Diagnostic performance of ultrasound and magnetic resonance imaging in ankle injuries: a retrospective cohort study

Abstract

Background

This study compares the diagnostic performance of ultrasound (US) and magnetic resonance imaging (MRI) in detecting abnormalities such as ligament tears, tendon injuries, or fractures in the ankle, while also evaluating their clinical utility and limitations.

Methods

This study involved 83 patients with first-time ankle injuries. Surgical exploration served as the reference standard to compare and analyze US and MRI, evaluating sensitivity, specificity, positive predictive value (PPV, %), negative predictive value (NPV, %), accuracy (%), and area under the curve (AUC).

Results

US demonstrated significant diagnostic advantages for most ankle injuries, achieving 100% across all metrics for anterior talofibular ligament (ATFL) injury (κ = 1, AUC = 1). Moreover, US exhibited significantly higher overall diagnostic performance (AUC) compared to MRI for the calcaneofibular ligament (CFL), deltoid ligament (DL), peroneus longus tendon (PLT), and fracture detection. Conversely, MRI showed superior performance for syndesmotic (SDM) injury compared to US, with sensitivity rates of 46.67% versus 33.33% (AUC: 0.658 vs. 0.591), while maintaining 100% specificity for both ATFL injury and fracture. Both modalities demonstrated high sensitivity (100%) and excellent diagnostic value (AUC > 0.9) for tibialis posterior tendon (TPT) injury.

Conclusion

This study confirms that US demonstrates superior performance in diagnosing superficial ligament injuries, specifically the ATFL and the CFL. In contrast, MRI is indispensable for evaluating deeper complex structures, such as the SDM and fracture lesions. Both modalities exhibit a complementary relationship in the diagnosis of ankle injuries.

Background

As one of the largest load-bearing joints in the human body, the ankle plays a crucial role in both daily activities and sports. Ankle injuries typically occur in sports that require running and jumping, such as basketball and football [1]. Approximately 70% of the population sustains ankle injuries at some point in their lifetime [2]. Ankle injuries commonly involve the anterior talofibular ligament (ATFL), calcaneofibular ligament (CFL), deltoid ligament (DL), syndesmosis (SDM), peroneus longus tendon (PLT), peroneus brevis tendon (PBT), and tibialis posterior tendon (TPT), or manifest as fractures [3, 4]. Approximately 70% of patients with acute ankle injuries may develop chronic ankle instability, which significantly impacts daily function [5]. Consequently, reliable and efficient diagnostic approaches are urgently needed to facilitate early diagnosis and initiate timely rehabilitation interventions.

Currently, ultrasound (US) and magnetic resonance imaging (MRI) are commonly employed to assist in the diagnosis of ankle injuries [6]. However, no established reference standard exists for the imaging diagnosis of certain ankle injuries, such as the CFL [7]. MRI exhibits high sensitivity for detecting soft tissue injuries, particularly demonstrating significant diagnostic sensitivity for injuries to the ATFL and CFL [8]. Nevertheless, its limitations, such as relatively high costs and susceptibility to motion artifacts, restrict its widespread application in clinical practice [9]. Conversely, US is anticipated to be a reliable tool for evaluating ankle injuries due to its advantages of convenient bedside operation and dynamic observation [10]. However, its diagnostic efficacy remains controversial, particularly due to its limited resolution for deep tissue structures.

This study aims to conduct a comparative analysis of the diagnostic accuracy of US and MRI for ankle injuries, using the results of surgical exploration as the gold standard.

Materials and methods

Patients

This study is a retrospective cohort study that systematically collects and analyzes clinical data from patients with ankle injuries treated at the 960th Hospital of the PLA Joint Logistics Support Force between December 2022 and May 2025. The inclusion criteria consisted of patients aged 18 to 40 years who presented with their first ankle injury. The study protocol was approved by Ethics Review Committee of the 960th Hospital of the PLA Joint Logistics Support Force and was conducted in accordance with the ethical principles of the Declaration of Helsinki. Informed consent was waived by the same Committee as this was a retrospective study of existing medical records.

