Changes in magnetic resonance imaging signal intensity in predicting complications during acute phase osteoporotic vertebral fractures

Abstract

Purpose

To demonstrate prediction of complications in osteoporotic vertebral fractures with magnetic resonance imaging (MRI) changes over time.

Methods

MRI signal intensities in osteoporotic vertebral fractures were investigated according to the interval between onset and imaging as follows: 0–10 days (early), 11–20 days (middle), and 21–30 days (later).

Results

The diffuse low pattern rates were 52%, 84%, and 95% and 20%, 24%, and 52% in the early, middle, and later periods on T1-WI and T2-WI, respectively.

Conclusions

The diffuse low pattern increased with time. The MRI prediction of complications depends on the time phase.

1. Introduction

Most osteoporotic vertebral fractures are usually treated conservatively and have good outcomes.¹ However, there are some cases in which conservative treatment fails, such as pseudarthrosis, spinal deformity, and spinal cord compression. If the risk of complications can be predicted at an early stage, minimally invasive surgical interventions such as kyphoplasty might be considered for alleviating pain and avoiding future complications.² There are several reports on detecting the risk factors of complications of conservative treatment, and magnetic resonance imaging (MRI) is one of the tools for predicting the prognosis.3, 4, 5, 6, 7, 8, 9, 10 However, these reports have not clearly defined the time of imaging. Osteoporotic vertebral fractures gradually collapse with time; therefore, the shape of the fracture and signal intensity on MRI may change correspondingly. Moreover, the prediction of prognosis by imaging may partly vary with the time of imaging. In many cases, MRI for detecting vertebral fracture is taken at the early stage after injury. It is important to know the natural course during the acute phase of a vertebral fracture for prognosis prediction using MRI signals.

Lateral dynamic radiography, a method of measuring vertebral instability, is also reported as a tool for prognosis prediction.⁸ The relation between the prognosis prediction by MRI signals and lateral dynamic radiography has not been elucidated. If the prediction by MRI signals is reflecting the same vertebral condition as that revealed by radiography, taking MRI to predict complications will not have much value.

We herein investigated the MRI signal changes of vertebral fractures during the acute phase to validate the universality of the prognosis prediction. In addition, we compared MRI signals and lateral dynamic radiography to investigate the relation between them.

2. Materials and methods

We retrospectively investigated the signal intensity of osteoporotic vertebral fractures in 100 consecutive patients who underwent MRI imaging studies within 1 month after onset in our institution between 2014 and 2018. This study was approved by the Kyushu University institutional review board (IRB approval no. 28–324). All patients were scanned with a 1.5-T MRI system (SIEMENS MAGNETOM ESSENZA): T1-WI with a slice thickness of 4 mm (repetition time [TR], 400–700 ms and time to echo, [TE] 13–14 ms), the same sequence with T2-WI (TR, 3500–4500 ms and TE, 85–110 ms), and short T1 inversion recovery [STIR] (TR, 4700–6000 ms and TE, 60–80 ms, and inversion time 170–180 ms). Written informed consent was not needed due to the retrospective nature of the study. Patients with pathological fractures, presence of neurological symptoms, or fractures requiring surgical intervention at the first diagnosis were excluded. Patients in whom the time of onset was unclear were also excluded from this study.

The patients were divided into three groups according to the interval between onset and imaging as follows: 0–10 days (early), 11–20 days (middle), and 21–30 days (later). The signal intensity of MRI followed standard classification: confined or diffuse low on T1-WI and normal, confined low, diffuse low, or confined high on T2-WI. Normal intensity referred to the signal intensity change detected on STIR images. The definition of MRI signal intensity in each group is shown in Fig. 1. Among the 20 patients who had MRIs twice within a 1-month-interval, the change of signal intensity was also investigated. Diffuse low was classified if at least two of three slices showed more than half low signal change. The vertebrae showing the diffuse low pattern containing a part of the high signal were classified as having a diffuse low pattern. Two spine surgeons who were blinded to the imaging times analyzed and classified the findings on the MRIs. The final assessment was made by based on a consensus view of all involved physicians.

Fig. 1.

Definition of magnetic resonance imaging (MRI) signals

The sites of fracture are shown by a white arrow.

