Diagnostic value of the axial view of magnetic resonance imaging to identify two-dimensional shapes of full-thickness rotator cuff tears

Abstract

Background

Coronal and sagittal views of magnetic resonance imaging (MRI) were used to determine rotator cuff tear size and fatty infiltration, but these images were not enough to identify the tear shape.

Purpose

To correlate the preoperative axial MRI views and arthroscopic surgical findings to identify the two-dimensional shapes in rotator cuff tears.

Material and Methods

This study included 166 patients who underwent arthroscopic repair between 2015 and 2018. Preoperative coronal, sagittal, and axial MRI views were evaluated for tear size and geographic configuration in axial sections, and the length and the width were measured and were matched with arthroscopic surgical views by lateral portals.

Results

The agreement of axial MRI views with the arthroscopic view was 88.0% in crescent, 97.2% in longitudinal, 78.6% in massive, and 100% in rotator cuff tear arthropathy. The mean agreement rate of axial MRI views with arthroscopic view was 81.9%. Mean mediolateral and anteroposterior tear sizes on axial MRI were 16.68 mm and 19.33 mm, respectively. Mean mediolateral and anteroposterior tear sizes by arthroscopic view were 21.49 mm and 21.04 mm, respectively. Tear sizes by MRI axial images were 71.3% of arthroscopic view. SST/IST degenerative changes were noted in most patients with massive tears and rotator cuff arthropathy (P = 0.001).

Conclusion

Rotator cuff tear shape on preoperative axial MRI view had close agreement (81.9%) with arthroscopic findings by lateral portal, and tear size by preoperative axial MRI views was 71.3% of that of arthroscopic view. Axial MRI views helped to predict the geometric tear shape of rotator cuff tears.

Keywords

Rotator cuff tear shape configuration size magnetic resonance imaging axial view

Introduction

Rotator cuff tears have various sizes and configurations. Making a precise definition of the pattern is an important initial step for surgical planning. Identification of tear patterns and anatomic repair can be easily performed for small to medium tears, but complex tear patterns cannot be easily distinguished and often require a surgical strategy for large to massive tears (1). Magnetic resonance imaging (MRI) is the best preoperative procedure for the evaluation of rotator cuff tendon tears, such as dimensions and extent of involved tendons, shape and configuration of tears, degree of muscle atrophy and fatty degeneration, and even reparability of large to massive tears (2,3). Preoperative MRI predicted the retraction and width of the tear with an accuracy of 80%, and positive consistency of the MRI findings has been reported with ultrasonographic and arthroscopic findings (4,5). These benefits present MRI as the essential diagnostic tool in preoperative planning. However, there were significant differences between the preoperative tear characteristics on MRI and the debrided tendon characteristics during surgery, which were extensive enough to classify the tear in a different category (6). Daniel et al. (7) also reported that MRI assessment of the rotator cuff was not accurate relative to arthroscopy and MRI diagnosis of partial tear was of little value.

DeOrio and Cofield used a one-dimensional description of the length of the greatest diameter of the tear based on the surgical findings and categorized tears into one of four groups: small; medium; large; and massive (8). Burkhart et al. (9) used the coronal and sagittal MRI views and arthroscopic findings and categorized tears into one of four groups: crescent; longitudinal (U-shaped and L-shaped); massive contracted; and rotator cuff tear arthropathy (9). The classification systems of Harryman et al. (10) and Gerber et al. (11) each characterized the status of the rotator cuff based on the number of torn tendons. However, these descriptions did not include tear shapes, and there might be differences between preoperative tear configuration on MRI and tear configuration during surgery.

MRI examinations provide basic three-dimensional information (coronal, sagittal, and axial images) on the tear shape. This information on the tear shape obtained from MRI provides guidance on decisive diagnosis and treatment plan. There may be adequate MRI views that represent the tear shape of the entire rotator cuff tendons more effectively. Some studies have described the tear shape based on coronal and sagittal MRI views, but the two-dimensional (2D) tear shape has not been fully evaluated (12,13). We hypothesized that the axial MRI views can identify the 2D shapes of rotator cuff tears. The aim of the present study was to identify the complete rotator cuff tear shape by evaluation of axial MRI views and to compare the tear shape with the arthroscopic findings.

