Utility of Zero Echo Time MRI for the Diagnosis and Characterization of Ankle Fractures

Abstract

Background: Zero echo time (ZTE) imaging is a relatively new magnetic resonance (MR) pulse sequence that provides bone-soft tissue contrast similar to that of computed tomography (CT). Purpose: We sought to (1) determine the accuracy of ZTE MRI for the diagnosis of common ankle fractures and (2) investigate whether ZTE imaging sequences are equivalent to the gold standard of CT for the characterization of fracture fragments. Methods: We conducted a prospective case series of 54 patients with acute ankle trauma, in whom ZTE MRI was performed, followed by surgical reduction. Fractures on the ZTE sequence were correlated with the operative report as the reference standard. Raw agreement (%) and correlation (κ) were calculated. Selected fracture fragments were measured in 2 dimensions (anterior-posterior and superior-inferior) on corresponding sagittal ZTE and CT images by 3 independent radiologists to determine reliability. Results: The ZTE sequence demonstrated 47 distal fibular, 17 medial malleolar, 24 posterior malleolar, 5 anterior talofibular ligament avulsion, and 4 distal tibial fractures on the 54 cases. Raw agreement with operative findings was 95% (range: 86%-100%) and correlation almost perfect (0.960 [0.926-0.995]). Fragment characterization was accurate and repeatable. Intraobserver and interobserver agreement was excellent. Conclusions: Our case series suggests that the use of the MRI ZTE sequence may provide images with CT-like contrast for characterizing acute ankle fractures.

Introduction

Ankle fractures represent 10% of all fractures and increase in incidence with age [15]. Magnetic resonance imaging (MRI) is the study of choice to evaluate injury to soft tissues, such as tendons, ligaments, and cartilage. On conventional magnetic resonance (MR) pulse sequences, all such structures are hypointense due to short T2 relaxation times. Information obtained indirectly, usually by the presence of fluid insinuating into a defect or tear, is visualized as contrasting signal hyperintensity. Small or subtle fracture fragments may be inconspicuous on MRI, so computed tomography (CT) is often used for fracture characterization due to its superior spatial resolution and contrast between bone and soft tissue [13].

Ultrashort time to echo (UTE) and zero echo time (ZTE) MRI sequences allow the visualization of structures with very rapid intrinsic T2 relaxation (≤10 ms), such as cortical bone [14], cartilage [3,4], and tendon [8,9,11]. ZTE imaging generally “flattens” the contrast for soft tissues while maintaining high contrast between bone and soft tissue, providing CT-like contrast when intensity is inverted [16]. It is now being used to study the musculoskeletal system, as recent innovations have decreased acquisition times and improved image contrast enough to be clinically useful [8]. Therefore, short time to echo sequences have the potential to characterize ankle injuries with adequate ligament and bony detail without the need for additional cross-sectional imaging and ionizing radiation exposures.

We therefore sought to determine whether fractures diagnosed with ZTE MRI would correlate with operative findings and whether acute ankle fracture characterization with ZTE MRI is equivalent to that of CT.

Methods

Institutional Review Board approval was obtained for this prospective case series. Patients were excluded if they had a history of ankle surgery, had contraindications to MRI, or were referred for nontraumatic indications. A total of 75 patients referred for ankle MRI in the setting of acute trauma between January 2016 and November 2022 provided informed consent for an additional ZTE sequence to be added to their standard-of-care MRI. Of these, 21 patients were excluded due to nonoperative management, leaving 54 cases, 27 men and 27 women, with an average age of 45.5 (range: 13–81) years. A total of 22 right and 32 left ankles were scanned. The average time between the MRI and surgery was 1.8 (SD ± 1.8) days. Preoperative CT was available for correlation in 6 cases. Time between the CT and MRI averaged 12 days (range: 0–53 days).

