Abstract
Background:
Lisfranc injuries pose diagnostic challenges, particularly in evaluating joint stability. Conventional weightbearing computed tomography (WBCT)-based distance measurements of the C1-M2 interval fail to account for the second metatarsal’s triplanar orientation, potentially leading to inaccuracies. This study introduces a new 3D-corrected triplanar measurement method correcting for axial, coronal, and sagittal alignment to improve diagnostic accuracy.
Methods:
In this retrospective study, 31 patients with acute Lisfranc injuries underwent bilateral WBCT. Injuries were defined based on radiographic findings in the first to third tarsometatarsal joints and the C1-M2 interval. Two fellowship-trained foot and ankle surgeons independently performed manual measurements using a previously described uniplanar method and a new triplanar technique, applied proximally and distally in the C1-M2 interval. Intra- and interrater reliability were assessed via intraclass correlation coefficients (ICCs), and side-to-side differences were compared using paired statistical tests.
Results:
The triplanar method demonstrated higher ICCs (intraobserver: 0.96-0.97; interobserver: 0.94-0.97) than the uniplanar method (intraobserver: 0.86-0.91; interobserver: 0.84-0.90), with distal measurements showing the highest reliability. Notably, the uniplanar method incorrectly measured the M1-M2 interval instead of the intended C1-M2 interval in 22.6% of injured feet. No such errors occurred in contralateral feet or with the triplanar method, which demonstrated 100% intra- and interobserver agreement. All 6 performed C1-M2 measurements showed significant differences between injured and contralateral feet (P < .05). The triplanar method applied distally in the coronal plane yielded the greatest absolute side-to-side difference (1.81 mm, SD 1.60).
Conclusion:
This study demonstrates excellent intra- and interobserver reliability for a novel WBCT-based method that realigns the measurement planes with the second metatarsal rather than the floor. This method improves measurement precision and prevents systematic errors observed with previous techniques, particularly the misidentification of the M1-M2 interval using uniplanar methods. Clinical validation studies correlating measurements with surgical outcomes are needed to establish diagnostic thresholds and confirm clinical utility.
Level of Evidence:
Level III, retrospective diagnostic study.
Introduction
Injuries to the tarsometatarsal (TMT) joint complex, commonly referred to as Lisfranc injuries, range from ligamentous, subtle ligamentous injuries to displaced fracture-dislocations. 21 Although the latter necessitate surgical intervention, the management of more subtle injuries depends on an assessment of joint stability to determine whether surgery is required or if nonoperative treatment is sufficient. 14 The primary ligamentous stabilizers for the Lisfranc joint complex include the interosseous ligament (“Lisfranc ligament”), which connects the medial cuneiform (C1) to the base of the second metatarsal (M2) as well as the plantar ligaments spanning from C1 to both M2 and the base of the third metatarsal (M3).7,13 Accordingly, special attention is paid to the C1-M2 interval when evaluating subtle instability and planning treatment. 17
Various assessment methods have been used over the past decades. Manual stress testing under fluoroscopy using an abduction-pronation maneuver can be performed either in the outpatient setting or under anesthesia.5,11,15 However, this method’s reliability is limited by the absence of standardized criteria and variability related to examiner technique. Bilateral weightbearing radiographs, commonly used to detect subtle changes via side-to-side comparison, have been advocated for assessing Lisfranc stability but may be limited by projectional overlap.1,5,6,12,17,18,23
Over the past decade, weightbearing computed tomography (WBCT) has emerged as a valuable 3D imaging modality for evaluating both acute and chronic foot and ankle conditions.2,10 In the context of Lisfranc injuries, WBCT enables distance,8,19 area,3,4,24 and volumetric measurements3,4,22 of the C1-M2 interval. However, most existing studies are based on healthy feet or cadaver models. To our knowledge, only 2 published studies have reported WBCT measurements in Lisfranc-injured feet, involving 14 and 15 feet, respectively.3,8 Among the proposed techniques, reproducible and time-efficient manual distance measurements currently appear to be the most suitable option for routine clinical application. Sripanich et al19,20 described a method to measure the C1-M2 distance with axial-plane correction only and demonstrated excellent intra- and interobserver agreement in healthy feet. Yet, because of the typical 25-30-degree craniocaudal orientation of the TMT joints and the second metatarsal, 16 failing to adjust the measurement plane may result in suboptimal intervals, bony overlap, and measurement errors in injured feet.