Exclusion criteria encompassed patients aged under 18 or over 40 years; individuals with prior ipsilateral ankle injuries, surgical interventions, or fractures; subjects with underlying conditions that could compromise ankle structure and function (e.g., osteoarthritis, cartilage damage, or rheumatoid arthritis); cases with missing or incomplete critical clinical documentation; and participants unable to fulfill study protocol requirements for various reasons. The recorded demographic and clinical parameters included patient gender, age, causes of ankle injury, timing of injury, duration of disease, and surgical outcomes.

US

All patients underwent a high-resolution ultrasound examination using a PHILIPS EPIQ7 with a linear array probe (5–12 MHz). The examination of the ankle joint was performed by an ultrasound specialist with over ten years of experience in musculoskeletal ultrasound, who strictly adhered to the diagnostic guidelines established by the European Society of Musculoskeletal Radiology (ESSR) [11]. All images were stored for further analysis.

MRI

All examined individuals were positioned supinely and scanned using a 3.0T MRI system (GE Discovery750) equipped with a dedicated multi-channel extremity coil. To enhance the visualization of the peroneal tendons, the ankle was aligned in a neutral position with approximately 20 degrees of plantar flexion [12]. The MRI results were diagnosed and interpreted by a certified musculoskeletal radiologist who was unaware of the patients’ clinical and ultrasound diagnoses, with strict adherence to the diagnostic guidelines established by the ESSR.

All patients in this study completed US and MRI within four days [4]. After confirming the surgical indications and completing preoperative preparations, all patients proceeded to surgical intervention. The procedures were performed by an attending surgeon with over ten years of specialized experience in ankle surgery. Surgical findings were documented after collaborative confirmation by two senior surgeons.

Statistical analysis

Data analysis was conducted using SPSS version 27.0. Categorical variables were reported as frequencies with corresponding percentages. The calculated metrics included sensitivity, specificity, positive predictive value (PPV, %), negative predictive value (NPV, %), accuracy (%), and the area under the curve (AUC). The Kappa statistic was applied to evaluate the consistency between the diagnoses derived from US and MRI modalities. Receiver Operating Characteristic (ROC) curves were generated for both US and MRI examinations, utilizing surgical outcomes as the gold standard. A P-value of less than 0.05 was considered statistically significant.

Results

A total of 83 patients were included in this study, with a mean age of 26.94 ± 5.58 years. Among the participants, 78 (94.00%) were male and 5 (6.00%) were female. The distribution of ankle injury duration at initial presentation was as follows: 54 cases (65.06%) had an injury duration of less than 7 days, 18 cases (21.69%) had an injury duration between 7 and 21 days, and 11 cases (13.25%) had an injury duration exceeding 21 days (Table 1).

Table 1.

Demographic characteristics and injury duration of patients with ankle injuries

Variable	n = 83
Age	26.94 ± 5.58 years
Gender
Female	5 (6.00%)
Male	78 (94.00%)
Disease duration
<7 days	54 (65.06%)
7–21 days	18 (21.69%)
>21 days	11 (13.25%)

Intraoperative confirmation revealed that the ATFL injury was the most prevalent pathology, affecting 93.98% of the enrolled patients. This was followed by SDM at 36.14%, CFL injury at 31.33%, fractures at 18.07%, and DL injury at 14.45%. The detailed numbers of positive cases identified through US and MRI are presented in Table 2.

Table 2.

