Foot & Ankle International http://fai.sagepub.com/
Tripod Index: A New Radiographic Parameter Assessing Foot Alignment Marut Arunakul, Annunziato Amendola, Yubo Gao, Jessica E. Goetz, John E. Femino and Phinit Phisitkul Foot Ankle Int published online 8 May 2013 DOI: 10.1177/1071100713488761 The online version of this article can be found at: http://fai.sagepub.com/content/early/2013/05/08/1071100713488761
Published by: http://www.sagepublications.com
On behalf of:
American Orthopaedic Foot & Ankle Society
Additional services and information for Foot & Ankle International can be found at: Email Alerts: http://fai.sagepub.com/cgi/alerts Subscriptions: http://fai.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> OnlineFirst Version of Record - May 8, 2013 What is This?
Downloaded from fai.sagepub.com at The University of Iowa Libraries on September 15, 2013
488761 research-article2013
FAIXXX10.1177/1071100713488761Foot & Ankle InternationalArunakul et al
Article
Tripod Index: A New Radiographic Parameter Assessing Foot Alignment
Foot & Ankle International XX(X) 1–10 © The Author(s) 2013 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1071100713488761 fai.sagepub.com
Marut Arunakul, MD1,2, Annunziato Amendola, MD1, Yubo Gao, PhD1, Jessica E. Goetz, PhD1, John E. Femino, MD1, and Phinit Phisitkul, MD1
Abstract Background: No single radiographic measurement takes into account complete foot alignment. We have created the Tripod Index (TI) to allow assessment of complex foot deformities using a standing anteroposterior (AP) radiograph of the foot. We hypothesized that TI would demonstrate good intraobserver and interobserver reliability and correlate with currently accepted radiographic parameters, in both flatfoot and cavovarus foot deformities. Methods: Three groups of patients were studied: 26 patients (30 feet) with flatfoot, 29 patients (30 feet) with cavovarus foot, and 51 patients (60 feet) without foot deformity as controls. Weight-bearing radiographs were obtained: foot AP with a hemispherical marker around the heel plus standard lateral and hindfoot alignment views. Radiographic measurements were made by 2 blinded investigators. Statistical analysis included intraclass correlation coefficients (ICCs), correlation of the TI with existing radiographic measurements using Pearson coefficients, and comparison between patient groups using analysis of variance. Results: Intraobserver and interobserver ICCs of TI (0.99 and 0.98, respectively) were excellent. In the flatfoot group, TI significantly correlated with AP talonavicular coverage angle (r = 0.43), medial cuneiform–fifth metatarsal height (r = -0.59), coronal plane hindfoot alignment (r = 0.53), and clinical hindfoot alignment (r = 0.39). In the cavovarus foot group, TI correlated significantly with AP talonavicular coverage angle (r = 0.77), calcaneal pitch angle (r = 0.39), medial cuneiform– fifth metatarsal height (r = -0.65), coronal plane hindfoot alignment (r = 0.55), and clinical hindfoot alignment (r = 0.61). Statistically significant differences between flatfoot-control and cavovarus foot–control were found in TI, AP talonavicular coverage angle, lateral talo–first metatarsal angle, calcaneal pitch angle, medial cuneiform–fifth metatarsal height, coronal plane hindfoot alignment, and clinical assessment of hindfoot alignment (all with P < .001). Conclusion: The TI was demonstrated to be a valid and reliable radiographic measurement to quantify the magnitude of complex foot deformities when evaluating flatfoot and cavovarus foot. Clinical Relevance: The TI may be helpful as an integrated assessment of complex foot deformities. Further clinical studies are recommended. Level of Evidence: Level III, retrospective comparative study. Keywords: Tripod Index, radiographic parameter, flatfoot, cavovarus foot
It is widely accepted that notable foot deformities can have a detrimental effect on the function of the foot and ankle.22,23,31,37 However, substantial controversy exists regarding clinical and radiological definitions of those deformities as well as the reliability and usefulness of standardized radiographic measurements in defining the deformities.9,38 Flatfoot and cavovarus foot alignment are examples of complex foot deformities commonly seen in clinical practice. The flatfoot deformity is characterized by a combination of collapse of the medial longitudinal arch, foot abduction at the talonavicular joint, and hindfoot valgus (subtalar joint eversion).4,15,18,20,29,32,34 Cavovarus foot is characterized by hindfoot varus (subtalar joint inversion),
midfoot cavus, plantar flexion of the first metatarsal, and forefoot adduction.26,30,47 Currently, several radiographic measurements are used to assess these deformities, but none of the radiographic parameters are able to integrate deformities at multiple planes and levels of the foot into a simple 1
Department of Orthopaedics and Rehabilitation, University of Iowa, Iowa City, Iowa, USA 2 Department of Orthopaedic Surgery, Thammasat University, Pathumthani, Thailand Corresponding Author: Phinit Phisitkul, MD, Department of Orthopaedics and Rehabilitation, the University of Iowa, 200 Hawkins Drive, Iowa City, IA 52242, USA. Email:
[email protected]
Downloaded from fai.sagepub.com at The University of Iowa Libraries on September 15, 2013
