Comparative Analysis of Real-Time Dynamic Ultrasound and Magnetic Resonance Imaging in the Diagnosis of Rotator Cuff Tear Injury.

OBJECTIVE: To explore the value of real-time dynamic ultrasound and magnetic resonance imaging (MRI) in the diagnosis of rotator cuff injury.
METHODS: From January 2020 to June 2021, the clinical data of 55 patients with rotator cuff injury were collected. All patients were examined by real-time dynamic ultrasound and MRI at the same time within two weeks before surgery, which was confirmed by arthroscopy. The value of real-time dynamic ultrasound, MRI, and the real-time dynamic ultrasound combined with MRI in the diagnosis of rotator cuff injury was evaluated.
RESULTS: Among the 55 patients with rotator cuff injury, real-time dynamic ultrasound showed full-thickness rotator cuff tear in 12 patients (21.82%), including type I in 2 patients (3.64%), type II in 6 patients (10.91%), and type III in 4 patients (7.27%), and partial rotator cuff tear in 31 patients (56.36%), including type IV in 16 patients (29.00%), type V in 5 patients (9.09%), and type VI in 10 patients (18.18%). MRI showed full-thickness rotator cuff tear in 12 patients (21.82%), including type I in 2 patients (3.64%), type II in 6 patients (10.91%), and type III in 4 patients (7.27%), and partial rotator cuff tear in 37 patients (67.27%), including type IV in 20 patients (36.36%), type V in 7 patients (12.73%), and type VI in 10 patients (18.18%). The combined examination showed full-thickness rotator cuff tear in 12 patients (21.82%), including type I in 2 patients (3.64%), type II in 6 patients (10.91%), and type III in 4 patients (7.27%), and partial rotator cuff tear in 39 cases (70.91%), including type IV in 21 cases (38.12%), type V in 7 patients (12.73%), and type VI in 12 patients (21.82%). The diagnostic accuracy of real-time dynamic ultrasound in complete tearing, partial tearing and the overall accuracy were 92.31%, 73.81% and 78.18%, respectively. The diagnostic accuracy of complete tear, partial tear and the overall accuracy of MRI were 92.31%, 88.00% and 89.09%, respectively. The diagnostic accuracy of complete tear and partial tear and overall accuracy in the real-time dynamic ultrasound combined with MRI were 92.31%, 95.24%, and 94.55%, respectively. The diagnostic accuracy and overall diagnostic accuracy of the combined examination of partial tears were higher than those of the real-time dynamic ultrasound and MRI examinations alone.
CONCLUSION: Real-time dynamic ultrasound and MRI have high application value in the diagnosis of rotator cuff injury, and the combined diagnosis is conducive to improving the diagnostic accuracy of patients with partial tear.

1. Introduction

The shoulder joint is the joint with the largest range of motion in the whole body, which mainly relies on the rotator cuff formed by the muscles around the shoulder joint to maintain the stability and activity of the shoulder joint. The rotator cuff is a sleeve-like structure composed of subscapularis muscle, supraspinatus muscle, infraspinatus, teres minor, and tendon wrapping humeral head, so as to maintain the stability of the shoulder joint. The incidence of rotator cuff injury caused by rotator cuff tendon degeneration increases with age. Rotator cuff tear is one of the most common causes of shoulder pain and activity limitation in patients. Epidemiology shows that rotator cuff tear accounts for 30%–70% of the causes of shoulder pain. The diagnosis of rotator cuff tear is complicated, and it is not uncommon to misdiagnose scapulohumeral periarthritis [1-3]. Rotator cuff tear injuries are often caused by subacromial impingement syndrome or trauma and are divided into partial tear and complete tear, which are difficult to repair after tear. If not treated properly or not repaired, some tear will often develop into complete tear. Therefore, accurate judgment of rotator cuff tear is of great significance for clinical selection of treatment options and prognosis judgment. Early detection of tear and surgical treatment can significantly shorten the lesion repair time and reduce the occurrence of postoperative re-tear [4, 5]. Arthroscopy is considered as the gold standard for the diagnosis of rotator cuff injury, but it is invasive and expensive and is not suitable for routine diagnosis of rotator cuff injury. At present, the main imaging examination methods of rotator cuff injury are magnetic resonance imaging (MRI) and real-time dynamic ultrasound. Magnetic resonance imaging (MRI) has become more and more mature in the diagnosis of rotator cuff injury because of its high resolution, high sensitivity, and high specificity [6]. Real-time dynamic ultrasound has been widely recognized by clinicians for its convenience, low cost, real-time dynamics, and non-invasive advantages. It has gradually become a commonly used method for clinical diagnosis of rotator cuff injury [7]. Previous studies have focused on the value of MRI and real-time dynamic ultrasound in the diagnosis of arthritis and joint lesions, but there is no definitive conclusion on the advantages and disadvantages of different imaging modalities for the diagnosis of rotator cuff injury [8]. On this basis, clinical data of 55 patients with rotator cuff injury were collected in this study to explore the application values of MRI and real-time dynamic ultrasound in the diagnosis of rotator cuff tear injury. The reports are as follows.