Material and Methods

Patient selection and evaluation

This retrospective study consisted of patients who underwent primary arthroscopic rotator cuff repair performed by a senior orthopedic surgeon (SWJ) between January 2015 and December 2018 in our institution. Inclusion criteria were full-thickness rotator cuff tendon tear that underwent primary arthroscopic rotator cuff repair and performance of an MRI scan of the operated shoulder <6 months before surgery. Exclusion criteria were partial thickness not easily determined in axial views, adhesive capsulitis, calcific tendinitis, tear associated with proximal humeral fracture, or inflammatory joint disease. The present study included patients with rotator cuff tear with adhesive capsulitis and excluded patients with adhesive capsulitis only.

Patient demographics (age and sex) were extracted from the medical charts. A total of 202 patients (131 men, 71 women) underwent arthroscopic surgery for full-thickness rotator cuff tear at our institution between January 2015 and December 2018. Of these, 166 met the inclusion criteria and were included in the study (106 men [63.9%], 60 women [36.1%]; mean age = 61.9 years; age range = 35–84 years). Right shoulders comprised 113 cases and left shoulders comprised 53 cases. The mean time from MRI examination to surgery was 1.7 months (range = 0–6 months) (Table 1).

Table 1.

Demographic characteristics of patients by tear types on MRI with arthroscopic verification.

	Non-visible (n = 14)	Crescent (n = 100)	Longitudinal (n = 36)	Massive (n = 14)	Rotator cuff arthropathy (n = 2)	P value
Sex						0.522
Male	11 (78.57)	64 (64.00)	22 (61.11)	7 (50.00)	2 (100.00)
Female	3 (21.43)	36 (36.00)	14 (38.89)	7 (50.00)	0 (0.00)
Side						0.205
Right	8 (57.14)	64 (64.00)	28 (77.78)	12 (85.71)	1 (50.00)
Left	6 (42.86)	36 (36.00)	8 (22.22)	2 (14.29)	1 (50.00)
Age (years)	56.79 ± 12.49	60.23 ± 8.56	65.39 ± 7.16	69.21 ± 7.92	65.00 ± 11.31	<0.001
Time from MRI to surgery (months)	4.86 ± 2.21	1.36 ± 0.99	1.64 ± 3.01	1.29 ± 0.73	1.00 ± 0.00	0.910
Follow-up (months)	27.14 ± 10.97	28.08 ± 11.96	22.14 ± 14.76	31.29 ± 14.84	27.5 ± 7.78	0.115

Values are given as n (%) or mean ± SD.

MRI, magnetic resonance imaging.

Diagnostic glenohumeral arthroscopy was performed through a standard posterior portal. After proper treatment of any intra-articular pathology, the arthroscope was placed into the subacromial space through the same posterior skin incision. A lateral portal was created, and the shape of the tear was fully evaluated from the lateral viewing portal. To allow complete visualization and classification of the rotator cuff tear, all bursal tissues were removed from the margins of the tear, and the tear was identified and measured by arthroscopic probe. After debridement of the degenerated tendon edges, tear size, retraction, delamination, and initial pattern were evaluated via the posterolateral and direct lateral portal. A probe with laser marks (5-mm increments) was routinely used for measuring the medial-to-lateral length and the anterior-to-posterior width of the tear by the senior surgeon, and these measurements were recorded in the operating room by the operating nurse. When the apex of the torn rotator cuff was located adjacent to the glenoid rim, the tear was considered to be U-shaped or L-shaped, and side-to-side sutures or side-to-end sutures were primarily performed. All patients underwent arthroscopic rotator cuff repair (modified Mason–Allen technique), which was sufficient to restore the footprint of the rotator cuff.

Burkhart et al. (9) categorized rotator cuff tears based on both coronal- and sagittal-oriented MRI images: Type 1 = crescent-shaped tears, relatively short and wide with a medial-to-lateral length less than the anterior-to-posterior width; Type 2 = longitudinal (U-shaped and L-shaped) tears, relatively long and narrow with a medial-to-lateral length greater than the anterior-to-posterior width; Type 3 = massive contracted tears, long and wide (both > 3 cm); and Type 4 = rotator cuff tear arthropathy, associated with significant glenohumeral arthrosis and complete loss of the acromiohumeral interspace and imaging shows articulation of the humeral head with the undersurface of the acromion and end-stage glenohumeral arthrosis.