2
Foot & Ankle International XX(X)
measurement. For example, high calcaneal pitch (dorsiflexion of calcaneus), a decreased talo–first metatarsal angle (plantar flexion of first metatarsal), increased distance of medial cuneiform–fifth metatarsal height (high arch of foot),3 and posterior position of the fibula relative to the tibia (external rotation of ankle axis)1 are all used to describe cavovarus foot deformity.30 Single measurements are not able to explain the additive effect of metatarsus adductus on a cavovarus deformity or the effect of forefoot abduction on a planovalgus deformity. The true understanding of compensatory deformity such as combined hindfoot varus and midfoot collapse is limited from commonly used linear or angular measurements. In addition, the influence of missing medial or lateral rays has never been accounted for when considering malalignment. The Tripod Index was created to fulfill this need by providing a single measurement that describes the additive or compensatory effects of deformities at different levels of the foot. We have created the Tripod Index to allow quantitative assessment of complex foot deformities on a standing anteroposterior (AP) radiograph using the relationship between the center of the talar head and the foundation of the tripod composed of the center of the heel and medial/lateral borders of the forefoot. The static “triangle of support,” which consists of center of heel, first metatarsal head, and fifth metatarsal head, has been known as the “foot tripod.”6,30 Any foot deformities could theoretically lead to imbalance of the foot tripod. Likewise, deformity correction surgeries, including Cotton osteotomy,6 dorsiflexion osteotomy of the first metatarsal bone,47 Dwyer osteotomy,14 and medial calcaneal slide osteotomy,33 have been described to correct imbalance of the foot tripod. The subtalar axis is the line that passes from the posterolateral corner of the calcaneus to the superomedial aspect of the talar neck with 42 ± 9 degrees upward inclination from the sagittal plane and 23 ± 11 degrees medial deviation from the axis of the foot passing through the second interdigital space.27 This axis runs closely through the center of the talar head, which is a consistent landmark on a normal weight-bearing AP radiograph.17,24 The proximity of the subtalar joint axis and the center of the talar head was found to be reproducible throughout the subtalar range of motion.45 The relationship between the foot tripod and the center of the talar head could theoretically demonstrate the effect of overall foot alignment on the subtalar joint.28 This simple and more readily measurable parameter could be developed to evaluate complex foot deformities. Because the foot tripod is determined by the landmarks from the hindfoot to the forefoot, this measurement has the potential to appreciate the summation of deformity including hindfoot, midfoot, and forefoot in multiple planes. We hypothesized that the Tripod Index would demonstrate good intraobserver and interobserver reliability, correlate with currently accepted radiographic parameters, and
Table 1. Diagnosis in Control Group. Diagnoses
No.
Morton’s neuroma Mononeuritis of peroneal nerve Noninsertional Achilles tendinopathy Insertional Achilles tendinopathy Central heel pain Hammer toe Mallet toe
10 1 31 8 3 5 2
reveal noteworthy differences between flatfoot, cavovarus foot, and control (neutral alignment) groups.
Methods The study was approved by our hospital’s institutional review board. Standard weight-bearing radiographs including foot AP with a hemispherical marker around the heel, lateral, and hindfoot alignment views as well as standardized physical examination were performed on the 3 groups of patients presenting to the foot and ankle clinic over a period of 12 months (June 2010 to May 2011). The first group consisted of 30 feet in 26 consecutive patients (average age, 42.1 ± 17.9 years; range, 17-84 years; 19 females and 7 males) with clinically apparent flatfoot deformity. The second group consisted of 30 feet in 29 consecutive patients (average age, 46.6 ± 15.4 years; range, 14-70 years; 20 females and 9 males) with clinically apparent cavovarus foot deformity. The third group (control) consisted of 60 feet in 51 consecutive patients (average age, 47.4 ± 12.4 years; range, 15-68 years; 34 females and 17 males) without clinically apparent foot and ankle deformity (Table 1). We identified and included all patients with symptomatic flatfoot deformity and cavovarus foot deformity from the patient databases of 3 senior foot and ankle orthopedic surgeons (P.P., J.E.F., and A.A.). All the patients either had undergone a reconstructive surgery or had been consented for surgery because of failure of nonoperative treatment. The control group was screened for the absence of congenital or acquired foot and ankle deformity. Patients who had a previous ipsilateral foot surgery were excluded. Demographic data were collected from patients’ charts. One of the senior authors (P.P., J.E.F., or A.A.) assessed the hindfoot alignment of each patient in standing position and categorized each as markedly valgus (+2), slightly valgus (+1), neutral (0), slightly varus (–1), or markedly varus (–2).
