How often do surgeons intervene on shoulder labral lesions detected at MR examination? A retrospective review of MR examinations correlated with arthroscopy

Abstract

Objective:

We report the prevalence of surgical intervention on shoulder labral lesions detected at MR examinations and how surgeons describe labral tears seen at MR examinations in their arthroscopy reports.

Methods:

A retrospective review of 100 consecutive patients aged 50 years or younger who had shoulder labral tears on MR and went on to have surgery performed. It was determined whether surgical intervention was performed on the MR lesions.

Results:

Of these 100 patients, 72 had superior labral anterior to posterior (SLAP) tears, 38 had posterior labral tears and 28 had anterior labral tears on MR examination. All 100 patients went on to arthroscopy. All lesions described on MRI were described on arthroscopy. Of the 72 SLAP tears, 64 were described as fraying on arthroscopy with 51 debrided. The remaining eight SLAP tears were tacked surgically. Of the 38 posterior labral tears, 36 were described as fraying on arthroscopy with 29 debrided and 2 had surgical tacking performed. Of the 28 anterior labral tears described on MR examination, 26 had surgical tacking performed and 2 were debrided. There were four SLAP tears, two anterior labral tears and three posterior labral tears seen on arthroscopy but not seen on MR examination.

Conclusion:

In this series, a high percentage of SLAP tears and posterior labral tears described on MR examination did not have surgical tacking. Most anterior labral tears had surgical tacking. Based on the above, our surgeons request we describe superior and posterior labral lesions as fraying and/or tearing, unless we can see a displaced tear. Most anterior labral lesions are treated with surgical tacking.

Advances in knowledge:

MRI allows for sensitive detection of labral tears. The tears often are not clinically significant.

Labral tears are common injuries that often require surgical intervention. In our practice, we commonly see labral tears on MR examination and report them to surgeons. With modern high-resolution MRI, as well as increased awareness of labral tears by radiologists, labral tears are commonly reported findings on MR examinations. Sensitivities and specificities for detection of labral tears as compared with those from arthroscopy at 3.0 T have been reported as follows: superior labral anterior to posterior (SLAP) tears (90% sensitive and 100% specific), anterior labral tears (89% sensitive and 100% specific) and posterior labral tears (86% sensitivity and 100% specific).¹ Sensitivities of 100% for anterior labral tears, 86% for superior labral tears and 74% for posterior labral tears as compared with those from arthroscopy have been reported using high-resolution 1.5 T conventional MR examinations.²

In our practice, we have been told by our surgeons that we describe some SLAP tears and posterior labral tears on MR that they find on arthroscopy to be degenerative fraying. For this reason, we sought to find (1) how often does a surgeon intervene on labral tears and (2) how do surgeons describe labral tears seen at MR examinations in their arthroscopy reports.

METHODS AND MATERIALS

100 consecutive conventional shoulder MR examinations performed between January 2012 and July 2012 on patients aged 50 years or younger who had labral tears described on MR examinations and who went on to arthroscopy were read retrospectively in consensus by two musculoskeletal radiologists with over 10 years' experience in reading musculoskeletal MRI. All patients were referred from one orthopaedic group. The three members of this orthopaedic group all had over 10 years' experience performing shoulder arthroscopy. The age range of the 100 patients was 18–49 years (mean, 37 years). All patients had arthroscopy performed within 45 days of the MR examination (range, 3–45 days); mean time between MR examination and arthroscopy was 21 days. Surgical reports were correlated with MR examinations. It was specifically determined whether surgical intervention (i.e. surgical tacking or debridement) was performed on the lesions described on MR examinations.

Scans were assessed for SLAP tears and anterior or posterior labral tears. The MR criterion used for the diagnosis of labral tear was an abnormality of the glenoid labrum morphology and/or signal intensity. Labral tears were diagnosed on MR examinations when there was high signal intensity on T₂ weighted images in the labrum even if there was not an irregular or a detached labrum. A SLAP was defined as a superior labral irregularity and/or high signal on T₂ weighted images within the superior labrum. SLAP tears that extended anteriorly were defined as SLAP tears and not as anterior labral tears. An anterior or a posterior labral tear was defined as an area of abnormal increased signal on T₂ weighted images in the labrum and/or an irregular or a detached labrum.