2. Data and Methods

2.1. General Information

From January 2020 to June 2021, the clinical data of 55 patients with rotator cuff injury were collected. There were 26 males and 29 females, and they were 21–70 years old, with the average age of (52.96 ± 8.29) years old, and the course of disease was from three days to 12 years, with the average course of disease of (32.49 ± 9.53) months. There were 42 patients with history of trauma, 3 patients with hypertension, and 6 patients with diabetes. Clinical manifestations include different degrees of shoulder pain and limited activity. This study was approved by the Hospital Ethics Committee, and informed consent was obtained from patients and their families.

Inclusion criteria were as follows: patients all confirmed with rotator cuff injury by arthroscopy; real-time dynamic ultrasound and MRI were performed; aged >18 years; and approved by the ethics committee of our hospital.

Exclusion criteria were as follows: contraindications to MRI or real-time dynamic ultrasound; other treatments before the examination; clinical data suggest space-occupying lesions in the shoulder; and incomplete imaging data.

2.2. Inspection Methods

All patients underwent simultaneous real-time dynamic ultrasound and MRI examinations within two weeks before surgery, which were finally confirmed by arthroscopy. Real-time dynamic ultrasound was performed using Philips IU22 color ultrasound instrument with C12-5 broadband linear array probe at the frequency of 5–15 MHz, and special software for skeletal muscle was used. When the patient is sitting, the examination sites mainly included supraspinatus tendon, infraspinatus tendon, subscapular tendon, teres minor tendon, and joint capsule. The transverse and longitudinal sections were examined, and the dynamic examination of the patient's shoulder joint was conducted when necessary. A diagnosis was made whether or not the shoulder joint is injured and the degree of injury, and the glenohumeral joint position was adjusted. The rotator cuff on the affected side was observed dynamically, and the shoulders on both sides were scanned by contrast.

MRI examination was performed using Philips Achieva 3.0 T TX superconducting MRI device from the Netherlands, using a coil dedicated to the shoulder joint. A cross-sectional scan was performed, the scanning plane was perpendicular to glenoid, and the scanning range was from acromion to humeral neck, with fast spin echo and T2 weighted sequence (TR/TE = 4000 ms/78 ms). An oblique coronal scanning was performed, and the slice was horizontally parallel to the long axis of the supraspinatus muscle, ranging from the lateral end of the clavicle to the shoulder peak. Spin-echo T1-weighted imaging (TR/TE = 450 ms/13 ms) and short-term inversion recovery sequence (TR/TE = 4000 ms/28 ms) were performed. An oblique sagittal scan was performed with the slice perpendicular to the long axis of the supraspinatus tendon, covering the humeral head and the entire glenoid, spin-echo T1-weighted imaging (TR/TE = 450 ms/13 ms), and short-term reversal recovery sequence (TR/TE = 4000 ms/28 ms). Scanning parameters were as follows: FOV = 24 cm × 24 cm, matrix = 640 × 640, layer thickness of 3.5 mm, and layer spacing of 1 mm. HPZ400 workstation is adopted for data processing with software.

2.3. Evaluation Method

The results of ultrasound and MRI were evaluated by two radiologists with rich clinical experience, respectively. In case of different opinions, the general practitioner made a consultation and finally reached an agreement. Finally, the results were compared with those of the arthroscopy.

