Advertisement

Arthroscopic Debridement for Refractory Lateral Epicondylitis Results for Substantial Improvement in Tendinosis Scores and Good Clinical Outcomes: Qualitative and Quantitative Magnetic Resonance Imaging Analysis

      Purpose

      To qualify and quantify the changes in magnetic resonance imaging (MRI) signals in the extensor tendons after arthroscopic debridement for lateral epicondylitis and evaluate the association between MRI findings and temporal clinical results by comparisons between recovered and unrecovered cases.

      Methods

      Thirty-four patients with refractory lateral epicondylitis treated with arthroscopic debridement were divided into recovered (n = 24) and unrecovered (n = 10) groups according to the Japanese Orthopaedic Association-Japan Elbow Society score. This study included any patients who underwent both the pre- and postoperative MRI and excluded patients with a previous history of any elbow surgery. Pre- and postoperative MRI findings were qualitatively categorized into 4 grades, quantified by measuring the percentage of tendinopathy area, and compared between the groups.

      Results

      Preoperatively, grading scores and percentages did not show significant differences between groups (P = .050 and .519). The respective numbers of patients with grades 1, 2, 3, and 4 were 1 (4%), 3 (13%), 10 (42%), and 10 (42%) in the recovered group; and 1 (10%), 2 (20%), 7 (70%), and 0 (0%) in the unrecovered group. The average percentages in the recovered and unrecovered groups were 42.3% (73.9 mm2/168.4 mm2); and 36.5% (50.5 mm2/131.0 mm2). However, postoperatively, they were significantly lower in the recovered group than in the unrecovered group (P = .007 and .014). The numbers and percentages in the recovered and unrecovered groups were 15 (63%), 8 (33%), 1 (4%), and 0 (0%) and 17.0% (28.6mm2/169.8mm2) and 2 (20%), 3 (30%), 5 (50%), and 0 (0%) and 30.5% (39.0 mm2/131.8 mm2).

      Conclusions

      Qualitative and quantitative MRI is useful for evaluating the progress of tendon healing after arthroscopic debridement. In the recovered and unrecovered groups, improvement of tendinopathy area were 60% versus 16%, indicating that postoperative MRI findings reflect clinical outcomes.

      Level of Evidence

      IV, case series with subgroup analysis.
      Lateral epicondylitis refers to enthesopathy of extensor carpi radialis brevis (ECRB) origin, with a reported incidence ranging from 1% to 3% in the general population.
      • Burn M.B.
      • Mitchell R.J.
      • Liberman S.R.
      • Lintner D.M.
      • Harris J.D.
      • McCulloch P.C.
      Open, arthroscopic, and percutaneous surgical treatment of lateral epicondylitis: A systematic review.
      ,
      • van Kollenburg J.A.
      • Brouwer K.M.
      • Jupiter J.B.
      • Ring D.
      Magnetic resonance imaging signal abnormalities in enthesopathy of the extensor carpi radialis longus origin.
      The majority of patients are successfully treated nonoperatively; however, 4% to 11% of patients require surgical treatment.
      • Qi L.
      • Zhang Y.D.
      • Yu R.B.
      • Shi H.B.
      Magnetic resonance imaging of patients with chronic lateral epicondylitis: Is there a relationship between magnetic resonance imaging abnormalities of the common extensor tendon and the patient's clinical symptom?.
      Arthroscopic debridement of the ECRB tendon is of interest as a treatment option for severe and refractory symptoms,
      • Kuklo T.R.
      • Taylor K.F.
      • Murphy K.P.
      • Islinger R.B.
      • Heekin R.D.
      • Baker Jr., C.L.
      Arthroscopic release for lateral epicondylitis: A cadaveric model.
      ,
      • Baker Jr., C.L.
      • Murphy K.P.
      • Gottlob C.A.
      • Curd D.T.
      Arthroscopic classification and treatment of lateral epicondylitis: Two-year clinical results.
      although there are no perceived advantages of arthroscopic surgery over open or percutaneous treatment.
      • Burn M.B.
      • Mitchell R.J.
      • Liberman S.R.
      • Lintner D.M.
      • Harris J.D.
      • McCulloch P.C.
      Open, arthroscopic, and percutaneous surgical treatment of lateral epicondylitis: A systematic review.
      Magnetic resonance imaging (MRI) plays as key a role in preoperative decision-making, similar to physical clinical examinations. There is a general consensus that the degree of signal increase within the ECRB tendon origin and tendon separation indicate a degenerative process of the lateral epicondylitis. In this regard, several researchers have reported the high reliability and practicality of MRI in preoperative disease diagnosis.
      • Jeon J.Y.
      • Lee M.H.
      • Jeon I.H.
      • Chung H.W.
      • Lee S.H.
      • Shin M.J.
      Lateral epicondylitis: Associations of MR imaging and clinical assessments with treatment options in patients receiving conservative and arthroscopic managements.
      • Aoki M.
      • Wada T.
      • Isogai S.
      • Kanaya K.
      • Aiki H.
      • Yamashita T.
      Magnetic resonance imaging findings of refractory tennis elbows and their relationship to surgical treatment.
      • Chen J.
      • Wang A.
      • Xu J.
      • Zheng M.
      In chronic lateral epicondylitis, apoptosis and autophagic cell death occur in the extensor carpi radialis brevis tendon.
      However, there is a paucity of studies investigating postoperative MRI changes after debridement of the ECRB tendon. Indeed, postoperative changes in the debrided tendons are poorly characterized, and their association with clinical outcomes remains unclear.
      Elucidation of postoperative changes in ECRB tendons on MRI can be helpful in predicting the healing progress of debrided tendons and determining appropriate management strategies. Therefore, this retrospective study is aimed to qualify and quantify the changes in MRI signals in the extensor tendons after arthroscopic debridement for lateral epicondylitis and evaluate the association between MRI findings and temporal clinical results by comparisons between recovered and unrecovered cases. The hypothesis was that debrided ECRB tendons that reorganized well during the postoperative period would be detectable on MRI and would be associated with better clinical outcomes (including pain relief and functional recovery) at the time of examination.

