Comparison and agreement analysis of Böhler and Gissane angles on radiographs and CT in dry human calcanei

Abstract

Purpose

This study aimed to compare the measurements of Böhler’s angle (BA) and the critical angle of Gissane (GA) between lateral radiographs and computed tomography (CT) scans in dry human calcanei, and to assess the inter- and intraobserver agreement for each angle across both imaging modalities.

Methods

A total of 160 dry human calcanei were examined. Both BA and AG were measured on digital radiographs and multiplanar CT reconstructions. Two independent observers - a radiologist and an orthopedist performed all measurements in two separate sessions. Mean values, standard deviations, and agreement metrics were calculated. The inter- and intraobserver reliability were assessed using intraclass correlation coefficients (ICC). Paired statistical tests were used to compare radiographic and tomographic measurements.

Results

The mean value of the BA was 31.8° ± 5.8° and 29.7° ± 5.6° on radiographies and CT-scans, respectively. The GA showed mean values of 108.7° ± 9.2° on radiographs and 125.7° ± 10.3° on CT. CT-based measurements of BA demonstrated the highest reproducibility (ICC > 0.90), with excellent inter- and intraobserver agreement. In contrast, GA showed poor agreement across all comparisons, particularly on radiographs (ICC < 0.30). Observer agreement was consistently higher for BA, and CT substantially reduced interobserver variability.

Conclusion

The BA demonstrated superior measurement reliability compared to the Gissane angle, particularly when assessed by CT. GA measurements demonstrated low reproducibility and should be interpreted with caution, especially when obtained from x-rays. These findings support the preferential use of CT-based BA for morphometric and clinical evaluations of the calcaneus.

Introduction

Calcaneal fractures are complex injuries that often result from high-energy trauma. They often involve the subtalar joint and thus can lead to significant disability and long-term functional limitations of the ankle [1, 2]. Calcaneal fractures are complex injuries frequently associated with long-term functional impairment, particularly when the subtalar joint is involved. Recent three-dimensional mapping and morphometric studies have demonstrated that fracture patterns, displacement vectors, and articular surface involvement are highly heterogeneous and closely related to clinical outcomes [3–5]. These findings reinforce the importance of reliable anatomical parameters in the radiological assessment of calcaneal fractures, as imaging-derived measurements continue to inform classification, surgical planning, and prognostic evaluation.

Lateral radiographs are commonly used to assess these injuries, and measurements of Böhler’s angle (BA) and the critical angle of Gissane (GA) can help evaluate the severity of displacement and joint involvement, particularly in intra-articular fractures [6, 7]. However, their reliability has been frequently questioned. Studies have reported high variability between observers, especially when fractures distort normal landmarks or when radiographic techniques introduce subtle misalignments [8, 9].

Computed tomography (CT) has emerged as a useful alternative, offering clearer visualization of bony contours and more consistent identification of key points used for angle measurement [10]. CT-based measurements of BA, including central and 3D variants, appear to yield more reproducible results - though the same cannot always be said for GA, which remains more susceptible to observer variation [11, 12].

Most previous studies have utilized clinical imaging, where measurement accuracy may be compromised by soft tissue interference, edema, and patient positioning variability. In this context, the use of dry bone specimens offers a controlled and standardized anatomical substrate, eliminating these confounding variables and allowing for a more precise evaluation of methodological consistency, although the literature lacks comparative studies in imaging exams in dry bones [1, 13, 14].

This study aims to compare the BA and GA angles taken from radiographs and CT scans of 160 dry human calcanei and assess their intra- and interobserver agreement.

Materials and methods

Sample

A total of 160 dry human calcanei from skeletally mature individuals were selected from the anatomical collection of the Laboratory of Anatomy, Department of Morphology, Universidade Federal Fluminense (UFF). All specimens were intact and well preserved, with identifiable articular contours and cortical surfaces.