Radiographic Studies Anteroposterior and lateral weight-bearing radiographs were obtained in a standardized manner. For AP images,
Downloaded from fai.sagepub.com at The University of Iowa Libraries on September 15, 2013
3
Arunakul et al
Figure 1. (A) Clinical photograph demonstrating posterior views of placement of both feet for AP radiograph with hemispherical markers around the heel. (B) Hemispherical marker.
patients were asked to stand with equal weight on both feet. Properly sized hemispherical markers, made from a metal plate padded with half-inch polyethylene foam, were put snugly on the backs of both heels (Figure 1). The beam was positioned 40 inches from the digital cassette and angled 15 degrees posteriorly toward the heels. Laser guides were used to align the center of the heel and second toe along the vertical axis of the cassette. For hindfoot alignment weightbearing, subjects stood on a radiolucent platform with equal weight on both feet. The platform consisted of a 0.5 × 20 × 18–inch sheet of Plexiglas mounted on a 6-inch-high metal frame. The frame included a slot for holding the x-ray cassette at a 20-degree angle from vertical. A 3 × 2 × 6–cm lead strip was placed tangent to the most posterior aspect of the heel and was oriented perpendicular to longitudinal axis of the foot. The x-ray tube was oriented 20 degrees from horizontal, so that it was perpendicular to the plane of the film. The beam was centered at the level of the ankle, and the field of exposure was from midshaft of the tibia to below the calcaneus. The source-to-film distance was 40 inches.5 The AP radiographs were examined for the AP talonavicular coverage angle40 as shown in Figure 2. The lateral radiographs were examined for the calcaneal pitch angle,43 the lateral talocalcaneal angle,10,48 the lateral talo–first metatarsal angle,21 and the medial cuneiform–fifth metatarsal height35 as shown in Figure 3. The hindfoot alignment radiographs were examined for the coronal plane hindfoot alignment39 (apparent moment arm) as shown in Figure 4. The Tripod Index was assessed on the AP radiograph as shown in Figures 5 and 6. All radiographic measurements were performed using iSite Enterprise 3.5 image analysis software (Philips Medical Systems Nederland B.V., Best, The Netherlands). The measurements were made by a foot and ankle orthopedic fellow (M.A.) and a biomechanical engineer (J.E.G.). They were trained to measure both the
standard and the new radiographic parameters by a senior foot and ankle orthopedic surgeon (P.P.). Intraobserver variability was determined by repeated measurement of the 7 radiographic parameters in 31 cases (25% of each group) performed by the foot and ankle orthopedic fellow at separate sessions 6 months apart in a blinded and randomized fashion. Interobserver variability was determined by comparing the measurements for 31 cases between the foot and ankle orthopedic fellow and the biomechanical engineer in a blinded and randomized fashion. The radiographic technique and the reliability of the hemispherical marker placement were tested in a pilot study that included 16 feet in 16 consecutive patients (average age, 49.9 ± 14.1 years; range, 17-70 years; 10 females and 6 males) with various degrees of deformity (5 symptomatic flatfoot, 2 symptomatic cavovarus foot, and 9 noninsertional Achilles tendinopathy). Standard foot AP weightbearing radiographs with a hemispherical marker around the heel were performed twice at 3 months apart. The AP radiographs were examined for the intermetatarsal angle (IMA)8,38,41 and the Tripod Index. With the numbers available, we found that radiographic measurements, including IMA and Tripod Index, did not have statistically significant differences between the first and the second radiographs (P = 1 and 0.96, respectively). The averaged IMAs were 8 ± 2.7 degrees and 8 ± 2.8 degrees, respectively. The averaged Tripod Indexes were 5% ± 47% and 5% ± 44%, respectively. The reliabilities of the first and second radiographs for IMA and Tripod Index were 0.99 and 0.99, respectively, which are considered excellent.
Statistical Analysis Data analysis was performed with the aid of a statistician using SAS software, version 9.2 (SAS Institute, Cary, NC).