Comparison of US, MRI, and surgical findings in ankle injuries

Site of Injury	US	MRI	Surgical diagnosis
ATFL	78 (93.98%)	60 (72.29%)	78 (93.98%)
CFL	23 (27.71%)	12 (14.46%)	26 (31.33%)
DL	9 (10.84%)	10 (12.05%)	12 (14.45%)
SDM	10 (12.05%)	14 (16.87%)	30 (36.14%)
PLT	4 (4.82%)	4 (4.82%)	7 (8.43%)
PBT	2 (2.41%)	2 (2.41%)	3 (3.61%)
TPT	2 (2.41%)	2 (2.41%)	2 (2.41%)
Fracture	11 (13.25%)	7 (8.43%)	15 (18.07%)

Tables 3 and 4 present the diagnostic performance of US and MRI in relation to surgical outcomes, encompassing sensitivity, specificity, PPV, NPV, accuracy and Kappa values. The consistency analysis between US diagnoses and surgical results revealed statistically significant Kappa values (P < 0.05) for the assessment of ATFL, CFL, DL, PLT, PBT, TPT, and fractures. However, the Kappa value for the assessment of SDM did not reach statistical significance. In contrast, MRI exhibited statistically significant Kappa values (P < 0.05) across all anatomical sites examined.

Table 3.

Diagnostic performance of US for ankle injuries

Site of Injury	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	Accuracy (%)	Kappa value	P value
ATFL	100.00	100.00	100.00	100.00	100.00	1	<0.001
CFL	88.46	92.98	85.19	94.64	91.57	0.806	<0.001
DL	75.00	95.77	75.00	95.77	92.77	0.708	<0.001
SDM	33.33	84.91	55.56	60.00	80.72	0.2	0.053
PLT	57.14	98.68	80.00	96.15	95.18	0.641	<0.001
PBT	66.67	95.00	33.33	98.10	93.98	0.416	<0.001
TPT	100.00	87.65	16.67	100.00	87.95	0.255	<0.001
Fracture	73.33	92.65	68.75	94.03	89.16	0.643	<0.001

Table 4.

Diagnostic performance of MRI for ankle injuries

Site of Injury	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	Accuracy (%)	Kappa value	P value
ATFL	76.92	100.00	100.00	78.26	78.31	0.287	<0.001
CFL	46.15	96.49	85.71	79.71	80.72	0.488	<0.001
DL	83.33	70.42	32.26	96.15	72.29	0.324	<0.001
SDM	46.67	84.91	63.64	73.77	71.08	0.335	0.002
PLT	57.14	84.21	80.00	95.52	83.13	0.261	0.008
PBT	66.67	86.25	15.38	98.57	85.54	0.203	0.013
TPT	100.00	90.12	20.00	100.00	90.36	0.305	0.001
Fracture	46.67	100.00	100.00	89.47	89.16	0.589	<0.001

Comparative diagnostic performance of US and MRI for ankle ligament injuries

Figure 1 illustrates the position of maximum ankle plantar flexion and the placement of the probe during the examination of the ATFL. The US image of a normal ATFL reveals densely arranged fibrous echoes, whereas an ATFL injury is characterized by thickened ligaments exhibiting low echo, a blurred or interrupted fibrous structure (Fig. 2). The normal ATFL on MRI is characterized by as a continuous band of uniform low-signal intensity. In instances of injury, the ligament appears thickened and irregularly shaped, exhibiting increased signal intensity or signs of rupture (Fig. 3), and may be surrounded by edema or effusion.

Fig. 1.

Ultrasound evaluation of the Anterior Talofibular Ligament (ATFL). The position and probe placement when the ankle is maximally plantar flexed

Fig. 2.

Ultrasound image patterns of both normal and abnormal Anterior Talofibular Ligament (ATFL). A The normal ATFL displays densely arranged fibrous echoes (white arrow). B The image of a thickened ATFL with low echo (white arrow), which suggests an injury. F. Fibula; Tal. Talus

Fig. 3.

Axial proton density-weighted MRI imaging of the left ankle reveals the normal and abnormal Anterior Talofibular Ligament (ATFL). A The normal ATFL is a thin and uniform band-like structure, presenting as a uniform low signal (yellow arrow). B Fat suppression imaging shows the course of the ATFL (yellow arrow), displaying a linear high signal that suggests a ligament injury. F. Fibula; Tal. Talus

In the diagnosis of ATFL injuries (Tables 3, 4 and 5), US demonstrated 100% performance across all diagnostic indicators, achieving perfect agreement with surgical findings (κ = 1) and an AUC of 1 (95% confidence interval [CI]: 1.000–1.000, P < 0.001). In contrast, while MRI also showed 100% specificity and PPV for ATFL diagnosis, its sensitivity significantly decreased to 76.92%, with accuracy dropping to 78.31% (κ = 0.287), and an AUC of 0.885 (95% CI: 0.802–0.968, P = 0.004). Figure 4A shows the ROC curve analyses of US and MRI for predicting ATFL injury.

Table 5.

Results of the ROC curve analyses of US and MRI for ankle injuries

Site of Injury	Indicator	AUC	95%CI	P value
ATFL	US	1.000	1.000–1.000	< 0.001
	MR	0.885	0.802–0.968	0.004
CFL	US	0.907	0.826–0.988	< 0.001
	MR	0.713	0.580–0.846	0.002
DL	US	0.854	0.705-1.000	< 0.001
	MR	0.769	0.630–0.908	0.003
SDM	US	0.591	0.460–0.722	0.169
	MR	0.658	0.530–0.786	0.017
PLT	US	0.779	0.547-1.000	0.015
	MR	0.707	0.481–0.933	0.071
PBT	US	0.808	0.484-1.000	0.071
	MR	0.765	0.445-1.000	0.121
TPT	US	0.938	0.870-1.000	0.035
	MR	0.951	0.892-1.000	0.03
Fracture	US	0.830	0.692–0.968	< 0.001
	MR	0.733	0.563–0.903	0.005

Fig. 4.

ROC curve analysis of US and MRI for ankle ligament injuries. A ATFL injury. B CFL injury. C DL injury. D SDM injury

In the diagnosis of CFL injury (Tables 3, 4 and 5), US demonstrated significantly higher sensitivity (88.46% vs. 46.15%), NPV (94.64% vs. 79.71%), and accuracy (91.57% vs. 80.72%) compared to MRI. Although US showed slightly lower specificity (92.98% vs. 96.49%), the PPVs were comparable (85.19% vs. 85.71%). The Kappa value for US was 0.806 (P < 0.001), with an AUC of 0.907 (95% CI: 0.826–0.988, P < 0.001). In contrast, MRI showed moderate agreement (κ = 0.488, P < 0.001), and an AUC of 0.713 (95% CI: 0.580–0.846, P = 0.002). Figure 4B shows the ROC curve analyses of US and MRI for the prediction of CFL injury.

In the diagnosis of DL injury (Tables 3, 4 and 5), US demonstrated significantly higher specificity (95.77% vs. 70.42%), PPV (75.00% vs. 32.26%), and accuracy (92.77% vs. 72.29%) compared to MRI. Conversely, MRI showed higher sensitivity (83.33% vs. 75.00%). The NPVs were comparable between US and MRI (95.77% vs. 96.15%). Furthermore, US exhibited superior interobserver agreement (κ = 0.708, P < 0.001) and diagnostic performance, with an AUC of 0.854 (95% CI: 0.705-1.000, P < 0.001). This performance surpassed that of MRI, which demonstrated an interobserver agreement of κ = 0.324 (P < 0.001) and an AUC of 0.769 (95% CI: 0.630–0.908, P = 0.003). Figure 4C shows the ROC curve analyses of US and MRI for the prediction of DL injury.

In the diagnosis of SDM injury (Tables 3, 4 and 5), US and MRI exhibited significant differences across key diagnostic metrics. Specifically, US demonstrated a significantly lower sensitivity compared to MRI (33.33% vs. 46.67%) while the specificity was identical for both modalities (84.91%). US showed lower PPV (55.56% vs. 63.64%) and NPV (60.00% vs. 73.77%) compared to MRI. Interestingly, the accuracy was marginally higher for US (80.72%) than for MRI (71.08%). However, the diagnostic consistency between US and MRI was low, with Kappa values of 0.335 (P = 0.002) for MRI and 0.200 (P = 0.053) for US. In terms of comprehensive diagnostic performance, MRI exhibited superior discriminative ability with an AUC of 0.658 (95% CI: 0.530–0.786, P = 0.017), whereas US’s AUC was 0.591 (95% CI: 0.406–0.722, P = 0.169). Figure 4D shows the ROC curve analyses of US and MRI for the prediction of SDM injury.

Comparative diagnostic performance of US and MRI for ankle tendon injuries

Figure 5 illustrates the positioning and probe placement for the foot and ankle in an inverted position with mild plantar flexion during the examination of the PLT. US imaging demonstrates the normal PLT as a fibrous cord exhibiting a uniform hyperechoic appearance and regular morphology, with no peritendinous effusion observed. In instances of injury, the tendon may exhibit signs of thickening (Fig. 6). MRI may reveal abnormal signal intensity within the affected region (Fig. 7).

Fig. 5.

Ultrasound evaluation of the Peroneus Longus Tendon (PLT). The position and probe placement for the foot and ankle in an inverted position with mild plantar flexion

Fig. 6.

Ultrasound image patterns of both normal and abnormal Peroneus Longus Tendons (PLT). A The long axis of the normal PLT exhibits a uniform fibrous high echogenic pattern (white arrow). B The tendon sheath of the PLT is observed to be thickened and displays low echogenicity (yellow star)

Fig. 7.

Axial proton density-weighted MRI imaging of the left ankle reveals the normal and abnormal Peroneus Longus Tendon (PLT). A The normal PLT shows uniform low signal intensity (yellow arrow). B Fat suppression imaging shows a high signal shadow around the PLT (yellow arrow). Tal. Talus; F. Fibula; PL. Peroneus longus tendon; PB. Peroneus brevis tendon

In the diagnosis of PLT injury (Tables 3, 4 and 5), US and MRI exhibit notable differences in key diagnostic indicators. US demonstrated significantly higher specificity (98.68% vs. 84.21%), NPV (96.15% vs. 95.52%), and accuracy (95.18% vs. 83.13%) compared to MRI. Both modalities achieved identical sensitivity (57.14%) and PPV (80%). The interobserver agreement for US was significantly superior (κ = 0.641, P < 0.001), whereas the interobserver agreement for MRI was low (κ = 0.261, P = 0.008). In terms of overall diagnostic performance, US achieved a higher AUC of 0.779 (95% CI: 0.547-1.000, P = 0.015) compared to MRI’s AUC of 0.707 (95% CI: 0.481–0.933, P = 0.071), indicating a trend toward significance. Figure 8A shows the ROC curve analyses of US and MRI for the prediction of PLT injury.

Fig. 8.

ROC curve analysis of US and MRI for ankle tendon injuries. A PLT injury. B PBT injury. C TPT injury

In the diagnosis of PBT injury (Tables 3, 4 and 5), both US and MRI showed identical sensitivity (66.67%). However, US exhibited significantly higher specificity (95.00% vs. 86.25%) and accuracy (93.98% vs. 85.54%) compared to MRI. Notably, US showed a higher PPV of 33.33% compared to MRI’s 15.38%. Additionally, US demonstrated moderate interobserver agreement, with a kappa statistic of 0.416 (P < 0.001), while MRI exhibited a low Kappa value of 0.203 (P = 0.013). Although US showed a marginally higher AUC of 0.808 (95%CI: 0.484-1.000, P = 0.071) compared to MRI’s AUC of 0.765 (95% CI: 0.445-1.000, P = 0.121), neither result reached statistical significance at the α = 0.05 threshold. Figure 8B shows the ROC curve analyses of US and MRI for the prediction of PBT injury.

In the diagnosis of TPT injury (Tables 3, 4 and 5), both US and MRI exhibited high sensitivity (100%) and NPV (100%), although they demonstrated low PPV (US: 16.67% vs. MRI: 20.00%). US showed slightly lower specificity (87.65%) compared to MRI (90.12%), while the accuracy was comparable (US: 87.95% vs. MRI: 90.36%). The interobserver agreement was low for both modalities, with a Kappa value of 0.255 for US and 0.305 for MRI, both of which were below 0.4 (both P < 0.05). US achieved an AUC of 0.938 (95% CI: 0.870-1.000, P = 0.035), while MRI yielded an AUC of 0.951 (95% CI: 0.892-1.000, P = 0.03), both indicating excellent diagnostic value (AUC > 0.9). Figure 8C shows the ROC curve analyses of US and MRI for prediction of TPT injury.

Comparative diagnostic performance of US and MRI for fracture

Figure 9 presents an US image of the distal fibula. The normal cortical bone of the distal fibula displays a smooth, continuous hyperechoic pattern; conversely, in the presence of a fracture, the cortical bone exhibits an abnormal discontinuity, as illustrated. MRI of the distal fibula shows that the normal cortical bone appears as a continuous, smooth, and uniformly hypointense signal. In contrast, when a fracture occurs, the continuity of the cortical bone is compromised (Fig. 10). In fracture diagnosis (Tables 3, 4 and 5), US and MRI demonstrated complementary diagnostic profiles. US exhibited a significantly higher sensitivity compared to MRI (73.33% vs. 46.67%), making it more effective for initial fracture detection. Conversely, MRI achieved 100% specificity and 100% PPV, ensuring that there are no false positives among positive diagnoses. Both modalities presented high NPV (US: 94.03% vs. MRI: 89.47%) and identical accuracy (89.16%). In terms of diagnostic reliability, substantial interobserver agreement was noted for both techniques (US: κ = 0.643, P < 0.001; MRI: κ = 0.589, P < 0.001). A comprehensive analysis of diagnostic performance revealed a higher AUC for US (0.830, 95% CI: 0.692–0.968, P < 0.001) compared to MRI (0.733, 95% CI: 0.563–0.903, P = 0.005), suggesting a superior overall diagnostic value for US. Figure 11 shows the ROC curve analyses of US and MRI for the prediction of fracture.

Fig. 9.

Ultrasound image patterns of both normal and fractured distal fibula. A The distal fibula shows a continuous, smooth and dense linear structure with high echogenicity. B Partial interruption of the cortical bone at the distal fibula (white arrow) accompanied by surrounding soft tissue edema (yellow star). F. Fibula

Fig. 10.

Axial proton density-weighted MRI imaging of the right ankle reveals the normal and fractured distal fibula. A The distal fibula exhibits a normal shape, characterized by a continuous and smooth bone cortex, which presents as a uniform low signal intensity. B Fat suppression imaging indicates a high-signal linear shadow in the distal fibula, highlighting a discontinuity in the continuity of the bone cortex (yellow arrow). Tal. Talus; F. Fibula

Fig. 11.

ROC curve analysis of US and MRI for fracture

Discussion

This comparative study of US and MRI in 83 patients with ankle injuries demonstrates the complementary roles of both imaging modalities in evaluating injuries to various anatomical structures. Notably, US demonstrates clear superiority in detecting injuries to superficial ligaments, such as ATFL and CFL, as well as tendons, including PLT and PBT. In contrast, MRI provides greater value in the assessment of deep structures, such as SDM and TPT.

This study demonstrates that the ATFL is the most commonly injured site in ankle injuries. US exhibits near-perfect diagnostic performance for ATFL injuries, achieving 100% sensitivity, specificity, PPV, NPV, and accuracy, which aligns fully with surgical findings. These results are consistent with recent research. A meta-analysis encompassing eight studies [13] involving 434 patients with acute ankle injuries indicates a high concordance between US and MRI in diagnosing acute lateral ankle ligament injuries. In the context of ATFL injuries, this analysis revealed that US has a sensitivity of 0.97 (95% CI: 0.89–0.99) and a specificity of 0.93 (95% CI: 0.84–0.97). Ergün et al. [14] further corroborated these findings, reporting a sensitivity of 95.4% for US in the detection of acute ATFL injuries. Collectively, this body of evidence supports the use of US as the first-line imaging modality for the evaluation of acute ATFL injuries.

In the assessment of CFL injury, this study showed a significant advantage of US over MRI. US exhibited a higher sensitivity (88.46% vs. 46.15%), a greater NPV (94.64% vs. 79.71%), and superior overall diagnostic accuracy (AUC: US 0.907 vs. MRI 0.713). Kappa value analysis further confirmed that US diagnostic results were highly reliable, showing excellent agreement with surgical findings (κ = 0.806), while MRI demonstrated only moderate agreement (κ = 0.488). These findings may be attributed to the unique anatomical structure of the CFL. Originating from the distal fibula and coursing obliquely to insert into the peroneal tubercle of the calcaneus, this orientation renders the CFL susceptible to partial volume effects in conventional MRI imaging planes [15]. In contrast, US dynamically overcomes this limitation through real-time scanning. Notably, employing an oblique coronal scanning plane can enhance the sensitivity and specificity of MRI [16]. Therefore, US serves as an efficient screening tool for CFL injury, particularly valuable for ruling out pathology. When combined with MRI using optimized scanning planes, these modalities provide more comprehensive diagnostic information.

In this study, US demonstrated high specificity, PPV, and interobserver consistency in diagnosing deep DL injuries. Consistent with recent studies, US is recognized as an effective assessment tool for DL injuries [17]. However, MRI exhibited greater sensitivity to deep ligament injuries, thereby enhancing its diagnostic capability. In the context of SDM injuries, MRI presented significant advantages over US, including higher sensitivity, PPV, NPV, stronger interobserver agreement, and a significantly greater AUC. The deep anatomical location of SDM injuries renders US vulnerable to acoustic shadowing and anatomical constraints. In contrast, the multiplanar imaging capability and high soft tissue contrast resolution of MRI endows it with high sensitivity for distinguishing normal from pathological tissues, thereby enabling more reliable detection of pathological alterations in deep ankle structures [18]. Furthermore, adherence to the ESSR guidelines reduce operator dependency and enhance the reproducibility of results. Therefore, the integration of both modalities in clinical practice optimizes diagnostic efficacy: US is suitable for initial screening and dynamic assessment, while MRI is more valuable for the definitive diagnosis of complex cases and surgical planning.

In the diagnosis of ankle tendon injuries, the diagnostic value of US and MRI varies due to their technical characteristics and the anatomical positioning of the tendons. For PLT injuries, US demonstrated exceptional diagnostic value owing to its high spatial resolution, dynamic real-time imaging capability, and reduced partial volume effect. Previous systematic reviews have confirmed its high sensitivity (94%) and specificity (94%) in detecting PLT tear, establishing it as a preferred diagnostic tool [19]. In the case of PBT injuries, the diagnostic performance of US and MRI exhibits complementarity. US is an effective initial screening tool for PBT injuries, boasting a high specificity of 95.00%. In contrast, MRI is a reliable method for exclusion, given its exceptional NPV of 98.57%. Nevertheless, both modalities exhibit limited sensitivity at 66.67%, which underscores the necessity for a tiered diagnostic approach that remains vigilant against the risks of missed diagnoses and false-positive results. Regarding the TPT, both US and MRI show high sensitivity and NPV; however, their relatively lower PPV indicates limitations in diagnostic specificity. Case reports suggest that significant inflammation around the TPT can interfere with MRI signals, leading to false-positive diagnoses of TPT injuries [20]. Therefore, the selection of clinical imaging methods should integrate anatomical characteristics while leveraging the distinct advantages of different imaging technologies.