Image acquisition

All imaging procedures were conducted at the Department of Radiology of the Antonio Pedro University Hospital (HUAP). Radiographic images were obtained in lateral view using a Philips Digital Diagnost system (ELEVA version 2.1). The calcanei were positioned with the medial side facing upward and aligned for standard lateral projection. Cardboard supports were used to maintain the orientation of the bones, compensating for the absence of soft tissues. Images were acquired in batches of 20 bones using a 35 × 43 cm digital detector, fine focus, 5 mAs, and 10 kV.

Computed tomography (CT) was performed using a Philips Brilliance 64-slice scanner. The bones were arranged on foam boards, 40 per scan, to minimize interference from the CT table and to simulate patient positioning. High-resolution images were acquired using the following parameters: collimation 16 × 0.625 mm, tube rotation time 0.5 s, field of view 174 mm, pitch 0.685, slice thickness 1 mm, reconstruction interval 0.8 mm, and bone reconstruction filter.

Angle measurement

The Böhler and Gissane angles were measured in radiographic and tomographic images based on standard anatomical landmarks (Fig. 1). Digital images were analyzed using the PAX^® software (Department of Radiology, HUAP), which allows precise definition of points and line angles within and across planes. In radiographs, angles were traced in a single sagittal view. In CT imaging, angle measurements were obtained using multiplanar reconstructions aligned with the sagittal plane of the calcaneus. When necessary, adjacent slices were reviewed to accurately identify the anatomical landmarks corresponding to the apex of the anterior process and the posterior articular facet (Fig. 2).

Fig. 1.

Lateral radiographic view of the ankle showing the Bohler angle (BA) and Gissane angle (GA). The BA is formed by a line drawn from the highest point of the anterior process of the calcaneus to the highest point of the posterior articular facet a and a line drawn from the highest point of the posterior articular facet to the highest point of the calcaneal tuberosity b. The GA is formed by a line drawn along the upward slope of the anterior calcaneal process a and a line drawn along the downward slope of the posterior facet, forming an apex at the lateral border of the posterior facet b.

Fig. 2.

CT-based measurement of calcaneal angles in sagittal multiplanar reconstruction. These illustrations demonstrate the angular definitions used in multiplanar CT analysis of dry calcanei

All measurements were independently performed by two evaluators: one radiologist and one orthopedic surgeon, both trained in musculoskeletal imaging.

Statistical analysis

Descriptive statistics were used to summarize the angle measurements and were presented in mean, standard deviation (±) or confidence interval. The Kolmogorov-Smirnov and Shapiro-Wilk tests were used to observe normal distribution of the data. When both compared distributions were normal, paired measurements were analyzed using the Student’s t-test. When normality was not confirmed, the Wilcoxon signed-rank test was applied. Homogeneity of variances was assessed using Levene’s test.

Agreement between radiographic and tomographic measurements, as well as between observers, was evaluated using the intraclass correlation coefficient (ICC), calculated using a two-way mixed-effects model. ICC values were interpreted as follows: poor (≤ 0.20), fair (0.21–0.40), good (0.41–0.60), very good (0.61–0.80), and excellent (≥ 0.81) [15]. Confidence intervals (CI 95%) were reported for all ICC values. Additionally, the proportion of exact numerical agreement (i.e., cases with identical paired values) was calculated for each comparison (D = 0).

All statistical analysis was conducted using SPSS version 22.0 (IBM Corp., Armonk, NY), adopting a significance level of 5%.

Results

A total of 160 calcanei were analyzed, and measurements of the Böhler and Gissane angles were obtained from both radiographs and computed tomography (CT) scans. The mean value of the Böhler angle was 31.8° ± 5.8° on radiographs and 29.7° ± 5.6° on CT. The Gissane angle showed mean values of 108.7° ± 9.2° on radiographs and 125.7° ± 10.3° on CT (Table 1; Figs. 3 and 4).

Table 1.