2.4. Evaluation Criteria

Diagnostic Criteria of Real-Time Dynamic Ultrasound [9]. The echo and fiber course in the affected shoulder were observed, together with whether there was fluid in the joint cavity and the glide sac of deltoid muscle, as well as hyperplasia and defect on the surface of humeral head, hyperplasia of synovial membrane, and abnormal changes in the tendon and tendon sheath of long head of biceps brachii. Normal rotator cuff showed that the echo inside rotator cuff was uniform and stronger than that of deltoid muscle. Large-area full-thickness rotator cuff injury (injury diameter > 3 cm) showed the missing rotator cuff or sudden thinning of rotator cuff. The middle and small full-thickness injuries of the rotator cuff (injury diameter ≤3 cm) showed that the hypoechoic area penetrating the whole thickness appeared in the middle of the uniform echo of rotator cuff muscle. Partial injury of rotator cuff joint surface showed that in the middle of the uniform echo of rotator cuff muscle, near humeral head side, and hypoechoic area which does not penetrate the whole layer appears. The acromion surface of the rotator cuff was damaged, and a hypoechoic area without penetrating the whole layer appeared near the acromion side in the myogenic homogeneous echo of the rotator cuff. Meanwhile, the long head tendon of biceps brachii and its tendon sheath were observed for abnormalities.

MRI Diagnostic Criteria [10]. Imaging showed high-signal TWI and STIR in our patient, which ran through the entire layer of supraspinatus tendon, indicating complete rotator cuff tear. In our patient, TWI and STIR had high signal intensity, which was shown on the same side of supraspinatus tendon. If the signal did not penetrate through the whole layer, it was shown as partial tear of rotator cuff.

Classification of Arthroscopic Rotator Cuff Injury [11]. Complete tears include type I: large fissure with tendon shortening; type II: small fissure without tendon shortening; and type III: no fissure, tendon swelling. Partial tears include type IV: inner layer (articular surface) tear; type V: outer layer (synovial surface) tear; and type VI: intra-tendon tear.

3. Statistical Methods

SPSS22.0 statistical software was used for data analysis. The measurement data were expressed as ( ± s). The t test or one-way analysis of variance was performed and the count data were expressed as (n(%)). The χ2 test or Fisher's exact probability test was performed, and P < 0.05 indicated that the difference was statistically significant.

4. Results

4.1. Arthroscopic Examination Results

In 55 patients with rotator cuff injury, there were 13 patients (23.64%) of full-thickness rotator cuff tear, including 2 patients (3.64%) of type I, 6 patients (10.91%) of type II, and 5 patients (9.09%) of type III, and 42 patients (76.36%) of partial rotator cuff tear, including 22 patients (40.00%) of type IV, 8 patients (14.54%) of type V, and 12 patients (21.82%) of type VI.

4.2. Real-Time Dynamic Ultrasonic Examination Results

In 55 patients with rotator cuff injury, real-time dynamic ultrasound showed full-thickness rotator cuff tear in 12 patients (21.82%), including type I in 2 patients (3.64%), type II in 6 patients (10.91%), and type III in 4 patients (7.27%), and partial rotator cuff tear in 31 patients (56.36%), including type IV in 16 patients (29.00%), type V in 5 patients (9.09%), and type VI in 10 patients (18.18%), as shown in Tables 1 and 2 and Figure 1.

Table 1

Result of real-time dynamic ultrasound in diagnosis of rotator cuff tear injury.

Group	n	Complete tear	Partial tear
Real-time dynamic ultrasound	43	12 (21.82%)	31 (56.36%)
Arthroscopy	55	13 (23.64%)	42 (76.36%)
χ 2 value	13.469
P value	≤0.001

Table 2

Classification result of real-time dynamic ultrasound in diagnosis of rotator cuff tear injury.

Group	Type I	Type II	Type III	Type IV	Type V	Type VI
Real-time dynamic ultrasound	2 (3.64%)	6 (10.91%)	4 (7.27%)	16 (29.09%)	5 (9.09%)	10 (18.18%)
Arthroscopy	2 (3.64%)	6 (10.91%)	5 (9.09%)	22 (40.00%)	8 (14.54%)	12 (21.82%)

Figure 1

Real-time ultrasound examination of typical patient pictures. Note: ultrasound showed that the left supraspinatus muscle was torn in full thickness, the subscapular tendon was torn, the capsule wall of the left deltoid muscle was thickened with effusion, and the infraspinatus tendon was injured.