Downloaded from fai.sagepub.com at The University of Iowa Libraries on September 15, 2013
4
Foot & Ankle International XX(X)
Figure 2. Measurement technique for the AP talonavicular coverage angle. Points a, b = medial and lateral margins (respectively) of the articular surface of the talus. Points c, d = medial and lateral margins (respectively) of the articular surface of the navicular. Line A is perpendicular to ab, and line B is perpendicular to cd. The AP talonavicular coverage angle is the angle created by the intersection of lines A and B.
Intraclass correlation coefficient (ICC) was calculated to determine intraobserver and interobserver reliability for all of the radiographic parameters measured. The following classification scheme was used for ICC: less than 0.40 was poor; 0.40-0.59 was fair; 0.60-0.74 was good; and greater than 0.74 was excellent.44 The correlations between the new parameter (Tripod Index) and the traditional radiographic parameters and physical examination were assessed with
Figure 3. Measurement technique for the lateral talo–first metatarsal angle, the calcaneal pitch, the lateral talocalcaneal angle, and the medial cuneiform–fifth metatarsal height. Lateral talo–first metatarsal angle: The longitudinal axis of the talus (line B1-B2) is established by placing a mark at the halfway point between the superior and inferior surfaces of the talus at the middle of the talus and the neck of the talus and connecting these 2 points. A similar method is used to determine the lateral longitudinal axis of the first metatarsal (line C1-C2), with mid-diaphyseal reference points used to form this axis. The angle formed by these 2 lines is the lateral talo–first metatarsal angle (angle A). This angle is considered negative when there is a cavus relationship between the axis of the talus and the first metatarsal. The calcaneal pitch is defined as the angle (angle D) formed by the intersection of a line drawn tangentially along the inferior aspect of the calcaneus (line E1-E2) and a line drawn along the plantar aspect of the soft tissue shadow of the hindfoot when the subject is weight-bearing (line F). The lateral talocalcaneal angle has the same talar axis line (line B1-B2) as described previously. A second line, E1-E2, is used to form the lateral talocalcaneal angle (angle G). Medial cuneiform–fifth metatarsal height: The most inferior surface of the distal aspect of the medial cuneiform is marked, and a line (line H) is extended from this point parallel to the floor. A similar line is marked (line I) on the most inferior portion of the base of the fifth metatarsal. The perpendicular distance between these 2 lines is measured with an electronic caliper and is designated the medial cuneiform–fifth metatarsal height (line J).
Pearson correlation coefficients. The r values range from –1 to +1, where 0 indicates the absence of any linear relationship and -1 and +1 indicate a perfect negative (inverse) and a perfect positive (direct) relationship, respectively.16 The differences in radiographic parameters and physical examination between flatfoot, cavovarus, and control group were assessed with Tukey-Kramer technique-adjusted analysis of variance. The reliability and the differences in radiographic parameters between the first and the second radiographs for pilot study were assessed by ICC and paired Student t test. P < .05 was considered statistically significance. For the diagnosis of symptomatic flatfoot and cavovarus foot, a post hoc analysis with a receiver operating characteristic
Downloaded from fai.sagepub.com at The University of Iowa Libraries on September 15, 2013
5
Arunakul et al
Figure 4. Measurement technique for the coronal plane hindfoot alignment (apparent moment arm). Line A is defined by bisecting the tibia 10 and 15 cm above the medial tibia plafond. Point B is the most inferior aspect of calcaneus. The apparent moment arm is determined by measuring the perpendicular distance between line A and point B (line BC). These values are positive if the weight-bearing axis of the leg (line A) is medial to point B (valgus calcaneus) and negative if line A is lateral to point B (varus calcaneus).
(ROC) curve was analyzed to determine a proper diagnostic threshold. The sensitivity and specificity of each radiographic parameter were determined.
Results The intraobserver reliability for both the traditional and new radiographic parameters was excellent, with correlation coefficients ranging from 0.83 to 0.99 (Table 2). The Tripod Index, the coronal plane hindfoot alignment, and the medial cuneiform–fifth metatarsal height demonstrated
Figure 5. Measurement technique for the Tripod Index. A is the center of the heel (center of the hemispherical marker), B is the medial edge of the medial sesamoid, C is the lateral edge of the fifth metatarsal head, D is the center of the talar head, E is the tripod angle created by the intersection of AB and AC, and F is the angle created by the intersection of AB and AD. Tripod Index (%) = (F/E) × 100. If the AD line is medial to the AB line, the index is positive; if the AD line is lateral to the AB line, the index is negative. The more positive Tripod Index, the more the center of the talar head is medial to the tripod of the foot as commonly seen in flatfoot deformity.
similarly high intraobserver reliability. The ICCs of all of these were higher than the ICCs for the AP talonavicular coverage angle, the calcaneal pitch angle, the lateral talo– first metatarsal angle, and the lateral talocalcaneal angle. The interobserver reliability for all radiographic parameters was good to excellent, with correlation coefficients ranging from 0.69 to 0.99 (Table 2). The Tripod Index, the calcaneal pitch angle, and the medial cuneiform–fifth metatarsal height demonstrated similar interobserver reliability.
