Abstract
Introduction
Hindfoot malalignment is linked to numerous pathologic conditions of the foot and ankle, such as progressive collapsing foot deformity, chronic ankle instability, and osteoarthritis.1 -3 Surgical treatment of these diseases, both joint sparing and joint replacing, usually aims to restore physiological alignment, generally characterized as a slight hindfoot valgus. 4 For both surgical planning and treatment evaluation, precise assessment of hindfoot alignment (HA) is crucial, as measurement errors may lead to inappropriate indication for procedures such as corrective osteotomy or fusion, or to inaccurate determination of the required correction. Despite growing interest and recognition of its importance, no standardized method or threshold value exists yet.
HA has traditionally been assessed both clinically and on weight-bearing standard radiographs. The hindfoot moment arm and hindfoot alignment angle (HAA) have been the most reported radiologic measurements for HA.5,6 However, malrotation, superimposition, and magnification are known limitations of these techniques.4,7,8 To counteract these effects, various alternative measurement methods on radiographs have been described, but could not attain sufficient reliability and precision.9 -12 With the increased accessibility of computed tomography (CT) and the introduction of weight-bearing computed tomography (WBCT), spatial information in the upright, weight-bearing position became available. Three-dimensional (3D) biometric methods that eliminate rotational bias have been described.13,14 However, these tools have not yet become standard in clinical practice because of limited availability and workflow integration challenges. Therefore, 2-dimensional (2D) measurements remain prevalent in routine care. 15 When performing 2D measurements in 3D WBCT data sets, image alignment and referencing using a coordinate system becomes critical, as rotational positioning significantly influences measurements.16,17 In CT imaging, HA is usually determined in the coronal plane, requiring both axial and sagittal alignment to ensure reproducible measurements. Axial referencing is straightforward as its plane can be oriented parallel to the floor. On the other hand, achieving proper sagittal plane alignment can be more challenging. To date, sagittal plane orientation relies predominantly on forefoot landmarks, with the second ray usually serving as reference. 15 However, different clinical scenarios may require different approaches to HA evaluation. Reference systems that incorporate forefoot position, such as the second ray, capture hindfoot-forefoot relationships necessary for comprehensive foot alignment evaluation. Conversely, when evaluating isolated hindfoot pathology or planning hindfoot-specific corrections, reference system independent of forefoot morphology may be more appropriate. Alternative anatomical landmarks based on more proximal structures, such as the ankle mortise, could provide relevant information in these specific clinical contexts.
The aim of this study was to compare HA measurements obtained using 3 different anatomical reference axes (second ray, ankle mortise, and forefoot midpoint) and to determine how the presence of foot deformities influences each measurement.
The hypotheses were that (1) reference systems incorporating forefoot landmarks would demonstrate significant correlations with 2D radiographic forefoot parameters, indicating systematic measurement bias in the presence of forefoot deformities, and that (2) the ankle mortise reference would show no significant correlations with these parameters, thus providing isolated hindfoot assessment independent of forefoot morphology.
Methods
This retrospective cohort study and multiplanar analysis of HA on all WBCT scans was performed at a single academic tertiary referral center. This study was approved by the cantonal ethics committee of Zurich (BASEC No.: 2024-00707). Informed consent for participation and publication was given by all patients. The study was conducted in accordance with the Declaration of Helsinki. Data analysis followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines. 18
Patient Population
Inclusion criteria comprised patients older than 18 years with available WBCT as well as dorso-plantar and lateral weight-bearing standard radiographs of the foot. Of 249 consecutive WBCT performed between its implementation on August 1, 2019, and March 31, 2024, 240 were conducted in patients who were at least 18 years old. WBCT with incomplete depiction of the foot and ankle joint (ie, distal tibia, entire hindfoot, midfoot and metatarsal bones), without corresponding dorso-plantar and lateral standard radiographs as well as with image quality deemed insufficient to determine anatomical landmarks (n = 87) were excluded. Partially amputated feet (n = 15) as well as repeated WBCTs of the same foot (n = 2) were also excluded. A total of 136 WBCTs of the foot and ankle of 126 patients were included in the study (Figure 1). Demographic data of each patient were collected, including age, gender, affected side, and reason for WBCT scans.
Flowchart presenting included and excluded patients. WBCT, weight-bearing computed tomography.
Radiological Data Collection
Radiographs were acquired with a digital radiography system (Ysio Max; Siemens Healthineers). For the dorso-plantar and the lateral images, the distance between the X-ray tube and the detector plate was 1.2 m, and the images were acquired with a 50 kV tube voltage and a 6.4 mAs tube current. For the dorso-plantar images, the X-ray beam was tilted 10 degrees.