The images were assessed for the presence or absence of anterior or posterior labral tears or SLAP tears and were not graded: that is, on retrospective review, images were not graded for high- or low-grade SLAP tears or detached or non-detached labral tears but rather simply whether such lesions were present or not. The institutional review board approval was obtained for this retrospective review and correlation with surgical records.

All patients underwent MRI of the shoulder in oblique coronal, oblique sagittal and axial planes on a 3.0 T GE Signa® scanner (General Electric Medical Systems, Waukesha, MI). Oblique coronal and sagittal fast spin-echo T₁ weighted [550/10, repetition time (TR)/echo time (TE), number of excitations (NEXs) of 2] sequences, oblique coronal and sagittal fast spin-echo fat-saturated T₂ weighted (3850/55, TR/TE, NEX of 4) sequences and fast spin-echo fat-saturated proton density axial (3250/55, TR/TE, NEX of 3) sequences with a field of view of 12 cm on all images were used. The slice thickness was 4 mm with a 10% interslice gap on all sequences except for the fast spin-echo proton density axial sequence, which had a 3-mm slice thickness. The echo train length was ten on all T₂ weighted and proton density sequences and three on the T₁ weighted sequences. The bandwidth was 31.25 kHz on all sequences. The imaging time for the oblique coronal and sagittal T₂ weighted sequences was 4 min and 43 s. The imaging time for the proton density axial sequences was 3 min and 26 s, whereas the imaging time for the T₁ weighted sequences was 2 min and 28 s. The matrix for all T₂ weighted sequences was 320 × 320, and the matrix for all T₁ weighted sequences was 320 × 256. A USA instruments three-channel phased array shoulder coil (USA instruments Inc., Aurora, OH) was used. One patient had an MR arthrogram performed in addition to the conventional MR examination of the shoulder.

RESULTS

Of these 100 patients, 72 had SLAP tears, 38 had posterior labral tears and 28 had anterior labral tears described on MR examination. Several patients had multiple labral tears.

All 100 patients went on to arthroscopy. All 72 SLAP tears, 38 posterior labral tears and 28 anterior labral tears described on MRI were described on arthroscopy. However, many of the lesions described as tears on MR reports were not described as tears on arthroscopy. Of the 72 SLAP tears, 64 were described as fraying on arthroscopy. Of these 64 cases, 51 were debrided. In 13 cases, no surgical intervention was performed. The remaining eight SLAP tears were described as SLAP tears on arthroscopy, and surgical tacking was performed. Of the 38 posterior labral tears, 36 were described as fraying on arthroscopy. Of these 36 patients, 29 were debrided. In seven cases, there was no surgical intervention. 2 of the 38 patients had surgical tacking performed arthroscopically. Of the 28 anterior labral tears described on MR examination, 26 had surgical tacking performed. 2 of the 28 patients had their anterior labrum debrided arthroscopically (Table 1, Figures 1–8).

Table 1.

Surgical intervention on labral tears described at MR examination

Position of labral tears	Tear on MR examination	Fraying on arthroscopy	Tear on arthroscopy	Surgical tacking	Surgical debridement	No surgical intervention
Anterior labrum	28	2	26	26	2	0
Posterior labrum	38	36	2	2	29	7
Superior labrum	72	64	8	8	51	13

Figure 1.

A 32-year-old male with shoulder pain. Proton density fat-saturated axial MR image (3250/55, repetition time/echo time) shows a posterior labral tear (arrow). Description on arthroscopy: a posterior labrum split tear with a flapped fragment; this was repaired.

Figure 8.

A 39-year-old male with shoulder pain. (a) T₂ weighted coronal MR image [3850/55, repetition time (TR)/echo time (TE)] demonstrates no definite superior labral tear (right arrow) and intact supraspinatus tendon (left arrow). (b) T₁ weighted fat saturated coronal MR arthrogram image (550/10, TR/TE) demonstrates superior labral tear (right arrow) and high-grade partial thickness supraspinatus tendon tear (left arrow). Arthroscopy report: superior labral anterior to posterior tear and supraspinatus tendon tear; these were repaired.

