Original Research
Musculoskeletal Imaging
June 2, 2021

FEVER: The Flexed Elbow Valgus External Rotation View for MRI Evaluation of the Ulnar Collateral Ligament in Throwing Athletes—A Pilot Study in Major League Baseball Pitchers

Abstract

To listen to the podcast associated with this article, please select one of the following: iTunes, Google Play, or direct download.
Please see the Author Video associated with this article.
BACKGROUND. Ulnar collateral ligament (UCL) injuries are common in throwing athletes owing to repetitive extreme valgus stress during overhead throwing maneuvers. Conventional positioning for elbow MRI provides suboptimal rendering of the UCL.
OBJECTIVE. The purpose of this prospective pilot study was to assess the effect of flexed elbow valgus external rotation (FEVER) on ulnotrochlear joint space measurement and reader evaluation of the UCL when the FEVER view is incorporated into standard elbow MRI of throwing athletes.
METHODS. A total of 44 Major League Baseball pitchers underwent elbow MRI including standard sequences and a coronal fat-saturated proton density–weighted sequence in the FEVER view. To achieve the FEVER view, specific positioning maneuvers are performed, and sandbags are placed to immobilize the elbow in valgus stress so that the UCL can be visualized parallel to its long axis. Patients recorded pain during FEVER on a scale of 0 (none) to 10 (maximal). Two radiologists independently evaluated standard and FEVER images to measure the ulnotrochlear joint space, assess confidence in UCL-related findings, and assess the UCL as normal or abnormal.
RESULTS. Pain during FEVER was rated 0 by 29 patients, 1–3 by 11 patients, 4–7 by four patients, and 8–10 by no patients. Intrareader agreement on ulnotrochlear joint space measurement was higher for FEVER (intraclass correlation coefficient [ICC], 0.92) than standard (ICC, 0.54) views. Averaged between readers, the mean increase in ulnotrochlear joint space in the FEVER compared with the standard view was 1.80 mm (95% CI, 1.58–2.03). Confidence was higher for the FEVER than for the standard view for reader 1 in assessment of the UCL as normal versus abnormal (mean increase in confidence, 0.40), intensity of abnormal signal (0.39), injury grade (1.04), and retraction (0.25) and for reader 2 in assessment of the UCL as normal versus abnormal (0.50), location of abnormal signal (0.46), intensity of abnormal signal (0.51), injury grade (0.96), and retraction (0.53). Readers 1 and 2 classified three and two additional UCLs as abnormal on FEVER view compared with standard view images; neither reader classified any UCL as abnormal on standard view but normal on FEVER view images.
CONCLUSION. The increased joint space width confirmed elbow valgus stress in the FEVER view. Diagnostic confidence increased, and additional UCLs were identified as abnormal.
CLINICAL IMPACT. Use of the FEVER view may improve MRI evaluation of the UCL in throwing athletes.