Mean values of the Bohler angle and Gissane angle by method and observer

Böhler angle	Radiography	CT scan	p value
Observer 1	31.5°±5.7°	31.2°±5.5°	0.275a
Observer 2	32.0°±5.9°	28.2°±5.3°.	< 0,0001b
Global	31.8°±5.8°	29.7 ± 5.6°	< 0.0001c
Gissane angle
Observer 1	107.4°±10.1°	126.5°±8.0°,	< 0,0001b
Observer 2	110.1°±8.0°	124.9°±12.2°	< 0,0001d
Global	108.7°±9.2°	125.7°±10.3°	< 0.0001c

^a = Radiography vs. CT scan (Observer 1), Wilcoxon test. ^b = Radiography vs. CT scan (Observer 2), Wilcoxon test. ^c = Radiography vs. CT scan (Global), Wilcoxon test. ^d = Radiography vs. CT scan (Observer 2), Paired t test

Fig. 3.

Comparison of mean Böhler’s angle between radiography and CT-scan measurements for each rater and global mean values. Bars represent mean ± standard deviation. Although both raters showed similar mean angles between modalities, CT-scan measurements tended to yield slightly lower values compared with radiographs

Fig. 4.

Comparison of mean Gissane’s angle between radiography and CT-scan measurements for each rater and global mean values. Bars represent mean ± standard deviation. CT-scan measurements consistently showed higher mean Gissane’s angle values than radiographs across both raters and global averages

All paired comparisons were performed using the paired Student’s t-test or the Wilcoxon signed-rank test, depending on the normality of the distributions.

For Observer 1, the Böhler angle measured 31.5° ± 5.7° on radiographs and 31.2° ± 5.5° on CT. For Observer 2, the respective values were 32.0° ± 5.9° and 28.2° ± 5.3°. For the Gissane angle, Observer 1 recorded radiographic values of 107.4° ± 10.1° and CT values of 126.5° ± 8.0°, while Observer 2 measured radiographic values of 110.1° ± 8.0° and CT values of 124.9° ± 12.2° (Table 1).

Statistical testing revealed no significant difference between radiographic and tomographic measurements of the Böhler angle for Observer 1 (p = 0.275). For Observer 2, the difference was significant (p < 0.001), as was the case in the global analysis (p < 0.001). For the Gissane angle, statistically significant differences between radiographic and tomographic values were found for both observers and in the global comparison (p < 0.001 in all cases) (Table 1).

Comparison between observers showed significant differences for the Böhler angle on both radiographs (p = 0.018, Wilcoxon test) and CT (p < 0.001, Wilcoxon test), and for the Gissane angle on radiographs (p < 0.001, Student’s t test) and CT (p = 0.015, Wilcoxon test).

The intraclass correlation coefficient (ICC) between radiographic and tomographic measurements of the Böhler angle was 0.875 (95% CI: 0.83–0.91) for Observer 1, 0.659 (0.56–0.74) for Observer 2, and 0.731 (0.68–0.78) in the global analysis. For the Gissane angle, the ICC values were 0.250 (0.10–0.39), 0.297 (0.15–0.43), and 0.262 (0.16–0.36), respectively (Table 2).

Table 2.

Agreement analysis between the angles measured on radiography (XR) and CT-scan (CT), for both observers

Observer	Böhler angle – XR vs. CT
	ICC (CI 95%)	Classification	D = 0
1	0.875 (0.83–0.91)	Excellent	18.13%
2	0.659 (0.56–0.74)	Very good	5.62%
Global	0.731 (0.68–0.78)	Very good	11.88%
Observer	Gissane angle – XR vs. CT
1	0.250 (0.10–0.39)	Poor	1.25%
2	0.297 (0.15–0.43)	Poor	1.88%
Global	0.262 (0.16–0.36)	Poor	1.56%

XR: Radiography; CT: CT-scan; ICC: intraclass coefficient; CI: confidence interval; D = 0: cases with identical paired values

Between observers, the ICC for the Böhler angle was 0.881 (0.84–0.91) on radiographs and 0.819 (0.76–0.86) on CT. For the Gissane angle, ICC values were 0.370 (0.23–0.50) on radiographs and 0.187 (0.03–0.33) on CT. Exact numerical agreement between radiographic and tomographic measurements of the Böhler angle occurred in 11.88% of cases, and for the Gissane angle in 1.56%. Agreement between observers occurred in up to 16.88% of Böhler angle cases and 7.50% of Gissane angle cases (Table 3).