Figure 2.

A 41-year-old male with shoulder pain. T₂ weighted coronal MR image (3850/55, repetition time/echo time) shows a superior labral tear (arrow). Description on arthroscopy: superior labral anterior to posterior tear; this was debrided.

Figure 3.

A 35-year-old male with shoulder pain. The proton density fat-saturated axial MR image (3250/55, repetition time/echo time) shows a displaced anterior labral tear (arrow). Arthroscopy findings: anterior labral tear; this was repaired.

Figure 4.

A 27-year-old male with shoulder pain. The proton density fat-saturated axial MR image (3250/55, repetition time/echo time) shows displaced anterior superior labral tear (arrow). Finding at arthroscopy: a tear to the superior labrum anterior to the biceps labral repair. This was considered a superior labral tear in our analysis.

Figure 5.

A 26-year-old male with shoulder pain. T₂ weighted coronal MR image (3850/55, repetition time/echo time) shows a superior labral anterior to posterior tear (arrow). Arthroscopy report: degenerative fraying of superior labrum; this was debrided.

Figure 6.

A 33-year-old female with shoulder pain. Proton density fat-saturated axial MR image (3250/55, repetition time/echo time) shows tearing of anterior (thick arrow) and posterior labra (thin arrow). There is also a joint effusion present. Arthroscopy report: degenerative tearing of labrum; this was debrided.

Figure 7.

A 29-year-old male with shoulder pain. Proton density fat-saturated axial MR image (3250/55, repetition time/echo time) shows tear of posterior labrum (arrow). Arthroscopy report: fraying and tearing posterior labrum; these were debrided.

There were four SLAP tears, two anterior labral tears and three posterior labral tears seen on arthroscopy but not seen on conventional MR examination. One of these SLAP tears was seen on MR arthrography despite not being seen on conventional MR examination (Figure 8). The four SLAP tears and two anterior labral tears had surgical tacking performed. The three posterior labral tears were debrided.

The surgeons described fraying as an irregularity of the labral surface at arthroscopy. The need for debridement was determined by the arthroscopist when he/she felt the degree of irregularity at the labral surface was such that debridement and smoothing of the surface might benefit the patient.

Sensitivity and specificity for detection of labral tears in this study were as follows: anterior labral tear, 93% sensitivity and 100% specificity; posterior labral tear, 93% sensitivity and 100% specificity; and superior labral tear, 95% sensitivity and 100% specificity (Table 2).

Table 2.

Sensitivity and specificity for detection of labral tears on MR examination as compared with arthroscopy

Position of labral tears	Tear on MR examination	Fraying/tear on arthroscopy	MR sensitivity compared with arthroscopy (%)	MR specificity compared with arthroscopy (%)
Anterior labrum	28	30	93	100
Posterior labrum	38	41	93	100
Superior labrum	72	76	95	100

DISCUSSION

In this study, there were a large number of labral tears described. There were a total of 138 labral tears described on MR examinations of these 100 patients who went on to surgery.

Of the 72 SLAP tears described on MR examinations, 64 were described as fraying at arthroscopy. Only eight cases were described as SLAP tears at arthroscopy with surgical tacking performed. There were 38 posterior labral tears described on MR examination. Of these, 36 were described as fraying on arthroscopy. In only two cases was surgical tacking performed. There were 28 anterior labral tears described on MR examination. 26 of these patients had surgical tacking performed.

In a previous surgical study by Hurley and Anderson,³ 100 shoulder arthroscopies were reviewed. In this study, a large number of glenoid labral lesions found at arthroscopy appeared to be associated with conditions other than instability. Many of these lesions were not treated surgically. In this study, there was a high correlation between tears in the anterior inferior glenoid labrum and anterior instability.

In another surgical study by Glasgow et al,⁴ it was noted on arthroscopy that there were glenoid labral injuries without associated instability in 72% of patients. Many of these lesions were not treated surgically. It was also found that in patients with anterior instability and labral tears, labral debridement was not a successful alternative to surgical tacking.

