Diagnostic sensitivity of ultrasound of the supraspinatus tendon when compared to magnetic resonance imaging prior to arthroscopy: A retrospective study

Ultrasound (USS) and magnetic resonance imaging (MRI) have been used as screening tools for rotator cuff tears with MRI reported as superior. When comparing USS and MRI for supraspinatus tears, less data is available.


| INTRODUCTION
Rotator cuff pathologies are a significant cause of shoulder pain. [1][2][3] Within the group of rotator cuff tendons, supraspinatus tendon tears represent the majority of injuries. [1][2][3][4] The shoulder's complexity of function with multidirectional movements makes it vulnerable to damage. 4 The supraspinatus tendon can tear partially or fully in thickness, tear size has a significant influence over repairability. Primary repair of a supraspinatus tendon may not be possible if the tendon is very large and has retracted. This differentiation in diagnosis is important to illustrate for the surgeon, as the path of treatment is different based on tear size and retraction. 4 A partial tear with less than 25% of the tendon involved will generally only have the damaged tissue removed. This process is called debridement. Partial tears greater than 25% of the tendon width and full thickness tears (FTT) will have the tendon reattached to the bone. [4][5][6][7][8][9] The two main diagnostic tools used to assess tendon tears prior to surgery are high resolution ultrasound (USS) and magnetic resonance imaging (MRI). 4,6 The advantages of USS include lower cost, higher accessibility, real time evaluation allowing for observation of tendon movement whilst performing the scan and in general better patient compliance as there is no confined space or limitations with movement. [1][2][3] Assessment and comparison of the opposite shoulder is also a distinct advantage with USS. 1,[4][5][6] The benefits of MRI include better tissue delineation, anatomy definition and the ability to interrogate bony degenerative changes and or fractures. [1][2][3] However, increased cost, time and patient intolerance are limitations to MRI's use. [1][2][3]7,10 USS and MRI have been used as screening tools for rotator cuff tears since 1984. 8 Subsequent research compared the accuracy of these two imaging modalities in the diagnosis of rotator cuff pathology, with the majority reporting MRI to be superior to USS. 2,[6][7][8]11 However, when the clinical question is focussed on the supraspinatus tendon, the argument to replace USS with the more expensive MRI is less persuasive as most recent literature demonstrates that there is little difference in the diagnostic superiority of MRI. 1,3,7,11 To date there is an abundance of data that has been published on rotator cuff tears. However, there is a paucity of concise information detailing USS role as a stand-alone diagnostic tool specifically for supraspinatus tears. Despite a comparative accuracy of 100% for FTT all studies were inconclusive in establishing the value of USS replacing MRI as the preferred pre-operative imaging tool when diagnosing partial thickness tears (PTT). [1][2][3]7,11 2 | AIM This study aims to investigate the sensitivity of USS compared to MRI in the pre-operative diagnosis of partial and full thickness supraspinatus tendon tears using shoulder arthroscopy as the gold standard. The study will assess if there is a difference in detection rates between the two imaging modalities. We hypothesised that USS can be used as a pre-operative screening tool for FTT and PTT without the need for MRI.

| METHOD
This project was a retrospective, quantitative study with data collected from participants' medical records. The information was collected over a period from 01/08/2020-01/08/2021. Data collected involved a number of general radiology practices. Ultrasounds were performed by general sonographers. Ultrasound machines involved in the study included Philips Epiq5, Canon Aplio 500, and General Electric Logiq E9.
These machines are equipped with high frequency linear probes in the range of 10-15 MHz appropriate for musculoskeletal scanning.
The MRI machines in this study included a Philips Ingenia 1. Comparisons were tabulated between the USS and MRI diagnostic sensitivity, with that of the gold standard finding of arthroscope assessing the results of full and partial supraspinatus thickness tears.
We detailed findings according to sensitivity using this formula; Sensitivity = [TP/TP + FN] x100 (TP = true positive and TP = false negative) categorising average sensitivity for FTT, average sensitivity for PTT and overall sensitivity for all tears (both PTT and FTT). [12][13][14] Tear sizes were classified using the Ellman classification system. 15 There was only one orthopaedic practice involved in this study and all MRI and ultrasound machines used were of similar standard, this endeavours to reflect the general imaging community.

