Measuring the Medial Clear Space in Uninjured Adult Ankles: Agreement With CT and Interobserver Reliability of Common Measurement Methods

Abstract

Background:

The medial clear space (MCS) is a radiographic proxy for the ankle mortise used to determine the need for surgical intervention for ankle fractures. Several measurement methods have been described but no consensus exists on which method is most appropriate to guide clinical care. Our objective was to assess commonly used measurement methods to determine which was most accurate and reliable for measuring the MCS on nonweightbearing and weightbearing studies.

Methods:

Imaging of 34 morphologically normal, uninjured patients was reviewed. Paired radiographs and computed tomography (CT) scans, 17 nonweightbearing and 17 weightbearing, were measured by 5 observers. Three different methods were used to measure the MCS on each radiograph and 1 method on CT. CT was used as the “true” value of the MCS. Two-sided P tests were used to compare each radiographic MCS measurement to the CT MCS value. Cronbach alpha was calculated for each MCS measurement method to determine the interobserver correlation.

Results:

Measuring the MCS halfway between the superior medial talar dome and the tip of the medial malleolus, perpendicular to the tibiotalar joint, was the most accurate on weightbearing imaging (no difference in radiographic method 2 vs CT, P = .09). There was a significant correlation between any nonweightbearing radiograph measurement and CT (all P < .001). Interobserver reliability was found to be good to excellent for nonweightbearing MCS measurement (ICC 0.82-0.97), whereas weightbearing imaging demonstrated good reliability (ICC 0.77-0.9).

Conclusion:

Any of the 3 tested measurement methods can be used when analyzing nonweightbearing radiographs. Measuring the MCS halfway between the superior medial talar dome and the tip of the medial malleolus, perpendicular to the tibiotalar joint, should be used when examining weightbearing imaging. To support standardization, this method may be reasonable to use, particularly on weightbearing radiographs.

Level of Evidence:

Level III, retrospective comparative study.

Introduction

Ankle fractures are common injuries that represent approximately 10% of all fractures treated by orthopaedic surgeons. ⁸ These injuries occur as a result of falls, sporting accidents, and direct trauma. ²³ The true incidence of ankle fractures is unknown, but it has recently been reported to be 4.22/10 000 person years in the United States, and as high as 16.8/10 000 person years over a 10-year period in the Dutch literature.^11,23

Treatment decisions depend on the congruency of the ankle mortise. As Ramsey and Hamilton ²¹ demonstrated in their seminal work, 1 mm of lateral talar displacement can alter the contact area of the joint by 42% and lead to unfavorable long-term sequelae such as arthritis and functional impairment.

The medial clear space (MCS), defined as the distance from the lateral edge of the medial malleolus to the medial edge of the talus, is used as a radiographic proxy for competency of the deltoid ligament and ankle congruency. Measurement of the MCS is most commonly performed on a mortise radiograph with or without a stress force applied. Additionally, its measurement has also been described on cross-sectional imaging using computed tomography (CT) and more recent studies have used volumetric measurements to assess the MCS.^3,7 Widening of the ankle mortise greater than 4 or 5 mm is considered to be pathologic and an indication for surgical fixation.^4,17

Despite the importance of the MCS measurement, there exists significant variability in the way that the MCS is measured with several descriptions in the literature. Most commonly, the MCS is measured between the posteromedial border of the talus and the posterolateral border of the medial malleolus 5 mm below the talar dome.^1,7,9,13,19 Other authors have also described a measurement between the medial malleolus and medial talar dome halfway between the top of the talar dome and the tip of the medial malleolus. Alternatively, a measurement at the level of the talar dome at an angle between the corner of the talus and the axilla of the medial plafond has also been described.^{2,4 -6,10,12,15,17,18,20,25} Although less commonly used, Schuberth et al ²⁴ described a more complex measurement involving the length of a line perpendicular to the slope of the medial talus at the midpoint of the talar dome. The 3 most common methods of MCS measurement are demonstrated in Figure 1.

Figure 1.

Radiographic measurements of the medial clear space. MCS No. 1: 5 mm inferior to the medial talar dome, perpendicular to the tibiotalar joint. MCS No. 2: Halfway between the superior medial talar dome and the tip of the medial malleolus, perpendicular to the tibiotalar joint. MCS No. 3: From the corner of the medial talar dome to the axilla of the medial plafond, oblique to the tibiotalar joint.