WBCT was acquired with a twin robotic X-ray system (Multitom Rax; Siemens Healthineers), with the patient standing on one leg. During the examinations of the foot, 434 projections were acquired with a 0.3-mm copper filter, with a 117 kV tube voltage, a 212 mAs total tube current, a scan time of 20 seconds, and a field of view of 420 mm.
Mid- and forefoot morphology was assessed using established radiographic parameters. The Meary angle was measured on lateral foot radiographs to characterize pes planus or cavus. 19 On dorso-plantar foot radiographs, the Sgarlato angle was measured to assess metatarsus adductus, with values above 20 degrees indicating deformity.20,21 Hallux valgus deformity was quantified using both the hallux valgus angle and the intermetatarsal I-II angle. The intermetatarsal I-V angle was measured to characterize splayfoot deformity. 22
To assess HA on WBCT, a new measurement method was developed, based on the approach described by Burssens et al. 15 In the original method, sagittal plane was achieved by aligning images according to the axis of the second ray, and the hindfoot angle was then measured in the resulting coronal plane as the angle between the tibial axis and a talocalcaneal axis connecting the calcaneal weight-bearing point to the middle of the upper border of the talus. The current study preserved this coronal plane measurement principle but systematically compared 3 different sagittal plane reference axes to evaluate how each influences HA measurements. While maintaining axial plane alignment parallel to the floor, the following reference axes based on anatomical landmarks were used for sagittal plane orientation:
Second ray axis (“MT2”): Axis parallel to the second metatarsal, as described by Burssens et al 15 (Figure 2A).
Ankle mortise axis (“ANKLE”): Axis parallel to a line connecting 2 points: (1) the midpoint between the medial and lateral malleoli at the anterior tibial plafond and (2) the corresponding midpoint at the posterior tibial plafond (Figure 2B).
Forefoot midpoint axis (“MIDPOINT”): Axis parallel to a line connecting the calcaneal weight-bearing point to the midpoint between the weight-bearing point of the first and fifth metatarsal heads (Figure 2C).
Notably, these references represented distinct anatomical regions. While the second ray axis depended solely on forefoot orientation and the ankle mortise axis on distal tibial anatomy, the forefoot midpoint axis incorporated both hindfoot and forefoot elements. Following sagittal plane alignment according to each of the 3 reference axes, 3 different coronal images perpendicular to the chosen sagittal orientations were generated. HA measurement was performed on each of these coronal views. Slice thickness was increased to simultaneously depict tibia, talus, and calcaneus. The tibial axis was defined as the line connecting the mediolateral midpoint of the tibia 10 cm proximal to the ankle joint to the mediolateral midpoint of the tibial plafond, following the principle of using distal tibial landmarks as described by Saltzman and El-Khoury. 5 To determine hindfoot axis, 2 lines were drawn at the medial and lateral border of the calcaneal tuberosity. The midpoints of these 2 lines were connected by a line, and the midpoint of this third line was marked as the calcaneal reference point. The hindfoot axis was then defined as the line connecting this calcaneal reference point to the mediolateral midpoint of the upper border of the talus. HAA was defined as the angle between the tibial axis and hindfoot axis, with positive values indicating valgus alignment and negative values indicating varus alignment (Figure 3). 15 To specify the reference system used to determine HAA, subscripts were applied: HAAMT2 for measurements using the second ray, HAAANKLE for the ankle mortise, and HAAMIDPOINT for the forefoot midpoint. Overall, the described measurement method differed from that of Burssens et al 15 in using alternative sagittal plane reference systems, assessing tibial axis on WBCT rather than full-length leg radiographs, and using different reference points on the calcaneus. All measurements were performed by 2 independent investigators, masked to the patient data, using the institutional PACS software (Phönix PACS).
Reference axes on weight-bearing computed tomography (WBCT) scan. Three different reference axes based on anatomical landmarks were used to orient the sagittal plane, generating corresponding coronal images perpendicular to the chosen sagittal orientations for hindfoot alignment measurement. (A) The second ray axis was aligned parallel to the second metatarsal. (B) The ankle mortise axis connected the midpoints between the medial and lateral malleoli at the anterior and posterior tibial plafond, whereas (C) the forefoot midpoint axis was aligned parallel to a line connecting the calcaneal weight-bearing point to the midpoint between the weight-bearing point of the first and fifth metatarsal heads.
