- Email: [email protected]
Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus
G Model FAS 1335 No. of Pages 9
Foot and Ankle Surgery xxx (2019) xxx–xxx
Contents lists available at ScienceDirect
Foot and Ankle Surgery journal homepage: www.elsevier.com/locate/fas
Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus Alessio Bernasconia,b,* , Lucy Coopera , Shirley Lylea , Shelain Patela , Nicholas Cullena , Dishan Singha , Matthew Welcka a b
Foot and Ankle Unit, Royal National Orthopaedic Hospital, Stanmore, United Kingdom Department of Public Health, Trauma and Orthopaedics, University of Naples Federico II, Naples, Italy
A R T I C L E I N F O
A B S T R A C T
Article history: Received 19 April 2019 Received in revised form 30 June 2019 Accepted 17 July 2019 Available online xxx
Introduction: Pes cavovarus is a three-dimensional (3D) foot deformity. New 3D semi-automatic measurements utilising weightbearing computerised topography (WBCT) images have recently been proposed to assess hindfoot alignment, but reliability in pes cavovarus has never been investigated. The aim of this study was to assess intraobserver and interobserver reliability of the foot ankle offset (FAO), calcaneal offset (CO) and hindfoot alignment angle (HAA) in pes cavovarus. Methods: Anonymised WBCT datasets from 51 feet (17 Charcot-Marie-Tooth related cavovarus, 17 idiopathic cavovarus and 17 controls) were retrospectively reviewed. Three observers (two senior foot and ankle fellows and one orthopaedic resident) independently measured FAO, CO and HAA using dedicated software, with measurements repeated two weeks apart. Subgroup analysis was performed to assess whether aetiology or severity of varus deformity and level of seniority affected reliability. Results: Mean values for intra and interobserver reliability for FAO (r = 0.98; ICC: 0.99), CO (r = 0.97; ICC: 0.98) and HAA measurements (r = 0.97; ICC: 0.98) were excellent. Subgroup analyses showed that FAO, CO and HAA's intra (r/r range, 0.77–0.95) and interobserver (ICC range, 0.88–0.98) reliability remained excellent in patients with Charcot-Marie-Tooth related cavovarus, idiopathic pes cavovarus and normal feet, regardless of the severity of deformity. No difference was found in FAO, CO and HAA mean values from three observers (p > 0.05 in all cases). Discussion: This study demonstrates that 3D semi-automatic measurements of WBCT images have excellent intra and interobserver reliability in the assessment of hindfoot alignment in pes cavovarus. Aetiology and severity of deformity, and level of seniority do not affect reliability of these measurements. Level of evidence: Level III, retrospective comparative study. © 2019 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved.
Keywords: Cavovarus Hindfoot High arch Charcot-Marie-Tooth Weightbearing CT Cone beam
1. Introduction Pes cavovarus is a three-dimensional (3D) complex foot deformity characterised by a varus hindfoot, high longitudinal arch and forefoot pronation. Although multiple pathologies may lead to its development, they can be broadly categorised into idiopathic [1–3] and acquired, whilst acquired can be sub-divided into neurological and nonneurological [4–7]. Charcot-Marie-Tooth (CMT) disease, anterior horn
* Corresponding author at: Foot and Ankle Unit, Royal National Orthopaedic Hospital, Middlesex, HA7 4LP, Stanmore, United Kingdom. E-mail addresses: [email protected] (A. Bernasconi), [email protected] (L. Cooper), [email protected] (S. Lyle), [email protected] (S. Patel), [email protected] (N. Cullen), [email protected] (D. Singh), [email protected] (M. Welck).
diseases, disorders of the central nervous system and poliomyelitis account for the most common neurological causes [4–7]. The diagnostic workup in cavovarus feet is traditionally based on thorough clinical assessment and investigations to establish aetiology followed by dedicated imaging, including weightbearing radiographs and computed tomography (CT) or magnetic resonance imaging [3–7]. Taking into account the 3D nature of cavovarus deformity and the necessity for multiplanar correction at different sites (hindfoot, midfoot and forefoot), detailed imaging of the foot during physiological weightbearing stance is extremely helpful for accurate surgical planning. Two-dimensional (2D) radiographs tend to be flawed by perspective, operator-related and superimposition biases [8–10], whilst standard CT, performed either non-weightbearing or with simulated weightbearing, does not correctly reproduce the standing foot posture [11–14]. These drawbacks may be overcome via 3D cone beam weightbearing
https://doi.org/10.1016/j.fas.2019.07.005 1268-7731/© 2019 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
2
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
Fig. 1. Flow chart of the 51 feet eligible for this study.
computed tomography (WBCT) which combines physiological weightbearing with lower doses of radiation than conventional CT [13,15]. The concept of 3D biometrics is a recent addition to the literature in order to study hindfoot alignment through the assessment of foot position in space and not only in single planes [16]. This is realised through the identification of 4 points on WBCT images to define a volume (rather than 3 points defining an angle) corresponding to a pyramid whose base is the traditional foot tripod and a vertex represented by the centre of the ankle joint. Dedicated software then processes the coordinates of these points allowing calculation of three parameters: foot and ankle offset (FAO), calcaneal offset (CO) and hindfoot alignment angle (HAA). Although from different perspectives, they all measure the
difference between the true hindfoot in space and a theoretical neutral one, enabling quantification of valgus or varus deformity and negating the aforementioned 2D imaging-related biases. To date, the most studied of the measurements is FAO, which corresponds to the lever arm generated at the ankle by two opposite forces, namely body weight (downwards) and the ground reaction force (upwards). Two previous studies have proven excellent intra and interobserver agreement for these 3D semiautomatic measurements [16,17]. The aim of this study was to assess intra and interobserver reliability of 3D biometrics on WBCT imaging in pes cavovarus patients diagnosed with CMT disease and idiopathic cavovarus. Our hypothesis was that FAO, CO and HAA were reliable measurements regardless of
Table 1 Patients’ demographics. Clinical characteristics of pes cavovarus patients (N 34) 45.5 16.1 (15–72) 31 (61) 21 (41) 30 6.7 (19–46)
Age (year) Male, N (%) Right side, N (%) BMI (kg/m2) Subgroups By aetiology
CMT (N 17)
Idiopathic (N 17)
Controls (N 17)
p-Value
Age (year) Male, N (%) Right side, N (%) BMI (kg/m2)
46.2 13 (39–53) 10 (58) 7 (42) 31.3 8 (27–35)
44 19 (33–54) 12 (70) 7 (42) 29.9 5 (27–32)
46.1 17 (37–55) 11 (64) 7 (42) 28.7 6 (25–32)
0.975* 0.932** 1** 0.738*
By severity
FAO > 0% (N 13)
FAO
Age (year) Male, N (%) Right side, N (%) BMI (kg/m2)
48.1 18.7 (37–58) 6 (46) 5 (38) 27 6.3 (23–30)
42.2 13 (34–49) 10 (77) 5 (38) 31.6 6.3 (28–35)
7% to 0% (N 13)
FAO
13% to
7% (N 12)
47.6 18.3 (37–58) 6 (50) 5 (41) 27.8 5.3 (24–31)
FAO < 13% (N 13) 44.2 17.5 (34–54) 9 (70) 6 (46) 32.8 7.5 (28–37)
BMI: Body Mass Index; CMT: Charcot-Marie-Tooth. Age and BMI are reported as: mean standard deviation; 95% confidence interval values in brackets. * Extended Fisher's exact test. ** Kruskal–Wallis test rank test.
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
3
Fig. 2. Illustration of the foot tripod with foot and ankle offset (FAO) calculation principle. Four landmarks that must be manually identified are the weightbearing point of the first metatarsal head (M1), the weightbearing point of the fifth metatarsal head (M5), the weightbearing point (lowest point) of the calcaneus (C) and the uppermost and central point of the talar dome (T) (left upper inset). Thereafter, the software will calculate the projection of T on the ground (T1) and the distance (d) between this latter and C1 (which corresponds to the point on the bisecting line of the triangle where the perpendicular to the bisecting line itself intersects T1). That distance given in percentage of the foot length (d/CH, where H is the midpoint of the M1–M5 line) will be FAO. Drawings by the authors superimposed to original image adapted from BoneBoxTM – Foot (©2014 iSO-FORM, LLC).
aetiology and severity of deformity. We also assessed differences in reliability for these measurements by observers with different degrees of seniority and different experience with WBCT imaging.
standards. The study was conducted following the STROBE guidelines. 2.2. Setting and participants
2. Methods 2.1. Study design This comparative retrospective study analysed existing data recorded as part of routine clinical care. All procedures were performed in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical
Cone beam WBCT datasets from 201 feet with a diagnosis of pes cavovarus imaged between May 2013 and June 2017 in the Foot and Ankle Unit of our institution were reviewed for this study. After removal of duplicated scans, 172 feet (100 patients) were available. All WBCT scans were performed as part of routine investigation. Initially, patients who met the following criteria were selected: datasets with no artefact and from skeletally mature patients with diagnosis of symptomatic pes cavovarus and CMT disease, with no
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
4
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
Fig. 3. Illustration of calcaneal offset (CO) and hindfoot alignment angle (HAA). The former is the distance between the true weightbearing point of the calcaneus (C) and the hypothetical weightbearing point (C2) if the bisecting line of the tripod passed through T1 (that is the projection on the ground of the centre of talar dome). HAA is estimative of the coronal angular alignment of the hindfoot, being measured as the angle formed by three points: centre of the talar dome projection on the ground (T1, as apex), the ideal position of the calcaneus (C2) and the actual position of the calcaneus (C). Drawings by the authors superimposed to original image adapted from BoneBoxTM – Foot (©2014 iSO-FORM, LLC).