All preoperative MRI examinations were performed in our institution on a 3.0-T MRI unit (Phillips, Ingenia, Philips Healthcare, Amsterdam, The Netherlands) with a dedicated eight-channel shoulder coil. Routine examination included coronal oblique T2-weighted (T2W), with fat saturation and T1-weighted (T1W) sequences (T2: TR = 4400 ms, TE = 54.91 ms; T1: TR = 300 ms, TE = 14 ms; slice thickness = 3 mm, field of view [FOV] = 150 mm), sagittal oblique proton density (PD) with fat saturation and T1W sequences (PD: TR = 2760 ms, TE = 26.54 ms; T1: TR = 300 ms, TE = 9 ms; slice thickness = 3 mm, FOV = 150 mm), and an axial PD sequence with fat saturation (TR = 2760 ms, TE = 26.54 ms, slice thickness = 3 mm, FOV = 160 mm). Preoperative MRI examinations of the affected shoulder were evaluated by two experienced orthopedic surgeons (DHK and HSK, with seven and five years of orthopedic experience, respectively) and one experienced musculoskeletal radiologist (YMK) who were blinded to both clinical findings, including the arthroscopic tear type and the original MRI report, and a consensus was reached. After one month, MRI examinations were re-evaluated for intra-observer reliability.

First, maximum mediolateral and anteroposterior (AP) tear sizes on axial MRI views were measured in mm × mm, high signal resonance with overlying or underlying low signal resonance suggested the partial tear and were not used in the measurement. The tear shape was estimated using the geometrical definition of rotator cuff tear based on the classification of Burkhart et al. (9). T2W axial images of adequate quality were evaluated to make the measurement of the rotator cuff tears shapes. T2W images were chosen because intact tendons appear dark and torn tendons appear bright. Multiple successive axial-oriented images views were initially assessed to better understand the tear shape in a 2D form; one or two best sections of axial images were used to assess the tear size. The shapes of rotator cuff tears involving the supraspinatus and infraspinatus tendons can be found at the level of scapular spine. Mediolateral tear size was measured in mm on coronal MRI views, and AP tear size was measured in mm on sagittal MRI views. Second, muscle mass atrophy was measured according to occupation ratio. The grading for muscle atrophy was as follows: grade 0 = no atrophy of muscle mass; grade 1 = 25% atrophy; grade 2 = 50% atrophy; grade 3 = 75% atrophy; and grade 4 = full atrophy of muscle mass (14). Fatty degeneration in the rotator cuff muscles were measured according to the Goutallier classification (15). Muscle fatty infiltration was evaluated on the most lateral oblique sagittal T1W images in which the scapular spine was seen in contact with the scapular body (the so-called Y-shaped view). A tangent line was drawn from the superior border of the scapular spine to the superior margin of the coracoid process (16). Third, parameters of operative notes were compared with MRI measurements for tear size and shape as visualized during surgery from both the posterior and lateral portals using Burkhart et al.’s classification as a reference (9). This study was approved by the Institutional Review Board of Samsung Medical Center (2019-SCMC-03-003).

Statistical analysis

Measurements are expressed as mean ± SD, unless otherwise stated. Comparisons of categorical variables between patients with different type of rotator cuff tears were performed using the Chi-square test. Comparisons of continuous variables between patients with different rotator cuff tear shapes were conducted using the Kruskal–Wallis test. Cohen’s Kappa were calculated to assess intra-observer reliability. A kappa value of 0.20 or less indicates a poor agreement; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, good; and 0.81–1.0, excellent agreement. Agreement was examined between the MRI-determined tear type and the arthroscopic-determined tear type. All calculations were performed using IBM SPSS 21.0 software (IBM Corp, Armonk, NY, USA). P values <0.05 were considered statistically significant.