Hindfoot alignment assessment with hindfoot alignment angle (HAA). HAA (*) was measured as the angle between the tibial axis (a) and the hindfoot axis (b) on a coronal view with increased slice thickness. The tibial axis was defined as the line connecting the mediolateral midpoint of the tibia 10 cm proximal to the ankle joint to the mediolateral midpoint of the tibial plafond. To determine the hindfoot axis, 2 lines were drawn at the medial and lateral border of the calcaneal tuberosity. The midpoints of these 2 lines were connected, and the midpoint of this third line served as the calcaneal reference point. The hindfoot axis was then defined as the line connecting this calcaneal reference point to the mediolateral midpoint of the upper border of the talus. Positive values indicated valgus alignment; negative values indicated varus alignment.
Primary and Secondary Outcomes
The primary outcome was HAA measured across 3 different reference systems (second ray, ankle mortise, and forefoot midpoint). Secondary outcomes were interrater reliability of HAA measurements and correlations between HAA and 2D foot morphology parameters (Meary angle, Sgarlato angle, hallux valgus angle, intermetatarsal I-II and I-V angles).
Statistical Analysis
Absolute values were presented as number (percentage) and continuous parameters as median (75% interquartile range [IQR]) because testing revealed non-normal distribution of variables (Shapiro-Wilk test,
Results
The median age at time of WBCT was 50.3 (IQR, 40.3, 58.4) years, with females comprising 64% of the cohort. The right foot was affected in 50% of cases. The most common reasons for WBCT were hallux valgus (41/136, 30.1%), followed by metatarsalgia (23/136, 16.9%) (Table 1). Radiographic parameters of the included feet were measured (Table 2). Among the cohort, 14 feet (10.3%) had metatarsus adductus, defined by Sgarlato angles exceeding 20 degrees.
Demographics (n = 136).
Abbreviations: OA, osteoarthritis; WBCT, weight-bearing computed tomography.
2D Standard Radiographic Parameters (n = 136).
Statistically significant differences in HAA values were observed across the 3 reference axes. The median values were 10.0 degrees for HAAMT2, 15.2 degrees for HAAANKLE, and 12.7 degrees for HAAMIDPOINT (all
3D Hindfoot Alignment Parameters With Different Reference Systems and Their Interobserver Reliability (n = 136).
Abbreviations: 3D, 3-dimensional; HAA, hindfoot alignment angle; HAAANKLE, HAA with ankle mortise reference; HAAMIDPOINT, HAA with forefoot midpoint reference; HAAMT2, HAA with second ray reference; ICC, intraclass correlation coefficient; SEm, standard error of measurement.
Friedman test
Impact of Sgarlato angle on hindfoot alignment measurements across 3 reference methods. Scatter plots show individual measurements with linear regression lines for visualization. Spearman rank correlations (ρ) quantify monotonic relationships. Note the statistically significant negative correlation for MT2 reference (ρ = −0.41,
Spearman Correlation Analysis Between Hindfoot Alignment and Foot Morphology Parameters. a
Abbreviations: HAA, hindfoot alignment angle; HAAANKLE, HAA with ankle mortise reference; HAAMIDPOINT, HAA with forefoot midpoint reference; HAAMT2, HAA with second ray reference.
Values are presented as Spearman ρ (
Discussion
This single-center retrospective cohort study and multiplanar analysis compared HA measurements obtained using 3 different anatomical reference axes and evaluated how foot deformities influence each measurement, finding that axis selection can significantly alter the clinical classification of alignment. Both hypotheses were confirmed. First, HAA measured with the second ray reference axis demonstrated a significant correlation with Sgarlato angle, indicating systematic measurement bias towards varus in the presence of metatarsus adductus. Second, ankle mortise referencing showed no correlation with any of the mid- and forefoot parameters other than Meary angle, thus suggesting a more isolated assessment of HA.