HIGHLIGHTS

Key Finding
Among 44 MLB pitchers undergoing elbow MRI, the FEVER view, compared with standard view, yielded increased ulnotrochlear joint space width (mean increase, 1.80 mm) and increased confidence for three of five UCL-related findings (mean increase, −0.14 to 0.98); two readers identified three and two additional UCLs as abnormal with FEVER.
Importance
FEVER achieves valgus stress as indicated by joint space widening and improves diagnostic confidence, potentially improving MRI evaluation of the UCL in throwing athletes.
The ulnar collateral ligament (UCL) spans the medial ulnohumeral joint and is critical for elbow stabilization and minimization of valgus stress, especially during overhead throwing motions [1]. The valgus-stabilizing portion of the UCL consists of anterior and posterior bundles. The anterior bundle is the stronger component and is more important for valgus stability at the elbow, contributing most of the resistance to valgus stress [2, 3]. The anterior and posterior bands of the anterior bundle resist valgus stress with elbow extension and flexion, respectively [4].
Acute and chronic UCL injuries are common in throwing athletes owing to the repetitive extreme valgus stress generated by overhead throwing maneuvers [5]. Overhead throwing places stress on the posterior band of the anterior bundle in the 60–120° flexed positions of late cocking and acceleration [4]. Repeated UCL tears can cause chronic valgus instability, pain, and dysfunction that necessitate surgical repair or reconstruction [6]. A study of over 5000 baseball athletes showed a 25% rate of UCL reconstruction among Major League Baseball (MLB) pitchers and a 15% rate among minor league pitchers [7, 8].
Physical examination maneuvers to achieve valgus stress of the elbow yield inconsistent results [9]. Alternatively, elbow radiography performed with stress maneuvers for throwing athletes has been reported [9]. A study of the use of a Telos radiographic stress device for imaging of asymptomatic professional baseball pitchers [10] showed greater medial joint space widening with stress in throwers' dominant elbow than in the nondominant elbow. A study of baseball players who underwent stress radiography after UCL reconstruction [11] similarly showed greater joint space opening in the dominant elbow.
Studies [12, 13] have also assessed stress evaluation of the UCL with ultrasound (US). One study [14] of a gravity stress method in collegiate baseball athletes showed significant medial joint space widening in the dominant compared with the nondominant elbow. Another US valgus stress study of asymptomatic baseball pitchers [12] showed ulnohumeral widening differences between dominant and nondominant elbows. A longitudinal study of gravity stress US in the assessment of medial elbow joint laxity in asymptomatic professional baseball pitchers [13] showed a significant increase in medial joint space widening in players who went on to have UCL injuries.
MRI is the reference standard for the diagnosis of UCL injury. MLB protocols often include MRI screening of pitchers to help predict future performance and risk of UCL injury, although such predictions from standard MRI have variable results [15, 16]. This difficulty relates in part to suboptimal rendering of the UCL in the imaging planes obtained with conventional patient positioning for elbow MRI. Combining MRI with stress evaluation by means of a different modality has been attempted. In a study of baseball players with medial elbow pain [17], accuracy for identifying UCL tear relative to surgical correlation was higher for the combination of MR arthrography and valgus stress US than for either modality alone. An additional study of baseball players with prior UCL injury showed highest accuracy in predicting rehabilitation outcomes from the combination of MRI and stress US [18].
To our knowledge, elbow valgus stress has not been described as part of an elbow MRI examination. Valgus stress during MRI would avoid the time and expense of combining MRI with stress radiography or US and also avoid potential interpretive challenges in comparing the modalities. This approach also combines the high spatial resolution of MRI with functional assessment of the ulnotrochlear joint. We developed the flexed elbow valgus external rotation MRI view (which we named the FEVER view) for assessment of the UCL in throwing athletes. The FEVER view is designed to produce valgus stress on the ulnotrochlear joint, minimize patient movement, and yield high-resolution long-axis parallel imaging of the UCL. In this pilot study, we evaluated the effect of use of the FEVER view on ulnotrochlear joint space measurement and reader evaluation of the UCL when it is incorporated into standard elbow MRI examinations of throwing athletes.

Methods

Patient Population

This HIPAA-compliant prospective study received institutional review board approval. Written and oral consent was obtained from each participant. The initial study eligibility criteria were referral of a professional baseball pitcher from one of three MLB teams for elbow MRI of the throwing arm between May and December 2019 for evaluation of medial elbow symptoms (acute and/or chronic pain or discomfort) or as part of a routine team physical evaluation. For study participants, elbow MRI included FEVER view imaging performed in addition to the standard sequences. Patients were eligible regardless of having undergone prior elbow surgery. Among 52 potentially eligible patients, five were excluded owing to age younger than 18 years and three because they chose not to participate. No patient was excluded because of body habitus. The final sample included 44 men (mean age, 23.1 years; range, 18–36 years). At MRI, 14 patients had no symptoms, and 30 had symptoms. Six patients had undergone UCL reconstruction before the MRI examination.

