithm. Then, the internal rotation, radial, and dorsal angulation deformities were measured. Given that the deformities were multi‐plane, the measurement methods were standardized as follows: (1) the internal rotation angle of radius (IRAR): given that the landmarks at the proximal ulnar part of the radius in CRUS patients disappeared due to the osseous fusion, IRAR was measured on the transverse plane of the healthy radius, and the angle between the perpendicular line of the projection of the cutting edge at the proximal part of the affected radius and the connecting line of the incisura ulnaris and the center of the radial head of the healthy radius was used as the IRAR; (2) radial angle of the radius (RAR): given that the distal articular surfaces were superimposed, the projection of the angle between the proximal axis of the affected radius and the healthy radius on the coronal plane of the healthy radius was used as the RAR; and (3) dorsal angle of the radius (DAR): given that the distal articular surfaces were superimposed, the projection of the angle between the proximal axis of the affected radius and the healthy radius on the sagittal plane of the healthy radius was used as the DAR (Figure 3).

FIGURE 3.

Measurement methods of radial deformities. (A) Distal radial articular surfaces were superimposed; (B–D) the arrow, star, and triangle represent the RAR, DAR, IRAR, respectively. (D) The rectangle and pentagon represent the projection of the cutting edge at the proximal part of the affected radius and the incisura ulnaris, respectively

Ulnar Deformities

Similar to the measurement methods of the radial deformities, the model of the healthy side of the patient or the modified model of the ulna of the healthy person was superimposed at the humeroulnar articular surface and olecranon using the interactive closest‐point algorithm. Given that the proximal parts were superimposed, the internal rotation, ulnar, and palmar angulation deformities of the ulna were measured using the following standardized methods: (1) internal rotation angle of the ulna (IRAU): the projection of the angle between the connecting line between the center of the ulnar head and ulnar styloid process of the affected ulna and the healthy ulna on the transverse plane of the healthy ulna was used as the IRAU; (2) ulnar angle of the ulna (UAU): the projection of the angle between the axis of the affected ulna and the healthy ulna on the coronal plane of the healthy ulna was used as the UAU; and (3) palmar angle of the ulna (PAU): the projection of the angle between the axis of the affected ulna and the healthy ulna on the sagittal plane of the healthy ulna was used as the PAU (Figure 4).

FIGURE 4.

Measurement methods of ulna deformities. (A) Humeroulnar articular surfaces and olecranon were superimposed; (B–D): the arrow, star, and triangle represent the UAU, PAU, and IRAU, respectively

Relative Length of Osseous Synostosis (RLOS)

The RLOS was measured as the ratio of the longest diameter of the fusion site of the affected ulna to the distance between the center of the ulnar head and the midpoint of the olecranon tip (Figure 5).

FIGURE 5.

Measurement method of the RLOS (D1/D2). (D1) the longest diameter of the fusion site (removed) of the affected ulna; (D2) the distance between the center of the ulna head and the midpoint of the olecranon tip

Proximal Metaphysis of Radius

Given that the cartilage of the radial head cannot be contoured clearly in the CT images, the relative transverse area of the proximal radial metaphysis (RAPAM) was chosen as the approximate indicator of the development of the radial head. Moreover, the RAPAM was measured as the ratio of the transverse area of the proximal metaphysis of the affected radius to the transverse area of the proximal metaphysis of the healthy radius (Figure 6).

FIGURE 6.

Measurement method of the RAPAM (A1/A2). (A1) The transverse area of the proximal metaphysis of the affected radius; (A2) the transverse area of the proximal metaphysis of the healthy radius

The Relative Distance of Ulnar Head Dislocation (RDUD)

Ulnar head dislocation can be classified as palmar and dorsal dislocations, and was observed and classified. The RDUD was measured as the ratio of the projection of the vertical distance between the palmar margin of the ulnar head and the extension line of the radial palmar margin to the projection of the length of the sigmoid notch on the transverse plane (Figure 7).

FIGURE 7.

Measurement method of the RDUD (L1/L2). (L1) The vertical distance between the palmar margin of the ulnar head and the extension line of the radial palmar margin; (L2) the length of the sigmoid notch; (L3) the extension line of the radial palmar margin; UH: ulnar head; R: radius

Statistical Analysis

IBM SPSS Statistics 26.0 (IBM Corp, America) was used for statistical analysis. The distribution of each deformity perimeter and the Pearson correlation among the perimeters of normal distribution was analyzed. And the coefficient of consistency was examined. The scatterplot between the FR and the other deformity perimeters was mapped to eliminate the index that was not linearly correlated with the FR. The analysis of variance (ANOVA) of each deformity index was performed to eliminate the index with unequal variances. The multivariate linear regression equation with FR as the dependent variable and the residual deformity perimeters as the influential factors was established to analyze how the independent perimeters related to the FR.