Ankle MRI was performed with an 8-channel foot/ankle coil (InVivo) in a 3.0T GE Discovery MR750 scanner. Our institution’s standard ankle MRI protocol was performed with a sagittal inversion recovery, as well as axial, sagittal, and coronal fast-spin echo pulse sequences with parameters as follows: 13 to 16 cm field of view, repetition time (TR) 4000 to 5000 ms, echo time (TE) 20 to 28 ms (effective), echo train length (ETL) 10 to 14, receiver bandwidth 50 to 62.5 kHz, number of excitations (NEX) 1 to 2, slice thickness 3.5 to 4 mm, and no interslice gap. The inversion time (TI) for the inversion recovery (IR) sequence was 190 ms. A vendor-supported (GE) prototype ZTE pulse sequence was used to acquire parasagittal images using frequency of 320 kHz, ±62.5 kHz bandwidth, 4 NEX averaging, and 256 spokes per segment. Field of view varied with patient size from 20 to 26 cm, and a slice thickness ranging from 1.0 to 1.6 mm was chosen to optimize coverage and scan time for each patient. Frequency encoding was performed in the superior-inferior direction.

To score each case, the ZTE sequence was prospectively reviewed by an experienced musculoskeletal radiologist blinded to other conventional sequences, the radiologists’ original report, and the operative report for the presence or absence of ankle fracture types the surgeon (DH) considered to be clinically significant. These consisted of lateral malleolar, medial malleolar, posterior malleolar, anterior talofibular avulsion, and other tibial fractures (distal shaft, pilon, etc). Comminuted fractures were classified as a single fracture.

For the 6 patients who also had CT scans of the ankle, 3 musculoskeletal radiologists independently measured the size of selected fracture fragments in 2 dimensions (anterior-posterior and superior-inferior) on corresponding sagittal ZTE and CT images. There was a 2-week interval between measurements for each modality.

Statistical Analysis

The operative report for each case was used as the reference standard. Fractures diagnosed on ZTE MRI were correlated with operative confirmation by retrospectively reviewing the operative report for the presence of fracture fixation on postoperative radiographs. Raw agreement (%) and the correlation coefficient (κ) with 95% confidence intervals were calculated [12].

Inter-rater reliability of fragment measurements on ZTE with CT was calculated with interclass correlation coefficients with 95% confidence intervals. Interpretation of correlation is as follows: <0.4, poor; 0.40 to 0.59, fair; 0.60 to 0.74, good; 0.75 to 1.00, excellent.

Results

In our 54 surgically managed acute ankle trauma patients scanned with the ZTE imaging sequence, a total of 97 different ankle fractures were diagnosed. These included 47 distal fibular, 24 posterior malleolar, 17 medial malleolar, 5 anterior talofibular ligament avulsion, and 4 distal tibial fractures. Anterior talofibular ligament avulsion fractures were scored separately from all other distal fibular fractures as they were often fixated with additional hardware. Three posterior malleolar fractures were missed on MR with ZTE sequence; these were found and repaired intraoperatively. One medial and 1 posterior malleolus fracture diagnosed on ZTE MRI were considered false positives as they were not surgically confirmed; the ankle was stable to stress in the operating room. The ZTE sequence alone provided the correct fracture diagnosis with 95% overall raw agreement and almost perfect correlation with the operative findings (Table 1). An ankle fracture as depicted with ZTE is seen in Fig. 1.

Table 1.

Agreement of MRI with a ZTE sequence with operative findings.

	ZTE +/OR +	ZTE +/ OR −	ZTE − / OR +	Raw agreement, %	Κ [95% CI]
Lateral malleolus	47	0	0	100	1.000
Medial malleolus	17	1	0	94	0.852 [0.713–0.991]
Posterior malleolus	24	1	3	86	0.958 [0.877–1]
ATFL avulsion	5	0	0	100	1.000
Distal tibia	4	0	0	100	1.000
Total	97	2	3	95	0.960 [0.926–0.995]

MRI magnetic resonance imaging, ZTE zero echo time, CI confidence interval, ATFL anterior talofibular ligament , OR odds ratio.

Fig 1.

A 15-year-old boy presented with pain 2 weeks after a wrestling injury. The sagittal proton density image (a) demonstrates a subtle posterior malleolus fracture (arrow). The sagittal ZTE sequence (b) renders the fracture more conspicuous (arrow). ZTE zero echo time.

Concurrently obtained CT was available for comparison in 6 cases (Fig. 2). Fragment measurements made on ZTE MRI and CT were not significantly different for any of the 3 radiologists (P > .597). There was excellent inter-rater agreement; measurements of the same fracture fragments by 3 radiologists were highly correlated both for CT and ZTE images (Table 2). There was excellent intraobserver agreement; each radiologist’s CT and ZTE measurements were highly correlated (Table 3). In addition to fracture size, other qualitative information about the fractures was well demonstrated on ZTE MRI, such as differentiation between cartilage and subchondral bone to better detect depression and gaps at the articular surface (Fig. 3).