Fig. 4. Example of identification of talar dome in severely dismorphic foot. In the upper row (A) the foot has been aligned along the second metatarsal bone (left upper quadrant) whilst in the middle row (B) along the first metatarsal one (left middle quadrant). In both cases the visualisation of the talar dome on the coronal (upper middle and central quadrants) and sagittal views (right upper and right middle quadrants) is unclear. In the lower row (C) the image is aligned along the bimalleolar axis (left lower quadrant), leading to a satisfactory visualisation of the ankle joint on both the coronal and sagittal plane (middle lower and right lower quadrant, respectively).
previous ipsilateral foot/ankle surgery, ability to heel weightbear (no severe fixed equinus deformity) and to stand unassisted for the time of a WBCT scan (approximately 50 s) (Fig. 1). Seventeen neurological pes cavovarus were enrolled and a further seventeen idiopathic cavovarus were matched by sex, side, age and BMI from the original pool (ratio 1:1). Seventeen datasets from normal feet (matched for the same variables) were selected as controls (Fig. 1). Overall, anonymised datasets from 51 feet were available (34 with cavovarus morphotype and 17 normal feet) (Table 1).
2.3. Variables Datasets were analysed using the TALASTM software plugin for CubeViewTM (CurveBeam©). The 3D coordinates of specific anatomical landmarks were required for the software to automatically process and FAO, CO and HAA, as described by Lintz et al. [16]. These coordinates included the most plantar and weightbearing voxel of the head of the 1st metatarsal, head of the 5th metatarsal
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
5
Fig. 5. Example of identification of weightbearing point of the calcaneus in severely dismorphic foot. In the axial view (left) there are multiple ‘lower’ points of the bone that may be misleading. On the coronal (centre) and sagittal plane (right) it is clear that one true ‘weightbearing point’ corresponds to the vertical axis of the bone, being the others osteophytes or calcifications.
Table 2 Interobserver reliability in the whole cohort and in subgroups. Measurement
Mean ICC (95% CI)
FAO CO HAA
0.99 (0.99–0.99)* 0.98 (0.96–0.99)* 0.98 (0.96–0.99)*
Subgroup analysis By aetiology FAO CO HAA
CMT cavovarus (N 17)
Idiopathic cavovarus (N 17)
*
*
0.98 (0.96–0.99) 0.98 (0.96–0.99)** 0.97 (0.94–0.99)*
0.97 (0.94–0.99)* 0.97 (0.94–0.99)* 0.97 (0.94–0.99)*
0.97 (0.94–0.99) 0.94 (0.89–0.96)* 0.84 (0.58–1.00)*
By severity
FAO > 0% (N 13)
FAO
FAO CO HAA
0.86 (0.73–1.00)* 0.86 (0.75–0.98)** 0.84 (0.64–1.00)*
0.84 (0.79–0.88)* 0.90 (0.81–0.99)* 0.88 (0.73–1.00)*
7% to 0% (N 13)
FAO
Controls (N 17)
13% to
7% (N 12)
0.91 (0.81–0.99)* 0.88 (0.78–0.95)* 0.92 (0.85–1.00)*
FAO < 13% (N 13) 0.95 (0.91–0.99)* 0.93 (0.88–0.96)* 0.94 (0.90–0.98)*
ICC: Intraclass Correlation Coefficient; 95% CI: 95% confidence interval; CMT: Charcot-Marie-Tooth; FAO: foot and ankle offset; CO: calcaneal offset; HAA: hindfoot alignment angle. * Normally-distributed variables. ** Nonnormally-distributed variables.
and calcaneal tuberosity, in addition to the most proximal and central aspect of the talar dome (Fig. 2) [16]. The FAO was described as a three-dimensional measurement of the torque acting in the ankle joint as result of body weight and ground reaction forces. It takes into consideration the relationship between the centre of gravity of the tripod of the foot and the centre of the ankle joint, represented by the apex of the talar dome. Negative measurements indicate a varus alignment, with the centre of the ankle lying laterally to the bisecting line of the foot tripod. Positive values represent a valgus alignment, with the centre of the ankle joint positioned medially to the foot line. Normative values reported by Lintz are 2.3% 2.9 in normal population, 11.6% 6.9 and 11.4% 5.7 in varus and valgus cases respectively [16]. The CO represents the distance (in mm) between a theoretically neutral position of the calcaneus and the true position of the calcaneus (Fig. 3). The HAA represents an estimate of the coronal angular alignment of the hindfoot and is measured as the angle formed by
three points: apex of the centre of the talar dome projected on the floor plane (as the vertex), the theoretical neutral position of the calcaneus and the actual position of the calcaneus (Fig. 3). A measurement protocol detailing comprehensive instructions was sent to three observers including two board-certified senior foot and ankle fellows (of which one had 3 years experience with measurements on WBCT images) and one orthopaedic resident. Of note, the protocol used was adopted from the original by Lintz et al. [16], with two further additions: (1) in order to achieve a clearer visualisation of the talar dome on coronal and sagittal views for the identification of the talar landmark, alignment of the horizontal floor with the bimalleolar axis on the axial plane was performed (Fig. 4); (2) in feet where multiple weightbearing points of the calcaneus were seen, the one corresponding to the vertical axis of the bone (on coronal view) was chosen (Fig. 5). After completing training with images from 3 pes cavovarus patients who were not included in the study, each investigator
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
6
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
Table 3 Intraobserver reliability in the whole cohort and in subgroups. Measurement
Mean r or r (95% CI)
FAO CO HAA
0.98 (0.98–0.98)* 0.97 (0.94–0.99)* 0.97 (0.92–1.00)*
Subgroup analysis By aetiology
CMT cavovarus (N 17)
Idiopathic cavovarus (N 17)
Controls (N 17)
FAO CO HAA
0.93 (0.89–0.97)* 0.93 (0.78–1.00)** 0.95 (0.88–1.00)*
0.93 (0.87–0.99)* 0.93 (0.82–1.00)* 0.93 (0.82–1.00)*
0.91 (0.69–1.00)* 0.89 (0.71–1.00)** 0.91 (0.72–1.00)*
By severity
FAO > 0% (N 13)
FAO
FAO CO HAA
0.81 (0.33–1.00)* 0.78 (0.30–1.00)** 0.85 (0.67–1.00)*
0.77 (0.70–0.85)* 0.86 (0.63–1.00)* 0.80 (0.55–1.00)*
7% to 0% (N 13)
FAO
13% to
7% (N 12)
0.88 (0.74–1.00)* 0.85 (0.64–1.00)* 0.81 (0.58–1.00)*
FAO < 13% (N 13) 0.86 (0.57–1.00)* 0.85 (0.62–1.00)* 0.92 (0.77–1.00)*
95% CI: 95% confidence interval; CMT: Charcot-Marie-Tooth; FAO: foot and ankle offset; CO: calcaneal offset; HAA: hindfoot alignment angle. * r = Pearson's correlation coefficient. ** r = Spearman's rank correlation coefficient.
independently measured FAO, CO and HAA on 51 feet. Each observer subsequently repeated all measurements following a two-week interval. Investigators were blinded to the patient's identification and measurements taken by the other investigators and the order of cases was randomised. 2.4. Statistical analysis Data were reported as mean value standard deviation (SD) and 95% confidence interval (95% CI). The Shapiro–Wilk test was utilised to assess normality of data distribution. Demographic (age, sex, side and BMI) differences between the three groups were assessed using an one-way analysis of variance (ANOVA) test for parametric continuous data, Kruskal–Wallis test for continuous non-parametric data and extended Fisher's exact test for categorical variables. Differences amongst observers were also investigated through Bonferroni correction. The Pearson or Spearman correlation test was used to assess intraobserver reliability, whilst Intraclass Correlation Coefficients (ICCs) and Bland–Altman plots with limits of agreement were calculated for interobserver reliability. Subgroup analysis was performed to assess whether clinical diagnosis (neurological cavovarus, nonneurological cavovarus, normal controls) or severity of varus deformity (4 groups based on FAOs 25th, 50th and 75th centiles) affected reliability of FAO measurements. Results were considered excellent if >0.74; good, 0.60–0.74; fair, 0.40–0.59; and poor, <0.40 [18]. Mean values for measurements taken by three observers were compared as overmentioned. p-value was set at 0.05. 3. Results 3.1. Overall Demographics for patients included are depicted in Table 1. Age and BMI were not distributed normally. No significant difference was found in three groups when comparing age, sex, laterality and BMI. Mean value for FAO SD was 12% 6.2 in pes cavovarus patients (including both CMT related and idiopathic ones) and 0.5% 3 in controls. Mean CO SD was 20.1 mm 9.5 and 0.9 mm 5.9, whilst mean HAA SD 38.3 19.1 and 1.8 9.5 respectively. Intra and interobserver reliability for FAO (r = 0.98; ICC: 0.98), CO (r = 0.97; ICC: 0.98) and HAA measurements (r = 0.97; ICC: 0.98) was excellent (Tables 2 and 3). Bland–Altman analysis for FAO found an average difference SD between observers of
0.06% 1 (Fig. 6A), whereas for CO and HAA average differences SD were 0.1 mm 2 and 0.2 2.8, respectively (Fig. 6B and C). 3.2. Subgroup analysis Subgroup analysis showed that FAO's intra and interobserver reliability remained excellent in CMT related cavovarus (r = 0.93; ICC: 0.98), idiopathic deformities (r = 0.93; ICC: 0.97) and normal feet (r = 0.91; ICC: 0.97). In four subgroups with different severity of varus, mean values for intraobserver (r range, 0.77–0.87) and interobserver reliability for FAO (ICC range, 0.84–0.95) were excellent regardless of the deformity (Tables 2 and 3). Similarly, subgroup analysis for CO and HAA demostrated an excellent intra and interobserver reliability for both measurements in all subgroups (Tables 2 and 3). 3.3. Differences according to level of expertise No difference was found in FAO, CO and HAA mean values from three observers, regardless of seniority status (p > 0.05 in all cases). The Bonferroni correction showed no differences between the first vs second, the first vs third and the second vs third observer (p 1.00 in all cases) (Table 4). 4. Discussion This study confirmed that 3D semi-automatic measurements used on weightbearing CT datasets show excellent intra and interobserver reliability for assessing hindfoot alignment in CMT related cavovarus, idiopathic pes cavovarus and normal controls. Furthermore, the aetiology (neurological or non-neurological) and severity of foot malalignment did not affect reliability of measurements. Mean values of measurements taken by observers with different level of seniority were also comparable. The introduction of WBCT has permitted collection of detailed foot imaging to be obtained in a physiological standing position [13,15]. It is established that the quality of images acquired through cone beam machines are comparable to traditional fan beam CT scanners [11,12], with lower radiation exposure for patients (0.01– 0.03 mSv vs 0.07 mSv) [13]. This has fostered enthusiasm regarding potential application of this new technology in the management of complex three-dimensional deformities including hallux valgus [19–21], hallux rigidus [22], syndesmotic injuries [14,23–25], chronic ankle instability [26] and pes planus [12,27–30].