Downloaded from fai.sagepub.com at The University of Iowa Libraries on September 15, 2013
6
Foot & Ankle International XX(X)
Figure 6. Case demonstration of the Tripod Index in flatfoot, cavovarus foot, and control groups. Table 2. Intraobserver and Interobserver Intraclass Correlation Coefficients for Radiographic Parameters. Radiographic Parameters Tripod Index AP talonavicular coverage angle Lateral talo–first metatarsal angle Calcaneal pitch angle Lateral talocalcaneal angle Medial cuneiform–fifth metatarsal height Coronal plane hindfoot alignment
Intraobserver
Interobserver
0.99 0.93
0.98 0.85
0.96
0.69
0.95 0.83 0.99
0.98 0.71 0.99
0.98
0.95
The ICCs of all of these were higher than the ICCs for the coronal plane hindfoot alignment, the AP talonavicular coverage angle, the lateral talo–first metatarsal angle, and the lateral talocalcaneal angle. In the flatfoot group, several statistically significant correlations were found. There was a correlation between the Tripod Index and the AP talonavicular coverage angle (r = 0.43, P = .017), the medial cuneiform–fifth metatarsal height (r = -0.59, P = .001), the coronal plane hindfoot alignment (r = 0.53, P = .003), and the clinical hindfoot alignment (r = 0.39, P = .032). The cavovarus foot group also demonstrated a statistically significant correlation between the Tripod Index and the AP talonavicular coverage angle (r = 0.77, P < .001), the calcaneal pitch angle (r = 0.39,
P = .035), the medial cuneiform–fifth metatarsal height (r = -0.65, P < .001), the coronal plane hindfoot alignment (r = 0.55, P = .002), and the clinical hindfoot alignment (r = 0.61, P < .001). Radiographic measurements and clinical assessment of hindfoot alignment that demonstrated significant differences between the flatfoot, cavovarus, and control groups included the Tripod Index, AP talonavicular coverage angle, lateral talo–first metatarsal angle, calcaneal pitch angle, medial cuneiform–fifth metatarsal height, coronal plane hindfoot alignment, and clinical hindfoot alignment (Table 3). Receiver operating characteristic curves demonstrated that the best cut points of the Tripod Index for diagnosing symptomatic flatfoot and cavovarus foot were 26% or more and –39% or less, respectively (Tables 4 and 5). When 26% or more of the Tripod Index was used as a threshold to diagnose symptomatic flatfoot, sensitivity and specificity were found to be 97% and 100%, respectively. When -39% or less of the Tripod Index was used as a threshold to diagnose symptomatic cavovarus foot, sensitivity and specificity were found to be 100% and 97%, respectively.
Discussion This study has shown that the Tripod Index has excellent reliability and correlates highly with currently used foot alignment radiographic parameters. The Tripod Index also demonstrated a statistically significant difference between flatfoot, cavovarus foot, and control groups.
Downloaded from fai.sagepub.com at The University of Iowa Libraries on September 15, 2013
7
Arunakul et al Table 3. Radiographic Parameters and Clinical Assessment for Control (Con), Flatfoot (Flat), and Cavovarus Foot (Cavo) Groupsa. Radiographic Parameters Tripod Index, % AP talonavicular coverage angle, degrees Lateral talo–first MT angle, degrees Calcaneal pitch angle, degrees Lateral talocalcaneal angle, degrees Medial cuneiform–fifth MT height, mm Coronal plane hindfoot alignment, mm Clinical assessment of hindfoot alignment
Control (n = 60)
Flatfoot (n = 30)
Cavovarus Foot (n = 30)
–4 ± 18
59 ± 21
–115 ± 53
9±5
23 ± 8
–10 ± 13
2±4
18 ± 8
–10 ± 5
23 ± 3
14 ± 4
27 ± 5
49 ± 5
49 ± 7
46 ± 5
10 ± 3
1±4
19 ± 6
–2 ± 4
14 ± 6
–15 ± 4
–0.1 ± 0.4
1.8 ± 0.4
–1.7 ± 0.5
P Value Flat-Con:
