Valgus Deviation of the Intersesamoid Crista in Hallux Valgus and Its Association With the Distal Metatarsal Articular Angle: A Pilot Study

Abstract

Background:

In hallux valgus (HV), the sesamoid bones embedded in the distal tendon of the flexor hallucis brevis and surrounding the tendon of the flexor hallucis longus are partially guided beneath the head of the first metatarsal (M1) by the intersesamoid crista. The distal metatarsal articular angle (DMAA) assesses the valgus deviation of M1 distal articular surface but is influenced by the pronation of the M1 head relative to the ground. Currently, it is unknown whether the intersesamoid crista itself deviates in valgus in association with the articular surface, and understanding this relationship may have clinical relevance for both the pathogenesis of hallux valgus and its surgical correction.

The aim of our study was to compare the angle between the longitudinal axis of the intersesamoid crista and the M1 longitudinal axis in patients with hallux valgus and control subjects and to evaluate its relationship with the DMAA.

Methods:

A retrospective study was conducted, including 10 HV and 10 matched controls. Weightbearing computed tomography (WBCT) images were automatically segmented with a dedicated software (Disior BoneLogic 2.0) and the angle between the longitudinal axes of the crista and M1 (Crista-M1-angle) as well as the 3d-DMAA (assessing the valgus deviation of the distal articular surface after computerized correction of M1 head pronation relative to the ground) were measured. However, after exclusions for image quality, 9 HV and 8 control feet were analyzed.

Results:

The mean Crista-M1 angle was deviated in valgus by 14.4 ± 8.7 degrees in 9 HV feet and by 5.5 ± 3.2 degrees in 8 control feet (P = .017). The median 3d-DMAA was deviated in valgus by 9.5 degrees (interquartile range 4.0) in the HV group and by 2.7 degrees (interquartile range 4.5) in controls (P < .001). A positive correlation was observed between Crista-M1 angle and 3d-DMAA (ρ = 0.57; r² = 0.328; P = .017).

Conclusion:

In our pilot study, the longitudinal axis of the intersesamoid crista tended to show greater valgus deviation in HV compared to controls, and this deviation appeared to be correlated with the valgus of M1 distal articular surface. These findings suggest a potential morphologic relationship between crista alignment and distal articular surface orientation. However, clinical implications, such as improved sesamoid tracking, remain speculative.

Level of Evidence:

Level III, retrospective comparative study.

Introduction

Hallux valgus (HV) is one of the most common painful pathologies of the foot, with numerous treatment options. ¹⁸ However, recurrences after surgical correction are frequent.^13,15 Those could be due to a lack of understanding of HV pathophysiology.^14,19

The head of the first metatarsal (M1) is composed of 2 plantar articular surfaces for the sesamoids, separated by an intersesamoid crista. This structure plays a biomechanical role by guiding the sesamoid bones, in close contact with surrounding soft tissues, particularly the tendons of the flexor hallucis brevis and longus.^1,8,24 In a cadaveric study, Breslauer and Cohen ² observed that the intersesamoid crista was completely eroded in cases of significant HV deformities, describing a possible inability of this ridge to contain soft tissues aligned with M1, thereby possibly contributing to the HV pathogenesis. However, this anatomical entity is rarely studied. Clarke et al, ³ using bone segmentation and semiautomatic measurements, reported a smaller volume of this crista in cases of HV, which could be indicative of erosion. They also found that this erosion was not associated with pronated malpositioning of the M1 head. But, to our knowledge, no data have been reported on malposition or deviation of this crista in HV population, particularly in relation to the longitudinal axis of M1.

