Imaging of the rotator cuff following repair: Human and animal models

Imaging of the rotator cuff following repair: Human and animal models Hollis G. Potter, MD,a Shari T. Jawetz, MD,b and Li Foong Foo, MD,c New York, NY

Imaging of the rotator cuff following repair may be challenging due to the type of fixation, surgical manipulation of the tissue and the presence of residual defects that may exist in the presence of good functional outcome. Both ultrasound and magnetic resonance (MR) imaging present unique advantages in evaluation of the postoperative tissue. MR imaging has superior soft tissue contrast and provides a more global shoulder assessment, including the degree of arthrosis, while ultrasound enables a more dynamic testing of the repaired tissue. Power Doppler ultrasound and new contrast agents provide insight into the degree of vascular recruitment following repair. (J Shoulder Elbow Surg 2007;16:134S-139S.) Challenges in imaging the postoperative rotator cuff

Accurate assessment of the postoperative rotator cuff is challenging due to several factors, including the size of the tear at the time of repair. In the setting of a massive rotator cuff repair, each tendon slip cannot always be reapproximated to the bony attachment. In that situation, there is a residual “defect” that is deemed of little consequence with regards to maintaining overall cuff function. As magnetic resonance (MR) imaging and ultrasound provide a sectional assessment of the cuff, evaluation in the postoperative setting may lead to factitious diagnosis of “re-tear of the previous repair” rather than a residual defect left at the time of surgery. It is therefore important that the clinician relate as much history as possible to the radiologist who images the repair tissue, including the size of the tear, any residual defect left at surgery and the use of any bioabsorbable fixation. In addition, the use of allografts, scaffolds or xenografts should be noted. Chief, Magnetic Resonance Imaging Hospital for Special Surgery Professor of Radiology Weill Medical College of Cornell University, bResearch Assistant, MRI Division, cResearch Fellow, MRI Division. Reprint requests: Hollis G. Potter, MD, MRI Division, Hospital for Special Surgery, 535 East 70th Street, New York, NY 10021 (E-mail: [email protected]). Copyright © 2007 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/2007/$32.00 doi:10.1016/j.jse.2007.02.114 a

134S

Additional factors include the degree of pre-existing tendinosis and the process of surgical manipulation, both of which can alter the underlying tendon morphology and signal characteristics or echogenicity on diagnostic MR imaging or ultrasound. Further morphologic changes induced by the repair include redirection of the tendon fibers at the reattachment site, which can affect collagen anisotropy, thereby altering the tendon’s signal characteristics and echogenicity at the attachment site. The presence of a surgical trough and/or fixation devices may also affect the tendon. Susceptibility artifact generated from metallic suture anchor fixation devices or inflammatory reactions surrounding bioabsorbable fixation may affect either MRI signal or ultrasound characteristics. Susceptibility artifact is encountered on MR imaging, and is denoted as the presence of frequency shifts generated by the easily magnetized paramagnetic or ferromagnetic metallic suture anchors compared to the less easily magnetized, diamagnetic soft tissue. This creates a focal area of signal void followed by a summation of frequency shifts, creating a bright area of signal hyperintensity that may be misinterpreted as a re-tear of the tendon (Figure 1). Slice averaging of tendon and inflammatory tissue surrounding bioabsorbable fixation devices may pose a further challenge when discerning cuff integrity (Figure 2). Signal hyperintensity or altered echogenicity may be misinterpreted as partial-thickness defects in the rotator cuff despite the presence of good functional outcome following cuff repair. In addition, persistent intrasubstance areas of delamination of the cuff that are not discerned at arthroscopy may account for some of the findings of altered cuff morphology in the absence of adverse functional outcome. Spielmann et al evaluated 30 rotator cuff tendons in 15 patients who had undergone clinically successful rotator cuff repair and had good to excellent scores on the Constant scale at follow-up evaluation.12 Using traditional MR imaging, clinically “silent” cuff tears were diagnosed in 11 out of 30 tendons that were noted to have either partial or full-thickness defects.12 No surgical correlation was provided in this study. Gaenslen et al also used MR imaging to assess 30 shoulders in twenty-nine symptomatic patients following rotator cuff surgery.2 Findings at MR imaging were com-

J Shoulder Elbow Surg Volume 16, Number 5S

Potter, Jawetz, and Foo

135S

Figure 1 45 year-old man two years following rotator cuff repair. Oblique coronal proton density MR images demonstrate increased signal adjacent to the metallic anchor (arrowhead in A) that reflects susceptibility artifact and should not be interpreted as a partial tendon tear. Note the signal normalizes on image (B), 4.6mm posterior to image (A).