The objective of this study was to introduce an improved triplanar method for measuring the C1-M2 distance by aligning the measurement planes with all 3 axial, sagittal, and coronal axes of the second metatarsal, and evaluate interclass correlation coefficients (ICCs) in both healthy and injured feet. Additionally, we aimed to compare this new method with the previously described uniplanar approach, which only provided axial alignment of M2.
We hypothesized that the new approach would demonstrate superior inter- and intraobserver reliability in both injured and contralateral feet, providing a more accurate representation of the true C1-M2 interval.
Material and Methods
This retrospective, institutional review board (IRB)-approved study (IRB number Pro00113556), included patients with acute Lisfranc injuries who underwent bilateral WBCT as part of their clinical evaluation. Patients were identified through an institutional database search using International Classification of Diseases, Tenth Revision (ICD-10), codes S92.2 (Fracture of other unspecified tarsal bone), S92.3 (Fracture of metatarsal bone), S92.7 (Multiple fractures of foot), S93.3 (Subluxation and dislocation of foot), and S93.6 (Sprain of foot), with WBCT imaging performed between February 2022 and February 2024 (Figure 1). The WBCT scans (CurveBeamAI, Hatfield, PA) were conducted with a voxel size of 0.37 mm, a 350-mm field-of-view diameter, a 200-mm field-of-view height, an exposure time of 9 seconds, and a total scan time of 54 seconds. All patients were encouraged to fully weightbear during the WBCT scan. The anatomical frame of reference used by the WBCT reconstruction engine is the RAS (Right, Anterior, Superior) orthonormal frame. The described method realigns all 3 planes with the second metatarsal, ensuring that measurements are independent of foot positioning during acquisition and the scanner’s anatomical reference frame. Therefore, any anatomical orthonormal frame of reference can be used for image acquisition.
Study inclusion flowchart.
Charts and scans were reviewed by 2 fellowship-trained orthopaedic foot and ankle surgeons with extensive experience in diagnosing and treating Lisfranc injuries and trained in WBCT imaging. Lisfranc injuries were defined based on WBCT findings of involvement of the first, second, or third TMT joints and the C1-M2 interval, typically characterized by avulsion fractures at key ligamentous stabilizers. Discrepancies between the surgeons were resolved through discussion. No intraoperative findings were used as a reference standard.
Injuries were initially diagnosed with radiographs or conventional CT, and WBCT was performed to confirm stability and aid treatment planning. Exclusion criteria included chronic Lisfranc injuries (defined as WBCT performed more than 6 weeks postinjury), absence of a contralateral WBCT scan, imaging acquired after surgical treatment, midfoot osteoarthritis, Charcot arthropathy, prior midfoot surgery, WBCT performed in a cast, or motion artifacts on the scan (Figure 1).
The 2 fellowship-trained orthopaedic foot and ankle surgeons independently performed the manual measurements using CubeVue version 4.2.0.1 (CurveBeamAI, Hatfield, PA). To assess intraobserver reliability, one of the 2 observers repeated the measurements after 2 weeks. At all times, observers were masked to each other’s measurements and to their own prior measurements.
Measurement Methods
Uniplanar method
Measurements of the C1-M2 distance in the axial and coronal planes were performed following the method described by Sripanich et al.19,20 The appropriate sagittal image displaying the entire second metatarsal (M2) and both the superior and inferior borders of its base was first identified. Using CubeVue’s one-third-distance marking tool, a line was drawn from the most superior-proximal to the most inferior-proximal point of the M2 base, defining its height. From the lower one-third point, a perpendicular line was drawn 6 mm distally along the proximal surface of the metatarsal base. This endpoint defined the measurement reference point, intersected by axial and coronal planes. Linear distances between C1 and M2 were then measured in each plane (Figure 2). Additionally, it was evaluated whether the reference point was correctly placed within the intended C1-M2 interval or if the method resulted in measurement of the M1-M2 interval instead (Figure 5).