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
Fig. 6. Bland–Altman analyses of agreement between observers for FAO (A), CO (B) and HAA (C). Each point represents the difference between two observers for a single patient (y axis) relative to the average value of these measurements (x axis). Overall, there were 153 points for each analysis (51 patients 3 observers). The even distribution around the bias line indicates repeatable measurements.
The majority of the aforementioned studies used twodimensional measurements, but in 2017 Lintz et al. proposed using FAO as a 3D biometric tool to take full advantage of the 3D images. Compared to traditional measurements, 3D biometrics reflect a more valid biomechanical approach taking into consideration the offset between ground reaction force and body weight, rather than simply angulation [16]. Authors reported excellent intra and interobserver reliability (0.97 and 0.99, respectively) on a retrospective cohort of 135 feet of which 57 normally aligned, 38 varus and 40 valgus feet [16]. More recently, Zhang confirmed high reliability (0.98 for both intra and interobserver analysis) in a prospective assessment of 249 feet (including 115 well aligned feet, 56 varus and 78 valgus)
7
[17]. A further study assessed FAO (together with other measurements) on 45 professional basketball athletes, with authors finding a tendency towards a varus hindfoot in that cohort [31]. Although not documented in previous studies, the use of TALASTM software (CubeViewTM, CurveBeam©) allows further calculations including CO and HAA as additional indicators of foot alignment (Fig. 3). Interestingly, measurements performed on WBCT scans from patients with acquired adult flatfoot (AAFD) have proven highly reliable [12,27,28], being superior to traditional CT in showing the ‘true’ osseous malalignment affecting these patients [12]. Although pes cavovarus is anatomically and clinically a completely distinct entity from AAFD, they both share a three-dimensional multiplanar complexity and a wide spectrum of clinical presentations (from asymptomatic to a highly stiff painful disabling foot) [5–7]. To the best of our knowledge, this is the first study to document the use of WBCT in pes cavovarus assessment with dedicated subgroup analysis. We showed that FAO, CO and HAA may be reliably used in more severe deformities, which is paramount in the assessment of extremely dysmorphic cavovarus feet. The FAO mean values have been shown to be highly consistent with the varus hindfoot population in the first study from Lintz ( 12% vs 11%, respectively) on a similar sample size (38 vs 34 feet), but somewhat different from Zhang's analysis who found a mean FAO at 6.6% on 56 varus feet. We believe that this discrepancy may be explained by the fact that ‘varus alignment’ is a generic definition encompassing a range from ‘subtle’ to ‘extreme’, therefore a comparison would be appropriate only between specific subgroups. In both studies the cause of varus hindfoot (neurological vs non-neurological) was not reported, and it is unclear as to whether both varus and cavus deformity were considered [16,17]. In the original study CO and HAA mean values were not reported [16], which we believe is a further strength of our study. The adoption of semi-automatic software may permit more reliable measurements to be obtained, reducing drawbacks related to potentially flawed interpretation of traditional radiographs in assessing cavovarus feet [8–10]. It will theoretically allow more accurate monitoring of disease progression and objective assessment of surgical correction. Although interobserver reliability is excellent for each of the three measurements (between 0.97and 0.98) with an even distribution of values around the bias line in Bland–Altman plots, FAO is superior to CO and HAA since in more severe deformities (at around 25 mm for CO and over 55 degrees for HAA) FAO values seem visually closer to limits of agreement of the plot with a lower number of outliers compared to CO and HAA (Fig. 6). No other studies in the literature have documented values for these three measurements, meaning no comparison is possible at this stage. Further studies are warranted to verify our findings. In this study we adopted a shared protocol to identify the landmarks necessary for 3D biometrics calculation, as per Lintz's original description [16]. An assumption was made that talar neck adduction and internal torsion may jeopardise the correct identification of the true coronal plane of the ankle. In other studies, axial plane measurements were performed by aligning the horizontal edge of the axial view with the axis of the first [12,28] or second metatarsal [32], however in severely dysmorphic feet this does not produce clear coronal and sagittal views of the talus (Fig. 4A and B). In our study, we have found that by selecting the axial plane along the bimalleolar axis, a more accurate representation of the talar dome on the other two planes was achieved (Fig. 4C). Similarly, longstanding calcaneal varus may cause reactive infero-lateral bony spurs which may confound the position of the most plantar based weightbearing point described by Lintz and so affect the FAO measurement. In this case, amongst
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
8
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
Table 4 Comparison of mean values for FAO, CO and HAA from three observers in our series.
FAO (%) Overall CMT cavovarus Idiopathic cavovarus Controls CO (mm) Overall CMT cavovarus Idiopathic cavovarus Controls HAA (degrees) Overall CMT cavovarus Idiopathic cavovarus Controls *
Board-certified senior foot and ankle fellow
Board-certified senior foot and ankle fellow
Orthopaedic resident
Mean
Mean
Mean
95% CI
95% CI
p-Value*
95% CI
7.8 14.1 10 0.7
10.1, 5.5 17.7, 10.6 12.1, 8 0.7, 2.1
7.6 13.6 9.8 0.5
9.8, 5.4 17.1, 10 12, 7.6 0.8, 1.9
7.9 14.1 10.1 0.4
10.2, 5.6 17.7, 10.5 12.3, 7.9 1, 1.9
0.97 0.96 0.97 0.96
13.2 23.6 17.2 1.1
16.9, 9.5 29.1, 18 20.1, 14.3 1.6, 4
12.7 22.5 16.5 0.8
16.3, 9.1 28.2, 16.8 19.6, 13.4 1.8, 3.5
13.5 23.3 17.8 0.6
17.2, 9.7 28.8, 17.7 21, 14.6 2.2, 3.6
0.95 0.95 0.82 0.96
24.8 44.9 31.9 2.3
31.9, 17.7 55.5, 34.3 38.4, 25.3 2.2, 6.8
24.7 44.2 31.7 1.8
31.8, 17.6 54.9, 33.6 38.7, 24.7 2.6, 6.3
25.2 45.5 31.7 1.3
32.4, 18.1 56.3, 34.6 38.3, 25.1 3.3, 6
0.99 0.98 0.99 0.95
ANOVA test.
multiple lowest points, for standardisation we recommend that the true weightbearing point will be in line with the vertical axis of the calcaneus (on coronal view) (Fig. 5). We believe that a more precise definition of these points has contributed to the excellent reliability shown in our series. We acknowledge limitations of this study. Firstly, the retrospective design and limited sample size. However, both points are mitigated as no other study exists investigating the use of WBCT measurements in pes cavovarus with different aetiologies. Secondly, in our series we have considered a limited number of measurements. Indeed, other authors have already reported excellent reliability of two-dimensional measurements in cavovarus feet [33,34], whereas our feeling is that research on threedimensional semi-automatic software is still lacking in the literature. Thirdly, despite being proven highly reliable, FAO may still be considered biased due to its semi-automatic nature and the variability in identifying landmarks, and formal validation is still lacking. Research is ongoing to provide fully validated automatic software, which is essential for efficiency and in order to minimise risk of bias, therefore facilitating use in a routine clinical setting. 5. Conclusion In the management of patients presenting with both CMT related or idiopathic pes cavovarus, 3D semi-automatic measurements such as foot and ankle offset, calcaneal offset and hindfoot alignment angle have shown excellent intra and interobserver reliability in the assessment of hindfoot alignment using cone beam weightbearing CT. No difference was shown with regard to the cause or severity of the condition. Measurements taken by surgeons with differing degrees of seniority and varied previous experience with WBCT images did not differ; inferring that 3D biometrics on WBCT images can be reliably utilised in severe neurological and non-neurological cavovarus feet to evaluate and monitor cavovarus deformity and to quantitatively assess the outcome of interventional procedures. Funding No funding has been provided for the current work. Conflict of interest N. Cullen and M. Welck declare to be minority shareholders for Weight-bearing CT Company.
A. Bernasconi, L. Cooper, S. Lyle, S. Patel and D. Singh declare that they have no conflict of interest directly or indirectly related to this work.