The distal metatarsal articular angle (DMAA), described by Piggott, is used to assess the valgus deviation of the distal articular surface of M1 from its longitudinal axis. ²⁰ This angle is, however, controversial. Several authors have reported a lack of reliability in its measurement, and other studies have shown that it is influenced by pronation of the M1 head.^21,25 Others go even further, claiming that this angle is strongly biased by the superimposition of the lateral plantar condyle (which is round) of the M1 head when pronated, creating an impression of roundness and valgus deviation of the articular surface.^7,23 However, even measurements taking this bias into account and correcting for M1 pronation report higher values of valgus deviation of the distal articular surface of M1 in HV. ¹⁰

Dedicated surgical techniques have been described to correct the DMAA, most consisting of distal osteotomies including a varization wedge to correct DMAA.^4,17 However, this type of surgical technique also modifies the position of the intersesamoid crista in the same way. To the authors' knowledge, we do not know whether the position of the intersesamoid crista is different in HV compared with controls and whether dysmorphia of the entire M1 head may exist, including valgus deviations of both its distal articular surface and the intersesamoid crista relative to the longitudinal axis of M1. This would be of interest, particularly in the case of surgical correction of the DMAA.

Our study therefore aimed to compare the angle between the longitudinal axis of the crista and the axis of the first metatarsal (M1) in hallux valgus and controls, and to assess its correlation with the DMAA. We hypothesized that the longitudinal axis of the crista would be deviated into valgus in the HV group, and that this deviation would be correlated with the DMAA.

Methods

In this institutional review board–approved retrospective case-control study (IRB 201904825), we included 10 patients with symptomatic HV from our institution's weightbearing computed tomography (CT) database (Curevebeam PedCat; Warrington, PA, USA). Ten controls were then selected to match the HV group in terms of age, body mass index (BMI), and gender. Patients with a history of foot surgery or fracture, patients with first metatarsophalangeal joint osteoarthritis (stages C and D according to the Stanmore classification), and congenital or juvenile HV were excluded.^6,26

Conventional radiograph measurements were performed using Carestream Vue PACS software. Weightbearing CT measurements were performed using CubeVue software. Automatic bone segmentation was performed using Disior BoneLogic 2.0 software.

Crista-M1 Angle

The shaft of the metatarsal was defined as the middle 60% of the longitudinal axis of the first metatarsal and was removed. The remaining 20% at both the proximal and distal ends were designated to be the base and head, respectively. The centroid of each of these 2 shapes was calculated. These were used as the endpoints for the vector defining the axis of the first metatarsal.

The intersesamoid crista was defined via selection of its surface using Geomagic Design X. Selections were performed by a board-certified orthopaedic surgeon and reviewed by a PhD-trained researcher. The axis of the intersesamoid crista was defined as the first principal component of the intersesamoid crista selection. The angle between the intersesamoid crista and first metatarsal was calculated as the difference between these 2 angles in the ground plane (Figure 1).

Figure 1.

Example of the Crista M1 angle on a superior view of the point clouds representing the left first metatarsal base and head. M1, first metatarsal.

DMAA

The DMAA was measured on conventional radiographs according to the angle between the distal articular surface of M1 and the orthogonal of longitudinal axis of M1, as previously described. ²⁰

3d-DMAA

The 3D-DMAA was measured on weightbearing CT after multiplanar correction of the first metatarsal alignment, according to a previously published protocol (Figure 2). Specifically, computerized adjustments were made to correct for M1 plantarflexion in the sagittal plane and coronal plane rotation, using the α angle. After aligning the axial and sagittal planes to the metatarsal axis, and rotating the sagittal plane according to the α angle, the distal articular surface was evaluated in the axial plane to measure the corrected DMAA. This method isolates the true valgus orientation of the M1 articular surface, independent of positional artifacts caused by metatarsal malalignment. ¹⁰

Figure 2.

3d-DMAA Measurement.

Statistical Methods

Normality of the distributions was assessed using the Shapiro-Wilk test. For variables with a normal distribution, we used a 2-sample Student t test. For non-normally distributed variables, we applied the Mann-Whitney U test. Correlations between DMAA, 3d-DMAA, and Crista-M1 angle were assessed using Pearson correlation coefficients. Statistical significance was set to P < .05.

Given the pilot nature of this study and the absence of existing data on the deviation angle of the crista relative to M1 longitudinal axis, a power calculation was performed on this parameter using the preliminary results of our study. The calculation has been conducted using a type 1 error rate of 1.0%, a type 2 error rate of 10.0%, and a 2-tailed test.