Description of the uniplanar method. (1) The appropriate sagittal image was selected to display the full second metatarsal (M2), including the superior and inferior borders of its base. A line was drawn using CubeVue’s one-third-distance tool from the most superior-proximal to the most inferior-proximal point of the M2 base, representing its height. From the lower one-third mark, a perpendicular (transverse) line was extended 6 mm distally along the proximal surface of the metatarsal base (orange line). The endpoint of this line served as the reference point for both coronal and axial reference planes. The C1-M2 distances were then measured in the coronal (2) and axial (3) views.
Triplanar measurement method for C1-M2 measurement
Before measuring the C1-M2 distance, all 3 anatomical planes were aligned with the second metatarsal (M2) (Figure 3). Alignment began in the axial plane by correcting axial rotation. A slab view was generated, and the midpoints of the proximal and distal diaphysis of M2 were marked. M2 was then rotated in the axial plane to align with these marks. After axial alignment, the slab was removed, and coronal alignment was performed by drawing a tangential line along the dorsal surface of the M2 base and adjusting the metatarsal in the coronal plane perpendicular to this line.
Triplanar alignment correction for the new method. Alignment began in the axial plane by creating a slab view, marking the midpoints of the proximal and distal diaphysis of M2, and rotating the bone accordingly (1). The coronal plane was aligned by drawing a tangential line along the dorsal surface of the M2 base (2A) and adjusting M2 perpendicular to this line (2B). Finally, sagittal alignment was achieved by marking midpoints of the proximal and distal M2 diaphysis in a sagittal slab view, aligning rotation by them (3A), and adjusting the reference line in the sagittal view just plantar to the plantar cortex of M2 (3B).
Sagittal alignment followed by generating a sagittal slab and marking midpoints of the proximal and distal M2 diaphysis. M2 was rotated in the sagittal plane to align with the connecting line. The reference line in the sagittal view (red line in Figure 3A) was then adjusted just plantar to the plantar cortex of M2.
With M2 fully aligned, measurement levels were defined in the axial plane. For the proximal measurement, the line was placed at the most proximal medial voxel of the M2 base within the C1-M2 interval, and for the distal measurement, at the most distal medial voxel of C1 within the C1-M2 interval. The C1-M2 distance was then measured in both axial and coronal planes using each method (Figure 4).
Description of proximal and distal C1-M2 measurements following triplanar alignment correction. For the proximal measurement, the reference line is placed at the most proximal medial voxel of the M2 base within the C1-M2 interval (blue line). For the distal measurement, it is placed at the most distal medial voxel of C1 in the same interval. Distances are measured in both the axial (1A, 2A) and coronal (2A, 2B) planes.
Statistics
Normality of continuous variables was assessed using the Shapiro-Wilk test. Paired comparisons between injured and uninjured sides were conducted using paired t tests for normally distributed variables and Wilcoxon signed-rank tests for non-normal variables. Interobserver agreement was evaluated using the ICC based on a 2-way mixed effects model for absolute agreement. ICCs were interpreted as follows: 0.81 to 0.99, almost perfect reliability; 0.61 to 0.80, substantial; 0.41 to 0.60, moderate; 0.21 to 0.40, fair; and ≤20, slight reliability. 9
Pearson or Spearman correlation coefficients (depending on normality) were calculated to assess internal consistency between axial and coronal measurements, as both represent the same distance between identical voxel pairs for each method and measurement location.
A P value <.05 was considered as significant. Statistical analyses were performed using Stata 18.0 (StataCorp, College Station, TX).
Results
A total of 31 patients with bilateral WBCT scans were included (11 male, 20 female). The mean age was 46.8 years (SD 16.5), and the mean BMI was 29.4 (SD 5.6). Injuries involved the right foot in 18 cases and the left foot in 13. WBCT imaging was obtained after a mean of 17 days (SD 11) postinjury.