MRI Acquisition and FEVER View

Elbow MRI was performed with a 3-T system (Magnetom Vida, Siemens Healthineers). Standard sequences were performed with a dedicated transmit and receive 16-channel elbow flex coil with the arm in the overhead elbow extended “superman” position. Acquired sequences of the elbow and the planes relative to the humerus included coronal T1-weighted (TR/TE, 600–700/10), fat-saturated (FS) proton density–weighted (PDW) (TR/TE, 3000/40–50), and FS T2-weighted (TR/TE, 4000/80–120) imaging; axial T1-weighted and FS PDW imaging; and sagittal T1-weighted and FS T2-weighted imaging. All sequences were performed at 2-to 3-mm slice intervals with 0.3-mm interslice gaps.
FEVER view acquisition (Fig. 1) was performed by trained technologists before or immediately after standard sequences with the same MRI equipment and elbow coil, though with the patient in the lateral decubitus position lying on the affected extremity. The shoulder was immobilized in external rotation (ER) in the ipsilateral decubitus position, and the shoulder and elbow were flexed to 90° with the palm up. The elbow was elevated approximately 4 inches (10.2 cm) above the table with two foam pads and a sandbag. Two 3.8-pound (1.7 kg) sandbags (7.6 pounds [3.4 kg] total) were placed on the distal forearm just proximal to the wrist to produce valgus stress and immobilization, resulting in approximately 10–20° elbow-to-forearm valgus angulation. The elbow was placed within the isocenter of the magnetic field, or as near as possible to the isocenter given body habitus considerations. The dedicated elbow flex coil was placed around the positioned elbow to optimize signal-to-noise ratio, contrast, and spatial resolution.
Fig. 1A —Flexed elbow valgus external rotation (FEVER) view acquisition.
A, Photograph shows volunteer lying on side for proper positioning for FEVER view. Flexed elbow is elevated, and sandbags are used to induce valgus stress. Elbow coil is not shown.
Fig. 1B —Flexed elbow valgus external rotation (FEVER) view acquisition.
B, 20-year-old male pitcher with normal elbow. Coronal fat-saturated proton density–weighted FEVER view MR image shows normal anterior bundle of ulnar collateral ligament (arrow) and 2.9-mm ulnotrochlear articular width (black line).
Fig. 1C —Flexed elbow valgus external rotation (FEVER) view acquisition.
C, Same pitcher as in B. Sagittal scout image shows flexed elbow position. Lines 1 and 2 denote approximately 35° scan angle for FEVER view.
An oblique coronal FS PDW sequence (TR/TE, 3000/40–50) was performed with a 256 × 256 matrix and 2-mm slice intervals with no interslice gap (Fig. 1B). Images were acquired through the UCL with an oblique angle of approximately 35°, approximating the course of the anterior bundle of the UCL in the flexed elbow position, as determined from the sagittal scout image (Fig. 1C). The FEVER view required approximately 10 minutes (4–7 minutes for positioning and scout imaging, 2–4 minutes for the single FS PDW sequence) of total additional table time.
After the MRI examination, patients were asked to report the level of pain or discomfort during FEVER on a scale of 0 (no pain) to 10 (maximal pain). Pain or discomfort was classified as mild (1–3), moderate (4–7), or severe (8–10).

MRI Analysis

Two musculoskeletal radiologists (M.S. and P.L.) with 26 and 32 years of experience independently reviewed all examinations. The radiologists first reviewed the standard images for an individual patient, including all sequences and planes, and then immediately reviewed the standard images together with the FEVER images for that patient. The radiologists were not blinded to patients' clinical information or the type of sequence, given the serial evaluation of standard and FEVER views. For each view, readers measured the ulnotrochlear joint space width between the subchondral bone surface of the ulna and the subchondral bone surface of the opposing humerus at the approximate mid-line of the ulnotrochlear joint on the slice located along the approximate course of the anterior bundle, as assessed on the scout image (Fig. 1C). The coronal FS PDW or FS T2-weighted sequence was used for this purpose for the standard views. In addition, for each view, the UCL was categorized as normal or abnormal.
A normal UCL was defined as the presence of a ligament with homogeneously low signal intensity. An abnormal UCL was defined as the presence of abnormal signal and/or various amounts of fiber obliteration. When the UCL was classified as abnormal in standard or FEVER views, the following additional findings were recorded for that view: periligamentous edema (present vs absent), location of abnormal signal (proximal vs distal), intensity of abnormal signal (high vs intermediate), injury grade (1–5 scale), and retraction (present vs absent). Injury grades were scored as 1, sprain (indistinct diffuse or localized intermediate or high signal); 2, low grade (localized intermediate or high signal involving 40% or less of UCL thickness); 3, intermediate grade (localized intermediate or high signal involving more than 40% but less than 60% of UCL thickness); 4, high grade (localized intermediate or high signal involving 60% or more of UCL thickness); or 5, complete tear (full-thickness UCL fiber obliteration on two or more consecutive coronal images). Finally, the readers recorded their confidence in whether the UCL was normal or abnormal for each elbow, confidence in the individual findings for elbows deemed abnormal, and an overall confidence score. The following 5-point Likert scale was used: 1, very low; 2, low; 3, moderate; 4, high; 5, very high.