Figure 2 59 year-old man following rotator cuff repair using bioabsorbable tacks, one of which was broken in the humeral head. Oblique coronal proton density MR images demonstrate an inflammatory reaction in the bursa and increased signal at the bursal margin of the cuff. Note the partially displaced head of one tack (arrow in B). Note that while the tissue immediately at the repair site is markedly inhomogeneous and more hyperintense compared to the lower signal intensity of the more medial aspect of the tendon, no re-tear or discernable gap is seen. The tissue between the tacks likely reflects residual tendon and inflammatory tissue in the adjacent bursa.

pared to surgical findings, disclosing sensitivity and specificity for full-thickness tears of 84% and 91%, respectively.2 MRI was noted to have 83% sensitivity

and specificity for detecting partial thickness tears.2 Magee et al evaluated 50 patients with persistent postoperative pain who underwent second-look ar-

136S

Potter, Jawetz, and Foo

throscopy.5 Using surgery as the standard, MR imaging was found to be 100% sensitive and 87% specific in revealing full or partial-thickness rotator cuff tears.5 Caution should be utilized in the MR evaluation of altered signal intensity in surgically repaired tendons. While the presence of a full thickness, discernible gap in the tendon is a good indicator of the presence of a full thickness defect, increased signal in a portion of the tendon may reflect tendinosis, artifact and/or postoperative remodeling that occurs following debridement or repair. The diagnosis of a partial tear is best made by the definition of a delaminated slip of tendon either on the articular or bursal margin of the cuff, rather than the presence of signal hyperintensity alone. Similar studies have also been performed using ultrasound. In the pre-operative setting, Teefey et al used ultrasound to evaluate 100 shoulders that subsequently underwent arthroscopy.13 As compared with arthroscopy, ultrasound was noted to have a sensitivity and specificity of 100% and 85%, respectively, for detecting full thickness rotator cuff tears.13 Fealy et al evaluated 50 patients who had undergone rotator cuff repair, including 28 mini-open, 14 open and 8 arthroscopic repairs, with 20 patients serving as controls.1 Ultrasound revealed that 48% of patients had a detectable rotator cuff repair defect postoperatively; these findings were not noted to correlate with either functional assessment or outcome at six months.1 Despite the lack of correlation of images with functional outcome, ultrasound has been determined to have an accuracy of 89% in assessment of postoperative cuff, based on a retrospective review of 44 consecutive patients.9 Ultrasound evaluation of the postoperative cuff can be an extremely effective primary means by which to image the tissue. It requires, however, a dedicated, high frequency (typically 10-12MHz) transducer suitable for musculoskeletal (as opposed to abdominal or pelvic) imaging, as well as an imager experienced in the operation of the ultrasound unit. Specific imaging techniques: Magnetic resonance imaging

The signal characteristics of tendon, as detected by MRI, are based on the relative mobility of free versus bound water, and the interaction of the excited hydrogen nuclei. Advantages of MR imaging over other imaging techniques in the assessment of the postoperative cuff include superior soft tissue contrast, the ability to provide a global assessment of the shoulder, including the cartilage and the labrum, and direct multiplanar capabilities. Following surgical manipulation of the cuff, there is increased propensity for mobility of water within the normally highly ordered architecture of rotator cuff tendons. This phenomenon

J Shoulder Elbow Surg September/October 2007

Figure 3 67 year-old man three months following rotator cuff repair. Oblique coronal proton density MR image shows a complete retear of the infraspinatus tendon (arrow). Note the intermediate signal intensity scar adjacent to the anchor in the greater tuberosity and the fluid intensity just lateral to the tendon (arrow).

creates areas of increased signal intensity within the cuff. The presence of partial and full-thickness re-tears of the cuff is best determined using high resolution imaging and small pixel size, which allows for superior depiction of small areas of re-tear and defects in the cuff. The use of imaging parameters that distinguish fluid intensity from scar tissue and the remaining normal tendon is also helpful (Figure 3). Fat suppression techniques “rescale” the contrast range, making fluid collection and gaps in the tendons more conspicuous. Modification of protocols in the presence of metallic fixation is necessary to reduce susceptibility artifact and yet still discern areas of partial cuff dehiscence (Figure 4). In the postoperative setting, fat suppression is best achieved by the use of a short tao inversion recovery (STIR) sequence, which provides more uniform fat suppression in the presence of an inhomogeneous magnetic field. Artifact from metallic fixation devices can be diminished by the use of a wide receiver bandwidth and altered protocol; these techniques have been shown to be effective in imaging the cuff surrounding shoulder arthroplasty as well.11 The increased signal intensity that may be seen following bioabsorbable fixation does not cause a similar disturbance of the magnetic field, but may create reactive inflammatory signal immediately adjacent to the tendon, as well as a regional synovitis or bursitis. A suggested protocol for imaging the postoperative cuff is provided in Table 1.