Intra- and interrater reliability results are presented in Tables 1 and 2. All measurements demonstrated substantial to almost perfect intrarater agreement, with the 4 triplanar measurements (proximal coronal, proximal axial, distal coronal, and distal axial) consistently achieving higher ICCs than the uniplanar axial and coronal measurements. Interrater ICCs demonstrated almost perfect agreement for all parameters in injured feet, with the new triplanar method consistently outperforming the original uniplanar technique. In contralateral feet, the original uniplanar measurements and the triplanar axial measurement at the proximal C1-M2 interval showed substantial agreement, whereas the triplanar coronal measurement at the proximal C1-M2 interval and all triplanar measurements at the distal C1-M2 interval achieved almost perfect reliability. Overall, ICCs were higher in injured feet than in contralateral feet for both intra- and interrater comparisons.
Intraclass Correlation Coefficients (ICCs) for Intrarater Reliability.
Intraclass Correlation Coefficients (ICCs) for Interrater Reliability.
Strong correlations were observed between axial and coronal measurements, with the new proximal and distal methods demonstrating greater internal consistency than the original method (Table 3).
Consistency Between Coronal and Axial Measurements.
Importantly, the original method led to incorrect measurement of the M1-M2 interval instead of the intended C1-M2 interval in 7 of the 31 injured feet (22.6%) (Figure 5). This mismeasurement was consistently identified with 100% interobserver and intraobserver agreement. No such errors occurred in contralateral, uninjured feet or when using the new methods.
Lisfranc case example showing mismeasurement with the uniplanar method resolved by the use of the new method. The original technique led to erroneous measurement between M1 and M2 rather than the intended C1-M2 interval (1A, 1B, 1C). In the same foot, both proximal and distal views using the new method correctly measure the true Lisfranc interval (2A, 2B, 2C, 2D).
Measurement results for the injured and contralateral sides are summarized in Table 4. All 6 variables demonstrated statistically significant side-to-side differences (P < .05), with the distal coronal measurement of the new method showing the largest absolute difference (1.81 mm, SD 1.60).
Measurement Results Comparing the Injured Side With the Uninjured, Contralateral Side.
A post hoc power analysis was conducted based on the interrater ICCs observed for all feet (injured and contralateral). The ICC for the uniplanar coronal measurements was 0.8950, and for the triplanar distal coronal measurements 0.9598 (Table 3). Using a 2-rater, 2-way random effects model with a significance level of .05, the minimum sample size required to detect this difference with 80% power was calculated to be 39 feet. With a total of 62 feet analyzed, the study was considered adequately powered to confirm the observed improvement in reliability.
Discussion
This study introduces a new triplanar measurement approach of the Lisfranc interval (C1-M2) on WBCT by correcting for rotational alignment in all 3 anatomical planes. Our results demonstrate overall excellent inter- and intraobserver reliability of the technique when applied in both the proximal and distal C1-M2 intervals, consistently achieving higher ICCs than the previously described method. Most importantly, this study identifies a critical systematic error in existing methodology: the original uniplanar method resulted in erroneous measurement of the M1-M2 interval instead of the intended C1-M2 interval in 22.6% of injured feet, while maintaining perfect accuracy in uninjured feet. This systematic measurement error, not previously reported, was eliminated with the triplanar approach.
In a cadaveric study, Sripanich et al 19 were the first to describe specific WBCT-based measurements of the Lisfranc interval, introducing the uniplanar method also used in our study. The same authors later applied this technique to a cohort of 96 uninjured feet. 20 They reported intrarater ICCs of 0.802 (axial) and 0.840 (coronal) and interrater reliabilities of 0.727 (axial) and 0.814 (coronal). Our results in uninjured feet are comparable, with intrarater ICCs of 0.81 for both the axial and coronal measurement, and interrater ICCs of 0.74 (axial) and 0.80 (coronal). However, the new measurement techniques outperformed the original method across nearly all comparisons, except for the proximal axial measurement in uninjured feet (ICC = 0.74). Interestingly, we found consistently higher ICCs in injured feet than in uninjured ones. This may be due to greater relative variability when measuring smaller, intact intervals, which can lower correlation values.