Statistical Analysis

Pain scores during FEVER were tabulated for symptomatic and asymptomatic patients. The association between the ulnotrochlear joint space measurement in the standard and FEVER views was estimated by means of linear regression analysis based on the means of the two readers' measurements. The ulnotrochlear joint space measurements were compared between FEVER and standard views in the full sample and among symptomatic patients and asymptomatic patients by means of paired t tests. The mean patient-level change in reader confidence between standard MRI and FEVER views was computed for presence of an abnormal UCL, other individual findings, and overall confidence. The frequency of the presence of an abnormal UCL and of individual findings was compared between standard and FEVER views. Single-rating absolute agreement two-way mixed model [19] intraclass correlation coefficients (ICCs) were used to estimate interreader agreement on ulnotrochlear joint space measurements. Weighted Cohen kappa coefficients were used to assess interreader agreement on binary and ordinal variables [20, 21]. Evidence of statistical significance was evaluated on the basis of p values and 95% CIs at an alpha value of .05. Analyses were performed with R software (version 3.6.1, R Foundation) [22], psych package [23].

Results

All 44 enrolled patients were able to complete the MRI examination, including standard and FEVER views. During FEVER, a total of 29 (65.9%) patients experienced no discomfort, 11 (25.0%) reported mild discomfort, four (9.1%) reported moderate discomfort, and none reported severe discomfort. Pain scores were similar between the symptomatic patients (no discomfort, nine [64.2%]; mild discomfort, four [28.6%]; moderate discomfort, one [7.1%]) and the asymptomatic patients (no discomfort, 20 [66.7%); mild discomfort, seven [23.3%]; moderate discomfort, three [10.0%]).
Interreader agreement on ulnotrochlear joint space measurement was significantly greater for the FEVER view (ICC, 0.92 [95% CI, 0.87–0.95]) than for the standard view (ICC, 0.54 [95% CI, 0.34–0.69]) (Table 1). The ulnotrochlear joint space on FEVER view images was linearly related to the ulnotrochlear joint space on the standard view MR images (regression slope, 1.8; standard error, 0.4; t = 4.9; p < .001) (Fig. 2), but the FEVER view accounted for less than one-half of the variation from the standard view (r2 = 0.346). The mean ulnotrochlear joint space values for readers 1 and 2 and for the mean of the two readers were 2.10 mm (95% CI, 2.00–2.20), 1.95 mm (95% CI, 1.85–2.05), and 2.03 (95% CI, 1.94–2.12) for the standard view and 3.90 mm (95% CI, 3.64–4.17), 3.76 mm (95% CI, 3.67–3.86), and 3.83 mm (95% CI, 3.57–4.09) for the FEVER view.
TABLE 1: Interreader Agreement of Two Readers on Findings on Standard View and Flexed Elbow Valgus External Rotation (FEVER) View MR Images
MRI FindingStandard ViewFEVER View
Ulnar collateral ligament (normal vs abnormal)0.68 (0.50–0.83)0.71 (0.54–0.85)
Periligamentous edema (present vs absent)0.70 (0.53–0.84)0.61 (0.41–0.79)
Location of abnormal signal (proximal vs distal)0.82 (0.68–0.92)0.77 (0.60–0.90)
Intensity of abnormal signal (high vs intermediate)0.61 (0.42–0.74)0.41 (0.18–0.60)
Injury grade (1–5)0.75 (0.61–0.84)0.79 (0.67–0.87)
Retraction (present vs absent)0.55 (0.33–0.75)0.63 (0.44–0.80)
Ulnotrochlear joint space measurement (mm)0.54 (0.34–0.69)0.92 (0.87–0.95)

Note—Results are weighted Cohen kappa correlation coefficient (aside from intraclass correlations for ulnotrochlear joint space measurement). Values in parentheses are 95% CIs. Bold type indicates the 95% CIs for the standard and FEVER views do not overlap each other, denoting a significant difference.