At present, there are no data to suggest that one method of MCS measurement is superior and the data on interobserver reliability of MCS measurement is scant and outdated.^4,22 As treatment decisions are often based on measurements in increments of millimeters, determining which method is most accurate is important. Furthermore, weightbearing imaging modalities are being increasingly used to assess stability in ankle fractures. As such, our primary objective was to assess commonly used measurement methods to determine which was most accurate and reliable for measuring the MCS. Our secondary objective was to determine the comparative influence of weightbearing vs nonweightbearing imaging on the primary objective.

Methods

After institutional review board (IRB) approval, the electronic medical record (EMR) system of a quaternary-care, academic medical center was reviewed for adult (age ≥18 years) patients seen in the orthopaedic surgery clinic with ankle imaging obtained. Patients with morphologically normal anatomy, no osseous injury, and paired radiograph and CT scan imaging (both being either nonweightbearing or weightbearing) were identified. In total, 17 patients with nonweightbearing-matched imaging and 17 patients with weightbearing-matched imaging were included.

Five observers with prior experience reading radiographic and CT studies reviewed each matched pair consisting of the mortise radiographic and CT imaging for each patient. Their backgrounds ranged from clinical researcher to physician fellow. The MCS was measured using 3 different methods on radiograph and then compared with MCS measurements on the patient’s corresponding CT imaging. CT imaging was used as the comparative true value of the distance between the medial talus and medial malleolus given its superior resolution and precise identification of osseous landmarks. Imaging was reviewed using a standard Picture Archiving and Communication System (PACS; eUnity Diagnostic Viewer [Mach 7, South Burlington, VT]). Data from each patient were collated in an encrypted spreadsheet (Microsoft Excel; Microsoft Corp, Redmond, WA). After all data were compiled, all identifiable patient information was removed. In total, 17 patients, divided equally between ankle laterality (right or left) and sex were reviewed for each group (Tables 1-3).

Table 1.

Demographic Data of Both the Nonweightbearing and Weightbearing MCS Measurement Cohorts Combined. ^a

Characteristics	Value
Age	51.69 (65.25-38.50)
BMI	30.031 (33.85-24.16)
Height	1.71 (1.80-1.63)
Laterality
Left	17/34 (50)
Right	17/34 (50)
Gender
Female	17/34 (50)
Male	17/34 (50)

Abbreviations: BMI, body mass index; MCS, medial clear space.

^aData are presented as mean IQR or number of patients and percentage of total cohort (%).

Table 2.

Demographic Data of the Nonweightbearing MCS Measurement Cohorts. ^a

Characteristics	Value
Age	52.65 (65.00-39.00)
BMI	30.891 (34.94-24.52)
Height	1.731 (1.83-1.64)
Laterality
Left	9/17 (52.94)
Right	8/17 (47.06)
Gender
Female	8/17 (47.06)
Male	9/17 (52.94)

Abbreviations: BMI, body mass index; MCS, medial clear space.

^aData are presented as mean IQR or number of patients and percent of total cohort (%).

Table 3.

Demographic Data of the Weightbearing MCS Measurement Cohorts. ^a

Characteristics	Value
Age	52.65 (71.50-33.00)
BMI	28.90 (33.17-24.10)
Height	1.69 (1.79-1.63)
Laterality
Left	8/17 (47.06)
Right	9/17 (52.94)
Gender
Female	8/17 (47.06)
Male	9/17 (52.94)

Abbreviations: BMI, body mass index; MCS, medial clear space.

^aData are presented as mean IQR or number of patients and percent of total cohort (%).

MCS Measurement Methods

Review of the literature yielded 3 commonly used methods of MCS measurement. More infrequently cited methods were excluded. Radiographs were not performed by a single technician at a single site, but rather varied by patient. The following methods were used to assess the medial clear space on radiograph in this investigation (Figures 1 and 2):

Figure 2.

Measurement of the medial clear space on CT scan. The coronal image is used to identify the level 5 mm below the superior aspect of the medial talar dome. The corresponding axial CT slice is then used to measure the MCS at its anterior-posterior midsubstance. CT, computed tomography; MCS, medial clear space.

• Method 1: 5 mm inferior from the highest point of the talar dome medially, parallel to the tibiotalar joint line.

• Method 2: halfway between the articular surface of the medial talar dome and the tip of the medial malleolus, parallel to the tibiotalar joint line (referred to as MCS parallel, or MCSp). ²

• Method 3: from the superomedial aspect of the talar dome to the superomedial axilla of the tibial medial plafond, obliquely (referred to as the MCS oblique, or MCSo). ¹⁷

Matched CT scans, obtained at the same time as the radiographs, were then reviewed for each patient. The coronal weightbearing CT slices were plotted parallel to the tibiotalar joint, and slices were obtained in 0.3-mm increments for each scan. Linked axial and coronal images for each ankle were visualized, and the MCS was measured in a similar manner to that described recently by Chien et al. ⁷

• CT Method 4: A distance of 5-mm inferior from the medial talar dome was measured, and the corresponding axial CT slice was obtained. The MCS at that level was measured on the axial CT scan image at its middistance anterior to posterior. (Figure 2).