J Shoulder Elbow Surg Volume 16, Number 5S

Potter, Jawetz, and Foo

137S

between fat and the surrounding muscle fascicles. The Goutallier classification of fatty infiltration was initially reported for computerized tomography (CT) but has been modified for MRI and utilizes a subjective assessment of the tomographic images of muscle to describe the degree of fatty infiltration.4 This classification system includes Stage 0 (completely normal muscle), Stage 1 (some fatty streaks), Stage 2 (still more muscle than fat), Stage 3 (equal amounts of fat and muscle) and Stage 4 (more fat than muscle).4 A more robust assessment of the degree of fatty infiltration requires direct quantitation of fat. This can be accomplished using direct proton MR spectroscopy, which directly quantifies the lipid content of the rotator cuff muscles.8 The latter technique requires special pulse sequences, high magnet homogeneity, specialized post-processing software and additional scan time. Specific imaging techniques: Ultrasound Figure 4 45 year-old man two years following multiple tendon rotator cuff repair and a positive lift-off test. Axial proton density MR image shows longitudinal stripping of the subscapularis tendon, with the bulk of the tendon (arrow) located well medial to the lesser tuberosity.

It is the experience of these authors that the use of gadolinium (either by intravenous or intra-articular injection) is not necessary to reliably assess the postoperative cuff, and that inadvertent extravasation from attempted MR arthrography may obscure regional pathology. The presence or absence of a “water-tight seal” is of less importance compared to careful evaluation of tendon continuity and tissue quality; simply assessing for a “leak” in a repair negates the value of the expanded soft tissue contrast afforded by more advanced imaging techniques. Longitudinal stripping of the tendons may, upon initial inspection, result in the appearance of an intact tendon-to-bone unit. However, careful inspection discloses the longitudinal stripping of the tendon, with progressive thinning as it approaches the tuberosity, accounting for the presence of rotator cuff dysfunction in the absence of a discernible “hole” on imaging. In the presence of a full-thickness cuff tear, scar tissue may also fill the gap; when organized, the scar may create the appearance of tendon-to-bone continuity. Gerber et al has denoted this as failure in continuity, initially described in a non-clinical model.3 In addition to assessing the tendon-to-bone continuity, it is important to assess the degree of fatty infiltration of regional muscle. This may be performed subjectively by inspecting the relative amount of fat on T1-weighted images. Fat yields high signal intensity on MR imaging due to its efficient energy transfer from the excited hydrogen nuclei to the surrounding molecular environment, allowing for high contrast

Advantages of ultrasound include the ability to routinely apply loads to the healing cuff and obtain immediate real time images. Power Doppler ultrasound imaging, an assessment of blood flow that demonstrates increased dynamic range compared to conventional Doppler techniques, yields improved sensitivity and is less subject to artifacts.6,7 Using power Doppler sonography to prospectively evaluate patients following rotator cuff repair, Fealy et al noted a predictable, significant decrease in vascular scores following rotator cuff repair over time as evaluated at six weeks, three months, and six months postoperatively.1 Of note, a vascular response that decreased with time was identified in the postoperative cohort but not in the asymptomatic controls.1 (Figure 5). This study gives insight into the biology of rotator repair, establishing that blood flow is a basic tenant of rotator cuff repair. Vascularity thus may serve as an important indicator of healing that complements the standardized assessment of tendon morphology. New ultrasound contrast agents have become available and allow for enhanced conspicuity of small vessels and visualization of capillary-level flow. A preliminary assessment of asymptomatic normal rotator cuffs using these contrast agents has shown regional variations of enhancement in the normal cuff with a persistent area of low vascularity; i.e., “critical zone” at the articular and medial margin of the supraspinatus footprint.14 Recent investigations have noted an age-related decrease in vascularity and increased perfusion in all regions of the cuff with activity.14 When applied to a healing rotator cuff model, these contrast agents may be of great interest in the assessment of vascular response to cuff repair.