To demonstrate the relationship between MRI signals and the condition of the fractured vertebra, we also investigated the degree of spinal instability in the patients who underwent radiography at the time of MRI scanning. Spinal instability was calculated by the changes in the angles of the endplates of fractured vertebra between the sitting and supine positions by X-rays. This X-ray study included 66 patients, and 34 patients were excluded because the data were lacking.

2.1. Statistical analysis

Categorical variables were compared by either a chi-square test or one-way analysis of variance (ANOVA) test. The TUKEY type post hoc analysis was performed to find the significance of the degree of difference when the difference between the groups was statistically significant. The statistical results were considered significant at p < 0.05. All statistical analyses were carried out using JMP pro 13 software (SAS Institute, Cary, NC).

3. Results

The details of the patients in the three groups are shown in Table 1. There were no significant differences concerning age, sex, and the site of injury. Five patients in the early group, two patients in the middle group, and one patient in the later group underwent balloon kyphoplasty (BKP) for persistent pain and vertebral instability. The rates of confined low signal intensity were 48% (27/56), 16% (4/25), and 5% (1/19), and those of diffuse low-signal intensity were 52% (29/56), 84% (21/25), and 95% (18/19) on T1-WI (early, middle, and later periods, respectively). The rates of the signal intensities on T2-WI in the early, middle, and later periods, respectively, were as follows: normal, 20% (11/56), 0% (0/25), and 0% (0/19); confined low, 39% (22/56), 64% (16/25), and 42% (8/19); confined high, 21% (12/56), 12% (3/25), and 5% (1/19); and diffuse low, 20% (11/56) including one patient with a high lesion, 24% (6/25) including three patients with high lesions, and 52% (10/19) including two patients with a high lesion. Diffuse low pattern occurrence increased with time on both T1-WI and T2-WI images (Fig. 2). MRI scans taken twice within 1 month were compared. The first imaging was performed at 11.0 ± 8.5 days and the second imaging was performed 30.3 ± 11.7 days after onset. Confined low signals were seen in 9 patients and diffuse low signals in 11 patients at the first MRI, whereas the confined low signal was noted in 1 patient and the diffuse low signal in 19 patients at the second MRI on T1-WI. A normal signal was noted in 4 and 0 patients, confined low in 9 and 6 patients, confined high in 5 and 2 patients, and diffuse low in 2 and 12 patients including 5 patients with high lesions, in the first and second MRIs on T2-WI, respectively. The proportion of diffuse low signals increased with time on both T1-WI and T2-WI images (Fig. 3, Fig. 4).

Table 1.

Characteristics of patients in each group.

	Early group (N = 56)	Middle group (N = 25)	Later group (N = 19)	P value
Age	77.7 ± 8.0	79.7 ± 7.0	82.1 ± 7.1	N·S.
Sex	M17,F39	M8,F17	M7,F12	0.87
Level	TL 31, Others 25	TL 11, Others 14	TL10, Others 9	0.64
Outcome	BKP 5	BKP 1	BKP 1

TL = thoracolumbar BKP = balloon kyphoplasty ± = standard deviation.

Fig. 2.

The rate of magnetic resonance imaging (MRI) signal intensity in each group on T1-WI and T2-WI

The rate of diffuse low signals increases with time on both T1-WI and T2-WI.

Fig. 3.

Magnetic resonance imaging (MRI) signals in two patients who underwent imaging twice within 1 month

Above (A) and below (B) images are the first lumbar fractures in the patients. The sites of fracture are shown by a white arrow. The above images (A) are taken 9 days (first) and 23 days (second) after onset. Below images (B) are taken one day (first) and 21 days (second) after onset. The MRI signals of both T1-WI and T2-WI become diffuse low with time.

Fig. 4.

Changes in magnetic resonance imaging (MRI) signals between the first and second MRI on T1-WI and T2-WI

The number of patients with MRI signal changes at the first and second MRI is shown in a table (left) and a bar graph (right). The rate of diffuse low signals increases with time on both T1-WI and T2-WI.

The degree of spinal instability was 1.5 ± 1.4° in normal (5 patients), 3.6 ± 3.1° in confined low (32 patients), 10.8 ± 5.3° in confined high (10 patients), and 9.4 ± 5.1° in diffuse low (19 patients) signals on T2-WI (Fig. 5). The confined high or diffuse low signals showed more spinal instability compared to normal and confined low signals (P < 0.01).