Statistical Analysis

The average age, BMI, and height was found along with the IQR (Microsoft Excel). A 2-sided P test was used to assess the statistical significance head-to-head between each of the measurement methods (radiograph and CT) and the remaining 3 (SPSS, IBM, Armonk, NY). The interobserver correlation coefficient (ICC) was found to assess the agreement between the five observers.

Results

Patient demographics of the nonweightbearing and weightbearing cohorts is included in Table 1, with the groups separately described in Tables 2 and 3. Laterality and sex were consistent between the 2 groups. Method 2 was found to be most accurate when using weightbearing imaging (no difference in radiographic vs CT measurement value, P = .09). There was no statistical difference in measurement methods when using nonweightbearing imaging (all 3 radiographic vs CT measurement values, P < .001). Interobserver reliability was found to be good to excellent for nonweightbearing MCS measurement (ICC 0.82-0.97), whereas measurements on weightbearing imaging demonstrated good reliability (ICC 0.77-0.9). Results are displayed in Table 4 for nonweightbearing measurements and Supplemental Table 1 for weightbearing measurement comparisons. ICC for each measurement is displayed in Supplemental Tables 2 and 3 for nonweightbearing and weightbearing, respectively.

Table 4.

Nonweightbearing 2-Sided P-Test Results for Statistical Significance Between Each of the 4 Measurement Methods (Radiography and CT Scan).

All Person Correlation	Two-Sided P	CI, Lower	CI, Upper	Std
MCS 1-2	.358	−0.111	0.425	0.149
MCS 1-3	<.001	−1.328	0.7496
MCS 1-4	<.001	−0.6615	−0.344	0.309
MCS 2-3	<.001	−1.2767	−0.732	0.529
MCS 2-4	<.001	−0.63394	−0.302	0.3218
MCS 3-4	<.001	0.2883	0.7837	0.4817

Abbreviations: CT, computed tomography; MCS, medial clear space; Std, standard deviation.

Discussion

The medial clear space is commonly used as a critical decision-making index for the treatment of ankle fractures. Despite its near ubiquitous use, there is no current consensus on the optimal method of measurement. In assessing the reliability of commonly measured angles on foot radiographs, Saltzman et al ²² reported that the intrinsic error in each measurement was due to landmark selection, line drawing, and goniometer use. They concluded that surgeons using radiographic indices to make treatment decisions should be wary of the fallibility of their assessments. This is especially pertinent to the discussion of the MCS given our imperfect understanding of its measurement in addition to the fact that surgical decision making is based on measurements in increments of millimeters.

There is relatively little data on the reproducibility of MCS measurement itself. Commenting on the reliability of radiographic measurements of injured and uninjured ankles, Brage et al ⁴ found that the MCS was not reliably measured across 4 different examiners on noninjury radiographs of the ankle. Initial radiograph review yielded an ICC of 0.1, with minimal improvement to 0.23 on secondary review by the examiners. Additionally, they found that intraobserver reliability increased with examiner experience. In conclusion, Brage et al stated that the decision for surgery should never be based solely on radiographs because of their inconsistent quality and low reliability on MCS measurement.

Metitiri et al ¹⁶ provided further insight into MCS measurement reliability in 2017. The group implanted spacers of known width into the MCS of 3 cadaveric specimens and then obtained radiographs. Measurements performed by orthopaedic surgeons at varying levels of training demonstrated poor accuracy and precision. In contradistinction to the results of Brage et al, ⁴ Metitiri et al ¹⁶ found that provider experience did not influence the accuracy or precision of measurements. Overall, there was a high degree of measurement error and lack of precision in MCS assessment.

The results of the current investigation put forth additional perspective. Historically there have been several different methods of MCS measurement described and commonly used in the literature.^{1,2,4,5,7,9,10,12,13,15,17 -20,24,25} Despite this, there is no comparative data between them and CT scan measurement of the MCS to support the optimal use of one method over another. The current investigation sought to clarify which measurement method was superior. Based on our results, it seems reasonable that any of the 3 tested measurement methods can be used when analyzing nonweightbearing radiographs given the lack of statistical difference and relative equivalence between methods. Measuring the MCS halfway between the superior medial talar dome and the tip of the medial malleolus, perpendicular to the tibiotalar joint, was found to be most accurate and therefore should be used when examining weightbearing imaging. This may be due to a normalized view of the ankle mortise with weightbearing that allows the anatomical landmarks used to make the measurement more reliably visualized. Similarly, this measurement is more patient specific than the others, as it is measured based on half the distance of each individual’s talar dome. To standardize one’s clinical practice, we advocate for its use in measuring the MCS for all imaging.