In terms of fracture diagnosis, the results of this study suggest that US and MRI technologies have complementary value. US exhibited a high sensitivity of 73.33%, making it suitable for rapid injury screening. With 100% specificity and PPV, MRI demonstrated its ability to accurately identify bone marrow edema and surrounding soft tissue injuries. Sadineni et al. pointed out that MRI can effectively monitor occult bone injuries, clearly illustrating the extent of injury and soft tissue involvement [21]. Although the Kappa values (US: 0.643, MRI: 0.589) indicated moderate to substantial consistency and the NPV was high (US: 94.03%, MRI: 89.47%), the AUC for US (0.830 vs. 0.733) further reinforces its advantages as a primary screening tool. Consistent with previous studies, US demonstrated both timeliness and sensitivity in acute ankle trauma [22], while MRI is more suitable for further evaluation and diagnosis of complex cases. In summary, the complementary efficacy of US and MRI provides an optimized pathway from initial screening to diagnosis in clinical practice.

This study confirms the advantages of US in the assessment of ankle injuries. The dynamic real-time imaging capability of US enhances its convenience in evaluating patients’ dynamic injuries and provides significant reference value for injuries such as the ATFL and CFL [4, 23]. Additionally, the affordability of US positions it as a preferred and reliable tool for emergency and primary care settings. However, US does have inherent limitations. Its diagnostic efficacy is compromised in the evaluation of SDM injuries due to restricted penetration depth and limitations in the acoustic window. Although the performance of MRI in diagnosing certain ligaments in this study is inferior to that of US, MRI possesses irreplaceable advantages, including superior performance in the global assessment of complex trauma and a comprehensive understanding of the anatomical details of the injury site prior to surgery. Nonetheless, MRI examinations also have drawbacks, such as prolonged examination times and patient discomfort. Patients with acute injuries often experience motion artifacts due to pain, which hinders their ability to maintain the necessary immobilized position during imaging. Consequently, clinicians must carefully select the appropriate imaging modality based on the suspected type of injury.

Limitations

This study has several limitations that must be acknowledged objectively. Firstly, this study analyzed diagnostic performance on a per-injury-type basis. However, we did not systematically assess the conditions of concomitant pathologies. This limits our ability to characterize the clinical injury complexes for which each imaging modality is indicated. Secondly, as all enrolled subjects were surgical patients, the absence of data on non-surgically managed ankle injuries hinders the assessment of the diagnostic performance of both imaging modalities for mild injuries. Thirdly, the relatively small sample sizes for injury types other than the ATFL and CFL could affect the statistical reliability of those findings.

Future research should focus on prospective, multi-center studies with larger cohorts to strengthen the evidence for US and MRI in diagnosing ankle injuries. A key goal should be to compare their accuracy for isolated versus concomitant injuries. From a technical perspective, the development of intelligent US image analysis systems could help reduce operator dependency. It is essential to prioritize the implementation of individualized comprehensive imaging assessment strategies based on anatomical localization and pathological characteristics to enhance diagnostic and therapeutic precision in the management of ankle injuries.

Conclusion

This study confirms that US demonstrates superior performance in diagnosing superficial ligament injuries, specifically the ATFL and the CFL. In contrast, MRI is indispensable for evaluating deeper complex structures, such as the SDM and fracture lesions. These two modalities exhibit a complementary relationship. In clinical practice, their application should be optimized based on the characteristics of the injury, and novel assessment approaches should be actively explored to achieve precise and efficient diagnosis and treatment of ankle injuries.

Authors’ contributions

WL: Writing - original draft, Acquisition, analysis, Visualization, Methodology, Investigation, Data curation. YHZ: Investigation, Data curation, Formal analysis, Methodology. YX: Writing - review & editing, Visualization, Supervision, Validation. YZ: Writing - review & editing, Project administration, Supervision, Validation.

Funding

Not applicable.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study protocol was approved by the Ethics Review Committee of the 960th Hospital of the PLA Joint Logistics Support Force (approval number: 2025 − 152) and was conducted in accordance with the ethical principles of the Declaration of Helsinki. Informed consent was waived by the same Committee as this was a retrospective study of existing medical records.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.