      Methods

      Study Setting

      This retrospective study was conducted under the approval of the institutional review board. The requirement for written informed consent from patients was waived due to the retrospective nature of the study.

      Patient Enrollment

      This study retrospectively examined 72 consecutive patients with lateral epicondylitis who underwent arthroscopic debridement at a single institution from August 2009 to August 2020. The inclusion criterion was any patients who underwent both the pre- and postoperative MRI. The exclusion criterion was patients with a previous history of any elbow surgery. A total of 34 patients were enrolled in this study.
      All the patients were diagnosed with refractory lateral epicondylitis based on treatment history, physical examination, and MRI findings, which were examined by experienced orthopaedic elbow surgeons (S.M., K.T. and K.S.) with >10 years of practice. The surgical indication for refractory lateral epicondylitis was intractable symptoms accompanied by signal changes on MRI after conservative treatment for at least 6 months. All the patients presented with tenderness in the lateral epicondyle over the extensor origin and pain during resistive dorsiflexion of the wrist with the elbow in full extension (provocation test for tennis elbow). Initially, all the patients received various nonsurgical treatments, including medications, steroid injections, activity modification, physical therapy, and the use of a brace, over a period of ≥6 months. MRI was performed pre- and postoperatively to evaluate ECRB tendon integrity and determine indications for arthroscopic debridement.

      Surgical Procedures

      Under general anesthesia, the patients were placed in the prone position, and a tourniquet was inflated around the upper arm. Lactated ringer’s solution (20 mL) was injected into the elbow joint from a “soft spot” in the triangle created among the lateral epicondyle, radial head, and olecranon. Arthroscopy was performed using anteromedial, anterolateral, soft spot, and direct lateral portals. The direct lateral portal was an additional portal that was slightly lateral to the soft spot portal. Posterior and posterolateral portals were established if necessary. A standard 4.0-mm, 30° wide-angle scope was used to visualize the radiocapitellar compartment in the anteromedial portal and access the ECRB tendon origin using the anterolateral portal. Then, the ECRB origin was debrided with a 3.5-mm full-radius shaver and released from the lateral epicondyle. The synovial plicae that was interposed in the radiocapitellar joint and covered the annular ligament was also debrided. Subsequently, the remaining plicae were resected in the posterior compartment using soft spots and direct lateral portals. Intraoperative findings of the capsular tear were evaluated according to the classification proposed by Baker et al.,
      • Baker Jr., C.L.
      • Murphy K.P.
      • Gottlob C.A.
      • Curd D.T.
      Arthroscopic classification and treatment of lateral epicondylitis: Two-year clinical results.
      and the presence of plicae was recorded. The Baker classification includes type I (intact capsule), type II (linear capsular tear), or type III (complete capsular tear) lesions. The portals were closed using a single suture. Postoperative treatment consisted 10 days of immobilization in loose dressing, followed by an active program under the supervision of an occupational therapist.

      Clinical Data Collection

      Pre- and postoperative data were collected using manual chart reviews. Postoperative data also were collected at the time of postoperative MRI evaluation. Demographic information, including age, sex, height, weight, body mass index, profession, dominant limb, and duration of symptoms, was recorded. The variables compared included tenderness over the extensor origin, positive provocation test for tennis elbow, elbow (flexion/extension) range of motion measured using a goniometer, and visual analog scale score for elbow pain. The Japanese Orthopaedic Association-Japan Elbow Society (JOA-JES) elbow function score for lateral epicondylitis also was recorded pre- and postoperatively. Based on the postoperative JOA-JES score, patients were divided into the recovered and unrecovered groups, with a cutoff of 60 points. In a previous study of postsurgical results of 20 elbows following arthroscopic debridement, all the participants reported improvement in their outcomes, with a minimum JOA-JES score of 58 points (average, 90 points).
      • Wada T.
      • Moriya T.
      • Iba K.
      • et al.
      Functional outcomes after arthroscopic treatment of lateral epicondylitis.
      Based on this, a JOA-JES score of > 60 points was considered to indicate a recovered tendon/joint in the present study. Furthermore, minimal clinically significant difference in the JOA-JES score was calculated to be 33 points. The JOA-JES scoring system is detailed in Table 1.
      Table 1JOA-JES Score for Lateral Epicondylitis
      I.Pain (30 points)NoneMildModerateSevere
      302520151050
      II.Function (20 points)
      ADL (12 points)EasyDifficultImpossible
       Lift a heavy object420
       Wring a wet towel420
       Pour a glass of water420
      MMT (8 points)543210
       Elbow flexion543210
       Elbow extension332100
      III.Lateral epicondylitis (50 points)-±+++
       Tenderness over the lateral epicondyle201050
       Pain with resisted wrist extension3020100
      NOTE. Possible score ranges from 0 to 100.
      ADL, Activities of Daily Living; JES, Japan Elbow Society; JOA, Japanese Orthopaedic Association; MMT, Manual Muscle Test.