4.3. MRI Results

In 55 patients with rotator cuff injury, MRI showed full-thickness rotator cuff tear in 12 patients (21.82%), including type I in 2 patients (3.64%), type II in 6 patients (10.91%), and type III in 4 patients (7.27%), and partial rotator cuff tear in 37 patients (67.27%), including type IV in 20 patients (36.36%), type V in 7 patients (12.73%), and type VI in 10 patients (18.18%), as shown in Tables 3 and 4 and Figure 2.

Table 3

Result of MRI in diagnosis of rotator cuff tear injury.

Group	n	Complete tear	Partial tear
MRI	49	12 (21.82%)	37 (67.27%)
Arthroscopy	55	13 (23.64%)	42 (76.36%)
χ 2 value	6.346
P value	0.012

Table 4

Classification result of MRI in diagnosis of rotator cuff tear injury.

Group	Type I	Type II	Type III	Type IV	Type V	Type VI
MRI	2 (3.64%)	6 (10.91%)	4 (7.27%)	20 (36.36%)	7 (12.73%)	10 (18.18%)
Arthroscopy	2 (3.64%)	6 (10.91%)	5 (9.09%)	22 (40.00%)	8 (14.54%)	12 (21.82%)

Figure 2

MRI oblique coronal pictures of typical patients. Note: rotator cuff complete tear type I. (a) MRI oblique coronal PDWI sequence shows that the high signal shadow runs through the whole layer of supraspinatus tendon, and the tendon is shortened (arrow). (b) MRI oblique coronal PDWl sequence showed that the high signal shadow ran through the whole layer of supraspinatus tendon, and the tendon had no shortening (arrow).

4.4. Diagnosis of Rotator Cuff Injury by the Combination of Real-Time Dynamic Ultrasound and MRI

The combined examination of real-time dynamic ultrasound and MRI showed that there were 12 patients (21.82%) of full-thickness rotator cuff tear, including 2 patients (3.64%) of type I, 6 patients (10.91%) of type II, and 4 patients (7.27%) of type III, and 39 patients (70.91%) of partial rotator cuff tear, including 21 patients (38.12%) of type IV, 7 patients (12.73%) of type V, and 12 patients (21.82%) of type VI, as shown in Tables 5 and 6.

Table 5

Result of joint inspection in diagnosis of rotator cuff tear injury.

Group	n	Complete tear	Partial tear
Joint inspection	52	12 (21.82%)	40 (72.73%)
Arthroscopy	55	13 (23.64%)	42 (76.36%)
χ 2 value	3.084
P value	0.079

Table 6

Classification result of joint inspection in diagnosis of rotator cuff tear injury.

Group	Type I	Type II	Type III	Type IV	Type V	Type VI
Joint inspection	2 (3.64%)	6 (10.91%)	4 (7.27%)	21 (38.12%)	7 (12.73%)	12 (21.82%)
Arthroscopy	2 (3.64%)	6 (10.91%)	5 (9.09%)	22 (40.00%)	8 (14.54%)	12 (21.82%)

4.5. Diagnostic Value of Single Examination and Combined Examination of Real-Time Dynamic Ultrasound and MRI in Rotator Cuff Injury

The diagnostic accuracy of real-time dynamic ultrasound for complete tear was 92.31%, while that for partial tear was 73.81%, with an overall accuracy of 78.18%. The diagnostic accuracy rate of MRI for complete tear was 92.31%, while that for partial tear was 88.00%, and the overall accuracy rate was 89.09%. The diagnostic accuracy of complete tear, partial tear, and overall tear in the combined examination was 92.31%, 95.24%, and 94.55%, respectively. The diagnostic accuracy and overall diagnostic accuracy of the combined examination of partial tears were higher than those of the real-time dynamic ultrasound and MRI examinations alone, as shown in Table 7.

Table 7

Diagnostic value of single examination and combined examination of real-time dynamic ultrasound and MRI in rotator cuff injury.