Results

Rotator cuff tears were classified on MRI as non-visible in 14 (8.4%) cases, crescent in 100 (60.2%) cases, longitudinal in 36 (21.7%) cases, massive in 14 (8.4%) cases, and rotator cuff arthropathy in 2 (1.2%) cases. The agreement of axial MRI views corresponding with the arthroscopic view was 88.0% (88/100 cases) in crescent, 97.2% (35/36 cases) in longitudinal, 78.6% (11/14 cases) in massive, and 100% (2/2 cases) in rotator cuff tear arthropathy. The mean agreement rate of axial MRI views with arthroscopic view was 81.9% (136/166 cases). There was a significant relation between classification of MRI measures and the arthroscopic findings. Cohen’s kappa values to assess the intra-observer reliability were 0.61 for crescent, 0.70 for longitudinal, 0.82 for massive, and 0.84 for rotator cuff arthropathy. Overall Cohen’s kappa value was 0.677, and a good reliability was noted consistently for interval reading.

The supraspinatus was the most commonly involved of all the rotator cuff tendons. Significantly, two more tendons were involved in massive and rotator cuff arthropathy tear types compared to crescent and longitudinal tear types (P = 0.001). Furthermore, muscle atrophy changes were observed in significantly higher frequencies in massive and rotator cuff tear arthropathy compared to crescent and longitudinal tears (P = 0.001) (Table 2).

Table 2.

Comparison of variables in arthroscopic findings and MRI views.

Arthroscopic findings	Crescent (n = 102)	Longitudinal (n = 49)	Massive (n = 12)	Rotator cuff arthropathy (n = 3)	P value
MRI tear type
Small	36 (35.29)	0 (0.00)	0 (0.00)	0 (0.00)
Medium	64 (62.75)	25 (51.02)	0 (0.00)	0 (0.00)
Large	2 (1.96)	21 (42.86)	7 (58.33)	1 (33.33)
Massive	0 (0.00)	3 (6.12)	5 (41.67)	2 (66.67)
Coronal mediolateral size (mm)	14.66 ± 5.68	29.89 ± 6.80	40.42 ± 11.96	46.67 ± 5.77	<0.001
Sagittal anteroposterior size (mm)	15.20 ± 5.93	27.98 ± 6.05	37.08 ± 10.97	46.67 ± 5.77	<0.001
Axial tear type
Non-visible	14 (13.73)	0 (0.00)	0 (0.00)	0 (0.00)
Crescent	88 (86.27)	12 (24.49)	0 (0.00)	0 (0.00)
Longitudinal	0 (0.00)	35 (71.43)	1 (8.33)	0 (0.00)
Massive	0 (0.00)	2 (4.08)	11 (91.67)	1 (33.33)
Rotator cuff arthropathy	0 (0.00)	0 (0.00)	0 (0.00)	2 (66.67)
Muscle atrophy
SSC	0.85 ± 0.67	1.37 ± 0.67	2.08 ± 1.08	3.00 ± 1.00	<0.001
SST	0.91 ± 0.73	2.08 ± 0.84	3.00 ± 0.74	3.67 ± 0.58	<0.001
IST	0.76 ± 0.55	1.45 ± 0.71	2.25 ± 1.06	2.67 ± 0.58	<0.001
Tangent sign	5 (4.90)	20 (40.82)	12 (100.00)	3 (100.00)	<0.001
Fatty change
SSC	0.81 ± 0.64	1.37 ± 0.64	2.08 ± 1.08	3.00 ± 1.00	<0.001
SST	0.91 ± 0.73	2.10 ± 0.82	3.08 ± 0.79	3.67 ± 0.58	<0.001
IST	0.75 ± 0.56	1.47 ± 0.68	2.25 ± 1.06	2.67 ± 0.58	<0.001
Cohen’s Kappa value	0.61 (0.47–0.72)	0.70 (0.55–0.81)	0.82 (0.71–0.90)	0.84 (0.73–0.92)	0.677

Values are given as n (%), mean ± SD, or median (range).

Non-visible tears on axial MRI were confirmed as small-sized or medium-sized tears on arthroscopic view in seven cases each.

IST, XX; MRI, magnetic resonance imaging; SSC, XX; SST, XX. [AQ: PLEASE EXPAND THE ACRONYMS IST, SSC, AND SST]

Mean mediolateral tear size on coronal MRI views was 20.38 ± 11.32 mm, mean anteroposterior tear size on sagittal MRI views was 19.83 ± 10.83 mm, and mean mediolateral and anteroposterior tear sizes on axial MRI were 16.68 ± 8.97 mm and 19.33 ± 9.22 mm, respectively. Mean mediolateral and AP tear sizes by arthroscopic view were 21.49 ± 11.43 mm and 21.04 ± 10.34 mm, respectively. Tear sizes by MRI axial images were 71.3% of those of arthroscopic view (Figs. 1 –4) (Table 3).