      MRI Analysis

      All the patients underwent pre- and postoperative MRI examinations using a 1.5-T or 3.0-T MRI scanner (Discovery MR750; GE Healthcare, Chicago, IL), according to a standard scanning protocol. The patients lay in the supine position, with their elbow at full extension and forearm in supination. The slice interval was set at 2 mm. The pre- and postoperative MRI findings were evaluated blindly to avoid any bias based on patient clinical outcomes. Evaluations were performed using a commercially available image-viewing software (Synapse Vincent; Fuji Film, Tokyo, Japan). Postoperative MRI was performed routinely in practice if consent was obtained. The average duration between postoperative MRI and surgery was 10.7 (range: 3-24) months.
      Signal changes at the ECRB origin were assessed pre- and postoperatively using T2-weighted fat-suppressed coronal MRI scans. Postoperative MRI scans were evaluated using slices corresponding to their preoperative counterparts by adjusting the cross-sectional alignment between slices using the Synapse Vincent software. This permitted direct comparison of the pre- and postoperative images. Pre- and postoperative ECRB tendon status was qualitatively evaluated using an MRI grading system based on the classification system reported by Walton et al.
      • Walton M.J.
      • Mackie K.
      • Fallon M.
      • et al.
      The reliability and validity of magnetic resonance imaging in the assessment of chronic lateral epicondylitis.
      for evaluation of the preoperative status of ECRB tendons. Information on tendon repair integrity for comparison with postoperative findings was also included, which was based on the Sugaya classification for the healing status of repaired rotator cuffs.
      • Sugaya H.
      • Maeda K.
      • Matsuki K.
      • Moriishi J.
      Functional and structural outcome after arthroscopic full-thickness rotator cuff repair: Single-row versus dual-row fixation.
      ,
      • Sugaya H.
      • Maeda K.
      • Matsuki K.
      • Moriishi J.
      Repair integrity and functional outcome after arthroscopic double-row rotator cuff repair. A prospective outcome study.
      MRI findings were classified into 4 categories based on the severity of tendinopathy, and the degree of separation of the tendon of the common extensor origin from the lateral epicondyle. The findings analyzed included intratendinous signal changes and morphologic alterations in normal uniform hypointense signals. MRI findings were classified as follows: grade 1 (normal), no signal changes and uniform hypointense tendon signal; grade 2 (mild), heterogeneous or focal hyperintense tendon signal; grade 3 (moderate), heterogeneous or focal hyperintense tendon signal with fluttering of the tendon; and grade 4 (severe), homogeneous or generalized hyperintense tendon signal with tendon separation (Fig 1).
      Figure thumbnail gr1
      Fig 1Qualitative magnetic resonance imaging grading system for lateral epicondylitis (T2 coronal images, right elbows). (A) Grade 1 (normal), no signal change or uniform hypointense tendon signal (arrow); (B) grade 2 (mild), heterogeneous or focal hyperintense tendon signal (arrow); (C) grade 3 (moderate), heterogeneous or focal hyperintense tendon signal with tendon fluttering (arrow); (D) grade 4 (severe), homogeneous or generalized hyperintense tendon signal with tendon separation (arrow).
      Signal alterations in the ECRB tendon on pre- and postoperative MRI were quantified by experienced orthopaedic elbow surgeon (S.M.) with >10 years of practice. The region of interest (ROI) was manually drawn using slices in qualitative assessment on T2-weighted fat-suppressed coronal MRIs. The ROI was located over the ECRB tendon origin, defined as the visible perimeter of the tendon and the border of the bone–tendon junction. The corresponding ROIs on both pre- and postoperative MRIs were used for evaluation based on careful segmentation and cross-referencing. Using this approach, the total area and signal intensity within the ROIs were measured. Local lesions exhibiting an increased signal in the tendon were quantified by calculating the percentage of the area with a signal intensity greater than quarter the intensity of the joint fluid, which produces a result equivalent to that of tendinopathy at the lateral epicondyle (Fig 2).
      Figure thumbnail gr2
      Fig 2Quantitative analysis of magnetic resonance imaging (MRI) signal changes in the extensor carpi radialis brevis tendon in adjusted cross-sectional coronal alignment between pre- and postoperative images (T2 coronal images, left elbows). (A) Preoperative MRIs with total tendinopathy area (light green) and increased signal area (dark green). (B) Postoperative MRIs with total tendinopathy area (light green) and increased signal area (dark green). High-signal intensity area in the postoperative image decreases compared with that in the preoperative image.