138S

Potter, Jawetz, and Foo

J Shoulder Elbow Surg September/October 2007

Table 1 Recommended protocol for MR imaging of the postoperative shoulder Recommended Protocol for MRI of Postoperative Shoulder at Hospital for Special Surgery Series I

Series II

Series III

Series IV

Oblique coronal fast spin echo TR 4000-6000 depending on patient size; TE 34; FOV 16 cm; matrix 512 ⫻ 384; slice thickness 3mm with no gap, bandwidth 31.25kHz; no phase wrap; NEX 2; ETL 8-14 Oblique coronal inversion recovery TR 5000-6000; TE 17; TI 150; FOV 16cm; matrix 256 ⫻ 192; slice thickness 3mm with no gap; bandwidth 31.25kHz; no phase wrap; NEX 2; ETL 6-8 Oblique sagittal fast spin echo TR 3500-4000; TE 34; FOV 16cm; matrix 512 ⫻ 224; slice thickness 4mm with 0.5mm gap; bandwidth 31.25kHz; no phase wrap; NEX 2; ETL 8-12 Axial fast spin echo TR 3500-5000; TE 34; FOV 15-16cm; matrix 512 ⫻ 384; slice thickness 3.5mm with no gap; bandwidth 31.25kHz; no phase wrap; NEX 2; ETL 8-12

BW, bandwidth**; TR, repetition time in msec; TI, inversion time in msec; ETL, echo train length; TE, echo time in msec; FOV, field of view; NEX, number of excitations. **Note that in the presence of metal fixation, a receiver bandwidth of at least 62.5kHz (for General Electric Systems) is recommended. To convert to Hz/pixel (Philips or Siemens systems), use the conversion formula Hz/pixel ⫽ (GE ⫻ 2000)/matrix in the frequency direction.

Figure 5 A, postoperative ultrasound performed at 1 month (time #1), demonstrating relatively high-flow state in an intact rotator cuff repair; B, postoperative ultrasound performed at 6 months (time #3) on the same patient, demonstrating relatively low-flow state in an intact rotator cuff repair. Scan was performed in the same arm position. This is a longitudinal image of the supraspinatus tendon and the color flow is demonstrated in the center of the more hyperechoic tendon. (Fealy S, Adler RS, Drakos MC, Kelly AM, Allen AA, Cordasco FA, Warren RF, and O’Brien SJ, American Journal of Sports Medicine 34(1):120-127 1 Copyright 2006 by American Orthopaedic Society for Sports Medicine, Reprinted by Permission of Sage Publications, Inc.).

Imaging of non-clinical models

Challenges in the assessment of a non-clinical rotator cuff model include the lack of standardized coils for the animal rotator cuff and altered joint morphology. In addition, the smaller tissue parts require both superior spatial and slice resolution. The use of imaging, however, is essential to provide an objective assessment of the entire tendon-to-bone interface, providing a macroscopic view of the surgically manipulated region. Rodeo et al used MRI to assess rotator cuff healing in a sheep model.10 In this study, 72 skeletally mature ewes underwent three treatments of the tendon-bone interface, including an osteoinductive bone protein extract and a type I collagen sponge carrier, collagen sponge carrier alone or no implant.

Imaging of the ewes was performed at six and 12 weeks, and disclosed a gap filled with repair tissue and new bone that consistently formed between the end of the repaired tendon and the bone (Figure 6). Of note, MR imaging showed that the volumes of newly formed bone and soft tissue in the tendon-tobone gap were greater in the growth factor-treated animals compared to the sponge-controlled group at the studied time points.10 In this study, imaging was performed with an in-plane resolution of 254 ␮m in the frequency direction by 369␮m in the phase direction, with a slice thickness of 1.6mm.10 This spatial resolution is much higher than is typically employed in the clinical assessment of rotator cuff and allows for high conspicuity of tendon-to-bone healing.10 The

Potter, Jawetz, and Foo

J Shoulder Elbow Surg Volume 16, Number 5S

139S

niques in the assessment of vascular recruitment, as well as more sophisticated MR techniques such as diffusion tensor imaging to assess collagen anisotropy and alignment in the remodeled tendon. REFERENCES

Figure 6 Axial MR image of a sheep 6 weeks following rotator cuff repair using an osteoinductive growth factor demonstrates intermediate signal intensity scar tissue between the lateral margin of the tendon (thick arrow) and the bone trough. Note the bone marrow edema at the original repair site and the suture that was located at the original repair site (thin arrows).