Table 3.

Agreement analysis between the angles measured by the two observers

Exam	Böhler angle – Observer 1 vs. Observer 2
	ICC (CI 95%)	Classification	D = 0
XR	0.881 (0.84–0.91)	Excellent	16.88%
CT	0.819 (0.76–0.86)	Very good	15.0%
Exam	Gissane angle – Observer 1 vs. Observer 2
XR	0.370 (0.23–0.50)	Fair	2.50%
CT	0.187 (0.03–0.33)	Poor	7.50%

XR: Radiography; CT: CT-scan; ICC: intraclass coefficient; CI: confidence interval; D = 0: cases with identical paired values

Discussion

The present study investigated the measurements of Böhler and Gissane angles using lateral-view radiographs and CT scans of dry human calcanei, aiming to assess their agreement across imaging modalities and evaluators. Our findings showed that the Böhler angle demonstrated significantly better agreement both between modalities and observers, while the Gissane angle showed lower reproducibility and was more susceptible to measurement variability. When comparing modalities, Böhler’s angle was on average 2.1° lower on CT than on radiographs, representing a relative difference of 6.6%. Conversely, the Gissane angle was 17.0° higher on CT, corresponding to a 13.5% relative difference. These values suggest that while both angles show structural discrepancies between imaging modalities, the disagreement is proportionally greater for GA.

In the overall analysis, disregarding observer identity, measurements of both Böhler and Gissane angles obtained from radiographs differed significantly from those obtained via CT (p < 0.001). On average and median comparisons, radiographic measurements of the Böhler angle were lower than CT-based values, suggesting an underestimation of this angle on radiographs. Conversely, the GA tended to be overestimated when measured on radiographs, showing higher values compared to CT.

Despite these differences, the agreement analysis showed excellent intraclass correlation between radiographic and CT measurements for the Böhler angle, which contrasts with the significance testing results. For the Gissane angle, however, both significance testing and ICC analysis pointed to poor agreement, reinforcing concerns about its measurement reliability.

When comparing measurements between the two observers, statistically significant differences were found for GA (p < 0.001, x-ray; p = 0.015, CT) and BA (p = 0.018, x-ray; p < 0.001, CT) across imaging modalities. Nonetheless, the ICC revealed very good to excellent interobserver agreement for the Böhler angle, while the Gissane angle again showed only poor to moderate agreement between raters.

The relevance of reproducible calcaneal measurements becomes particularly evident when considered alongside recent three-dimensional and outcome-based studies. Yu et al. (2022) [3] demonstrated that specific regions of fracture involvement and displacement patterns, identified through fracture mapping, are associated with prognosis and functional outcomes. Similarly, Lu et al. (2024) [4] emphasized that three-dimensional morphometric parameters contribute meaningfully to prognostic stratification in calcaneal fractures. In this context, angular measurements that lack reproducibility - such as the GA observed in the present study - may offer limited value when used in isolation for clinical decision-making, particularly in complex fracture patterns.

This discrepancy between statistical significance and ICC has been previously highlighted by Bartko (1994) [16], who recommended prioritizing paired significance tests over ICC when interpreting agreement. Overall, the results underscore that angle measurements differ meaningfully between imaging modalities and are also influenced by observer expertise - a factor particularly relevant given the different clinical backgrounds of the raters involved in this study.

These results are in line with earlier studies reporting higher interobserver reliability for Böhler angle measurements compared to the Gissane angle, particularly when assessed using CT. For instance, Eelsing et al. (2024) [1] found that central and three-dimensional reconstructions of the Böhler angle on CT showed excellent agreement, outperforming both medial and lateral reconstructions as well as traditional radiographs. Similarly, De Boer et al. (2020) [10] highlighted the influence of radiographic obliquity on angle distortion, underscoring the need for alternative imaging planes to improve agreement.