References [1] Mosca VS. The cavus foot. J Pediatr Orthop 2001;21:423–4. [2] Helliwell TR, Tynan M, Hayward M, Klenerman L, Whitehouse G, Edwards RH. The pathology of the lower leg muscles in pure forefoot pes cavus. Acta Neuropathol 1995;89:552–9. [3] Manoli A, Graham B. The subtle cavus foot, “the underpronator”. Foot Ankle Int 2005;26:256–63, doi:http://dx.doi.org/10.1177/107110070502600313. [4] Burns J, Landorf KB, Ryan MM, Crosbie J, Ouvrier RA. Interventions for the prevention and treatment of pes cavus. Cochrane Database Syst Rev 2007; CD006154, doi:http://dx.doi.org/10.1002/14651858.CD006154.pub2. [5] Wicart P. Cavus foot, from neonates to adolescents. Orthop Traumatol Surg Res 2012;98:813–28, doi:http://dx.doi.org/10.1016/j.otsr.2012.09.003. [6] Maynou C, Szymanski C, Thiounn A. The adult cavus foot. EFORT Open Rev 2017;2:221–9, doi:http://dx.doi.org/10.1302/2058-5241.2.160077. [7] Kaplan JRM, Aiyer A, Cerrato RA, Jeng CL, Campbell JT. Operative treatment of the cavovarus foot. Foot Ankle Int 2018;39:1370–82, doi:http://dx.doi.org/ 10.1177/1071100718798817. [8] Willauer P, Sangeorzan BJ, Whittaker EC, Shofer JB, Ledoux WR. The sensitivity of standard radiographic foot measures to misalignment. Foot Ankle Int 2014;35:1334–40, doi:http://dx.doi.org/10.1177/1071100714549188. [9] Barg A, Amendola RL, Henninger HB, Kapron AL, Saltzman CL, Anderson AE. Influence of ankle position and radiographic projection angle on measurement of supramalleolar alignment on the anteroposterior and hindfoot alignment views. Foot Ankle Int 2015;36:1352–61, doi:http://dx.doi.org/10.1177/ 1071100715591091. [10] Baverel L, Brilhault J, Odri G, Boissard M, Lintz F. Influence of lower limb rotation on hindfoot alignment using a conventional two-dimensional radiographic technique. Foot Ankle Surg 2017;23:44–9, doi:http://dx.doi. org/10.1016/j.fas.2016.02.003. [11] Richter M, Seidl B, Zech S, Hahn S. PedCAT for 3D-imaging in standing position allows for more accurate bone position (angle) measurement than radiographs or CT. Foot Ankle Surg 2014;20:201–7, doi:http://dx.doi.org/10.1016/j. fas.2014.04.004. [12] de Cesar Netto C, Schon LC, Thawait GK, da Fonseca LF, Chinanuvathana A, Zbijewski WB, et al. Flexible adult acquired flatfoot deformity. J Bone Joint Surg 2017;99:e98, doi:http://dx.doi.org/10.2106/JBJS.16.01366. [13] Barg A, Bailey T, Richter M, de Cesar Netto C, Lintz F, Burssens A, et al. Weightbearing computed tomography of the foot and ankle: emerging technology topical review. Foot Ankle Int 2018;39:376–86, doi:http://dx.doi. org/10.1177/1071100717740330. [14] Malhotra K, Welck M, Cullen N, Singh D, Goldberg AJ. The effects of weight bearing on the distal tibiofibular syndesmosis: a study comparing weight bearing-CT with conventional CT. Foot Ankle Surg 2018, doi:http://dx.doi.org/ 10.1016/j.fas.2018.03.006. [15] Lintz F, Cesar Netto C de, Barg A, Burssens A, Richter M, Weight Bearing CT International Study Group. Weight-bearing cone beam CT scans in the foot and ankle. EFORT Open Rev 2018;3:278–86, doi:http://dx.doi.org/10.1302/20585241.3.170066. [16] Lintz F, Welck M, Bernasconi A, Thornton J, Cullen NP, Singh D, et al. 3D biometrics for hindfoot alignment using weightbearing CT. Foot Ankle Int 2017;38:684–9, doi:http://dx.doi.org/10.1177/1071100717690806.
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx [17] Zhang JZ, Lintz F, Bernasconi A, Zhang S. 3D biometrics for hindfoot alignment using weightbearing computed tomography. Foot Ankle Int 2019;40:720–6, doi:http://dx.doi.org/10.1177/1071100719835492. [18] Senn S. Review of Fleiss, statistical methods for rates and proportions. Res Synth Methods 2011;2:221–2, doi:http://dx.doi.org/10.1002/jrsm.50. [19] Kimura T, Kubota M, Taguchi T, Suzuki N, Hattori A, Marumo K. Evaluation of first-ray mobility in patients with hallux valgus using weight-bearing CT and a 3-D analysis system. J Bone Joint Surg 2017;99:247–55, doi:http://dx.doi.org/ 10.2106/JBJS.16.00542. [20] Kimura T, Kubota M, Suzuki N, Hattori A, Marumo K. Comparison of intercuneiform 1-2 joint mobility between hallux valgus and normal feet using weightbearing computed tomography and 3-dimensional analysis. Foot Ankle Int 2018;39:355–60, doi:http://dx.doi.org/10.1177/ 1071100717744174. [21] Welck MJ, Singh D, Cullen N, Goldberg A. Evaluation of the 1st metatarsosesamoid joint using standing CT—the Stanmore classification. Foot Ankle Surg 2018;24:314–9, doi:http://dx.doi.org/10.1016/j.fas.2017.03.005. [22] Cheung ZB, Myerson MS, Tracey J, Vulcano E. Weightbearing CT scan assessment of foot alignment in patients with hallux rigidus. Foot Ankle Int 2018;39:67–74, doi:http://dx.doi.org/10.1177/1071100717732549. [23] Osgood GM, Shakoor D, Orapin J, Qin J, Khodarahmi I, Thawait GK, et al. Reliability of distal tibio-fibular syndesmotic instability measurements using weightbearing and non-weightbearing cone-beam CT. Foot Ankle Surg 2018, doi:http://dx.doi.org/10.1016/j.fas.2018.10.003. [24] Krähenbühl N, Bailey TL, Weinberg MW, Davidson NP, Hintermann B, Presson AP, et al. Impact of torque on assessment of syndesmotic injuries using weightbearing computed tomography scans. Foot Ankle Int 2019, doi:http:// dx.doi.org/10.1177/1071100719829720 107110071982972. [25] Burssens A, Vermue H, Barg A, Krähenbühl N, Victor J, Buedts K. Templating of syndesmotic ankle lesions by use of 3D analysis in weightbearing and nonweightbearing CT. Foot Ankle Int 2018;39:1487–96, doi:http://dx.doi.org/ 10.1177/1071100718791834.
9
[26] Lintz F, Bernasconi A, Baschet L, Fernando C, Mehdi N, de Cesar Netto C. Relationship between chronic lateral ankle instability and hindfoot varus using weight-bearing cone beam computed tomography. Foot Ankle Int 2019, doi:http://dx.doi.org/10.1177/1071100719858309 107110071985830. [27] Probasco W, Haleem AM, Yu J, Sangeorzan BJ, Deland JT, Ellis SJ. Assessment of coronal plane subtalar joint alignment in peritalar subluxation via weightbearing multiplanar imaging. Foot Ankle Int 2015;36:302–9, doi:http://dx.doi. org/10.1177/1071100714557861. [28] de Cesar Netto C, Shakoor D, Roberts L, Chinanuvathana A, Mousavian A, Lintz F, et al. Hindfoot alignment of adult acquired flatfoot deformity: a comparison of clinical assessment and weightbearing cone beam CT examinations. Foot Ankle Surg 2018, doi:http://dx.doi.org/10.1016/j.fas.2018.10.008. [29] Jeng CL, Rutherford T, Hull MG, Cerrato RA, Campbell JT. Assessment of bony subfibular impingement in flatfoot patients using weight-bearing CT scans. Foot Ankle Int 2018;40:, doi:http://dx.doi.org/10.1177/1071100718804510 1071100718804510. [30] Bernasconi A, de Cesar Netto C, Barg A, Burssens A, Richter A, Weight Bearing CT International Study Group LF. AAFD: conventional radiographs are not enough! I need the third dimension. Tech Foot Ankle Surg 2019 April 10. [31] de Cesar Netto C, Bernasconi A, Roberts L, Pontin PA, Lintz F, Saito GH, et al. Foot alignment in symptomatic national basketball association players using weightbearing cone beam computed tomography. Orthop J Sport Med 2019, doi:http://dx.doi.org/10.1177/2325967118824149 Feb 21. [32] Burssens A, Peeters J, Buedts K, Victor J, Vandeputte G. Measuring hindfoot alignment in weight bearing CT: a novel clinical relevant measurement method. Foot Ankle Surg 2016;22:233–8, doi:http://dx.doi.org/10.1016/j. fas.2015.10.002. [33] Arunakul M, Amendola A, Gao Y, Goetz JE, Femino JE, Phisitkul P. Tripod index. Foot Ankle Int 2013;34:1411–20, doi:http://dx.doi.org/10.1177/1071100713488761. [34] Podolnick JD, Donovan DS, DeBellis N, Pino A. Is pes cavus alignment associated with lisfranc injuries of the foot? Clin Orthop Relat Res 2017;475:1463–9, doi:http://dx.doi.org/10.1007/s11999-016-5131-6.