investigators further noted that upon histologic analysis, a well organized fibrous tissue was seen connecting tendon to bone, which easily could have been misinterpreted as an intact repair.10 The gap in the tendon-to-bone healing was only evident by the macroscopic view afforded by the imaging techniques. The increased signal intensity in the reparative tissue reflected the increased mobility of water and was indicative of poorly organized matrix. Specific MRI protocols and surface coils for animal studies will vary depending on the model studied. High in plane resolution (less than 300 ␮m in the frequency direction) and thin slices (ⱕ 2mm) are recommended, with an intermediate echo time (TE) that allows for differential contrast between fluid, articular cartilage and the repaired tendon. Similar to clinical studies, the addition of a fat-suppressed sequence is of use in discerning small fluid collection, inflammatory reaction and bone marrow edema at the enthesis. Conclusion

Challenges exist for both ultrasound and MR imaging in the assessment of the postoperative rotator cuff. Careful attention to technique is important and a detailed history provided by the surgeon is very helpful. Imaging, however, provides an important objective assessment of tendon-to-bone healing, regardless of subjective patient functional status. Future directions include refinement of ultrasound contrast tech-

1. Fealy S, Adler RS, Drakos MC, Kelly AM, Allen AA, Cordasco FA, et al. Patterns of vascular and anatomical response after rotator cuff repair. Am J Sports Med 2006;34(1):120-127. 2. Gaenslen ES, Satterlee CC, Hinson GW. Magnetic resonance imaging for evaluation of failed repairs of the rotator cuff. Relationship to operative findings. J Bone Joint Surg 1996;78(A): 1391-1396. 3. Gerber C, Schneeberger AG, Perren SM, Nyffeler RW. Experimental rotator cuff repair. J Bone Joint Surg 1999;81A(9):12811290. 4. Goutallier D, Postel JM, Bernageau J, Lavau L, Voisin MC. Fatty muscle degeneration in cuff ruptures. Pre- and postoperative evaluation by CT scan. Clin Orthop Rel Res 1994;304:78-83. 5. Magee TH, Gaenslen ES, Seitz R, Hinson GA, Wetzel LH. MR imaging of the shoulder after surgery. Am J Roentgenol 1997; 168:925-928. 6. Newman JS, Adler RS, Rubin JM. Power Doppler sonography: Use in measuring alterations in muscle blood volume after exercise. Am J Roentgenol 1997;168:1525-1530. 7. Newman JS, Adler RS, Bude RO, Rubin JM. Detection of soft tissue hyperemia; Value of power Doppler sonography. Am J Roentgenol 1994;163:385-389. 8. Pfirrmann CWA, Schmidt MR, Zanetti M, Jost b, Gerber C, Hodler J. Assessment of fat content in supraspinatus muscle with proton MR spectroscopy in asymptomatic volunteers and patients with supraspinatus tendon lesions. Radiology 2004;232(3):709716. 9. Prickett WD, Teefey SA, Galatz LM, Calfee RP, Middleton WD, Yamaguchi K. Accuracy of ultrasound imaging of the rotator cuff in shoulders that are painful postoperatively. J Bone Joint Surg 2003;85A(6):1084-1089. 10. Rodeo SA, Potter HG, Kawamura S, Turner AS, Kim HJ, Atkinson B. Augmentation of rotator cuff tendon healing using an osteoinductive growth factor. American Academy of Orthopaedic Surgeons – 69th Annual Meeting Proceedings. Dallas, TX. February 13-17, 2002. 11. Sperling JW, Potter HG, Craig EV, Flatow E, Warren RF. MRI of the painful shoulder arthroplasty. J Shoulder Elbow Surg 2002; 11(4):315-321. 12. Spielmann AL, Forster BB, Kokan P, Hawkins RH, Janzen DL. Shoulder after rotator cuff repair: MR imaging findings in asymptomatic individuals – Initial experience. Radiology 1999;213: 705-708. 13. Teefey SA, Hasan SA, Middleton WD, Patel M, Wright RW, Yamaguchi K. Ultrasonography of the rotator cuff: A comparison of ultrasonographic and arthroscopic findings in one hundred consecutive cases. J Bone Joint Surg 2000;82A(4):498-504. 14. Warren RF, Rudzki JR, Adler RS, Kadrmas WR, Verma N, Pearle A, et al. Contrast-enhanced ultrasound characterization of the vascularity of the rotator cuff tendon: Age- and activity-related changes in the intact asymptomatic rotator cuff. American Shoulder and Elbow Surgeons (ASES) – 23rd Annual Meeting Proceedings. Chicago, IL. September 13-15, 2006.