      Statistical Analysis

      Statistical analyses were performed using IBM SPSS Statistics for Windows, version 24.0 (IBM Corp., Armonk, NY). Statistical significance was set at P < .05 for all statistical analyses.
      Intra- and interobserver reliabilities for qualitative MRI grading scores and quantitative measurements were assessed using the intraclass correlation coefficient (ICC). The interobserver reliability of 4 observers (3 experienced orthopaedic elbow surgeons [S.M., K.T., and K.S.] with >10 years of practice and 1 orthopaedic fellow [S.M.]) was determined. The intraobserver reliability was assessed by analyzing 2 different measurements by a single experienced orthopaedic elbow surgeon (S.M.) with >10 years of practice. The agreement was classified as poor (ICC = .01 to .20), fair (ICC = .21 to .40), moderate (ICC = .41 to .60), substantial (ICC = .61 to .80), or excellent (ICC = .81 to 1.00).
      • Landis J.R.
      • Koch G.G.
      The measurement of observer agreement for categorical data.
      Pre- and postsurgical MRI findings (grading scores and percentages of the tendinopathy area) were compared using the Wilcoxon signed rank test. For further analysis, pre- and postsurgical MRI findings were compared in both recovered and unrecovered groups using the Wilcoxon signed rank test.
      To evaluate the association between MRI findings and temporal clinical results, in pre- and postoperative period, the grading scores and percentages of the tendinopathy area were compared between recovered and unrecovered groups using the Mann–Whitney U test.
      Patient demographics and pre-/intra-/postoperative findings were compared between the recovered and unrecovered groups using the Mann–Whitney U test for continuous values or the χ2 test for categorical values.
      Multinomial logistic regression analysis was performed to identify the independent association of explanatory variables with the postoperative improvement in JOA-JES scores. Explanatory variables, including pre- and postoperative percentages of the tendinopathy area and the changes between them, were entered into the multinomial logistic regression model, controlling multicollinearity using the Spearman correlation analysis.
      The associations between intraoperative Baker classifications and preoperative MRI findings, including the grading scores and percentages of the tendinopathy area, were analyzed using Spearman correlation coefficients (R). The strength of correlations was classified as slight (R < .2), low (R = .2 to .4), moderate (R = .4 to .7), or high (R > .7).
      • Guilford J.P.
      Fundamental statistics in psychology and education.
      A priori power analyses (α = .05, 1 – β = .9, 2-tailed) were conducted using G∗Power (University at Kiel, Kiel, Germany) to identify improvements in MRI grading scores of 1 ± 1.5. The minimum sample size for identifying meaningful differences was 26 patients.