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
Foot and Ankle Surgery xxx (2019) xxx–xxx
Contents lists available at ScienceDirect
Foot and Ankle Surgery journal homepage: www.elsevier.com/locate/fas
Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus Alessio Bernasconia,b,* , Lucy Coopera , Shirley Lylea , Shelain Patela , Nicholas Cullena , Dishan Singha , Matthew Welcka a b
Foot and Ankle Unit, Royal National Orthopaedic Hospital, Stanmore, United Kingdom Department of Public Health, Trauma and Orthopaedics, University of Naples Federico II, Naples, Italy
A R T I C L E I N F O
A B S T R A C T
Article history: Received 19 April 2019 Received in revised form 30 June 2019 Accepted 17 July 2019 Available online xxx
Introduction: Pes cavovarus is a three-dimensional (3D) foot deformity. New 3D semi-automatic measurements utilising weightbearing computerised topography (WBCT) images have recently been proposed to assess hindfoot alignment, but reliability in pes cavovarus has never been investigated. The aim of this study was to assess intraobserver and interobserver reliability of the foot ankle offset (FAO), calcaneal offset (CO) and hindfoot alignment angle (HAA) in pes cavovarus. Methods: Anonymised WBCT datasets from 51 feet (17 Charcot-Marie-Tooth related cavovarus, 17 idiopathic cavovarus and 17 controls) were retrospectively reviewed. Three observers (two senior foot and ankle fellows and one orthopaedic resident) independently measured FAO, CO and HAA using dedicated software, with measurements repeated two weeks apart. Subgroup analysis was performed to assess whether aetiology or severity of varus deformity and level of seniority affected reliability. Results: Mean values for intra and interobserver reliability for FAO (r = 0.98; ICC: 0.99), CO (r = 0.97; ICC: 0.98) and HAA measurements (r = 0.97; ICC: 0.98) were excellent. Subgroup analyses showed that FAO, CO and HAA's intra (r/r range, 0.77–0.95) and interobserver (ICC range, 0.88–0.98) reliability remained excellent in patients with Charcot-Marie-Tooth related cavovarus, idiopathic pes cavovarus and normal feet, regardless of the severity of deformity. No difference was found in FAO, CO and HAA mean values from three observers (p > 0.05 in all cases). Discussion: This study demonstrates that 3D semi-automatic measurements of WBCT images have excellent intra and interobserver reliability in the assessment of hindfoot alignment in pes cavovarus. Aetiology and severity of deformity, and level of seniority do not affect reliability of these measurements. Level of evidence: Level III, retrospective comparative study. © 2019 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved.
Keywords: Cavovarus Hindfoot High arch Charcot-Marie-Tooth Weightbearing CT Cone beam
1. Introduction Pes cavovarus is a three-dimensional (3D) complex foot deformity characterised by a varus hindfoot, high longitudinal arch and forefoot pronation. Although multiple pathologies may lead to its development, they can be broadly categorised into idiopathic [1–3] and acquired, whilst acquired can be sub-divided into neurological and nonneurological [4–7]. Charcot-Marie-Tooth (CMT) disease, anterior horn
* Corresponding author at: Foot and Ankle Unit, Royal National Orthopaedic Hospital, Middlesex, HA7 4LP, Stanmore, United Kingdom. E-mail addresses: [email protected] (A. Bernasconi), [email protected] (L. Cooper), [email protected] (S. Lyle), [email protected] (S. Patel), [email protected] (N. Cullen), [email protected] (D. Singh), [email protected] (M. Welck).
diseases, disorders of the central nervous system and poliomyelitis account for the most common neurological causes [4–7]. The diagnostic workup in cavovarus feet is traditionally based on thorough clinical assessment and investigations to establish aetiology followed by dedicated imaging, including weightbearing radiographs and computed tomography (CT) or magnetic resonance imaging [3–7]. Taking into account the 3D nature of cavovarus deformity and the necessity for multiplanar correction at different sites (hindfoot, midfoot and forefoot), detailed imaging of the foot during physiological weightbearing stance is extremely helpful for accurate surgical planning. Two-dimensional (2D) radiographs tend to be flawed by perspective, operator-related and superimposition biases [8–10], whilst standard CT, performed either non-weightbearing or with simulated weightbearing, does not correctly reproduce the standing foot posture [11–14]. These drawbacks may be overcome via 3D cone beam weightbearing
https://doi.org/10.1016/j.fas.2019.07.005 1268-7731/© 2019 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
2
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
Fig. 1. Flow chart of the 51 feet eligible for this study.
computed tomography (WBCT) which combines physiological weightbearing with lower doses of radiation than conventional CT [13,15]. The concept of 3D biometrics is a recent addition to the literature in order to study hindfoot alignment through the assessment of foot position in space and not only in single planes [16]. This is realised through the identification of 4 points on WBCT images to define a volume (rather than 3 points defining an angle) corresponding to a pyramid whose base is the traditional foot tripod and a vertex represented by the centre of the ankle joint. Dedicated software then processes the coordinates of these points allowing calculation of three parameters: foot and ankle offset (FAO), calcaneal offset (CO) and hindfoot alignment angle (HAA). Although from different perspectives, they all measure the
difference between the true hindfoot in space and a theoretical neutral one, enabling quantification of valgus or varus deformity and negating the aforementioned 2D imaging-related biases. To date, the most studied of the measurements is FAO, which corresponds to the lever arm generated at the ankle by two opposite forces, namely body weight (downwards) and the ground reaction force (upwards). Two previous studies have proven excellent intra and interobserver agreement for these 3D semiautomatic measurements [16,17]. The aim of this study was to assess intra and interobserver reliability of 3D biometrics on WBCT imaging in pes cavovarus patients diagnosed with CMT disease and idiopathic cavovarus. Our hypothesis was that FAO, CO and HAA were reliable measurements regardless of
Table 1 Patients’ demographics. Clinical characteristics of pes cavovarus patients (N 34) 45.5 16.1 (15–72) 31 (61) 21 (41) 30 6.7 (19–46)
Age (year) Male, N (%) Right side, N (%) BMI (kg/m2) Subgroups By aetiology
CMT (N 17)
Idiopathic (N 17)
Controls (N 17)
p-Value
Age (year) Male, N (%) Right side, N (%) BMI (kg/m2)
46.2 13 (39–53) 10 (58) 7 (42) 31.3 8 (27–35)
44 19 (33–54) 12 (70) 7 (42) 29.9 5 (27–32)
46.1 17 (37–55) 11 (64) 7 (42) 28.7 6 (25–32)
0.975* 0.932** 1** 0.738*
By severity
FAO > 0% (N 13)
FAO
Age (year) Male, N (%) Right side, N (%) BMI (kg/m2)
48.1 18.7 (37–58) 6 (46) 5 (38) 27 6.3 (23–30)
42.2 13 (34–49) 10 (77) 5 (38) 31.6 6.3 (28–35)
7% to 0% (N 13)
FAO
13% to
7% (N 12)
47.6 18.3 (37–58) 6 (50) 5 (41) 27.8 5.3 (24–31)
FAO < 13% (N 13) 44.2 17.5 (34–54) 9 (70) 6 (46) 32.8 7.5 (28–37)
BMI: Body Mass Index; CMT: Charcot-Marie-Tooth. Age and BMI are reported as: mean standard deviation; 95% confidence interval values in brackets. * Extended Fisher's exact test. ** Kruskal–Wallis test rank test.
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
3
Fig. 2. Illustration of the foot tripod with foot and ankle offset (FAO) calculation principle. Four landmarks that must be manually identified are the weightbearing point of the first metatarsal head (M1), the weightbearing point of the fifth metatarsal head (M5), the weightbearing point (lowest point) of the calcaneus (C) and the uppermost and central point of the talar dome (T) (left upper inset). Thereafter, the software will calculate the projection of T on the ground (T1) and the distance (d) between this latter and C1 (which corresponds to the point on the bisecting line of the triangle where the perpendicular to the bisecting line itself intersects T1). That distance given in percentage of the foot length (d/CH, where H is the midpoint of the M1–M5 line) will be FAO. Drawings by the authors superimposed to original image adapted from BoneBoxTM – Foot (©2014 iSO-FORM, LLC).
aetiology and severity of deformity. We also assessed differences in reliability for these measurements by observers with different degrees of seniority and different experience with WBCT imaging.
standards. The study was conducted following the STROBE guidelines. 2.2. Setting and participants
2. Methods 2.1. Study design This comparative retrospective study analysed existing data recorded as part of routine clinical care. All procedures were performed in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical
Cone beam WBCT datasets from 201 feet with a diagnosis of pes cavovarus imaged between May 2013 and June 2017 in the Foot and Ankle Unit of our institution were reviewed for this study. After removal of duplicated scans, 172 feet (100 patients) were available. All WBCT scans were performed as part of routine investigation. Initially, patients who met the following criteria were selected: datasets with no artefact and from skeletally mature patients with diagnosis of symptomatic pes cavovarus and CMT disease, with no
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
4
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
Fig. 3. Illustration of calcaneal offset (CO) and hindfoot alignment angle (HAA). The former is the distance between the true weightbearing point of the calcaneus (C) and the hypothetical weightbearing point (C2) if the bisecting line of the tripod passed through T1 (that is the projection on the ground of the centre of talar dome). HAA is estimative of the coronal angular alignment of the hindfoot, being measured as the angle formed by three points: centre of the talar dome projection on the ground (T1, as apex), the ideal position of the calcaneus (C2) and the actual position of the calcaneus (C). Drawings by the authors superimposed to original image adapted from BoneBoxTM – Foot (©2014 iSO-FORM, LLC).