Fig. 2.

A 38-year-old woman presented with ankle pain after a fall 2 weeks prior to imaging. A sagittal fast-spin echo image (a) demonstrates an oblique fracture of the distal fibula and a tear of the anterior tibiofibular ligament. The ZTE sequence (b) better delineates the fracture and allows differentiation between the ligament fibers and the cortical avulsion fragment. A sagittal CT image from the same patient (c), also demonstrating the avulsion fragment, is shown for comparison. Red arrows, distal fibular fracture; solid yellow arrow, anterior tibiofibular ligament tear; dashed yellow arrow, fibular avulsion fragment. ZTE zero echo time, CT computed tomography.

Table 2.

Inter-rater reliability.

Modality	N	ICC	95% confidence interval
ZTE	16	0.995	0.986-0.998
CT	16	0.997	0.991-0.999

Correlation of measurements from the same modality between 3 different radiologists.

ICC interclass correlation coefficient, ZTE zero echo time, CT computed tomography.

Table 3.

Intra-rater reliability.

Radiologist	N	ICC	95% confidence interval
Rater 1	16	0.995	0.985-0.998
Rater 2	16	0.999	0.998-1.000
Rater 3	16	0.999	0.994-1.000

Correlation of measurements from sagittal ZTE and corresponding sagittal CT images made by 3 different radiologists.

ICC interclass correlation coefficient, ZTE zero echo time, CT computed tomography.

Fig. 3.

A 63-year-old woman presented for evaluation after a chair fell on her left foot 2 days prior to imaging. A sagittal fast-spin echo image (a) demonstrates a mildly comminuted fracture of the posterior malleolus (red arrow). The corresponding ZTE sequence (b) better delineates the fragment (red arrow), without signal from the adjacent cartilage, revealing mild subchondral depression at the articular surface (yellow arrow). ZTE zero echo time.

Discussion

For orthopedic surgeons treating ankle injuries, accurate characterization of fractures is essential. Fracture size and configuration as well as articular congruency are important factors to consider when determining whether surgery is needed, the correct approach, and predicting outcomes [10]. Our study sought to determine whether the short time to echo ZTE sequence could diagnose ankle fractures, using operative findings and CT as reference standards. Fractures shown on the ZTE sequence had high agreement with the operative findings, with 95% raw agreement and almost perfect correlation (κ 0.960 [0.926–0.995]) overall. Fracture fragment characterization was equivalent to CT, showing excellent intra-rater and inter-rater reliability.

Our study has limitations. Surgically explored fractures may have been correctly diagnosed with MRI, but if they were not operatively recognized, they would have been counted as a false-positive interpretation. In addition, since only ZTE was analyzed, it is likely many of the fractures would have been visible on conventional MRI alone. The CT was available for comparison in a small minority of cases (6 of 54), precluding its use as an imaging standard. As reported by our referring physicians, CT scans were often deemed unnecessary as the high-contrast bone images provided by the ZTE sequence were adequate for fracture diagnosis and surgical planning.

Other groups have recently validated short time to echo MR pulse sequences as a useful technique for fracture evaluation. The MRI with UTE had a sensitivity of 93.3% and accuracy of 98.7% in the diagnosis of pediatric skull fractures, using CT as the gold standard [18]. It is as accurate for fracture characterization as CT, displaying similar information about fragment depression and displacement [4,6]. The ZTE MRI has also delineated osseous detail similar to CT in the shoulder [6], hip [7], knee [5], cervical spine [1], and sacroiliac joints [17]. It has the added advantage of avoiding ionizing radiation and providing information about the supporting ligaments, tendons, and articular cartilage [2]. This sequence is especially beneficial to pediatric and pregnant patients, who are more sensitive to the effects of ionizing radiation.

In conclusion, the findings of this case series suggest that ZTE imaging may be a useful technique for diagnosing and characterizing fractures in the setting of ankle trauma. The addition of ZTE to standard MRI pulse sequences provides a comprehensive evaluation of the injured ankle and may obviate the need for CT in many cases, avoiding its attendant ionizing radiation and additional cost.