Fig. 1.

(a) Axial MRI view of small sized rotator cuff tearswith a crescent-shape in a 75-year-old man (10 × 15 mm). (b) Arthroscopic view by lateral portal (15 × 20 mm).

Fig. 2.

(a) Axial MRI view of medium sized rotator cuff tears with a longitudinal-shape in a 65-year-old woman (20 × 15 mm). (b) Arthroscopic view by lateral portal (30 × 20 mm).

Fig. 3.

(a) Axial MRI view of massive rotator cuff tears of a 69-year-old man (40 × 30 mm). (b) Arthroscopic view by lateral portal (40 × 40 mm).

Fig. 4.

(a) Axial MRI view of rotator cuff tear arthropathy of a 72-year-old man (40 × 40 mm). (b) Arthroscopic view by lateral portal (50 × 50 mm).

Table 3.

Measurement of MRI views and arthroscopic findings.

	Overall (n = 166)	Crescent (n = 100)	Longitudinal (n = 36)	Massive (n = 14)	Rotator cuff arthropathy (n = 2)
Coronal tear size (mm)	20.38 ± 11.32	16.15 ± 7.65	28.58 ± 5.54	38.29 ± 4.81	42.50 ± 3.54
Sagittal tear size (mm)	19.83 ± 10.83	15.54 ± 6.45	27.31 ± 7.81	35.86 ± 7.79	45.00 ± 7.07
MRI axial view
Mediolateral size (mm)	16.68 ± 8.97	12.54 ± 3.73	21.47 ± 6.00	37.64 ± 5.89	37.50 ± 3.54
Anteroposterior size (mm)	19.33 ± 9.22	16.16 ± 6.08	23.58 ± 7.05	37.71 ± 5.27	37.50 ± 3.54
Arthroscopic view
Mediolateral size (mm)	21.49 ± 11.43	16.85 ± 7.24	30.88 ± 7.12	39.29 ± 12.54	45.00 ± 7.07
Anteroposterior size (mm)	21.04 ± 10.34	16.85 ± 6.65	29.26 ± 5.38	37.50 ± 10.88	45.00 ± 7.07
Matching between axial view and arthroscopic view (%)	71.30	70.90	56.00	96.20	69.40
Matching between coronal view and arthroscopic view (%)	94.83	95.85	92.55	97.45	94.44
Matching between sagittal view and arthroscopic view (%)	94.25	92.23	93.36	95.63	100

Values are given as mean ± SD unless specified otherwise.

MRI, magnetic resonance imaging.

Discussion

The most important finding is that a best axial view indicates the shape of the rotator cuff tears. One or two axial view can be inferred in two dimensional shapes of the rotator cuff tears. Until now, researches have been one-dimensional measurements of the tear shape in the coronal or sagittal MRI views. One or two axial images can be inferred in 2D shapes of the rotator cuff tears. As we hypothesized, the present study demonstrated that characterization of rotator cuff tears on preoperative MRI is significantly correlated with the arthroscopic findings. Axial MRI and arthroscopic tear shapes were consistent with 88%, 97%, 79%, and 100% of crescent, longitudinal, massive, and rotator cuff arthropathy, respectively, in the present study.

MRI is a reliable diagnostic tool in detecting and characterizing rotator cuff tears. It is well established that there is a strong correlation between complete tear as seen on MRI and that found during surgery (17). Iannoti et al. (18) reported a 100% sensitivity and 95% specificity for diagnosing full-thickness tears and defining their size on MRI when correlated to arthroscopic findings. Teefey et al. (19) reported that MRI predicted retraction and width of a tear with an accuracy of 63% and 80%, respectively. However, Bryant et al. (5) reported a difference in size and concluded that ultrasound, MRI, and arthroscopy underestimated tear size by 33%, 30%, and 12%, respectively. Consistency of MRI findings has been reported against both ultrasonographic and arthroscopic findings. Reliability of MRI in interpretation of rotator cuff tendon problems has been studied. Eren et al. (6) reported that radiologists showed higher agreement on MRI views compared with surgeons. Surgeons measured a larger tear length and width compared to radiologists, which could be related to tear enlargement during surgery and the surgeon’s tendency to ignore debrided tissue. There were significant differences between the preoperative tear characteristics on MRI and the debrided tendon characteristics during surgery, which were extensive enough to classify the tear in a different category.