Fig. 4. Example of identification of talar dome in severely dismorphic foot. In the upper row (A) the foot has been aligned along the second metatarsal bone (left upper quadrant) whilst in the middle row (B) along the first metatarsal one (left middle quadrant). In both cases the visualisation of the talar dome on the coronal (upper middle and central quadrants) and sagittal views (right upper and right middle quadrants) is unclear. In the lower row (C) the image is aligned along the bimalleolar axis (left lower quadrant), leading to a satisfactory visualisation of the ankle joint on both the coronal and sagittal plane (middle lower and right lower quadrant, respectively).
previous ipsilateral foot/ankle surgery, ability to heel weightbear (no severe fixed equinus deformity) and to stand unassisted for the time of a WBCT scan (approximately 50 s) (Fig. 1). Seventeen neurological pes cavovarus were enrolled and a further seventeen idiopathic cavovarus were matched by sex, side, age and BMI from the original pool (ratio 1:1). Seventeen datasets from normal feet (matched for the same variables) were selected as controls (Fig. 1). Overall, anonymised datasets from 51 feet were available (34 with cavovarus morphotype and 17 normal feet) (Table 1).
2.3. Variables Datasets were analysed using the TALASTM software plugin for CubeViewTM (CurveBeam©). The 3D coordinates of specific anatomical landmarks were required for the software to automatically process and FAO, CO and HAA, as described by Lintz et al. [16]. These coordinates included the most plantar and weightbearing voxel of the head of the 1st metatarsal, head of the 5th metatarsal
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
5
Fig. 5. Example of identification of weightbearing point of the calcaneus in severely dismorphic foot. In the axial view (left) there are multiple ‘lower’ points of the bone that may be misleading. On the coronal (centre) and sagittal plane (right) it is clear that one true ‘weightbearing point’ corresponds to the vertical axis of the bone, being the others osteophytes or calcifications.
Table 2 Interobserver reliability in the whole cohort and in subgroups. Measurement
Mean ICC (95% CI)
FAO CO HAA
0.99 (0.99–0.99)* 0.98 (0.96–0.99)* 0.98 (0.96–0.99)*
Subgroup analysis By aetiology FAO CO HAA
CMT cavovarus (N 17)
Idiopathic cavovarus (N 17)
*
*
0.98 (0.96–0.99) 0.98 (0.96–0.99)** 0.97 (0.94–0.99)*
0.97 (0.94–0.99)* 0.97 (0.94–0.99)* 0.97 (0.94–0.99)*
0.97 (0.94–0.99) 0.94 (0.89–0.96)* 0.84 (0.58–1.00)*
By severity
FAO > 0% (N 13)
FAO
FAO CO HAA
0.86 (0.73–1.00)* 0.86 (0.75–0.98)** 0.84 (0.64–1.00)*
0.84 (0.79–0.88)* 0.90 (0.81–0.99)* 0.88 (0.73–1.00)*
7% to 0% (N 13)
FAO
Controls (N 17)
13% to
7% (N 12)
0.91 (0.81–0.99)* 0.88 (0.78–0.95)* 0.92 (0.85–1.00)*
FAO < 13% (N 13) 0.95 (0.91–0.99)* 0.93 (0.88–0.96)* 0.94 (0.90–0.98)*
ICC: Intraclass Correlation Coefficient; 95% CI: 95% confidence interval; CMT: Charcot-Marie-Tooth; FAO: foot and ankle offset; CO: calcaneal offset; HAA: hindfoot alignment angle. * Normally-distributed variables. ** Nonnormally-distributed variables.
and calcaneal tuberosity, in addition to the most proximal and central aspect of the talar dome (Fig. 2) [16]. The FAO was described as a three-dimensional measurement of the torque acting in the ankle joint as result of body weight and ground reaction forces. It takes into consideration the relationship between the centre of gravity of the tripod of the foot and the centre of the ankle joint, represented by the apex of the talar dome. Negative measurements indicate a varus alignment, with the centre of the ankle lying laterally to the bisecting line of the foot tripod. Positive values represent a valgus alignment, with the centre of the ankle joint positioned medially to the foot line. Normative values reported by Lintz are 2.3% 2.9 in normal population, 11.6% 6.9 and 11.4% 5.7 in varus and valgus cases respectively [16]. The CO represents the distance (in mm) between a theoretically neutral position of the calcaneus and the true position of the calcaneus (Fig. 3). The HAA represents an estimate of the coronal angular alignment of the hindfoot and is measured as the angle formed by
three points: apex of the centre of the talar dome projected on the floor plane (as the vertex), the theoretical neutral position of the calcaneus and the actual position of the calcaneus (Fig. 3). A measurement protocol detailing comprehensive instructions was sent to three observers including two board-certified senior foot and ankle fellows (of which one had 3 years experience with measurements on WBCT images) and one orthopaedic resident. Of note, the protocol used was adopted from the original by Lintz et al. [16], with two further additions: (1) in order to achieve a clearer visualisation of the talar dome on coronal and sagittal views for the identification of the talar landmark, alignment of the horizontal floor with the bimalleolar axis on the axial plane was performed (Fig. 4); (2) in feet where multiple weightbearing points of the calcaneus were seen, the one corresponding to the vertical axis of the bone (on coronal view) was chosen (Fig. 5). After completing training with images from 3 pes cavovarus patients who were not included in the study, each investigator
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
6
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
Table 3 Intraobserver reliability in the whole cohort and in subgroups. Measurement
Mean r or r (95% CI)
FAO CO HAA
0.98 (0.98–0.98)* 0.97 (0.94–0.99)* 0.97 (0.92–1.00)*
Subgroup analysis By aetiology
CMT cavovarus (N 17)
Idiopathic cavovarus (N 17)
Controls (N 17)
FAO CO HAA
0.93 (0.89–0.97)* 0.93 (0.78–1.00)** 0.95 (0.88–1.00)*
0.93 (0.87–0.99)* 0.93 (0.82–1.00)* 0.93 (0.82–1.00)*
0.91 (0.69–1.00)* 0.89 (0.71–1.00)** 0.91 (0.72–1.00)*
By severity
FAO > 0% (N 13)
FAO
FAO CO HAA
0.81 (0.33–1.00)* 0.78 (0.30–1.00)** 0.85 (0.67–1.00)*
0.77 (0.70–0.85)* 0.86 (0.63–1.00)* 0.80 (0.55–1.00)*
7% to 0% (N 13)
FAO
13% to
7% (N 12)
0.88 (0.74–1.00)* 0.85 (0.64–1.00)* 0.81 (0.58–1.00)*
FAO < 13% (N 13) 0.86 (0.57–1.00)* 0.85 (0.62–1.00)* 0.92 (0.77–1.00)*
95% CI: 95% confidence interval; CMT: Charcot-Marie-Tooth; FAO: foot and ankle offset; CO: calcaneal offset; HAA: hindfoot alignment angle. * r = Pearson's correlation coefficient. ** r = Spearman's rank correlation coefficient.
independently measured FAO, CO and HAA on 51 feet. Each observer subsequently repeated all measurements following a two-week interval. Investigators were blinded to the patient's identification and measurements taken by the other investigators and the order of cases was randomised. 2.4. Statistical analysis Data were reported as mean value standard deviation (SD) and 95% confidence interval (95% CI). The Shapiro–Wilk test was utilised to assess normality of data distribution. Demographic (age, sex, side and BMI) differences between the three groups were assessed using an one-way analysis of variance (ANOVA) test for parametric continuous data, Kruskal–Wallis test for continuous non-parametric data and extended Fisher's exact test for categorical variables. Differences amongst observers were also investigated through Bonferroni correction. The Pearson or Spearman correlation test was used to assess intraobserver reliability, whilst Intraclass Correlation Coefficients (ICCs) and Bland–Altman plots with limits of agreement were calculated for interobserver reliability. Subgroup analysis was performed to assess whether clinical diagnosis (neurological cavovarus, nonneurological cavovarus, normal controls) or severity of varus deformity (4 groups based on FAOs 25th, 50th and 75th centiles) affected reliability of FAO measurements. Results were considered excellent if >0.74; good, 0.60–0.74; fair, 0.40–0.59; and poor, <0.40 [18]. Mean values for measurements taken by three observers were compared as overmentioned. p-value was set at 0.05. 3. Results 3.1. Overall Demographics for patients included are depicted in Table 1. Age and BMI were not distributed normally. No significant difference was found in three groups when comparing age, sex, laterality and BMI. Mean value for FAO SD was 12% 6.2 in pes cavovarus patients (including both CMT related and idiopathic ones) and 0.5% 3 in controls. Mean CO SD was 20.1 mm 9.5 and 0.9 mm 5.9, whilst mean HAA SD 38.3 19.1 and 1.8 9.5 respectively. Intra and interobserver reliability for FAO (r = 0.98; ICC: 0.98), CO (r = 0.97; ICC: 0.98) and HAA measurements (r = 0.97; ICC: 0.98) was excellent (Tables 2 and 3). Bland–Altman analysis for FAO found an average difference SD between observers of
0.06% 1 (Fig. 6A), whereas for CO and HAA average differences SD were 0.1 mm 2 and 0.2 2.8, respectively (Fig. 6B and C). 3.2. Subgroup analysis Subgroup analysis showed that FAO's intra and interobserver reliability remained excellent in CMT related cavovarus (r = 0.93; ICC: 0.98), idiopathic deformities (r = 0.93; ICC: 0.97) and normal feet (r = 0.91; ICC: 0.97). In four subgroups with different severity of varus, mean values for intraobserver (r range, 0.77–0.87) and interobserver reliability for FAO (ICC range, 0.84–0.95) were excellent regardless of the deformity (Tables 2 and 3). Similarly, subgroup analysis for CO and HAA demostrated an excellent intra and interobserver reliability for both measurements in all subgroups (Tables 2 and 3). 3.3. Differences according to level of expertise No difference was found in FAO, CO and HAA mean values from three observers, regardless of seniority status (p > 0.05 in all cases). The Bonferroni correction showed no differences between the first vs second, the first vs third and the second vs third observer (p 1.00 in all cases) (Table 4). 4. Discussion This study confirmed that 3D semi-automatic measurements used on weightbearing CT datasets show excellent intra and interobserver reliability for assessing hindfoot alignment in CMT related cavovarus, idiopathic pes cavovarus and normal controls. Furthermore, the aetiology (neurological or non-neurological) and severity of foot malalignment did not affect reliability of measurements. Mean values of measurements taken by observers with different level of seniority were also comparable. The introduction of WBCT has permitted collection of detailed foot imaging to be obtained in a physiological standing position [13,15]. It is established that the quality of images acquired through cone beam machines are comparable to traditional fan beam CT scanners [11,12], with lower radiation exposure for patients (0.01– 0.03 mSv vs 0.07 mSv) [13]. This has fostered enthusiasm regarding potential application of this new technology in the management of complex three-dimensional deformities including hallux valgus [19–21], hallux rigidus [22], syndesmotic injuries [14,23–25], chronic ankle instability [26] and pes planus [12,27–30].