      Results

      Five patients with inaccessible MRI findings and 30 patients who did not undergo postoperative MRI were excluded. Three patients were excluded because of a history of elbow trauma surgeries. The mean duration of symptoms before surgery was 25.4 months, and the mean follow-up period was 27.4 months.
      There were 24 patients in the recovered group and 10 patients in the unrecovered group. The number of patients with minimal clinically significant difference (>33 points) was 22 (92%) in the recovered group and 10 (100%) in the unrecovered group. This also validates the cutoff set at 60 points. No significant differences were observed in age (52.5 years vs 51.5 years, P = .564), sex (13 males and 11 females vs 7 males and 3 females, P = .393), and profession (10 physical and 14 non-physical workers vs 5 physical and 5 nonphysical workers, P = .656) between the 2 groups. Table 2 presents the additional comparisons of patient demographics. As with the patient demographics, pre- and intraoperative findings of the recovered and unrecovered groups were similar. In contrast, significant differences were observed in postoperative physical findings between the 2 groups, with the exception of range of motion at the elbow (Table 3).
      Table 2Patient Demographics
      VariableTotal (N = 34)Recovered Group (N = 24)Unrecovered Group (N = 10)P Value
      Demographics
       Age, y52.2 ± 10.052.5 ± 10.051.5 ± 10.4.564
       Sex, n; %.393
      Male20; 59%13; 54%7; 70%
      Female14; 41%11; 46%3; 30%
       Height, m1.64 ± 0.081.64 ± 0.071.66 ± 0.12.669
       Weight, kg63.7 ± 12.263.2 ± 12.764.8 ± 11.7.589
       Body mass index23.6 ± 4.223.6 ± 4.823.4 ± 2.2.515
       Profession, n; %.656
      Physical work15; 44%10; 42%5; 50%
      Nonphysical work19; 56%14; 58%5; 50%
       Sporting or music activity13; 38%8; 33%5; 50%.362
       Lesion on dominant limb, n; %26; 77%18; 75%8; 80%.754
       Duration of symptom, mo25.4 ± 15.923.8 ± 15.529.0 ± 17.0.305
      Table 3Pre-, Intra- and Postoperative Findings
      VariableTotal (N = 34)Recovered Group (N = 24)Unrecovered Group (N = 10)P Value
      Preoperative findings
       Extensor origin tenderness, n; %34; 100%24; 100%10; 100%>.999
       Positive provocation test for tennis elbow, n; %34; 100%24; 100%10; 100%>.999
       Range of motion,°
      Flexion136.9 ± 6.5136.9 ± 6.7137.0 ± 6.3.985
      Extension–2.4 ± 8.9–2.7 ± 8.1–6.3 ± 11.1.564
       VAS score for elbow pain, points7.5 ± 0.77.5 ± 0.77.6 ± 1.0.867
       JOA-JES score, points36.0 ± 4.036.2 ± 4.035.4 ± 4.0.341
       Grading score of tendinopathy area, n; %.050
      Grade 12; 6%1; 4%1; 10%
      Grade 25; 15%3; 13%2; 20%
      Grade 317; 50%10; 42%7; 70%
      Grade 410; 30%10; 42%0; 0%
      Intraoperative findings
       Baker classification, n; %
      Type I19; 56%14; 58%5; 50%.826
      Type II11; 32%7; 29%4; 40%
      Type III4; 12%3; 13%1; 10%
       Presence of plicae, n; %28; 82%20; 83%8; 80%.816
      Postoperative findings
       Extensor origin tenderness, n; %12; 35%5; 21%7; 70%.006
       Positive provocation test for tennis elbow, n; %8; 24%2; 8%6; 60%.001
       Range of motion,°
      Flexion139.2 ± 3.3139.8 ± 3.2138.0 ± 3.5.287
      Extension0.5 ± 4.71.6 ± 4.0-2.0 ± 5.4.133
       VAS score for elbow pain, points3.7 ± 3.32.0 ± 2.17.6 ± 2.0<.001
       JOA-JES score, points76.9 ± 19.886.0 ± 13.755.0 ± 14.3<.001
       Grading score of tendinopathy area, n; %<.001
      Grade 117; 50%15; 63%2; 20%
      Grade 211; 32%8; 33%3; 30%
      Grade 36; 18%1; 4%5; 50%
      Grade 40; 0%0; 0%0; 0%
      NOTE. Bold type indicates significant P values (P < .05).
      JES, Japan Elbow Society; JOA, Japanese Orthopaedic Association; VAS, Visual Analog Scale.
      Postoperative grading scores and percentages of the tendinopathy area on MRI were significantly better than the preoperative values (P < .001 for both): the respective numbers of patients classified as having grades 1, 2, 3, and 4 changed from 2 (6%), 5 (15%), 17 (50%), and 10 (30%) to 17 (50%), 11 (32%), 6 (18%), and 0 (0%) (Fig 3A); the average percentage of the high-signal intensity area changed from 40.5% (66.4 mm2/156.3 mm2) to 21.1% (31.7 mm2/158.3 mm2) (Fig 4A).
      Figure thumbnail gr3
      Fig 3Qualitative grading scores of tendinopathy area based on pre- and postoperative magnetic resonance imaging findings. (A) Total patients, (B) recovered group, and (C) unrecovered group. Postoperative grading scores are significantly improved when compared with preoperative values. In the recovered group, postoperative scores significantly improved; in contrast, in the unrecovered group, no significant differences are observed between the pre- and postoperative scores.
      Figure thumbnail gr4