Results

Two patients in the control group and 1 patient in the HV group were excluded because of insufficient quality images. There was no significant difference between the 2 study groups in terms of demographics (Table 1).

Table 1.

Demographics.

	Hallux Valgus Group (n = 9)	Control Group (n = 8)	P Value
Body mass index, mean (SD)	28.1 (5.6)	32.3 (8)	.22
Age, y, mean (SD)	39 (7.5)	37.9 (12.1)	.53
Gender: female, n (%)	6 (75)	5 (55.6)	.62

Crista-M1 Angle

The mean Crista-M1 angle presented a significantly greater valgus deviation in the HV group (14.4 ± 8.7 degrees, 95% CI 7.7-21.1) compared with the control group (5.5 ± 3.2 degrees, 95% CI 2.9-8.2) (P = .017) (Figure 3).

Figure 3.

Comparison of the Crista-M1 angle in hallux valgus (HV) and control groups.

DMAA

On conventional radiographs, the mean DMAA was significantly higher in the HV group (25.1 ± 7.9 degrees, 95% CI 19-31.2) compared with controls (7.4 ± 2.9 degrees, 95% CI 5-9.8) (P < .001) (Figure 4).

Figure 4.

(A) Comparison of the DMAA in hallux valgus (HV) and control groups. (B) Comparison of the 3d-DMAA in hallux valgus (HV) and control groups.

3d-DMAA

On weightbearing CT, after computerized correction of the M1 head pronation relative to the ground, the mean 3d-DMAA was significantly higher in the HV group (9.5 degrees [interquartile range 4.0], 95% CI 8.1-16.8) compared to controls (2.7 degrees [interquartile range 4.5], 95% CI 1.1-5) (P < .001) (Figure 4).

Correlations

A non-significant positive correlation was found between the Crista-M1 angle and the DMAA (ρ = 0.44, r² = 0.193, P = .08). A significant positive correlation was observed between the Crista-M1 angle and the 3d-DMAA (ρ = 0.57, r² = 0.33, P = .017) (Figure 5).

Figure 5.

(A) Correlation between the Crista-M1 angle and the DMAA. (B) Correlation between the Crista-M1 angle and the 3d-DMAA.

Power Calculation

Based on these results, assuming a mean crista deviation of 14.4 degrees in the HV group and 5.5 degrees in the control group, with a standard deviation of 8.7 degrees, we estimated that a sample size of 29 feet per group would be required.

Discussion

This pilot study found that the intersesamoid crista had greater valgus deviation in HV cases than in controls, with a moderate correlation to the valgus alignment of the M1 distal articular surface.

In severe HV, a metatarsosesamoid subluxation can be observed. Although this may appear to be a lateral dislocation of the sesamoids, it is more likely to be due to a medial displacement of the M1 head over the sesamoids. ²⁴ Thus, the intersesamoid crista may play a role in stabilizing the M1 head, preventing it from rolling off the sesamoid. ¹⁴ When surgically correcting HV, the goal is to realign the M1 head over the sesamoids. Other authors also aim to realign the distal articular surface by varizing it to correct the DMAA and create better balance in the first metatarsophalangeal joint.^4,17 Performing this correction on a crista that is not deviated in valgus would create a mismatch between the M1 distal articular surface and the longitudinal axis of the intersesamoid crista. Following our study findings, the valgus deviation appeared to be correlated with the longitudinal axis of the crista. This supports the argument for correcting the DMAA when present (while also considering the pronation of the M1 head relative to the ground) as it would, at the same time, realign the crista and potentially improve the alignment and tracking of the sesamoids around the crista, thereby enhancing the balance of the first metatarsophalangeal joint (Figure 6).

Figure 6.

Sesamoid tracking concept.