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
Fig. 6. Bland–Altman analyses of agreement between observers for FAO (A), CO (B) and HAA (C). Each point represents the difference between two observers for a single patient (y axis) relative to the average value of these measurements (x axis). Overall, there were 153 points for each analysis (51 patients 3 observers). The even distribution around the bias line indicates repeatable measurements.
The majority of the aforementioned studies used twodimensional measurements, but in 2017 Lintz et al. proposed using FAO as a 3D biometric tool to take full advantage of the 3D images. Compared to traditional measurements, 3D biometrics reflect a more valid biomechanical approach taking into consideration the offset between ground reaction force and body weight, rather than simply angulation [16]. Authors reported excellent intra and interobserver reliability (0.97 and 0.99, respectively) on a retrospective cohort of 135 feet of which 57 normally aligned, 38 varus and 40 valgus feet [16]. More recently, Zhang confirmed high reliability (0.98 for both intra and interobserver analysis) in a prospective assessment of 249 feet (including 115 well aligned feet, 56 varus and 78 valgus)
7
[17]. A further study assessed FAO (together with other measurements) on 45 professional basketball athletes, with authors finding a tendency towards a varus hindfoot in that cohort [31]. Although not documented in previous studies, the use of TALASTM software (CubeViewTM, CurveBeam©) allows further calculations including CO and HAA as additional indicators of foot alignment (Fig. 3). Interestingly, measurements performed on WBCT scans from patients with acquired adult flatfoot (AAFD) have proven highly reliable [12,27,28], being superior to traditional CT in showing the ‘true’ osseous malalignment affecting these patients [12]. Although pes cavovarus is anatomically and clinically a completely distinct entity from AAFD, they both share a three-dimensional multiplanar complexity and a wide spectrum of clinical presentations (from asymptomatic to a highly stiff painful disabling foot) [5–7]. To the best of our knowledge, this is the first study to document the use of WBCT in pes cavovarus assessment with dedicated subgroup analysis. We showed that FAO, CO and HAA may be reliably used in more severe deformities, which is paramount in the assessment of extremely dysmorphic cavovarus feet. The FAO mean values have been shown to be highly consistent with the varus hindfoot population in the first study from Lintz ( 12% vs 11%, respectively) on a similar sample size (38 vs 34 feet), but somewhat different from Zhang's analysis who found a mean FAO at 6.6% on 56 varus feet. We believe that this discrepancy may be explained by the fact that ‘varus alignment’ is a generic definition encompassing a range from ‘subtle’ to ‘extreme’, therefore a comparison would be appropriate only between specific subgroups. In both studies the cause of varus hindfoot (neurological vs non-neurological) was not reported, and it is unclear as to whether both varus and cavus deformity were considered [16,17]. In the original study CO and HAA mean values were not reported [16], which we believe is a further strength of our study. The adoption of semi-automatic software may permit more reliable measurements to be obtained, reducing drawbacks related to potentially flawed interpretation of traditional radiographs in assessing cavovarus feet [8–10]. It will theoretically allow more accurate monitoring of disease progression and objective assessment of surgical correction. Although interobserver reliability is excellent for each of the three measurements (between 0.97and 0.98) with an even distribution of values around the bias line in Bland–Altman plots, FAO is superior to CO and HAA since in more severe deformities (at around 25 mm for CO and over 55 degrees for HAA) FAO values seem visually closer to limits of agreement of the plot with a lower number of outliers compared to CO and HAA (Fig. 6). No other studies in the literature have documented values for these three measurements, meaning no comparison is possible at this stage. Further studies are warranted to verify our findings. In this study we adopted a shared protocol to identify the landmarks necessary for 3D biometrics calculation, as per Lintz's original description [16]. An assumption was made that talar neck adduction and internal torsion may jeopardise the correct identification of the true coronal plane of the ankle. In other studies, axial plane measurements were performed by aligning the horizontal edge of the axial view with the axis of the first [12,28] or second metatarsal [32], however in severely dysmorphic feet this does not produce clear coronal and sagittal views of the talus (Fig. 4A and B). In our study, we have found that by selecting the axial plane along the bimalleolar axis, a more accurate representation of the talar dome on the other two planes was achieved (Fig. 4C). Similarly, longstanding calcaneal varus may cause reactive infero-lateral bony spurs which may confound the position of the most plantar based weightbearing point described by Lintz and so affect the FAO measurement. In this case, amongst
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
8
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx
Table 4 Comparison of mean values for FAO, CO and HAA from three observers in our series.
FAO (%) Overall CMT cavovarus Idiopathic cavovarus Controls CO (mm) Overall CMT cavovarus Idiopathic cavovarus Controls HAA (degrees) Overall CMT cavovarus Idiopathic cavovarus Controls *
Board-certified senior foot and ankle fellow
Board-certified senior foot and ankle fellow
Orthopaedic resident
Mean
Mean
Mean
95% CI
95% CI
p-Value*
95% CI
7.8 14.1 10 0.7
10.1, 5.5 17.7, 10.6 12.1, 8 0.7, 2.1
7.6 13.6 9.8 0.5
9.8, 5.4 17.1, 10 12, 7.6 0.8, 1.9
7.9 14.1 10.1 0.4
10.2, 5.6 17.7, 10.5 12.3, 7.9 1, 1.9
0.97 0.96 0.97 0.96
13.2 23.6 17.2 1.1
16.9, 9.5 29.1, 18 20.1, 14.3 1.6, 4
12.7 22.5 16.5 0.8
16.3, 9.1 28.2, 16.8 19.6, 13.4 1.8, 3.5
13.5 23.3 17.8 0.6
17.2, 9.7 28.8, 17.7 21, 14.6 2.2, 3.6
0.95 0.95 0.82 0.96
24.8 44.9 31.9 2.3
31.9, 17.7 55.5, 34.3 38.4, 25.3 2.2, 6.8
24.7 44.2 31.7 1.8
31.8, 17.6 54.9, 33.6 38.7, 24.7 2.6, 6.3
25.2 45.5 31.7 1.3
32.4, 18.1 56.3, 34.6 38.3, 25.1 3.3, 6
0.99 0.98 0.99 0.95
ANOVA test.
multiple lowest points, for standardisation we recommend that the true weightbearing point will be in line with the vertical axis of the calcaneus (on coronal view) (Fig. 5). We believe that a more precise definition of these points has contributed to the excellent reliability shown in our series. We acknowledge limitations of this study. Firstly, the retrospective design and limited sample size. However, both points are mitigated as no other study exists investigating the use of WBCT measurements in pes cavovarus with different aetiologies. Secondly, in our series we have considered a limited number of measurements. Indeed, other authors have already reported excellent reliability of two-dimensional measurements in cavovarus feet [33,34], whereas our feeling is that research on threedimensional semi-automatic software is still lacking in the literature. Thirdly, despite being proven highly reliable, FAO may still be considered biased due to its semi-automatic nature and the variability in identifying landmarks, and formal validation is still lacking. Research is ongoing to provide fully validated automatic software, which is essential for efficiency and in order to minimise risk of bias, therefore facilitating use in a routine clinical setting. 5. Conclusion In the management of patients presenting with both CMT related or idiopathic pes cavovarus, 3D semi-automatic measurements such as foot and ankle offset, calcaneal offset and hindfoot alignment angle have shown excellent intra and interobserver reliability in the assessment of hindfoot alignment using cone beam weightbearing CT. No difference was shown with regard to the cause or severity of the condition. Measurements taken by surgeons with differing degrees of seniority and varied previous experience with WBCT images did not differ; inferring that 3D biometrics on WBCT images can be reliably utilised in severe neurological and non-neurological cavovarus feet to evaluate and monitor cavovarus deformity and to quantitatively assess the outcome of interventional procedures. Funding No funding has been provided for the current work. Conflict of interest N. Cullen and M. Welck declare to be minority shareholders for Weight-bearing CT Company.
A. Bernasconi, L. Cooper, S. Lyle, S. Patel and D. Singh declare that they have no conflict of interest directly or indirectly related to this work.