      Fig 4Quantitative percentages of the tendinopathy area based on pre- and postoperative magnetic resonance imaging findings. (A) Total patients, (B) recovered group, and (C) unrecovered group. Postoperative percentages of tendinopathy area are significantly improved when compared with preoperative values. In the recovered group, postoperative findings significantly improved; in contrast, in the unrecovered group, no significant differences are observed between the pre- and postoperative percentages.
      In the recovered group, postoperative findings revealed significant improvements in grading scores and percentages of the tendinopathy area (P < .001 for both): the respective numbers of patients classified as having grades 1, 2, 3, and 4 changed from 1 (4%), 3 (13%), 10 (42%), and 10 (42%) to 15 (63%), 8 (33%), 1 (4%), and 0 (0%) (Fig 3B): the average percentage of the high-signal intensity area changed from 42.3% (73.9 mm2/168.4 mm2) to 17.0% (28.6 mm2/169.8 mm2), which was an improvement of 59.8% (Fig 4B).
      In contrast, no significant differences were observed between the pre- and postoperative scores and percentages in the unrecovered group (P = .564 and .114, respectively): the respective numbers of patients classified as having grades 1, 2, 3, and 4 changed from 1 (10%), 2 (20%), 7 (70%), and 0 (0%) to 2 (20%), 3 (30%), 5 (50%), and 0 (0%) (Fig 3C); the average percentage of the high-signal intensity area changed from 36.5% (50.5 mm2/131.0 mm2) to 30.5% (39.0 mm2/131.8 mm2), which was an improvement of 16.4% (Fig 4C).
      On preoperative MRI, no significant differences were observed in the grading scores and percentages of the tendinopathy area (P = .050 and .519, respectively) between the recovered and unrecovered groups, despite a trend for lower grading scores in the unrecovered group; the respective numbers of patients classified as having grades 1, 2, 3, and 4 were 1 (4%), 3 (13%), 10 (42%), and 10 (42%) in the recovered group versus 1 (10%), 2 (20%), 7 (70%), and 0 (0%) in the unrecovered group (Fig 5A); the average percentage of the high-signal intensity area were 42.3% (73.9 mm2/168.4 mm2) in the recovered group versus 36.5% (50.5 mm2/131.0 mm2) in the unrecovered group (Fig 6A).
      Figure thumbnail gr5
      Fig 5Qualitative grading scores of the tendinopathy area on magnetic resonance imaging (MRI) findings in the recovered and unrecovered groups. (A) Before surgery and (B) after surgery. On preoperative MRI, no significant differences are observed in the grading scores between the recovered and unrecovered groups; in contrast, on postoperative MRI, the grading scores are significantly lower in the recovered group than in the unrecovered group.
      Figure thumbnail gr6
      Fig 6Quantitative percentages of the tendinopathy area on magnetic resonance imaging (MRI) findings in the recovered and unrecovered groups. (A) Before surgery and (B) after surgery. On preoperative MRI, no significant differences are observed in the percentages of the tendinopathy between the recovered and unrecovered groups; in contrast, on postoperative MRI, the percentages of the tendinopathy area are significantly lower in the recovered group.
      In contrast, on postoperative MRI, the grading scores and percentages of the tendinopathy area were significantly lower (P = .007 and .014, respectively) in the recovered group than in the unrecovered group; the respective number of patients classified as having grades 1, 2, 3, and 4 was 15 (63%), 8 (33%), 1 (4%), and 0 (0%) in the recovered group and 2 (20%), 3 (30%), 5 (50%), and 0 (0%) in the unrecovered group, respectively (Fig 5B); the average percentage of the high-signal intensity area was 17.0% (28.6 mm2/169.8 mm2) in the recovered group versus 30.5% (39.0 mm2/131.8 mm2) in the unrecovered group (Fig 6B).
      Multinomial regression analysis revealed that the changes between the pre- and postoperative percentages of the tendinopathy area were independently associated with poor postoperative outcomes of JOA-JES scores (odds ratio = .763; P = .018). There was no multicollinearity among the explanatory variables.
      Correlation analyses revealed that intraoperative findings were moderately correlated with the percentages of the tendinopathy area (R = .457, P = .010).
      The ICCs for intraobserver reliability of pre- and postoperative MRI findings were 0.747 and 0.811 for grading scores and 0.815 and 0.841 for quantitative measurements, respectively (P < .001 for all), indicating excellent agreement with the exception of substantial agreement in preoperative grading scores. The ICCs for interobserver reliability of pre- and postoperative MRI findings were 0.646 and 0.782 for grading scores and 0.731 and 0.832 for quantitative measurements, respectively (P < .001 for all), indicating substantial agreement in preoperative findings and excellent agreement in postoperative findings.