Fig. 2 —Scatterplot shows relation between ulnotrochlear joint space measurements in standard and flexed elbow valgus external rotation (FEVER) views. Observations are coded on basis of two readers' evaluations of ulnar collateral ligament: abnormal (both views, both readers), n = 34; normal (both views, both readers), n = 3; normal in standard view (both readers), abnormal in FEVER (both readers), n = 2; readers disagreed, n = 5. Diagonal indicates least-squares regression line.
According to the mean of the measurements by the two readers, the mean increase in the ulnotrochlear joint space on the FEVER view images compared with the standard view was 1.80 mm (95% CI, 1.58–2.03; t = 15.9; df = 43; p < .001). Also according to the mean of the measurements by the two readers, the mean ulnotrochlear joint space on the standard view images was 2.06 mm (95% CI, 1.92–2.21) in symptomatic patients and 2.01 mm (95% CI, 1.90–2.12) in asymptomatic patients. Furthermore, the mean ulnotrochlear joint space on the FEVER view images was 3.80 mm (95% CI, 3.49–4.11) in symptomatic patients and 3.84 mm (95% CI, 3.49–4.19) in asymptomatic patients. The mean increase in the ulnotrochlear joint space on the FEVER view compared with the standard view images was not significantly different (t = −1.2; df = 42; p = .25) between symptomatic (mean difference, 2.00 mm [95% CI, 1.85–2.14]) and asymptomatic (mean difference, 1.72 mm [95% CI, 1.61–1.82]) patients.
Table 2 compares reader confidence for the two views. For reader 1, confidence was significantly higher (based on nonoverlap of 95% CIs with zero) for the FEVER than for the standard view for assessment of the UCL as normal versus abnormal (mean increase in confidence, 0.40 (95% CI, 0.19–0.63]), intensity of abnormal signal (0.39 [95% CI, 0.17–0.72]), injury grade (1.04 [95% CI, 0.76–1.29]), retraction (0.25 [95% CI, 0.01–0.45]), and overall confidence (1.09 [95% CI, 0.90–1.29]). For reader 2, confidence was significantly higher for the FEVER than for the standard view for assessment of the UCL as normal versus abnormal (0.50 [95% CI, 0.31–0.69]), location of abnormal signal (0.46 [95% CI, 0.13–0.70]), intensity of abnormal signal (0.51 [95% CI, 0.28–0.66]), injury grade (0.96 [95% CI, 0.68–1.11]), retraction (0.53 [95% CI, 0.30–0.75]), and overall confidence (0.78 [95% CI, 0.54–1.01]).
TABLE 2: Reader Confidence in Findings on Standard View and Flexed Elbow Valgus External Rotation (FEVER) View MR Images and Mean Within-Patient Differences in Confidence Between the Two Views
MRI FindingReader 1Reader 2
Mean Confidence (Standard)Mean Confidence (FEVER)Mean Difference (FEVER - Standard)Mean Confidence (Standard)Mean Confidence (FEVER)Mean Difference (FEVER - Standard)
Ulnar collateral ligament (normal vs abnormal)4.55 (4.30–4.79)4.95 (4.89–5.02)0.40 (0.19–0.63)4.41 (4.20–4.62)4.91 (4.82–5.00)0.50 (0.31–0.69)
Periligamentous edema4.68 (4.47–4.88)4.54 (4.31–4.77)−0.14 (−0.41 to 0.10)4.51 (4.26–4.77)4.71 (4.51–4.91)0.20 (−0.13 to 0.45)
Location of abnormal signal4.81 (4.64–4.98)4.87 (4.74–5.00)0.06 (−0.11 to 0.22)4.49 (4.19–4.79)4.95 (4.87–5.02)0.46 (0.13–0.70)
Intensity of abnormal signal4.46 (4.22–4.7)4.85 (4.73–4.96)0.39 (0.17–0.72)4.41 (4.19–4.62)4.92 (4.83–5.01)0.51 (0.28–0.66)
Injury grade3.70 (3.43–3.97)4.74 (4.52–4.96)1.04 (0.76–1.29)3.86 (3.67–4.06)4.82 (4.69–4.94)0.96 (0.68–1.11)
Retraction4.49 (4.23–4.74)4.74 (4.55–4.94)0.25 (0.01–0.45)4.41 (4.18–4.63)4.97 (4.92–5.03)0.53 (0.30–0.75)
Overall confidence3.80 (3.58–4.01)4.89 (4.79–4.98)1.09 (0.90–1.29)4.02 (3.86–4.19)4.80 (4.55–5.04)0.78 (0.54–1.01)