The valgus deviation of the crista could be explained by 2 main hypotheses: a developmental origin, where an intrinsic malformation of the first metatarsal head leads to crista deviation from an early stage (in utero or postnatal growth), or an acquired process, where progressive erosion and degeneration occur over time due to mechanical stress and altered biomechanics in the metatarsosesamoid complex. On one hand, the Clarke et al ³ study supports the erosion theory, as they observed a reduced crista volume in HV patients. On the other hand, the fact that in our study, both the crista and the deviation of the distal articular surface deviated in the same way in valgus may support a developmental origin. However, our study did not enable determination of whether the valgus deviation of the crista is primarily due to erosion or developmental factors, which will require further research.

Metatarsosesamoid arthritis also appears to be related to M1 head pronation relative to the ground. Soares et al ²² found that the presence of intersesamoid crista osteoarthritis correlated with greater HV severity (higher HVA and IMA), but with a lower degree of M1 head pronation. This aligns with other studies suggesting that a decrease in M1 head pronation relative to the ground may be one of the key triggering factors for the M1 head to roll out of the sesamoids (through a combined supination and abduction motion) while the sesamoids remain attached to the second metatarsal by the deep transverse ligament.^5,11,12,14 Based on these findings, the crista may function as a stabilizing structure that helps maintain the M1 head pronated over the sesamoids in the early stages of HV, until it becomes eroded and loses its supportive role. This loss of structural integrity could contribute to the disconnection between the M1 head and the sesamoids, ultimately leading to the development of HV deformity. ¹⁴

Maddocks et al ¹⁶ performed a cadaveric study evaluating the relationship between the intersesamoid crista and the DMAA. Their findings did not reveal any significant association between crista erosion and an increase in DMAA, suggesting that these 2 parameters may remain independent. These findings could align with our results. When the DMAA and the longitudinal axis of the crista are well aligned, sesamoid tracking remains smooth, reducing the likelihood of metatarsosesamoid osteoarthritis. Rather, the issue would arise when a mismatch occurs between these 2 parameters. In such cases, improper alignment may disrupt normal sesamoid tracking, increase joint stress, and contribute to metatarsosesamoid degeneration.

This pilot study has several limitations that must be considered. First, its retrospective design and very small sample size (n = 17 total) substantially limit the generalizability and robustness of the findings. However, even with a small number of patients, our study reports significant and consistent differences between groups, suggesting that the association may be strong. Second, while we have found a significant relationship between the valgus deviation of the crista and the M1 distal articular surface, our study design does not allow us to determine causality. However, these findings provide valuable insights and perspectives for future research. Third, our study measured the angle between the M1 axis and crista axis, but we do not yet understand the positional variation of the crista (whether it is more medial or lateral). Fourth, we did not report the height or degree of erosion of the crista, which may influence the results of this study. We partially addressed this limitation by excluding patients with advanced degenerative changes of the first metatarsophalangeal joint (stages C and D according to the Stanmore classification). ²⁶ Nevertheless, future studies should include a more robust assessment of the crista, including its height and volume to better control this factor. Fifth, although our protocol combined both manual and automated steps, we did not perform formal intra or interrater reliability assessment of the segmentation and definition of the crista axis within this pilot study. However, it is important to note that the segmentation process itself was based on software and methodologies with previously established reliability and reproducibility. ⁹ Importantly, this study was not designed as a methodological validation paper, but rather as an initial morphologic exploration of intersesamoid crista orientation in HV, an aspect that has not been previously described. We believe these preliminary findings provide a valuable foundation for future, more robust investigations. These future studies should incorporate reliability assessment and align with ongoing international efforts to standardize 3D measurements in foot and ankle imaging. ⁹ Finally, because our study did not include dynamic imaging, we can only speculate about the functional behavior of the metatarsosesamoid joint. However, it does highlight the importance of sesamoid tracking, a concept that is not yet widely recognized and whose role is possibly underestimated.

Conclusion

In our study, the longitudinal axis of the intersesamoid crista tended to show greater valgus deviation in HV compared to controls, and this deviation appeared to be correlated with the valgus of M1 distal articular surface. These findings suggest a possible morphologic link between the distal M1 articular surface and crista orientation in HV. However, clinical or surgical implications, such as improved sesamoid tracking, remain speculative and warrant further investigation.