In both studies by Sripanich et al,19,20 no instances of measuring the incorrect M1-M2 interval were reported. In contrast, we observed this error in 22.6% of injured feet but in none of the contralateral, uninjured feet. This limitation likely went unnoticed in earlier studies involving only cadavers or healthy subjects. This discrepancy may be caused by medial opening of the second TMT joint under weightbearing in injured feet, which can result in abduction of the second metatarsal. Because the uniplanar axial and coronal measurements are performed perpendicular to M2 in the axial plane, they are prone to angular distortion in injured configurations. Furthermore, visualization of the entire second metatarsal—required to identify the reference point in the uniplanar method—does not necessarily imply true axial alignment, potentially introducing measurement error. Notably, intra- and interobserver agreement for this mismeasurement was perfect, suggesting the error is systematic rather than random. Additionally, using a fixed offset of 6 mm distal to the joint line across all feet, rather than anatomical reference points, may further compromise accuracy. Finally, medial column subluxation or proximal displacement of C1 could also lead to selection of an incorrect interval, although no such overt cases were observed in our cohort.
Another limitation of the uniplanar method is its reliance on identifying the full plantar margin of the M2 base in the sagittal plane. In the setting of Lisfranc injuries, this can be challenging due to frequent plantar avulsion fractures of M2 (Figure 6).
Avulsion of the M2 base may lead to improper measurement using the uniplanar method. The avulsion of the M2 base results in difficulties identifying the correct point for measurement, as described in the original method. This anatomical disruption interferes with an accurate identification of the M2 base.
Notably, the triplanar measurements performed distally in the C1-M2 interval yielded the greatest side-to-side differences (Table 4), suggesting their potential value in detecting clinically relevant widening at the Lisfranc interval.
Other studies have also explored C1-M2 measurements on WBCT. Falcon et al 8 assessed 112 feet, including 15 with Lisfranc injuries, and reported a mean C1-M2 distance of 3.66 mm (SD 0.72) in healthy feet, falling between our original axial (4.35 mm) and new distal axial (3.06 mm) measurements. However, differences in measurement technique (eg, selecting the axial slice at the middle of point of best-viewed articulation) limit direct comparison. Bhimani et al 3 proposed an axial C1-M2 measurement performed 10 mm below the dorsal surface of C1 but did not specify whether the measurements were performed proximal, central, or distal in the Lisfranc interval. They reported a mean distance of 2.1 mm (SD 1.1) in 72 healthy feet and an interobserver ICC of 0.96. However, this ICC was based on only 4 cases, which limits its generalizability.
This study has several limitations. First, this was a single-center study using 1 WBCT scanner model, which may limit generalizability to other imaging systems or protocols. Full weightbearing during WBCT acquisition was neither confirmed nor documented for each patient, and individual variability (eg, guarding or antalgic posture) may have influenced widening of the C1-M2 interval. Second, the triplanar alignment method requires manual operator adjustment and visual estimation, introducing potential subjective variability that may affect reproducibility in broader clinical settings. Although our study focused on detecting differences in reproducibility between methods, diagnostic accuracy (sensitivity, specificity, or predictive values) was not assessed. Moreover, the observed variability in uninjured feet raises the potential for misinterpretation of side-to-side differences. Our injury cohort was heterogeneous, including various patterns of TMT joint involvement, and we did not stratify results by injury severity or specific anatomical patterns. Establishing normative reference values and clinical decision thresholds correlated with surgical outcomes may be a useful future step toward improving diagnostic confidence. Finally, the clinical significance of the measured differences (maximum 1.81 mm) remains unclear without correlation to treatment outcomes.