Fig. 5.

Spinal instability in each group

Confined high and diffuse low signals show more spinal instability compared to normal and confined low signals. * = P < 0.05.

4. Discussion

Several risk factors of osteoporotic vertebral fracture complications have been reported. The fracture site, fracture shape, and MRI signal intensity were reportedly related to complications of vertebral fractures. High or diffuse low signals on T2-WI are risk factors for the delayed union, possibly caused by bleeding and accumulation of fluid and the wide range of destruction of bony trabecula.^3,4,7 MRI signals are often used for the prediction of complications of vertebral fractures; however, MRI signals that show the condition of the vertebra are believed to vary according to the imaging time. In this study, we showed that the MRI signals of the fractured vertebra were different based on the time after onset.

Predicting the prognosis of osteoporotic vertebral fracture has recently come into focus because minimally invasive surgeries such as kyphoplasty have been reported to provide better clinical outcomes than conservative treatment.^11,12 Some studies suggested that early intervention could be an option in the patients with risk factors of complications.² Early intervention appears feasible if the prognosis of vertebral fractures could be predicted because highly invasive surgery is needed in the case of pseudarthrosis and neurological deficits. Among several risk factors, MRI signal intensity may be an easy-to-use tool as several institutions have reported the reliability of predicting the prognosis with this method. However, according to Takahashi et al., the problem of prediction by MRI was that the imaging times were unclear, and they reported that the sensitivity and specificity of the delayed union in T2-W1 (combined diffuse low or high) were 70.0% and 83.7%, respectively, within 2 weeks after onset; by 1 month after enrollment, sensitivity and specificity were 93.1% and 71.2%, respectively [9]. They suggested the need for sequential MRIs at two different time points but did not mention details on the causes of the changes in reliability with time. When we grouped the patients by the time after onset, the rate of diffuse low signal, the worse outcome predictor of the fractured vertebra, tended to increase with time. The increasing rate of diffuse low signal intensity is thought to contribute to the difference in sensitivity and specificity with time by the prediction of MRI signal intensity change.

The delayed union that is more evident in confined high or diffuse low signals on T2-WI is natural because the signals are showing the poorer condition of the fractured vertebra. When we investigated the spinal instability related to each MRI signal, the instability was more severe in high or diffuse low signals on T2-WI. The MRI signals during the acute phases are thought to reflect the condition of the fractured vertebra and spinal instability at the time. Therefore, even in the cases of vertebral fractures showing good clinical outcomes on MRI at the early phase after injury, follow-up of patients is necessary because some osteoporotic fractured vertebrae worsen and spinal instability develops with time.

Our study had a few limitations. There was a variation in the number of patients in each group, and the number in the later group was few. In addition, we could not show the outcome of each group and the reliability of the MRI signal intensities because we did not conduct sufficient follow-ups. We cannot deny the possibility that the diffuse low signal rate was higher in the later group due to the involvement of patients with worse outcomes. To overcome the limitation, we investigated 20 patients longitudinally and showed that the signal changed with time. In the transverse and longitudinal comparisons, the later the imaging, the more the fractured vertebra showing diffuse low signals increased. If we could identify the patients who might shift to diffuse low signal, minimally invasive surgical interventions before them shifting to a stage of worse signals would be more effective because performing kyphoplasty for a vertebra with higher instability leads to a higher incidence of adjacent fractures.¹³ Elucidating risk factors associated with not only the vertebral condition but also general conditions such as age, past medical history, osteoporosis, spinal alignment, back muscle, and others, makes early intervention possible after prognosis prediction before the vertebral condition worsens. Further investigations are needed to study this aspect.

In summary, MRIs performed later tended to show diffuse low signals on both T1-WI and T2-WI. As the signal intensity on MRI differs over time in the acute phase, we need to carefully follow-up patients with osteoporotic vertebral fracture regardless of the prediction by MRI.

Authors’ contributions

KI contributed to the conception and design of the study, acquisition of the data, analysis and interpretation of the data, and drafting the manuscript. HK and TS aided in the acquisition of the data. KH participated in the design and coordination and helped in drafting the manuscript. All authors read and approved the final manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of competing interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.