Note—Values in parentheses are 95% CIs. Positive mean difference values (all except periligamentous edema for reader 1) denote higher confidence in FEVER view. Sample sizes are as shown in Table 3, aside from overall confidence, which was scored for all patients. Confidence was scored from 1 (least confident) to 5 (most confident). Bold type denotes CIs of the differences between views that do not overlap zero, indicating a statistically significant difference in in confidence between standard and FEVER views.

The two readers agreed that the UCL was abnormal in both standard and FEVER views of 34 (77.3%) patients and that the UCL was normal in both standard and FEVER views of three (6.8%) patients. For two (4.5%) patients, both readers classified the UCL as normal in the standard view but abnormal in the FEVER view. For the other five (11.4%) patients, readers disagreed on at least one view. Overall, using the FEVER view, readers 1 and 2 scored three and two additional UCLs as abnormal. Figure 2 shows a scatterplot comparing ulnotrochlear joint space measurements between standard and FEVER views. Each point is coded in terms of whether the readers deemed the UCL normal or abnormal in each view.
Table 3 summarizes the frequency of the individual findings in each view for each reader. Periligamentous edema was considered present by reader 1 in 19 (48.6% [95% CI, 32.2–65.3]) instances in the standard view and in 16 (41.0% [95% CI, 26.0–57.8]) in the FEVER view and by reader 2 in 19 (51.4% [95% CI, 34.7–67.8]) instances in standard view and in 14 (36.8% [95% CI, 22.3–54.0]) in the FEVER view. Retraction was considered present by reader 1 in eight (21.6% [95% CI, 10.4–38.7) instances in the standard view and in seven (17.9% [95% CI, 8.1–34.1]) in the FEVER view and by reader 2 in 11 (29.7% [95% CI, 16.4–47.2]) instances in the standard view and in five (13.2% [95% CI, 4.9–28.9]) in the FEVER view. However, differences between views for these features and for location of abnormal signal, intensity of abnormal signal, and injury grade were not different between views on the basis of overlapping 95% CIs. Figures 35 show examples of differences in UCL evaluation between the standard and FEVER views.
TABLE 3: Comparison of Standard and Flexed Elbow Valgus External Rotation (FEVER) Views in Terms of Frequency of MRI Findings by Each Reader
MRI FindingReader 1Reader 2
Standard ViewFEVER ViewStandard ViewFEVER View
No.%No.%No.%No.%
Ulnar collateral ligament abnormal37/4484.1 (69.3–92.8)39/4488.6 (74.6–95.7)37/4484.1 (69.3–92.8)38/4486.4 (72.0–94.3)
Periligamentous edema present19/3748.6 (32.2–65.3)16/3941.0 (26.0–57.8)19/3751.4 (34.7–67.8)14/3836.8 (22.3–54.0)
Abnormal signal distal7/3718.9 (8.6–35.7)8/3920.5 (9.9–36.9)7/3718.9 (8.6–35.7)9/3823.7 (12.0–40.6)
Abnormal signal hyperintense19/3751.4 (34.7–67.8)22/3956.4 (39.8–71.8)20/3754.1 (37.1–70.2)20/3852.6 (36.0–68.7)
Injury grade > 230/3781.1 (64.3–91.4)31/3979.5 (63.1–90.1)21/3756.8 (39.6–72.5)20/3852.6 (36.0–68.7)
Retraction present8/3721.6 (10.4–38.7)7/3917.9 (8.1–34.1)11/3729.7 (16.4–47.2)5/3813.2 (4.9–28.9)

Note—Values in parentheses are 95% CIs. Findings other than an abnormal ulnar collateral ligament (UCL) were classified only for patients with an abnormal UCL. Overlapping 95% CIs between views indicate the absence of a significant difference.

