Non-traumatic rotator cuff tear is the most common cause of the should pain which occurs frequently in the elder[5, 6]. Compared the conservative treatment, arthroscopic rotator cuff repair presented good rehabilitation[7]. However, it is still difficult to determine or predict whether the rotator cuff can be repaired and the type of tear in the preoperative planning[4, 8, 9].
Classification of rotator cuff tear (RCT) using shoulder magnetic resonance (MR) imaging is traditionally based on tear width and thickness, not arthroscopic classification. However, for surgical RCT repair, arthroscopic classification is more useful because the surgical method depends on the classification type. Compared with a radiologist’s interpretation of knee MRI, in which the shape of a meniscal tear is regularly included in the description, the shape of a rotator cuff tendon tear remains a relatively difficult and misunderstood concept that has not taken a prominent place in the lexicon of the musculoskeletal radiologist. Some scholars have reported that axial MRI images are helpful for correctly dividing crescent tear and longitudinal tear [10], but how to identify L-shaped tear in longitudinal tear has not been reported. Some scholars presume that 3D reconstructions of rotator cuff tears would improve accuracy of the tear shape characterization [11], However, post-processing software limits its application. Other scholars have studied the diagnostic value of three-dimensional shoulder magnetic resonance arthrography in the classification of rotator cuff tear; however, it was found that the diagnostic efficiency of L-type tear was low [12], and the invasive characteristics of arthrography limit its application. In this study, we use the signs of conventional MRI sequence to interpret L-type tear, which is easier to implement in operation, so it has great application value.
Preoperative MRI presents pivotal role in assessing the type of tear. Although some study reported that preoperative MRI exhibited low efficiency in determining the type of tear or contributing to the preoperative planning[9]. Our research found that the scintillation sign could be well interpreted to distinguish the L-shaped tears. There are also some previous studies believed that after the scar is removed from the stump of the rotator cuff during surgery, the tear shape will change significantly. Therefore, limited sense was regarded in guiding the preoperative plan for the tear classification of preoperative MRI examination [3]. From our study, we found that the intraoperative removal of the scar from the stump may enlarge the tear, for the patients with longitudinal tears at least, but it did not change the tear type. Therefore, we believe that the tear classification of preoperative MRI examination is still instructive for surgery.
Preoperative MRI is often applied to classify rotator cuff tears. According to the geometric classification of rotator cuffs [4], non-large rotator cuff tears can be divided into crescent-shaped (L<W, W<2cm) and longitudinal tear (L>W, W<2cm). Longitudinal tear can be further divided into U-shaped and L-shaped. For L-shaped tears, correct identification and inflection point are the prerequisites for successful repair, whilst crescent tears require only direct tendon-bone sutures. U-shaped tears require more symmetrical side-to-side sutures, then the tendon-bone sutures. The purpose of our research is to identify L-shaped tears more accurate before surgery, to make adequate preoperative plan and increase the efficiency of successful repair. If we lack a correct understanding of L-shaped tears, due to the existing scars, the contracture and remodelling of the tear, it is difficult to correctly identify the inflection points of the transverse and longitudinal fractures of the L-shaped tear to achieve ideal intraoperative repairs [13]. The misalignment will also lead to failed repair of some rotator cuff, namely anatomical reduction, and low-tension repair.
When we found that the scintillation sign combined with the geometric classification of the rotator cuff tear can efficiently identify the L-shaped tear. The principle of the scintillation sign was analysed as follows: from the starting point of anatomy, the L-shaped tear is the longitudinal crack characterized by extremely steep edges on one side and extremely asymmetrical shape. The apex is extremely biased to one side, and the other side of the apex is a relatively less steep gradient shape (Figure 4A). Although the edges of the U-shaped tear are relatively steep, the shape is relatively more symmetrical. As a result, the middle of the crack is more similar to the arc-shaped gradual shape (Figure 4B). The scanning spacing (SP) of our current MRI scan is 4.4mm (4mm for the scanning layer and 0.4 mm for the layer spacing), the apex or sub-apex of the L-shaped tear appears directly in the next frame of the continuous rotator cuff tissue is highly correlated with a visual sign of sudden apex flash, while we were reviewing the oblique coronal scan (Figure 4A). However, the apex of the U-shaped tear often appears in the 2 to 3 frames after the tear occurs due to the above-mentioned anatomical characteristics, and then gradually changes to continuous rotator cuff tissue (Figure 4B). Therefore, the existence of the scintillation sign is based on the anatomical characteristics of the L-shaped tear and the thin thickness of the existing MRI.
Due to the morphological characteristics of L-type tear, theoretically, the probability of observing scintillation sign in oblique coronal position should be approximately 100%. However, our observation results are not so ideal when judged by the gold standard of the observation results of surgical records. We tried to analyse the possible causes. Firstly, for narrow L-tear, the partial volume effect of MRI may affect the interpretation accuracy of scintillation sign. Therefore, we speculate that using smaller slice thickness and spacing for scanning may improve the sensitivity and specificity of interpretation; Secondly, due to the deviation of intraoperative observation of the operator, some U-shaped tears may be recorded as L-shaped tears, or some L-shaped tears may be recorded as U-shaped tears. Finally, our sample size is small, which may bring statistical deviation. The above conditions will reduce the accuracy of scintillation sign. If there is a larger sample size and further prospective research can be done to better control the above variables, the accuracy of scintillation sign in the diagnosis of L-type tear may be better.
In summary, combined with the MRI geometric classification of rotator cuff tears, the scintillation sign is highly specific and sensitive in identifying L-shaped tears, which contributes to the preoperative identification. The accuracy of the scintillation sign needs further research and observation with a larger sample size to determine.