In a surgical study by Mohtadi et al⁵ comparing MR examinations with those of arthroscopy, it was concluded that MRI is a useful tool in the identification of shoulder pathology but the clinical correlation of this information at arthroscopy and the assessment of outcomes remain unknown.

In a surgical study by Van der Veen et al,⁶ it was noted that there were many lesions shown on MRI that could not be confirmed by arthroscopy and therefore did not have therapeutic consequences. The study also noted some labral lesions that needed surgical treatment not detected on MRI. The study concluded that, at their institution, MRI had limited value prior to arthroscopic treatment of post-traumatic shoulder instability.⁶ Some of the discrepancies in the study may be owing to differences in descriptions of labral pathology between the radiologist and the surgeon.

Connolly⁷ performed a retrospective surgical study assessing the accuracy of conventional MRI for detection of SLAP type 2 tears. He concluded that conventional MRI is not reliable for detection of type 2 SLAP tears in the community setting. In this study, sensitivity for detection of SLAP type 2 tears was 38%. This study was performed on a variety of MR units ranging from 0.3 to 3.0 T.

The above studies in the surgical literature^3–7 describe a higher prevalence of discrepancies in the descriptions of labral pathology than that reported in the radiology literature.^1,2 In particular, there were a number of labral lesions described at arthroscopy that were not felt to be associated with instability. In two of the studies,^5,6 it was questioned whether MRI provided clinically useful information for arthroscopy.

Some of the discrepancies of findings in the surgical literature vs the radiology literature may be owing to semantic differences in descriptions of labral pathology. In particular, in the study by Van der Veen et al,⁶ it was stated that there were many lesions shown on MRI that could not be confirmed by arthroscopy. Some of this discrepancy may be owing to differences in description. If the surgeon sees a lesion that is not considered clinically significant to tack or debride, he/she might not describe the lesion thoroughly in the arthroscopy report. In addition, some lesions described as tears on the MR report, may be described as fraying in the arthroscopy report. In surgical studies, these findings may be considered discrepant findings between the arthroscopy report and the radiology report.

There were four SLAP tears, two anterior labral tears and three posterior labral tears seen on arthroscopy but not seen on conventional MR examination in the current study. The sensitivity and specificity for detection of labral tears in this study were as follows: anterior labral tear, 93% sensitivity and 100% specificity; posterior labral tear, 93% sensitivity and 100% specificity; and superior labral tear, 95% sensitivity and 100% specificity (Table 2).

Sensitivities and specificities for detection of labral tears as compared with those from arthroscopy at 3.0 T have been reported as follows: SLAP tears (90% sensitive and 100% specific), anterior labral tears (89% sensitivitive and 100% specific) and posterior labral tears (86% sensitivity and 100% specific).¹ Sensitivities of 100% for anterior labral tears, 86% for superior labral tears and 74% for posterior labral tears as compared with arthroscopy have been reported using high-resolution 1.5 T conventional MR examinations.² Sensitivities and specificities for detection of labral tears as compared with those from arthroscopy in the current study compares favourably with previously published studies.

In the current study, conventional MR was highly accurate in detecting labral pathology. However, the description of such pathology was often described slightly differently by the surgeon in the arthroscopy report. In our practice, the surgeons tend to describe abnormal labra that are not unstable as fraying rather than tearing.

In our practice, labral tears are commonly seen and described on MR examination. High-resolution MR scanning allows for very sensitive detection of such tears. However, some of these tears might not be considered clinically significant by a surgeon. In particular, a high percentage of SLAP and posterior labral tears described on MR examination were described as fraying at arthroscopy. Most of these tears did not have surgical tacking performed.

In most cases, anterior labral tears did have surgical tacking performed. Anterior labral tears are commonly associated with anterior dislocations and often occur in a younger population. In addition, anterior labral tears are often found to be unstable at arthroscopy.

Posterior labral tears and SLAP tears can often be degenerative in nature and often are not unstable at arthroscopy. These tears will occur in a variety of age groups. The main criterion used by our surgeons for surgical tacking of these lesions is instability at arthroscopy.

Based on the above results, our surgeons request we describe superior and posterior labral lesions as fraying and/or tearing unless we can see a displaced tear. Most anterior labral lesions are treated with surgical tacking.