References [1] Mosca VS. The cavus foot. J Pediatr Orthop 2001;21:423–4. [2] Helliwell TR, Tynan M, Hayward M, Klenerman L, Whitehouse G, Edwards RH. The pathology of the lower leg muscles in pure forefoot pes cavus. Acta Neuropathol 1995;89:552–9. [3] Manoli A, Graham B. The subtle cavus foot, “the underpronator”. Foot Ankle Int 2005;26:256–63, doi:http://dx.doi.org/10.1177/107110070502600313. [4] Burns J, Landorf KB, Ryan MM, Crosbie J, Ouvrier RA. Interventions for the prevention and treatment of pes cavus. Cochrane Database Syst Rev 2007; CD006154, doi:http://dx.doi.org/10.1002/14651858.CD006154.pub2. [5] Wicart P. Cavus foot, from neonates to adolescents. Orthop Traumatol Surg Res 2012;98:813–28, doi:http://dx.doi.org/10.1016/j.otsr.2012.09.003. [6] Maynou C, Szymanski C, Thiounn A. The adult cavus foot. EFORT Open Rev 2017;2:221–9, doi:http://dx.doi.org/10.1302/2058-5241.2.160077. [7] Kaplan JRM, Aiyer A, Cerrato RA, Jeng CL, Campbell JT. Operative treatment of the cavovarus foot. Foot Ankle Int 2018;39:1370–82, doi:http://dx.doi.org/ 10.1177/1071100718798817. [8] Willauer P, Sangeorzan BJ, Whittaker EC, Shofer JB, Ledoux WR. The sensitivity of standard radiographic foot measures to misalignment. Foot Ankle Int 2014;35:1334–40, doi:http://dx.doi.org/10.1177/1071100714549188. [9] Barg A, Amendola RL, Henninger HB, Kapron AL, Saltzman CL, Anderson AE. Influence of ankle position and radiographic projection angle on measurement of supramalleolar alignment on the anteroposterior and hindfoot alignment views. Foot Ankle Int 2015;36:1352–61, doi:http://dx.doi.org/10.1177/ 1071100715591091. [10] Baverel L, Brilhault J, Odri G, Boissard M, Lintz F. Influence of lower limb rotation on hindfoot alignment using a conventional two-dimensional radiographic technique. Foot Ankle Surg 2017;23:44–9, doi:http://dx.doi. org/10.1016/j.fas.2016.02.003. [11] Richter M, Seidl B, Zech S, Hahn S. PedCAT for 3D-imaging in standing position allows for more accurate bone position (angle) measurement than radiographs or CT. Foot Ankle Surg 2014;20:201–7, doi:http://dx.doi.org/10.1016/j. fas.2014.04.004. [12] de Cesar Netto C, Schon LC, Thawait GK, da Fonseca LF, Chinanuvathana A, Zbijewski WB, et al. Flexible adult acquired flatfoot deformity. J Bone Joint Surg 2017;99:e98, doi:http://dx.doi.org/10.2106/JBJS.16.01366. [13] Barg A, Bailey T, Richter M, de Cesar Netto C, Lintz F, Burssens A, et al. Weightbearing computed tomography of the foot and ankle: emerging technology topical review. Foot Ankle Int 2018;39:376–86, doi:http://dx.doi. org/10.1177/1071100717740330. [14] Malhotra K, Welck M, Cullen N, Singh D, Goldberg AJ. The effects of weight bearing on the distal tibiofibular syndesmosis: a study comparing weight bearing-CT with conventional CT. Foot Ankle Surg 2018, doi:http://dx.doi.org/ 10.1016/j.fas.2018.03.006. [15] Lintz F, Cesar Netto C de, Barg A, Burssens A, Richter M, Weight Bearing CT International Study Group. Weight-bearing cone beam CT scans in the foot and ankle. EFORT Open Rev 2018;3:278–86, doi:http://dx.doi.org/10.1302/20585241.3.170066. [16] Lintz F, Welck M, Bernasconi A, Thornton J, Cullen NP, Singh D, et al. 3D biometrics for hindfoot alignment using weightbearing CT. Foot Ankle Int 2017;38:684–9, doi:http://dx.doi.org/10.1177/1071100717690806.
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005
G Model FAS 1335 No. of Pages 9
A. Bernasconi et al. / Foot and Ankle Surgery xxx (2019) xxx–xxx [17] Zhang JZ, Lintz F, Bernasconi A, Zhang S. 3D biometrics for hindfoot alignment using weightbearing computed tomography. Foot Ankle Int 2019;40:720–6, doi:http://dx.doi.org/10.1177/1071100719835492. [18] Senn S. Review of Fleiss, statistical methods for rates and proportions. Res Synth Methods 2011;2:221–2, doi:http://dx.doi.org/10.1002/jrsm.50. [19] Kimura T, Kubota M, Taguchi T, Suzuki N, Hattori A, Marumo K. Evaluation of first-ray mobility in patients with hallux valgus using weight-bearing CT and a 3-D analysis system. J Bone Joint Surg 2017;99:247–55, doi:http://dx.doi.org/ 10.2106/JBJS.16.00542. [20] Kimura T, Kubota M, Suzuki N, Hattori A, Marumo K. Comparison of intercuneiform 1-2 joint mobility between hallux valgus and normal feet using weightbearing computed tomography and 3-dimensional analysis. Foot Ankle Int 2018;39:355–60, doi:http://dx.doi.org/10.1177/ 1071100717744174. [21] Welck MJ, Singh D, Cullen N, Goldberg A. Evaluation of the 1st metatarsosesamoid joint using standing CT—the Stanmore classification. Foot Ankle Surg 2018;24:314–9, doi:http://dx.doi.org/10.1016/j.fas.2017.03.005. [22] Cheung ZB, Myerson MS, Tracey J, Vulcano E. Weightbearing CT scan assessment of foot alignment in patients with hallux rigidus. Foot Ankle Int 2018;39:67–74, doi:http://dx.doi.org/10.1177/1071100717732549. [23] Osgood GM, Shakoor D, Orapin J, Qin J, Khodarahmi I, Thawait GK, et al. Reliability of distal tibio-fibular syndesmotic instability measurements using weightbearing and non-weightbearing cone-beam CT. Foot Ankle Surg 2018, doi:http://dx.doi.org/10.1016/j.fas.2018.10.003. [24] Krähenbühl N, Bailey TL, Weinberg MW, Davidson NP, Hintermann B, Presson AP, et al. Impact of torque on assessment of syndesmotic injuries using weightbearing computed tomography scans. Foot Ankle Int 2019, doi:http:// dx.doi.org/10.1177/1071100719829720 107110071982972. [25] Burssens A, Vermue H, Barg A, Krähenbühl N, Victor J, Buedts K. Templating of syndesmotic ankle lesions by use of 3D analysis in weightbearing and nonweightbearing CT. Foot Ankle Int 2018;39:1487–96, doi:http://dx.doi.org/ 10.1177/1071100718791834.
9
[26] Lintz F, Bernasconi A, Baschet L, Fernando C, Mehdi N, de Cesar Netto C. Relationship between chronic lateral ankle instability and hindfoot varus using weight-bearing cone beam computed tomography. Foot Ankle Int 2019, doi:http://dx.doi.org/10.1177/1071100719858309 107110071985830. [27] Probasco W, Haleem AM, Yu J, Sangeorzan BJ, Deland JT, Ellis SJ. Assessment of coronal plane subtalar joint alignment in peritalar subluxation via weightbearing multiplanar imaging. Foot Ankle Int 2015;36:302–9, doi:http://dx.doi. org/10.1177/1071100714557861. [28] de Cesar Netto C, Shakoor D, Roberts L, Chinanuvathana A, Mousavian A, Lintz F, et al. Hindfoot alignment of adult acquired flatfoot deformity: a comparison of clinical assessment and weightbearing cone beam CT examinations. Foot Ankle Surg 2018, doi:http://dx.doi.org/10.1016/j.fas.2018.10.008. [29] Jeng CL, Rutherford T, Hull MG, Cerrato RA, Campbell JT. Assessment of bony subfibular impingement in flatfoot patients using weight-bearing CT scans. Foot Ankle Int 2018;40:, doi:http://dx.doi.org/10.1177/1071100718804510 1071100718804510. [30] Bernasconi A, de Cesar Netto C, Barg A, Burssens A, Richter A, Weight Bearing CT International Study Group LF. AAFD: conventional radiographs are not enough! I need the third dimension. Tech Foot Ankle Surg 2019 April 10. [31] de Cesar Netto C, Bernasconi A, Roberts L, Pontin PA, Lintz F, Saito GH, et al. Foot alignment in symptomatic national basketball association players using weightbearing cone beam computed tomography. Orthop J Sport Med 2019, doi:http://dx.doi.org/10.1177/2325967118824149 Feb 21. [32] Burssens A, Peeters J, Buedts K, Victor J, Vandeputte G. Measuring hindfoot alignment in weight bearing CT: a novel clinical relevant measurement method. Foot Ankle Surg 2016;22:233–8, doi:http://dx.doi.org/10.1016/j. fas.2015.10.002. [33] Arunakul M, Amendola A, Gao Y, Goetz JE, Femino JE, Phisitkul P. Tripod index. Foot Ankle Int 2013;34:1411–20, doi:http://dx.doi.org/10.1177/1071100713488761. [34] Podolnick JD, Donovan DS, DeBellis N, Pino A. Is pes cavus alignment associated with lisfranc injuries of the foot? Clin Orthop Relat Res 2017;475:1463–9, doi:http://dx.doi.org/10.1007/s11999-016-5131-6.
Please cite this article in press as: A. Bernasconi, et al., Intraobserver and interobserver reliability of cone beam weightbearing semi-automatic three-dimensional measurements in symptomatic pes cavovarus, Foot Ankle Surg (2019), https://doi.org/10.1016/j.fas.2019.07.005