Results

The Overall Introduction

The data of each deformity perimeter were normally distributed. The ICC was 0.68, which shows a moderate consistency considering the observers' bias.

The average age and the median age of samples was 9.9 ± 4.9 years old and 8.2 years old, respectively, only eight samples were acquired from patients over 13 years old and two of them were over 20 years old.

The FR was 79.72°on average, and was similar to the pervious study. ⁵ For the radial deformities, the IRAR had the most severe deformity degree compare to RAR and DAR, with 82.69° on average. For the ulnar deformities, the deformity degree of IRAU was 56.33° on average, and was the worst compared to UAU and PAU. The RLOS was 0.21 on average, which may suggest that the synostosis always stop at the proximal 1/5 of ulna. The RAPAM was less than half of the normal area with 0.40 on average. Regarding the ulnar head dislocation and the RDUD at DRUJ, all of the ulnar heads showed a dorsal dislocation, and the RDUD was 0.53 on average (Table 1).

TABLE 1.

Average data of each perimeter (N = 48)

Perimeter	Average	Maximum	Minimum
FR	79.72°	120.11°	21.55°
RAR	13.20°	33.93°	0.55°
DAR	21.69°	43.24°	0.02°
IRAR	82.69°	176.76°	27.71°
UAU	5.46°	18.48°	−7.68°
PAU	10.20°	40.64°	−13.82°
IRAU	56.33°	159.04°	10.15°
RLOS	0.21	0.48	0.14
RAPAM	0.40	0.70	0.180
RDUD	0.53	1.08	0.11

Perimeters with Strong Pearson Correlation with FR

Pearson correlation analysis results show that the FR had a strong correlation with the DAR (0.601, p < 0.01), a moderate correlation with the IRAR (0.552, p < 0.01) and IRAU (0.421, p < 0.01), and a low correlation with the PAU (0.286, p = 0.048), whereas the FR had no correlation with the RAR, UAU, RLOS, RAPAM, and RDUD (Table 2).

TABLE 2.

Pearson correlation analysis between FR and other perimeters (N = 48)

Perimeter	FR
	r	p
DAR	0.601	<0.001
IRAR	0.552	<0.001
IRAU	0.421	0.003
PAU	0.286	0.048
RLOS	0.259	0.076
UAU	0.230	0.116
RAR	0.040	0.789
RAPAM	−0.276	0.058
RDUD	0.238	0.103

Perimeters with Low/Moderate Correlation

The IRAR had a low correlation with the DAR (0.347, p = 0.016). The UAU had a moderate correlation with the PAU (9.543, p < 0.01), whereas the PAU had a low correlation with the IRAU (0.378, p < 0.01).

The Multiple Linear Regression Equation of CRUS

A scatterplot between each deformity and the FR was mapped (Figure 8). It can be seen that there was no linear relationship between the RAPAM and the FR. The homogeneity of variance between each perimeter and the FR was tested, and the RDUD was eliminated. Taking the FR as the dependent variable and the other residual deformity perimeters as the influencing factors, the multiple linear regression equation was established (Table 3). The results showed that the forearm deformity equation was as follows: FR = 35.896 + 0.989 DAR + 0.271 IRAR. In our opinion, this equation suggested that the deformities on the coronal and transverse plane may be the primary cause of this complex radius shape. And maybe the DAR and IRAR should be corrected in the first place during the operation.

FIGURE 8.

Scatterplot between each deformity perimeter and the FR. Y axis in each scatterplot was the FR. (A) The blue dots, the green dots, the red dots represented the UAU, PAU, IRAU, respectively. (B) The blue dots, the green dots, the red dots represented the RAR, DAR, IRAR, respectively. (C–E) X axis in each scatterplot was the RLOS, RAPAM, RDUD, respectively

TABLE 3.

Multiple linear regression analysis of the forearm deformity equation

	Unstandardized coefficients		t	P
	β	SE
Constant	35.896	7.065	5.081	0.000
DAR	0.989	0.240	4.125	0.000
IRAR	0.271	0.078	3.454	0.001

Note: F = 22.064, R = 0.704, adjusted R² = 0.473 (p < 0.01).