Precise information about HA is paramount to evaluate, plan, and perform surgical procedures. Nevertheless, a gold standard and subsequently threshold value to aid surgeons in the decision-making process have yet to be established. Currently, the hindfoot angle described by Burssens et al 15 is one of the most widely used methods in the literature. It measures the angle between the tibial and talocalcaneal axes on WBCT. The approach offers several advantages. It is rotationally stable, applicable to any hindfoot configuration, and highly repeatable. 3D computer-aided design analysis demonstrated a strong correlation, further validating this method. 14 However, some limitations need to be addressed. While using the weight-bearing point of the calcaneus to determine HA demonstrated high reproducibility, this method consistently yielded lower HA values compared with traditional techniques that rely on the calcaneal axis, 27 potentially underestimating the extent of deformity. Using the second metatarsal as reference for sagittal plane alignment can significantly affect HA measurements. Zhu et al 28 assessed HA using a modified calcaneal moment arm on coronal WBCT images and demonstrated how rotational changes in image positioning can distort measurements. Their study showed that a neutral hindfoot was shifted into varus with external rotation of the foot and into valgus with internal rotation. These considerations should be contextualized within the broader transition toward 3D measurement techniques.13,14,29 -31 Biometrics such as the Foot and Ankle Offset (FAO) introduced by Lintz et al 13 use a true 3D reference frame, which fundamentally resolves the rotational bias that affects 2D imaging by eliminating the need for arbitrary plane selection. However, despite their theoretical advantages, clinical implementation remains limited because of specialized software requirements, increased processing time and workflow integration challenges. Until 3D methods become standard practice, optimizing and understanding the limitations of 2D methods remain essential for clinical decision making.
The current study revealed the following key findings regarding reference axes in HA measurements. First, it confirmed the statistically significant variability in HA measurements depending on the chosen reference axis. As pointed out by Zhu et al, 28 the variability of HA measurements can be attributed to the fact that, despite the availability of precise 3D data from WBCT, measurement methods based on 2D concepts remain susceptible to the inherent limitations of 2D imaging, such as the need for appropriate rotation and referencing. Second, statistically significant negative correlation was confirmed between Meary angle and HAA across all reference methods. This pattern across all 3 reference axes demonstrated that each measurement method could reliably detect the expected anatomical relationship between foot arch morphology and HA, with a trend toward valgus HA with pes planus and varus HA with pes cavus. Third, it revealed that the presence of metatarsus adductus substantially alters HA measurements when using the second ray as reference axis. The negative Spearman correlation between HAAMT2 and Sgarlato angle indicates a shift towards varus alignment with more marked metatarsus adductus. The moderate strength of this correlation suggests that metatarsus adductus is a contributing factor to measurement differences but does not fully explain the observed variability between reference systems. Nevertheless, this finding indicates that relying on second ray alignment in patients with metatarsus adductus could lead to underestimation of valgus deformity or even misclassification of alignment. In contrast, HAA measurements referenced to the ankle mortise or the forefoot midpoint provided results that were unaffected by the presence of metatarsus adductus (Figure 5). Further analysis of forefoot parameters showed weak but statistically significant correlation between HAA with forefoot midpoint reference and intermetatarsal I-II angle. A true anatomical relationship cannot be excluded, as previous studies already reported associations between hallux valgus and hindfoot valgus. 32 However, the current study found no correlations between hallux valgus angle and HAA measurements using any of the reference axes. Finally, evaluation of absolute differences between reference methods revealed that 20.6% of feet had an HAA delta greater than 10 degrees between second ray and ankle mortise referencing, indicating that in approximately 1 in 5 patients the choice between these 2 reference methods could result in a clinically meaningful difference in HA assessment.
Example of the impact of metatarsus adductus on hindfoot alignment measurements across 3 reference methods. The left panel displays a superimposed axial CT image with reference lines illustrating the second ray (MT2, red), ankle mortise (ANKLE, blue), and forefoot midpoint axis (MIDPOINT, green). This patient presented with marked metatarsus adductus (Sgarlato angle 28.6 degrees). The corresponding coronal CT images (right panels) demonstrate the dramatic impact of reference axis selection on hindfoot alignment measurements. Although second ray referencing indicates hindfoot varus (HAAMT2: −7.6 degrees), both ankle mortise and forefoot midpoint referencing show valgus alignment (HAAANKLE: 16.2 degrees, HAAMIDPOINT: 12.8 degrees) demonstrating how the presence of metatarsus adductus systematically skews second ray referenced measurements towards varus alignment.
Overall, the ankle mortise reference had distinct characteristics: while maintaining correlation with the Meary angle, thus reflecting true hindfoot anatomy, it demonstrated no statistically significant correlations to any other forefoot parameters, therefore qualifying as isolated HA assessment method. This pattern aligns with the anatomical basis of each reference method. The ankle mortise reference axis only depends on the anatomy of the distal tibia and remains independent of foot morphology distal to the ankle joint. Isolated HA assessment with the ankle mortise reference can be particularly suitable for preoperative planning or postoperative evaluation of hindfoot procedures. However, a proximal reference landmark may fail to account for forefoot deformities and the effect they have on HA. In contrast, second ray referencing depends exclusively on forefoot orientation. This approach captures the mechanical relationship between forefoot position and HA, while remaining susceptible to the effects of forefoot deformities. The forefoot midpoint reference axis uses a reference point between the first and fifth metatarsal heads rather than a single anatomical structure. This way it partially mitigates but cannot eliminate the influence of localized mid- and forefoot deformities. These findings emphasize that, when using 2D measurement methods, reference system selection should be guided by the specific clinical question and surgical planning needs, rather than based on a universal standard.