      Discussion

      During the postoperative period, the debrided ECRB tendons showed homogenization of MRI signals and continuity at the extensor origin, resulting in substantial improved tendinosis scores. Furthermore, comparisons between recovered and unrecovered groups revealed that the degree of tendon integrity was in accordance with clinical recovery at the time of examination.
      Based on the current consensus, MRI is used to assess the progression of refractory lateral epicondylitis before surgical or conservative treatment. A limited number of studies have investigated postoperative MRI changes in debrided tendons. As such, the tendon healing process remains poorly described, and its association with clinical outcomes remains unclear. Postoperative tendon healing in rotator cuff repair has been well studied using MRI. Sugaya et al.
      • Sugaya H.
      • Maeda K.
      • Matsuki K.
      • Moriishi J.
      Functional and structural outcome after arthroscopic full-thickness rotator cuff repair: Single-row versus dual-row fixation.
      ,
      • Sugaya H.
      • Maeda K.
      • Matsuki K.
      • Moriishi J.
      Repair integrity and functional outcome after arthroscopic double-row rotator cuff repair. A prospective outcome study.
      classified repaired tendons into 5 categories based on repair integrity in direct relation to clinical outcomes. They concluded that shoulders with insufficient healing on MRI demonstrated inferior functional outcomes. A more recent study introducing a clinically compatible MRI acquisition method (namely, ultrashort echo time–T2∗ mapping) has quantified postoperative rotator cuff healing and suggested a relationship between healing of the tendons and clinical outcomes.
      • Xie Y.
      • Liu S.
      • Qu J.
      • Wu P.
      • Tao H.
      • Chen S.
      Quantitative magnetic resonance imaging UTE-T2∗ mapping of tendon healing after arthroscopic rotator cuff repair: A longitudinal study.
      In accordance with previous literature, this study showed that debrided ECRB tendons were detectable on MRI, and the tendon repair integrity reflected clinical outcomes at the time of examination. Nevertheless, longitudinal studies are warranted to gain further insight into the chronological changes in debrided tendons and to identify prognostic factors.
      There is no way of knowing whether the decrease in T2 high signals could be attributed to bulk debridement or to reorganization of the tendon. However, one possible piece of evidence that the T2 signal changes might be less involved with bulk debridement is the fact that the total overall area of tendon did not change before and after surgery (156.3 mm2 vs 158.3 mm2). Debrided tendons might reorganize via formation of normal tendinous tissues or scar formation, and the size of tendon might have returned to what it was before, even if the ECRB origin was released from the lateral epicondyle. Edematous or angiogenic tendinous changes at the origin of the extensor tendons are presumed to lead to signal changes on MRI.
      • Nirschl R.P.
      Elbow tendinosis/tennis elbow.
      Postoperatively, these changes are thought to be underpinned by tissue replacement with neogenetic tissues that are stimulated or regenerated by the debridement of granulation tissues. It is noted that intraoperative findings did not correlate well with preoperative MRI findings. One possible explanation for this is that the Baker classification does not take into account the aforementioned intratendinous changes.
      Previous studies have demonstrated several predictors of postoperative functional outcomes, including a high baseline pain score,
      • Haahr J.P.
      • Andersen J.H.
      Prognostic factors in lateral epicondylitis: A randomized trial with one-year follow-up in 266 new cases treated with minimal occupational intervention or the usual approach in general practice.
      presence of persistent pain,
      • Jeon J.Y.
      • Lee M.H.
      • Jeon I.H.
      • Chung H.W.
      • Lee S.H.
      • Shin M.J.
      Lateral epicondylitis: Associations of MR imaging and clinical assessments with treatment options in patients receiving conservative and arthroscopic managements.
      symptom duration >12 months, and history of previous injection
      • Knutsen E.J.
      • Calfee R.P.
      • Chen R.E.
      • Goldfarb C.A.
      • Park K.W.
      • Osei D.A.
      Factors associated with failure of nonoperative treatment in lateral epicondylitis.
      ; however, the conclusions still vary across reports. Moreover, little was known regarding the association of outcomes with postoperative imaging. In the unrecovered group, reduced improvements on MRI were associated with poor temporal outcomes. Nearly one-third of the total patients, more than that reported previously, were not fully recovered at the time of examination, which raises the possibility that the healing process in these patients may have been delayed or disrupted. In particular, the 2 patients in whom the postoperative MRI was obtained 3 months after surgery were categorized into the unrecovered group, thus supporting this theory. Otherwise, there may have been no changes in the ECRB tendon compared with the baseline values. In this regard, patients exhibiting an absence (or attenuation) of signal changes on preoperative MRI may not exhibit improvements in tendinopathy, and they should thus be considered for surgery only with caution. This was supported by the fact that the “changes” in the high-signal intensity area between pre- and postoperative measurements were independent factors for poor outcomes. Moreover, patients exhibiting residual symptoms and reduced improvement after surgery may require further follow-up or additional surgical intervention. If discrepancies are noted between postoperative symptoms and MRI findings, other factors, such as the presence of radiocapitellar instability due to iatrogenic over-debridement or psychogenic factors, may need to be considered.
      The strength of this study lies in investigating and analyzing postoperative MRI findings of the ECRB tendons after arthroscopic debridement. To date, to the best of our knowledge, there has been a lack of a uniform grading system for lateral epicondylitis that considers both pre- and postoperative findings. The proposed grading system reflects the status of tendon healing, and the high reliability of this system supports its utility. Further, this is supported by the quantitative analyses of MRI signal changes. Moreover, this study demonstrates that improvements in the MRI parameters are associated with clinical outcomes after arthroscopic debridement, which may help guiding treatment decisions before and after surgery. The current findings indicate that postoperative evaluation using MRI findings facilitates the management of patients whose conditions are not improving clinically to visualize tendon healing status and ascertain whether the pain can be attributed to something else. The present study results might show that we can refrain from unnecessary MRIs.

      Limitations

      This study has multiple limitations. First, the study was inherently limited by the retrospective nature of chart reviews, which could have introduced selection bias. Moreover, retrospective studies can only determine association and correlation but not causation. Second, there were potential errors in evaluating tendinopathy lesions arising from region-specific MRI usage, such as misclassifying lesions and dispersions during manual drawing; however, ICC analysis revealed high intra- and inter-observer agreement with respect to these. Third, there may be a referral bias, as the patients in this study sought care at a tertiary care center. Fourth, the scoring system used in this study is based on physician-reported outcomes and not the outcome directly reported by patients. As with many scoring systems that have been used for elbow disorders,
      • Longo U.G.
      • Franceschi F.
      • Loppini M.
      • Maffulli N.
      • Denaro V.
      Rating systems for evaluation of the elbow.
      it has not been validated, and there is a risk of it assessing only a few aspects of elbow function. Finally, in multinomial regression analysis, it was changes of the tendinopathy area that associated with postoperative outcomes; however, it will be very hard to judge the relative improvements in clinical settings.

      Conclusions

      Qualitative and quantitative MRI is useful for evaluating the progress of tendon healing after arthroscopic debridement. In the recovered and unrecovered groups, improvement of tendinopathy area was 60% versus 16%, indicating that postoperative MRI findings reflect clinical outcomes.