Group	Overall accuracy	Complete tear	Partial tear
Real-time dynamic ultrasound	43/55 (78.18%)	12/13 (92.31%)	31/42 (73.81%)
MRI	49/55 (89.09%)	12/13 (92.31%)	37/42 (88.00%)
Joint inspection	52/55 (94.55%)	12/13 (92.31%)	40/42 (95.24%)

5. Discussion

The rotator cuff tear injury mainly involves the supraspinatus muscle, and the injury is mainly located in the ischemic risk area about 1 cm from the supraspinatus muscle stop point. The tears of infraspinatus, subscapularis muscle, and teres minor alone are very rare because the supraspinatus muscle is located at the intersection of forces around the shoulder joint, and thus it is extremely vulnerable. Studies have shown that the main pathogenesis of rotator cuff tear injury is long-term strain and tendon degeneration, combined with impact and compression under shoulder. Acute or repeated trauma and shoulder instability are also important causes of the disease. The main symptoms of rotator cuff injury are pain and limited movement. Clinically, there are many related complaints about shoulder diseases, and clinicians easily misdiagnose due to lack of understanding of rotator cuff injury [9-11]. At present, arthroscopy is the gold standard for diagnosing rotator cuff injury, and physical examination is the general investigation method. However, the former cannot be used as an extensive screening method due to its invasive nature, while the latter is relatively subjective and has a great relationship with the degree of cooperation of patients and physicians' experience, making it difficult to be used as the main diagnostic method. Therefore, exploring the diagnostic value of real-time dynamic ultrasound, MRI, and other non-invasive imaging in rotator cuff injury is of great significance to fill the gap in clinical application, but the related research in China is rare.

In the 55 patients with rotator cuff injury, real-time dynamic ultrasound showed full-thickness rotator cuff tear in 12 patients (21.82%), including type I in 2 patients (3.64%), type II in 6 patients (10.91%), and type III in 4 patients (7.27%), and partial rotator cuff tear in 31 patients (56.36%), including type IV in 16 patients (29.00%), type V in 5 patients (9.09%), and type VI in 10 patients (18.18%). MRI showed full-thickness rotator cuff tear in 12 patients (21.82%), including type I in 2 patients (3.64%), type II in 6 patients (10.91%), and type III in 4 patients (7.27%), and partial rotator cuff tear in 37 patients (67.27%), including type IV in 20 patients (36.36%), type V in 7 patients (12.73%), and type VI in 10 patients (18.18%). On real-time dynamic ultrasound, partial tear of the supraspinatus tendon was characterized by a decrease or sometimes increase in irregular echo reflections at the tendon edge in the focal area. On MRI, the partial tear was characterized by focal areas in weighted image with discontinuous fibers filled with fluid signals. T2-weighted image of the fat suppression sequence shows lesions more clearly because it better showed that the tendon tear is full of liquid. In addition to the focal tendon defect, there was also wear on the tendon surface or a change in the tendon diameter. A full-thickness tear means that the continuity of the rotator cuff tendon fibers is completely interrupted, which leads to the communication between glenohumeral joint and subacromial space [12]. On real-time dynamic ultrasound, the direct signs of a complete tear were the absence of rotator cuff tissue, local focal tendon defects, and tendon retraction, while the indirect signs were flattening of the tendon bursa surface, thinning of the rotator cuff, roughened cortex of the greater tubercle, glenohumeral effusion, and the acromion-deltoid glide capsule [13]. On MRI, a complete rotator cuff tear presented as an irregular margin of fluid signal intensity on T2-weighted image, involving the entire thickened tendon. However, in approximately 10% of the lacerations, the tendon discontinuity appeared hypointense on T2-weighted images due to chronic cicatrix formation. Fluid was also visible in the acromion-deltoid sliding bag and glenohumeral joint. The anterior-posterior tear of the supraspinatus muscle is greater than 2.5 cm, and the infraspinatus tendon or subscapularis muscle building is often involved in the tear. Huge rotator cuff tears usually involve supraspinatus muscle, infraspinatus muscle, and subscapular muscles at the same time, with tendon retraction [14, 15].