Following limitations should be considered when interpreting the current findings. This study was conducted at a single academic tertiary referral center and WBCT were only performed in symptomatic patients, which may limit the applicability of the findings to other settings and populations. The exclusion of 87 scans for image quality may have introduced selection bias if complex deformities were systematically harder to landmark. No formal sample size calculation was performed. Although correlations were found between HA measurements and mid- and forefoot parameters, true anatomical relationships between HA and other foot deformities cannot be definitively excluded. However, the magnitude of the correlation between Sgarlato angle and HAA with second ray referencing strongly suggests this relationship is mostly caused by geometric distortion, rather than anatomical compensation. The tibial axis was determined on WBCT by assessing its distal 10 cm and using the mediolateral midpoint of the plafond as the distal reference point. This approach may be more susceptible to rotational variability compared to using 2 midshaft reference points and precluded the ability to detect abnormalities in the proximal tibia. Nevertheless, this offers practical advantages, as full-length leg radiographs are not routinely obtained in patients with foot and ankle pathologies. Intrarater reliability was not assessed, limiting the evaluation of measurement consistency. The imaging workflow required standard radiographs before WBCT, as the latter is reserved for specific indications, resulting in slightly increased radiation exposure compared to WBCT alone with digitally reconstructed radiographs. Finally, HA measurements were not correlated to clinical outcome parameters or surgical results. Therefore, clinical implications cannot be demonstrated but only postulated, as the absence of a clinical gold standard prevents identifying the most accurate reference axis. In the absence of clinically validated outcome thresholds, the current findings indicate systematic measurement differences rather than identify a clinically superior reference method. Future studies should aim to investigate the impact of different HA measurement methods on surgical planning and determine which method best predicts surgical success.
Conclusion
This study demonstrated that 2D HA measurements on WBCT are significantly influenced by the choice of reference system. The second ray reference incorporates forefoot morphology but appeared particularly problematic in the presence of metatarsus adductus, with systematic measurement bias towards varus alignment. Such measurement artifacts could alter the process of surgical decision making, ultimately compromising patient outcomes. In contrast, the ankle mortise reference could provide isolated measurement of HA regardless of concurrent forefoot deformities. When using 2D measurement methods, reference system selection could be tailored to whether surgical planning requires comprehensive foot alignment assessment or isolated hindfoot evaluation.
Supplemental Material
sj-pdf-1-fao-10.1177_24730114261421600 – Supplemental material for Exploratory Analysis of the Effect of Foot Deformities on Hindfoot Alignment Measurement: A Comparison of Second Ray, Ankle Mortise, and Forefoot Alignment on Weight-bearing CT
Supplemental material, sj-pdf-1-fao-10.1177_24730114261421600 for Exploratory Analysis of the Effect of Foot Deformities on Hindfoot Alignment Measurement: A Comparison of Second Ray, Ankle Mortise, and Forefoot Alignment on Weight-bearing CT by Alberto Pedrazzini, Pascal R. Furrer, Flamur Zendeli, Arnd F. Viehöfer, Stephan H. Wirth, Reto Sutter and Silvan Beeler in Foot & Ankle Orthopaedics
Footnotes
Ethical Considerations
This study was approved by the cantonal ethics committee of Zurich (BASEC No.: 2024-00707). Informed consent for participation and publication was given by all patients. The study was conducted in accordance with the Declaration of Helsinki.
Author Contributions
A. Pedrazzini: conceptualization, acquisition of data, analysis and interpretation of data, drafting the manuscript. P. R. Furrer: conceptualization, methodology, analysis and interpretation of data. F. Zendeli: acquisition of data. A. F. Viehöfer: conceptualization, methodology. S. H. Wirth: conceptualization, methodology. R. Sutter: conceptualization, methodology. S. Beeler: idea, conceptualization, methodology, analysis and interpretation of data, drafting the manuscript. All authors: revision of the manuscript, final approval of the version to be published.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Reto Sutter, MD, reports that Balgrist University Hospital has an academic research agreement with Siemens Healthineers and Bayer. Disclosure forms for all authors are available online.
Data Availability Statement
The data that support the findings for this study are available to other researchers from the corresponding author upon reasonable request.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