      Supplementary Data

      References

        • Burn M.B.
        • Mitchell R.J.
        • Liberman S.R.
        • Lintner D.M.
        • Harris J.D.
        • McCulloch P.C.
        Open, arthroscopic, and percutaneous surgical treatment of lateral epicondylitis: A systematic review.
        Hand (N Y). 2018; 13: 264-274
        • van Kollenburg J.A.
        • Brouwer K.M.
        • Jupiter J.B.
        • Ring D.
        Magnetic resonance imaging signal abnormalities in enthesopathy of the extensor carpi radialis longus origin.
        J Hand Surg Am. 2009; 34: 1094-1098
        • Qi L.
        • Zhang Y.D.
        • Yu R.B.
        • Shi H.B.
        Magnetic resonance imaging of patients with chronic lateral epicondylitis: Is there a relationship between magnetic resonance imaging abnormalities of the common extensor tendon and the patient's clinical symptom?.
        Medicine (Baltimore). 2016; 95e2681
        • Kuklo T.R.
        • Taylor K.F.
        • Murphy K.P.
        • Islinger R.B.
        • Heekin R.D.
        • Baker Jr., C.L.
        Arthroscopic release for lateral epicondylitis: A cadaveric model.
        Arthroscopy. 1999; 15: 259-264
        • Baker Jr., C.L.
        • Murphy K.P.
        • Gottlob C.A.
        • Curd D.T.
        Arthroscopic classification and treatment of lateral epicondylitis: Two-year clinical results.
        J Shoulder Elbow Surg. 2000; 9: 475-482
        • Jeon J.Y.
        • Lee M.H.
        • Jeon I.H.
        • Chung H.W.
        • Lee S.H.
        • Shin M.J.
        Lateral epicondylitis: Associations of MR imaging and clinical assessments with treatment options in patients receiving conservative and arthroscopic managements.
        Eur Radiol. 2018; 28: 972-981
        • Aoki M.
        • Wada T.
        • Isogai S.
        • Kanaya K.
        • Aiki H.
        • Yamashita T.
        Magnetic resonance imaging findings of refractory tennis elbows and their relationship to surgical treatment.
        J Shoulder Elbow Surg. 2005; 14: 172-177
        • Chen J.
        • Wang A.
        • Xu J.
        • Zheng M.
        In chronic lateral epicondylitis, apoptosis and autophagic cell death occur in the extensor carpi radialis brevis tendon.
        J Shoulder Elbow Surg. 2010; 19: 355-362
        • Wada T.
        • Moriya T.
        • Iba K.
        • et al.
        Functional outcomes after arthroscopic treatment of lateral epicondylitis.
        J Orthop Sci. 2009; 14: 167-174
        • Walton M.J.
        • Mackie K.
        • Fallon M.
        • et al.
        The reliability and validity of magnetic resonance imaging in the assessment of chronic lateral epicondylitis.
        J Hand Surg Am. 2011; 36: 475-479
        • Sugaya H.
        • Maeda K.
        • Matsuki K.
        • Moriishi J.
        Functional and structural outcome after arthroscopic full-thickness rotator cuff repair: Single-row versus dual-row fixation.
        Arthroscopy. 2005; 21: 1307-1316
        • Sugaya H.
        • Maeda K.
        • Matsuki K.
        • Moriishi J.
        Repair integrity and functional outcome after arthroscopic double-row rotator cuff repair. A prospective outcome study.
        J Bone Joint Surg Am. 2007; 89: 953-960
        • Landis J.R.
        • Koch G.G.
        The measurement of observer agreement for categorical data.
        Biometrics. 1977; 33: 159-174
        • Guilford J.P.
        Fundamental statistics in psychology and education.
        3rd ed. McGraw-Hill Book Company, New York1956
        • Xie Y.
        • Liu S.
        • Qu J.
        • Wu P.
        • Tao H.
        • Chen S.
        Quantitative magnetic resonance imaging UTE-T2∗ mapping of tendon healing after arthroscopic rotator cuff repair: A longitudinal study.
        Am J Sports Med. 2020; 48: 2677-2685
        • Nirschl R.P.
        Elbow tendinosis/tennis elbow.
        Clin Sports Med. 1992; 11: 851-870
        • Haahr J.P.
        • Andersen J.H.
        Prognostic factors in lateral epicondylitis: A randomized trial with one-year follow-up in 266 new cases treated with minimal occupational intervention or the usual approach in general practice.
        Rheumatology (Oxford). 2003; 42: 1216-1225
        • Knutsen E.J.
        • Calfee R.P.
        • Chen R.E.
        • Goldfarb C.A.
        • Park K.W.
        • Osei D.A.
        Factors associated with failure of nonoperative treatment in lateral epicondylitis.
        Am J Sports Med. 2015; 43: 2133-2137
        • Longo U.G.
        • Franceschi F.
        • Loppini M.
        • Maffulli N.
        • Denaro V.
        Rating systems for evaluation of the elbow.
        Br Med Bull. 2008; 87: 131-161