| Variables and demographics
Data collected included patient age reported as mean (standard deviation) (m (SD)), gender as a ratio of male: female and degree of supraspinatus tear. [12][13][14] Data were analysed reporting on number of supraspinatus tears reported as sensitivity (true positive case) for USS and MRI and categorised into 1. number and sensitivity of partial thickness tears <than 5 mm 2. number and sensitivity of partial thickness tears >than 5 mm

number and sensitivity of full thickness tears
Exclusion criteria were those who had a longer than 3-month period between diagnosis and arthroscope, and whose pathology was unrelated to the supraspinatus tendon and participants aged less than 18 years.
Participants signed a consent for de-identified data to be used for research purposes. This consent was provided to each participant by the surgeon as part of a non-obligatory pre-surgery protocol.
All participants in this study were recruited and identified using Medicare numbers, this allowed the capture of only those participants who had an USS, MRI and arthroscopic repair. Central Queensland Ethics Committee provided ethics approval (ethics number 2021-010). 16

| RESULTS
One hundred and three participants who had arthroscopic shoulder repair had their data reviewed comparing the diagnostic accuracy of pre-operative USS and MRI with arthroscopy used as gold standard.
Overall, there were 63 FTT and 40 PTT diagnosed by arthroscopy.
The ratio of male: female participants were 63:40 with a mean age of 64 (10.8) years. Data reporting sensitivity for USS and MRI for FTT and PTT and overall sensitivity for all tears is listed in Table 1.
Specificity could not be calculated as people with negative results did not present for surgery.
Comparable sensitivities were seen with the two modalities of USS and MRI with MRI having slightly higher sensitivity overall.
The sensitivity of MRI and USS was the same for FTT. For PTT greater than 5 mm, MRI had a 13.1% improvement in diagnosis compared to USS. For PTT less than 5 mm USS had a 5% improvement in diagnosis compared to MRI. Six participants were under reported on USS where the tear size was greater than 5 mm on arthroscopy, however in tear sizes smaller than 5 mm, USS and MRI results were similar. However, due to the small sample size, it is difficult to conclude whether this is clinically significant. Overall, the sensitivity of USS and MRI were similar.