The results of this study also demonstrate that the majority of MCS measurement methods had good or excellent interobserver correlation within our 5-member group, who represented a broad range of clinical experience. For the nonweightbearing MCS measurements, there was an ICC between 0.82 and 0.97, indicating good to excellent reliability. ¹⁴ The weightbearing MCS measurements had an ICC between 0.77 and 0.9, demonstrating good reliability. One could reasonably expect that weightbearing measurements would provide better interobserver reliability because the ankle mortise is more neutrally aligned and not affected by varying amounts of plantarflexion as it is in nonweightbearing radiographs, but this was not the case. Therefore, our results are aligned with prior studies demonstrating that observer level of experience does not influence the reliability of measurements, in this case on either nonweightbearing or weightbearing imaging.

There are several important limitations of this study that bear mentioning when interpreting our results. First, measurement of the MCS is highly dependent on rotation of the radiographs. The quality of each radiograph being reviewed varied based on several factors including the position of the beam vis-a-vis the patient’s ankle, the amount of weightbearing applied at the time of imaging, and the viewing software the images are ultimately accessed through. Despite this, all radiographs reviewed were deemed acceptable—meaning there was a congruent joint space around the entirety of the ankle joint space—and we feel this variability is typical of clinical practice.

Although not a limitation per se, the difference in representation of the MCS on radiography vs CT is also a particular consideration worth noting. MCS was originally described as a measurement based on plain radiographs with its inherent limitations previously reported in the literature (some of which were discussed in this investigation). However, radiograph is a 2-dimensional representation of a 3-dimensional space as is the MCS on axial CT. We used the MCS measurement on axial CT as the “true” value given its demonstrated accuracy. ⁷ Although there exists predicate literature examining ankle fractures that have used this technique, we recognize that the MCS was originally described as a radiographic parameter on radiograph that has now been extrapolated to an anatomic distance measurement independent of imaging modality used.

Conclusions

Measurement of the MCS is important in surgical decision making, and we provide further insight on its accuracy and reliability. In clinical practice, any of the 3 tested measurement methods can be used when analyzing nonweightbearing radiographs given the lack of statistical difference and relative equivalence between methods. Measuring the MCS halfway between the superior medial talar dome and the tip of the medial malleolus, perpendicular to the tibiotalar joint, should be used when examining weightbearing imaging. To facilitate standardizing one’s clinical practice, we advocate for its use in measuring the MCS for all imaging.

Supplemental Table 1.

Weightbearing 2-Sided P-Test Results for Statistical Significance Between Each of the 4 Measurement Methods (Radiograph and CT scan).

All Person Correlation	Two-Sided P	CI, Lower	CI, Upper	Std
MCS 1-2	.214	–0.0345	0.142	0.1723
MCS 1-3	<.001	–1.0132	–0.4201	0.5768
MCS 1-4	.024	0.0328	0.4004	0.3574
MCS 2-3	<.001	–1.071	–0.47044	0.5842
MCS 2-4	.09	–0.0284	0.353	0.3714
MCS 3-4	<.001	0.7156	1.151	0.4235

Abbreviations: CT, computed tomography; MCS, medial clear space; Std, standard deviation.

Supplemental Table 2.

ICCs for Nonweightbearing MCS measurements.

MCS 1 ICC
Cronbach alpha	on std items
0.891	0.907
MCS 2 ICC
Cronbach alpha	on std items
0.871	0.883
MCS 3 ICC
Cronbach alpha	on std items
0.972	0.973
MCS 4 ICC
Cronbach alpha	on std items
0.827	0.837

Abbreviations: ICC, interobserver correlation coefficient; MCS, medial clear space; std, standard.

Supplemental Table 3.

ICCs for Weightbearing MCS Measurements.

MCS 1 ICC
Cronbach alpha	on std items
0.852	0.857
MCS 2 ICC
Cronbach alpha	on std items
0.766	0.778
MCS 3 ICC
Cronbach alpha	on std items
0.891	0.901
MCS 4 ICC
Cronbach alpha	on std items
0.791	0.804

Abbreviations: ICC, interobserver correlation coefficient; MCS, medial clear space; std, standard.