While the literature often supports the reliability of Böhler’s angle, questions persist regarding its consistency across different populations. Ma et al. (2019) [17] evaluated normal values in a Chinese cohort and reported a BA of 31.6º, while Sengondan et al. (2012) [12] showed a BA of 30.6º and a GA of 126.7º in an Indian cohort, and Seyahi et al. (2009) [18] observed a mean BA of 33.8º and a mean GA of 115º, thus suggesting ethnic influences. Our use of dry calcanei, without complete demographic data, limits our ability to assess such variability, but the findings presented herein still contribute normative anatomical data for morphometric applications regarding Brazilian cohorts.

In contrast to the reliability of the BA, the GA in our study exhibited notably lower ICC values across all comparisons, reaffirming its reputation in the literature as a less reliable morphometric parameter. Prior analyses have identified challenges in consistently defining the apex of the calcaneal notch, especially in disrupted or remodeled bone surfaces, even when CT is used [6, 8]. Eelsing et al. (2024) [1] likewise reported that GA measurements yielded only moderate interobserver agreement even under 3D conditions, which they attributed to the angle’s susceptibility to subtle spatial variations and complex surface geometry.

The consistently poor inter- and intraobserver agreement observed for the Gissane angle across both radiographic and CT modalities raises important questions about its ongoing utility in clinical practice. While GA remains a historically recognized parameter in calcaneal fracture evaluation, our findings suggest that its inherent susceptibility to anatomical ambiguity and imaging variability significantly undermines its reliability. In clinical settings where precise morphometric assessment is critical - such as preoperative planning or outcome evaluation - GA should be interpreted with considerable caution, and ideally not be used in isolation [1, 12, 19]. Its role may still be justified as part of a broader assessment framework when combined with fracture classification systems or axial reconstructions. However, based on our results and prior literature, Böhler’s angle offers a more dependable reference and should be prioritized when assessing calcaneal morphology and displacement.

Our findings also showed that radiographs tended to overestimate both angles compared to CT, a tendency observed in both clinical and experimental studies. Ramachandran and Shetty (2015) [20] emphasized the influence of radiographic technique on angular measurement variability, while Seyahi et al. (2009) [18] demonstrated that standardization of projection alone is insufficient to eliminate distortion. In this regard, Labronici et al. (2019) [7] offered valuable clinical insight by comparing pre- and postoperative radiographs, showing that radiographic BA is not only affected by projection but also by surgical correction, further highlighting the dynamic nature of the angle depending on imaging and intervention.

The difference in angle reproducibility between evaluators observed in our study further highlights the subjective element inherent in morphometric analysis. This aligns with findings by Sayed-Noor et al. (2011) [9], who noted that interobserver agreement remains limited for several fracture classification systems, including those incorporating angular metrics. Our results also reflect the concerns raised by Knight et al. (2006) [19], who questioned the diagnostic reliability of Böhler and Gissane angles in emergency settings. Still, De Boer et al. (2020) [10] and Barroco et al. (2021) [6] suggest that training and standardization can mitigate some of these discrepancies.

While 3D-CT has been proposed as a more accurate tool for anatomical reconstruction and classification, its role in angular measurement remains contested. Veltman et al. (2014) [2] argued that 3D CT, despite improving visualization, may not offer a substantial advantage in fracture classification or angle-based evaluation when compared to conventional CT or radiographs. Our study supports this position, as we found improved reproducibility for Böhler’s angle with CT but did not find enough benefit to fully endorse 3D CT as essential in all cases.

We recognize that segmented three-dimensional imaging may offer specific advantages for angle measurements. By allowing more consistent identification of anatomical landmarks in three-dimensional space, segmentation could reduce variability, particularly for Gissane’s angle, which is highly sensitive to slice orientation. Nevertheless, this approach requires specialized software, technical expertise, and longer processing time, limiting its applicability in routine clinical workflows. Although not used in the present study, segmented 3D reconstructions may serve as a valuable tool for future investigations aiming to reduce observer subjectivity and increase standardization in morphometric analysis.