Among the missed diagnosis results of real-time ultrasound examination, there was one patient with partial tear, which was mainly caused by lack of experience and the patient's pain could not guarantee the shoulder joint internal rotation and abduction position. The other 3 patients were torn inside tendon, and no obvious lesions were found by ultrasound. By analyzing all the missed diagnosis results, it can be found that the accuracy of ultrasound diagnosis of complete rotator cuff tear is higher, but it is relatively lower for partial rotator cuff tear. MRI showed that 2 patients with type VI rotator cuff injury were undetected, and the diagnosis of type VI was difficult. When effusion appeared in the joint cavity and the acromion glide sac, the intra-tendon signal would decrease due to the partial volume effect, which was similar to the moderate signal of tendon degeneration, which easily caused missed diagnosis.

The results of this study showed that the combined examination showed full-thickness rotator cuff tear in 12 patients (21.82%), including type I in 2 patients (3.64%), type II in 6 patients (10.91%), and type III in 4 patients (7.27%), and partial rotator cuff tear in 39 patients (70.91%), including type IV in 21 patients (38.12%), type V in 7 patients (12.73%), and type VI in 12 patients (21.82%). MRI examination has the advantages of non-invasiveness, non-radioactivity, anatomical repeatability, and excellent tissue contrast. It can perform multi-angle and multi-plane scanning imaging and can provide normal and abnormal imaging images of multiple tissue structures at the same time. Rotator cuff injury is mainly the morphological change of tendon to different degrees. After the rotator cuff is torn, there is liquid infiltration in the split position or fissure, which manifests as the morphological and signal abnormalities of tendon on MRI image. The liquid signal shows low signal on TWI and high signal on TWI. When the inner layer of the supraspinatus tendon is torn, fluid will penetrate, and the torn tendon will have obvious high signal. Therefore, MRI has a high accuracy in the diagnosis of type IV rotator cuff injury [16, 17]. When the frequency of ultrasound increases obviously, the ultrasound images will be clearer, the articular surface, synovium, and tendon can be clearly observed, and the moving tendon can be dynamically checked in real time, so that the tendon and its continuity can be displayed more clearly and completely. Interrupted rotator cuff continuity, partial tendon defect, and anechoic or non-uniform echo in the fissure can be observed in real-time dynamic ultrasound, and the degree of rotator cuff injury can be further evaluated according to the morphological changes of the tendon and the echo condition at the laceration [18].

The diagnostic accuracy and the overall diagnostic accuracy of the combined real-time dynamic ultrasound and MRI examination of partial tears are obviously improved as compared with those of the separate dynamic ultrasound and MRI examination. For real-time dynamic ultrasound, intra-tendon tear, tendonitis, tendon degeneration, and granulation tissue proliferation have similar ultrasound image manifestations. Therefore, simple abnormal ultrasound needs to be combined with the accompanying signs around the shoulder joint and even requires the patient to perform multi-axial exercise, while the examiner should perform multi-sectional examination. However, for MRI-induced injury to type V rotator cuff, especially when effusion is present in the glide sac of acromion or the bursa of deltoid muscle, the lesion is more likely to be displayed with the infiltration of effusion. However, a long course of disease, especially accompanied by thickening and ossification of the coracoid-shoulder arch, hyperplasia, and fibrosis of the glide sac at the acromion, may easily result in misdiagnosis [19]. Benedetto et al. [20] thought that the fat layer of the inferior acromion capsule was discontinuous, that is, there was V-type rotator cuff injury, and the lesion was easy to show. However, patients with a long course of disease, especially with thickening and ossification of coracoid arch, hyperplasia, and fibrosis of acromion glide sac, will easily lead to misdiagnosis. As far as clinical practicability is concerned, real-time dynamic ultrasound has the advantages of real-time dynamics, repeatability, and low cost, so it is suggested as a routine diagnostic method. At the same time, based on the conclusion that the combined diagnosis of real-time dynamic ultrasound and MRI in this study was effective, MRI could be used as a complementary diagnosis of real-time dynamic ultrasound to improve the accuracy of the diagnosis in patients with partial tear [21].

In summary, real-time dynamic ultrasound and MRI have high application value in the diagnosis of rotator cuff injury, and the combined diagnosis is conducive to improving the diagnostic accuracy of patients with partial tear.