Measuring muscle mass with cross-sectional area provides important information on muscle strength, and preoperative fatty degeneration is a prognostic factor affecting the anatomical and functional outcomes of rotator cuff surgery (20). Changes in muscle atrophy and fatty infiltration in the present study were observed in significantly higher frequencies in massive and rotator cuff arthropathy tear groups. Meyer et al. (21) have shown that preoperative fatty muscle infiltration according to the Goutallier staging is an important predictor for repair failure. Kim et al. (22) have shown that development of fatty infiltration in the supraspinatus and infraspinatus muscles is strongly associated with both tear size and location. Because fatty infiltration was associated with increased re-tear rates and poorer functional outcomes, surgery should be performed before fatty infiltration development. Tear size on MRI axial images was significantly increased in longitudinal and massive groups compared to the crescent group, and muscle atrophy, tangent sign, and fatty change were also significantly increased in our cases.

The question remains as to whether one or two axial images can represent the entire tear shape. As rotator cuff tears were often covered by the acromion or the humeral head, axial images of the MRI were not able to show the whole shape of the rotator cuff tears. As rotator cuff tears have different sites, a single technique of MRI examination was inadequate to show the shapes of all rotator cuff tears. The total number of non-visible groups was 14 cases and, 11 cases of these were confirmed as small-sized tears and three cases of these were confirmed as medium-sized tears in the arthroscopic views. There is thought to be a hidden lesion at the MRI axial cutting interval. Moreover, one slice section near the greater tuberosity has the potential to underestimate the shape of rotator cuff tears due to retraction in the musculotendinous junction when the rotator cuff muscle is injured (23). When comparing MRI coronal and sagittal tear size and MRI axial tear size, the tear size measured in the axial view was smaller than that measured in coronal and sagittal views. The average tear length and width on coronal and sagittal views (20.38 and 19.83 mm) was significantly higher than those measured on axial views (16.68 and 19.33 mm).

We confirmed the predictive value and usefulness of the conventional one to two slices of preoperative axial MRI images. The axial view represented 81.9% of the shape of rotator cuff tears in the present study, and specific longitudinal tears were evaluated accurately in the axial images (97.2%). The differences did not affect the operative findings. Therefore, this method may be regarded as a useful and reliable option for predicting the shapes of rotator cuff tears.

One of the advantages of this study was the surgical confirmation of the shape of rotator cuff tears during arthroscopic surgery. Great care was taken to acquire the quantitative measurement of the sizes and shapes by two orthopedic surgeons and one radiologist, and the best possible conclusion made.

However, this study has three limitations. First, this was a retrospective case series, and additional prospective studies will be required to validate these findings. Larger cohorts in prospective and longer follow-up studies are warranted. Second, the longest period between MRI and surgery is six months, during which the tear size and shape could change with time. Finally, this study does not compare the shape of rotator cuff tear of the axial view with that of the coronal and sagittal views.

In conclusion, rotator cuff tear shapes on preoperative axial MRI views had close agreement (81.9%) with arthroscopic findings based on lateral portal, and tear sizes on preoperative axial MRI views had an agreement of 71.3% of those on arthroscopic views. The axial MRI views helped to predict the geometric tear shape of rotator cuff tears.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Sung-Weon Jung

Min-Jae Jang

References

Dwyer

Razmjou

Henry

, et al. Association between pre-operative magnetic resonance imaging and reparability of large and massive rotator cuff tears. Knee Surg Sports Traumatol Arthrosc 2015; 23:415–422.

Yoo

Ahn

Yang

, et al. Correlation of arthroscopic repairability of large to massive rotator cuff tears with preoperative magnetic resonance imaging scans. Arthroscopy 2009; 25:573–582.

Kim

Park

Rhee

YG.