Recent technological advances have introduced more complex WBCT-based assessments, including area3,4,24 and volumetric measurments.3,4,22 Although one study reported higher diagnostic accuracy using volumetric analysis, 3 such methods remain time-consuming and are not yet practical for routine clinical decision making, leaving linear distance measurements as the most feasible tool in everyday practice for now.
This study includes the largest cohort to date of Lisfranc injuries evaluated with WBCT. The new triplanar alignment method realigns all 3 WBCT planes with the second metatarsal, allowing for improved measurement accuracy by eliminating distortions from the floor-based alignment typically used in uniplanar techniques. In this study, we found that this approach had excellent reliability. We believe it will be practical to implement in clinical workflows, as the alignment of M2 might be achievable even through visual estimation. The proposed triplanar method is intended for all types of Lisfranc injuries that undergo weightbearing CT. However, in cases with clear malalignment, manual measurements may be unnecessary for surgical decision making. Based on our findings, we recommend applying the new approach in the distal C1-M2 interval to better highlight side-to-side differences between injured and uninjured feet. However, in cases where avulsion fractures within the Lisfranc interval interfere with the distal measurement, the proximal method may serve as a valuable alternative. However, no minimal detectable change or clinical cutoff has been established to date. Identifying such thresholds would require correlating WBCT-based measurements with intraoperative findings or clinical outcomes. Prospective studies defining thresholds predictive of the need for operative intervention are needed to enhance clinical applicability.
Conclusion
This study demonstrates excellent intra- and interobserver reliability of a novel WBCT-based method for measuring the Lisfranc interval. The new triplanar technique, which reorients the measurement planes to align with the second metatarsal rather than the floor, improves measurement precision and eliminates systematic errors that occur with uniplanar methods in injured feet. Most significantly, this technique prevents the misidentification of measurement intervals that occurred in nearly one-quarter of injured feet using previous methods. The measurements in the distal C1-M2 interval revealed more pronounced side-to-side differences, suggesting greater clinical utility. However, clinical validation studies correlating these measurements with surgical findings and patient outcomes are essential to establish diagnostic thresholds and confirm the clinical utility of this technique for guiding treatment decisions.
Supplemental Material
sj-pdf-1-fao-10.1177_24730114251372593 – Supplemental material for Three-Plane Alignment of the Second Metatarsal Improves Reliability of Weightbearing CT Measurements in Lisfranc Injury Assessment
Supplemental material, sj-pdf-1-fao-10.1177_24730114251372593 for Three-Plane Alignment of the Second Metatarsal Improves Reliability of Weightbearing CT Measurements in Lisfranc Injury Assessment by Wolfram Grün, Pierre-Henri Vermorel, Emily J. Luo, Daniel Yang, Enrico Pozzessere, Grayson M. Talaski, Francois Lintz and Cesar de Cesar Netto in Foot & Ankle Orthopaedics
Footnotes
Acknowledgements
The authors wish to thank Antoine Acker, MD, for inspiring the present study.
ORCID iDs
Ethical Approval
Our Institutional review board (IRB) approved this study (IRB number Pro00113556).
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: Emily J. Luo, MHSc, reports general disclosures of Sana Biotechnology (shareholder). Grayson M. Talaski, BSE, reports general disclosures of Restore3d (consultant). Francois Lintz, MD, PhD, reports general disclosures of Paragon28 (consultant, shareholder), CurvebeamAI (consultant, shareholder), Newclip Technics (consultant, royalties), Podonov (consultant, royalties), LINNOV (founder, shareholder), Followinvest (shareholder), International WBCT Society (co-founder, past president). Cesar de Cesar Netto, MD, PhD, reports general disclosures of Paragon28 (consultant, medical advisory board, royalties), CurvebeamAI (consultant, shareholder), Ossio (consultant), Zimmer (consultant), Stryker (consultant), International WBCT Society (co-founder, president), Exactech (consultant), Arthrex (consultant), Tayco Brace (shareholder), Extremity Medical (consultant), AOFAS (committee member), Foot Ankle Clinics (editor in chief). Disclosure forms for all authors are available online.
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.