While our findings support the use of CT-based BA as a more reliable morphometric parameter, it is important to recognize that routine CT may not be feasible in all clinical settings due to higher cost, reduced accessibility, and increased radiation exposure. Therefore, although CT can enhance measurement accuracy, especially in complex or preoperative cases, radiographs remain a practical and widely available tool. In such contexts, reliance on Böhler’s angle - given its superior agreement across modalities - may help offset the inherent limitations of radiographic imaging. In contrast, Gissane angle measurements, due to their low reproducibility, should be interpreted with caution and not used in isolation for clinical decision-making.

In such contexts, reliance on Böhler’s angle, given its superior agreement across modalities, may help offset the inherent limitations of radiographic imaging. In contrast, the Gissane angle, due to its low reproducibility, should not be used in isolation for clinical decision-making, particularly in surgical planning or prognostic modeling. These findings reinforce recent recommendations to complement angular measurements with fracture classification systems and axial reconstructions, rather than relying solely on lateral angle values [1, 10].

In this study, the use of dry human calcanei provided standardized conditions that eliminated confounding factors such as soft tissue interference, patient movement, and variable positioning. This approach allowed for precise morphometric comparisons between radiographic and CT measurements, focusing exclusively on the intrinsic anatomical variability of the calcaneus. However, this methodological advantage also represents a limitation, as clinical imaging of calcaneal fractures is influenced by factors absent in dry specimens including swelling, soft-tissue density, and projection differences related to patient positioning. Consequently, the present findings should be interpreted as reflecting ideal measurement conditions rather than direct clinical performance. Nonetheless, these results establish a reproducible baseline for understanding angular variability, which can guide interpretation and optimization of imaging protocols in fracture assessment and surgical planning. Future studies involving cadaveric or in vivo imaging are warranted to validate these observations under realistic clinical condition.

Additionally, the sex and age of the calcanei were unknown, although multiple studies have shown that neither sex nor age significantly affects Böhler or Gissane angle values in skeletally mature individuals, suggesting that population variability in these parameters is low once skeletal development is complete [13, 14, 21].

However, studies regarding different ethnicities showed significant angular variation between them [18]. Without this information, our findings should be interpreted as population-agnostic, and caution is advised when extrapolating these normative values to specific demographic contexts.

Finally, while using two trained evaluators from distinct specialties offers a realistic assessment of clinical variability, a larger panel of raters could provide more robust estimates of interobserver agreement, as emphasized by Sayed-Noor et al. (2011) [9] and Otero et al. (2015) [8] thus we lay this as a suggestion for future studies.

Conclusion

This study found that Böhler’s angle demonstrates acceptable intra- and interobserver agreement, particularly when measured on CT. In contrast, the Gissane angle showed poor reproducibility across both imaging modalities and observers. Although CT improved consistency for both angles, radiographs remain a practical and accessible tool. Given its higher reliability, Böhler’s angle should be prioritized in morphometric assessments of the calcaneus, whereas the Gissane angle should not be used in isolation for clinical decision-making.

Author contributions

Conception and design: TN, EC, AdSAnalysis and interpretation: TN, AdS, LP, MB, VGData collection: TN, EC, AdS, RFWriting the article: TN, YZ, RF, LP, MBCritical revision of the article: TN, EC, AdS, YZ, RF, LP, MB, VGFinal approval of the article*: TN, EC, AdS, YZ, RF, LP, MB, VGStatistical analysis: TN, EC, AdS, LP, MBOverall responsibility: TN, VG.

Funding

No funding.

Data availability

Not applicable.

Declarations

Ethical approval and consent to participate

This study was approved by the Ethics Committee of the Universidade Federal Fluminense under protocol number CAAE: 31326514.9.0000.5243. All procedures were conducted in accordance with the Declaration of Helsinki (2000).

Consent to publish

Not applicable.

Competing interests

This study did not receive any financial support from public, commercial, or non-profit sources. The authors have no competing interests to declare that are relevant to the content of this article.

Informed consent

Informed consent was obtained for the writing of this article.