Can preoperative magnetic resonance imaging predict the reparability of massive rotator cuff tears?

Am J Sports Med 2017; 45:1654–1663.

Sela

Eshed

Shapira

, et al. Rotator cuff tears: correlation between geometric tear patterns on MRI and arthroscopy and pre- and postoperative clinical findings. Acta Radiol 2015; 56:182–189.

Bryant

Murrell

A comparison of clinical estimation, ultrasonography, magnetic resonance imaging, and arthroscopy in determining the size of rotator cuff tears. J Shoulder Elbow Surg 2002; 11:219–224.

Eren

Ozben

Gunerbuyuk

, et al. Rotator cuff tear characteristics: how comparable are the pre-operative MRI findings with intra-operative measurements following debridement during arthroscopic repair? Int Orthop 2019; 43:1171–1177.

Daniel

Mark

Bryon

, et al. Magnetic resonance imaging assessment of the rotator cuff: is it really accurate? Arthroscopy 1997; 13:710–719.

DeOrio

Cofield

RH.

Results of a second attempt at surgical repair of a failed initial rotator-cuff repair.

J Bone Joint Surg Am 1984; 66:563–567.

Burkhart

Sanders

Denard

, et al. MRI and arthroscopy correlations of the shoulder: a case-based approach. Instr Course Lect 2012; 61:185–200.

10.

Harryman

2nd Mack

Wang

, et al. Repairs of the rotator cuff. Correlation of functional results with integrity of the cuff. J Bone Joint Surg Am 1991; 73:982–989.

11.

Gerber

Fuchs

Hodler

The results of repair of massive tears of the rotator cuff. J Bone Joint Surg Am. 2000; 82:505–515.

12.

Lee

Lucas

Lansdown

, et al. Magnetic resonance rotator cuff fat fraction and its relationship with tendon tear severity and subject characteristics. J Shoulder Elbow Surg 2015; 24:1442–1451.

13.

Nitin

B. Jain, MD

Jamie

Collins

, et al. Reliability of magnetic resonance imaging assessment of rotator cuff: The ROW Study PMR 2015; 7:245–254.

14.

Warner

Laurence

Parsons

, et al. Diagnosis and treatment of anterosuperior rotator cuff tears. J Shoulder Elbow Surg 2001; 10:37–46.

15.

Goutallier

Postel

Bernageau

, et al. Fatty muscle degeneration in cuff ruptures. Pre- and postoperative evaluation by CT scan. Clin Orthop Relat Res 1994; 304:78–83.

16.

Kissenberth

Rulewicz

Hamilton

, et al. A positive tangent sign predicts the repairability of rotator cuff tears. J Shoulder Elbow Surg 2014; 23:1023–1027.

17.

Sasiponganan

Dessouky

Ashikyan

, et al. Subacromial impingement anatomy and its association with rotator cuff pathology in women: radiograph and MRI correlation, a retrospective evaluation. Skeletal Radiol 2019; 48:781–790.

18.

Iannotti

Deutsch

Green

, et al. Time to failure after rotator cuff repair: a prospective imaging study. J Bone Joint Surg Am 2013; 95:965–971.

19.

Teefey

Rubin

Middleton

, et al. Detection and quantification of rotator cuff tears. Comparison of ultrasonographic, magnetic resonance imaging, and arthroscopic findings in seventy-one consecutive cases. J Bone Joint Surg Am 2004; 86:708–716.

20.

Davidson

Burkhart

Richards

, et al. Use of preoperative magnetic resonance imaging to predict rotator cuff tear pattern and method of repair. Arthroscopy 2005; 21:1428.

21.

Meyer

Wieser

Farshad

, et al. Retraction of supraspinatus muscle and tendon as predictors of success of rotator cuff repair. Am J Sports Med 2012; 40:2242–2247.

22.

Kim

, et al. The relationship between rotator cuff tear and four acromion types: cross-sectional study based on shoulder magnetic resonance imaging in 227 patients. Acta Radiol 2019; 60:608–614.

23.

Knudsen

Gelineck

Søjbjerg

, et al. Functional and magnetic resonance imaging evaluation after single-tendon rotator cuff reconstruction. J Shoulder Elbow Surg 1999; 8:242–246.