Discussion

Aims of Study

In this study, we explored the weighting perimeter of the major influential factors related to the severity of forearm rotation deformity, finding out that the DAR had the strongest correlation with FR (0.601, p < 0.01), it was further confirmed by the regression foundation (FR = 35.896 + 0.989 DAR + 0.271 IRAR), which suggested that the DAR and IRAR should be corrected in the first place during the operation. These results may be helpful for further improvement of the treatment of CRUS.

Quantify the Individual Perimeter in CRUS

The digital reconstruction technology was used to reconstruct the osseous structure of the forearms of 48 patients with type 3 deformities, and the results showed that all the patients had radial and ulnar deformities, osseous synostosis at PRUJ, and deformities at the proximal and distal part of the radius and ulna, which were consistent with the results of previous research. ⁴ , ⁵ , ¹⁷ , ¹⁸ , ¹⁹

To investigate the relationship of the deformity perimeters and the FR, the Pearson correlation test among the deformity perimeters was performed, which showed that the FR had a strong correlation with the DAR (0.601, p < 0.01), moderate correlation with the IRAR (0.552, p < 0.01) and IRAU (0.421, p < 0.01), and low correlation with the PAU (0.286, p = 0.048). The DAR had a low correlation with the IRAR (0.347, p = 0.016) and IRAU (0.369, p = 0.01). The results suggested that for the FR, the most influential deformity was the DAR, the IRAR and IRAU also had much correlation with the FR and they were not correlated with each other, the other deformities seemed to have little correlation with the FR.

Clarify the Quantitive Relationship between Severity and Individual Perimeters

Then, to qualify this relationship, the multiple linear regression analysis of each deformity perimeter was performed, which established the relationship equation between the forearm deformity severity and other independent deformity perimeters (forearm deformity equation): FR = 35.896 + 0.989 DAR + 0.271 IRAR. The results suggest that the DAR and IRAR may greatly contribute to the severity of forearm rotation deformity, compared to other independent deformities. And during the reconstruct surgery, the dorsal angulation has the priority to be corrected, followed by the inner rotation angulation of radius. Besides, for patients who are not suitable to perform reconstruction surgery (for example, the RLOS longer than 0.5), this result can also help surgeons to determine the osteotomy site during derotational osteotomy, to minimize the influence of deformity.

The Analysis of IRAU's Result

The IRAU had a moderate correlation with the FR but was excluded in the forearm deformity equation; thus, the IRAU had little correlation with the FR, which was not entirely consistent with previous research. ¹⁹ Thus, the moderate Pearson correlation between these two deformities may be a result of confounding factors.

The Analysis of RAPAM's Result

In this study, the average RAPAM was 0.4, which was less than half of the normal area. Moreover, given that the shape and development of the radial heads are far from normal in type 3 cases, which are often in the shape of a sharp cone, it is difficult to have a radial head in a correct shape after surgery. This may affect the postoperative alignment of the humeroradial joint, thereby affecting the outcomes of forearm rotation and elbow stability. This may be one of the reasons that caused the poor postoperative forearm rotation function and complications.

Strengths and Limitations

In previous studies, researchers described and analyzed the deformities of CRUS, and the complexity was already known to us. But the quantitative analysis of this deformity has not been reported. A clear numerical description can give more confidence and help the orthopaedic surgeons during operation, which is what we found in this study. Also, this study still has several limitations. First, given that the disease is a rare congenital malformation, the sample size of this study was small, the results may be biased. Second, all the cases in this study were classified as type 3, which makes this study of little importance to the other types of CRUS. Third, the age of the patients ranged from 1 to 28 years old, and the data were not grouped and analyzed by age. Due to the disease characteristic, we could not collect more related patients, and although the majority of the sample's age was as steady as possible, the results may be biased due to a large range of ages.

In summary, among the deformities caused by the CRUS, the DAR and IRAR may be the most important indicators affecting the severity of forearm rotation deformities, whereas the IRAU and PAU did not highly correlate with FR. The forearm deformity equation was as follows: FR = 35.896 + 0.989 DAR + 0.271 IRAR. The results of this study may suggest that the dorsal deformity of radius is the one most in need of correction, and might play some role in designing the treatment strategy in the future.

Authors' Contributions

Chen Yang and Lu Liu are responsible for the article writing and data analysis, Shan‐lin Chen is responsible for check and approve. Qi‐pei Wei and Fan Bai are responsible for the data collection.

Conflict of Interest Statement

The authors declare that they have no competing interests.

Availability of Data and Material

The datasets used or analyzed during the current study are available from the corresponding author on reasonable request.