Background
Partial-thickness rotator cuff tears (PTRCTs) are a common cause of adult shoulder pain, with a relatively high prevalence ranging from 13% to 32% [
1]. However, diagnosis and treatment remain controversial. Based on cadaver and clinical studies, most PTRCTs involve the supraspinatus tendon, and articular-side tears are two to three times more common than bursal-side tears [
2‐
5]. Isolated PTRCTs are rare, except those occurring in supraspinatus tendons [
3]. Herein, we report three cases of isolated bursal-side infraspinatus tears that were difficult to detect by magnetic resonance imaging (MRI). In these cases, computed tomography (CT) bursography helped identify the lesion.
Discussion and conclusions
The cases presented demonstrate that an isolated PTRCT can occur on the bursal-side tendon of the infraspinatus. Regarding the epidemiology of rotator cuff tears, a cadaver study of 249 supraspinatus tendons revealed that 13% had PTRCTs; of these, 55% were intratendinous tears, 27% were articular-side tears, and 18% were bursal-side tears [
3,
8]. In addition, clinical studies have also shown that articular-side PTRCTs are more common than bursal-side tears [
9,
10]. An arthroscopic examination of 41 partial-thickness supraspinatus tears showed bursal-side tears in 20% and articular-side tears in 80% [
10]. From these reports in the literature, isolated bursal-side PTRCT of the infraspinatus is unusual.
The bursal-side infraspinatus tear may be identified as a longitudinal lesion by CT bursography due to the morphological characteristics of the infraspinatus. The infraspinatus has a long tendinous portion in the superior half of the muscle, which curves anteriorly and extends to the anterolateral area of the highest impression of the greater tuberosity [
11]. Kato
et al. [
12] reported that the infraspinatus is composed of a transverse part and an oblique part in accordance with muscle fiber direction. Both parts have partially independent morphology, and the transverse part adjoins the posterior surface of the main tendinous portion of the oblique part as a thin tendinous membrane that may be fragile [
12]. Therefore, longitudinal rotator cuff tears may occur when the infraspinatus is torn from the adjoining area of the tendon to the gap between the transverse and oblique muscles. This area may also become more fragile due to tendon degeneration, especially in older patients, as seen in our cases. To our knowledge, the present report is the first to document these unusual rotator cuff tears.
MRI is commonly accepted as one of the best noninvasive procedures for evaluating full-thickness rotator cuff tears and has a sensitivity and specificity of >90% [
13‐
15]. However, the sensitivity of MRI for PTRCTs is often lower than that for full-thickness tears. A meta-analysis of 29 articles regarding the detection of rotator cuff tears demonstrated that MRI has an overall sensitivity of 67% and specificity of 94% for PTRCTs [
14]. Therefore, fat suppression in MRI has been used to increase the detection of PTRCTs. Singson
et al. [
16] compared T2-weighted MRI findings with and without fat suppression in the detection of PTRCTs and found that the sensitivity of fat-suppressed MRI was 92%, versus 67% without fat suppression. Another study on the detection of PTRCTs with fat-suppressed MRI also showed a high sensitivity of 82% [
17]. However, Xiao
et al. [
18] reported that fat-suppressed MRI for bursal-sided PTRCTs had a lower sensitivity of 74.3%. Hence, even with fat-suppressed MRI, it may be difficult to diagnose a bursal-side PTRCT correctly.
Shoulder arthrography is also an accepted imaging technique for evaluating rotator cuff tears. CT arthrography has a high sensitivity and specificity for the diagnosis of articular-side PTRCTs and full-thickness tears but cannot detect bursal-side PTRCTs [
19,
20]. In comparison, subacromial bursography is useful for identifying bursal-side PTRCTs and full-thickness tears [
21‐
23]. Schneider
et al. [
23] examined 17 patients with a bursal-side PTRCT and found that these were detected by MRI and ultrasonography with a sensitivity of 64% and 41%, respectively, while bursography had a higher sensitivity of 82%.
Three methods of bursography combined with other imaging examinations have been reported. First, CT bursography can visualize superficial rotator cuff lesions [
24,
25]. Fermand
et al. [
25] combined bursography with CT arthrography to examine 33 patients with shoulder pain. They found that 15 of these patients had radiological abnormalities of the deep surface of the subacromial bursa, including a thumbprint-like notch in the rotator cuff muscle, serrated unevenness near the lateral edge of the rotator cuff muscle, a linear fissure in the rotator cuff tendon, unevenness of the superficial cuff surface with loose bodies in the bursa, partial avulsion of the rotator cuff tendon, and an ulcer on the superficial cuff surface. Second, ultrasonographic bursography has also been reported and noted to be more informative for rotator cuff tears than plain ultrasonography [
26,
27]. Cheng
et al. [
26] used percutaneous ultrasound-guided subacromial bursography to examine 63 shoulders and found that ultrasonographic bursography correctly identified all full-thickness tears but missed one in five patients who were misdiagnosed with a bursal-side PTRCT by plain ultrasonography. Finally, MR bursography is also a possible imaging technique. However, whether it is reliable for detecting bursal-side PTRCTs remains controversial owing to the difficulty of injecting gadopentetate dimeglumine, which is not a radiocontrast agent, accurately into the subacromial bursa [
28].
There are several limitations to this report to consider. First, subacromial bursography requires injection of contrast medium and therefore is invasive. Furthermore, CT exposes the patient to radiation. Second, the infraspinatus degenerative tear could be an incidental finding, which may not be the source of the patient’s symptoms and might not need isolated repair or debridement in the patients. Third, we did not demonstrate the accuracy of CT bursography in the diagnosis of bursal-side PTRCTs. However, the cases presented do provide a basis for further studies regarding bursography and its usefulness and accuracy, especially in patients with ongoing pain but normal findings on standard imaging.
In conclusion, the patients in the three cases presented had isolated bursal-side infraspinatus tears, and CT bursography was useful for identifying them. Therefore, patients with shoulder pain refractory to conservative management but normal MRI findings should be evaluated with CT bursography. Further studies are required to evaluate and compare the diagnostic accuracy of bursography combined with CT with ultrasonography and MRI for bursal-side PTRCTs.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.