| DISCUSSION
Our results demonstrate the sensitivity of USS and MRI are comparable in diagnosing full thickness and partial thickness supraspinatus tears. Since USS is less expensive and more available, it could be considered as the screening tool of choice when supraspinatus tendon integrity is the only question. [1][2][3]7,11 The incidence of rotator cuff tears increases with age. 17,18 FTT are reported in 25% of individuals over the age of 60 and 50% of individuals over 80 years of age. 19 The age and gender demographics in our study are consistent with the literature that indicates older age and male gender are associated with a higher number of supraspinatus tears. [1][2][3]19 Supraspinatus PTT and FTT may be part of a degenerative process where the tendon progressively wears out due to repetitive stress or due to an acute injury. [1][2][3]12,13 A PTT is defined as focal discontinuation without continuation of the tear to the opposing tendon surface. 4,12 These tears are often treated conservatively; sometimes with physiotherapy or corticosteroid injections. In cases where this treatment fails to help, debridement under arthroscope is used. 6,8 Partial tears can be described as bursal, articular surface tears and intrasubstance tears. 4,[20][21][22] The articular side of the supraspinatus tendon is commonly torn with a critical hypo vascular zone existing 10-15 mm proximal to the supraspinatus tendon insertion onto the humeral head making it vulnerable to damage. 4,[20][21][22] The bursal sided tears are more likely due to subacromial impingement and are less common than articular and intrasubstance tears. 4,20-22 USS is useful for detection of impingement due to the ability to scan dynamically which has advantage over the stationary nature of MRI.
The intrasubstance tear is confined to the interstitial layers of the tendon, they are also known as concealed interstitial delamination tears. Previous studies indicate that PTT's on the articular surface are the most common supraspinatus pathology diagnosed. [1][2][3][4]23 However in these studies the participants were from a broader cohort, including those that did not require surgical intervention. In our study there was a higher number of FTT (n = 63) than PTT (n = 40). The captured participants in our study all needed surgical intervention giving a higher prevalence of significant tears that could not be managed conservatively. It is important to correctly diagnose the location and thickness of tears as it directly impacts on management. 4,8,11,17 Our study shows good correlation between MRI and USS giving confidence in USS ability to correctly define tendon tear location and size.
Arthroscopic repair for a FTT involves reattachment of the torn tendon onto the bone and may involve acromioplasty, where the underside of the acromion is shaved to allow for better tendon sliding due to an increase in the subacromial space. 4   MRI, although more expensive and less well tolerated, still remains the preferred imaging choice prior to surgery, as it is generally considered to have a higher sensitivity overall when diagnosing PTT. [1][2][3] Our study involved a group of generalist sonographers using differing USS machines intending to reflect a less specialised group. Therefore, our results may be more indicative of a broader cross section of examinations available to the general community. According to Cole et al. 26 rotator cuff tear sensitivity using USS is operator and machine dependant. 26 USS performed by experienced specialised sonographers improves tear detection rates. 26 Sensitivity rate using USS will continue to further improve with better sonography training and continued improvement of spatial resolution, ultimately allowing greater confidence in the use of USS as a stand-alone diagnostic tool for supraspinatus tears. [1][2][3][4] Shoulder arthroscope is the gold standard for diagnosis of intraarticular pathologies. 3 However, because arthroscopy involves an anaesthetic and a surgical procedure, it is not appropriate to be used as the first line of assessment. 9 Although MRI remains the preferred imaging modality for shoulder assessment, 1-3 USS is emerging as a cost-effective, reliable, and convenient alternative due to continued improvement in transducer resolution and consistent education. 24,26,27 While the results of our research may impact on the diagnostic imaging choices prior to supraspinatus arthroscopy with USS being used as an alternative to MRI, our study suggests, practices should audit USS to see if these results concur generally.
Arthroscopic technological advances have allowed for better surgical intervention. Arthroscope has become less invasive and less expensive. Intervention for moderate PTT is therefore now more likely. 6,8,17,28,29 The evolution of surgical intervention makes it important for USS to detect a small partial thickness tear because these tears are now more likely in the future to be referred to surgery rather than being treated conservatively. 6,8,28 PTT, if left untreated can enlarge and propagate into FTT, early intervention prevents tear progression and protects cuff elasticity. 24 Lower grade tears do well with arthroscopic debridement alone while large tears need a repair. 4,23,24 As USS resolution and user proficiency continues to improve in detecting PTTs USS role will become more important for early intervention diagnosis, perhaps in the future replacing MRI.
Recent meta-analysis published in 2020 and 2021, 27,30 found good correlation between MRI and USS for FTT though a lower sensitivity for both modalities compared to arthroscope for characterising PTT. 27,30 These results appear to be broadly consistent with earlier systemic reviews. With advances in technologies, MRI and USS have together shown significant improvements in the detection of both PTT and FTT. 31,32 Our study also reported an improvement in USS sensitivity when compared with reports of detection of PTT. 29 Our study did not assess accuracy though Malik et al. 1  tears. This steady improvement in USS sensitivity highlights that significant improvement due to improved spatial resolution, technological advances and user proficiency over time will make USS more likely to be used as a stand-alone diagnostic tool in the future. [1][2][3]26,34,35 However, in cases where underlying significant glenohumeral joint osteoarthritis is present, MRI will always have superior diagnostic capabilities compared to USS.

| STRENGTHS AND LIMITATIONS
The retrospective study was chosen as an efficient way to report on data with no inconvenience to participants and eliminating loss to follow up. Retrospective research analyses in this study used preexisting results in the form of a written report. This allowed for less risk of data manipulation and misclassification bias as all measurements were taken prior to the research. 12,13 Researcher bias was also reduced due to the quantitative rather than a qualitative nature of this study, allowing for less observer

| CONCLUSION
Advances in USS shoulder imaging due to increased spatial resolution has revolutionised shoulder imaging over the last 20 years. This study indicates substantial agreement for supraspinatus tear diagnosis when comparing MRI to USS performed in general imaging departments.
Given USS is less expensive and more available, it could be considered as a first line screening tool when the main question is one of tendon integrity. However, where the patient has significant underlying osteoarthritis, MRI should be included in the imaging workup prior to surgery.
As technology continues to improve with ultra-high frequency ultrasound probes, further studies to assess accuracy of supraspinatus tear detection should be evaluated, the assumption is that in future, USS will have the capacity to serve as a stand-alone diagnostic tool in accurately assessing supraspinatus tears prior to arthroscopy.

ACKNOWLEDGMENTS
Open access publishing facilitated by Central Queensland University, as part of the Wiley -Central Queensland University agreement via the Council of Australian University Librarians.

CONFLICT OF INTEREST STATEMENT
Ann Quinton is an editorial board member for Sonography and a coauthor on this article. Other declare no conflict of interest.