Fig. 3A —20-year-old male pitcher with acute medial elbow pain after pitching.
A, Coronal standard (A) and flexed elbow valgus external rotation (FEVER) (B) MR images show low-grade partial ulnar collateral ligament tear (arrow, B), which is more evident on FEVER image. Periligamentous edema (arrowhead) is evident in both views. FEVER view shows mild ulnotrochlear joint space widening to 4.3 mm (black line, B).
Fig. 3B —20-year-old male pitcher with acute medial elbow pain after pitching.
B, Coronal standard (A) and flexed elbow valgus external rotation (FEVER) (B) MR images show low-grade partial ulnar collateral ligament tear (arrow, B), which is more evident on FEVER image. Periligamentous edema (arrowhead) is evident in both views. FEVER view shows mild ulnotrochlear joint space widening to 4.3 mm (black line, B).
Fig. 4A —32-year-old male pitcher with medial elbow pain and decreased pitching velocities for 6 months.
A, Coronal standard (A) and flexed elbow valgus external rotation (FEVER) (B) MR images both show central high signal (arrow) in proximal ulnar collateral ligament (UCL). FEVER view shows marked ulnotrochlear joint space widening to 6.2 mm (black line, B). Valgus instability was confirmed intraoperatively during UCL reconstruction.
Fig. 4B —32-year-old male pitcher with medial elbow pain and decreased pitching velocities for 6 months.
B, Coronal standard (A) and flexed elbow valgus external rotation (FEVER) (B) MR images both show central high signal (arrow) in proximal ulnar collateral ligament (UCL). FEVER view shows marked ulnotrochlear joint space widening to 6.2 mm (black line, B). Valgus instability was confirmed intraoperatively during UCL reconstruction.
Fig. 5A —23-year-old male pitcher with acute medial elbow pain.
A, Coronal standard MR image shows high-intensity signal distortion, retraction, and periligamentous edema of proximal ulnar collateral ligament (UCL), indicating high-grade partial tear (arrow).
Fig. 5B —23-year-old male pitcher with acute medial elbow pain.
B, Coronal flexed elbow valgus external rotation (FEVER) MR image shows more localized high-signal tear (arrow) of proximal UCL, less retraction, and less periligamentous edema with moderate ulnotrochlear joint space widening (to 4.9 mm; black line) than standard image does.
For normal versus abnormal UCL, presence of periligamentous edema, location of abnormal signal, intensity of abnormal signal, injury grade, and presence of retraction, interreader agreement ranged from a kappa value of 0.55 to 0.82 for the standard view and from 0.41 to 0.79 for the FEVER view (Table 1). There was no evidence that interreader agreement was significantly different between the two views for any of these features on the basis of overlapping 95% CIs.

Discussion

In our study of MLB pitchers undergoing MRI, FEVER achieved elbow valgus stress, as confirmed by the significant increase in the ulnotrochlear joint space in this view. Elbow valgus stress during MRI has not been previously shown, to our knowledge. Valgus stress during diagnostic testing has particular value for overhead throwing athletes [5]. Though valgus stress has been used for evaluating the UCL with radiography [10, 24] and US [12, 13, 18], the clinical utility of achieving valgus stress with these modalities remains substantially limited owing to patient and observer variability and suboptimal spatial and soft tissue contrast resolution of the small, inherently low contrast UCL. The FEVER view for MRI is intended to yield reproducible elbow valgus stress, which facilitates functional UCL evaluation of throwing athletes. When the FEVER view was used in our sample, confidence scores increased for both readers for differentiation of normal from abnormal UCLs and for a range of findings, including intensity of abnormal signal, injury grade, and presence of retraction.
For acquisition of FEVER view images, the patient is placed in the ipsilateral decubitus position with the elbow and shoulder in fixed 90° flexion with the shoulder in ER with the palm up. This position approximates the arm position in the late cocking phase of pitching within the constraints of the MRI gantry and provides a flat base to accommodate the sandbags used to produce valgus stress. Although it is difficult to apply variable shoulder ER for this view owing to the substantial space and geometric constraints of the MRI system, the immobilized position of the shoulder ER may facilitate achieving elbow valgus stress.
The effect of shoulder rotation on the elbow in a sample of high school throwing athletes who did not have symptoms has been described [25]. Bilateral shoulder ER angles were compared between the elbow in fixed extension and 90° flexion to determine elbow valgus laxity compared with standard valgus stress during US. In that study, shoulder ER angle differences were significantly associated with US evaluation of valgus laxity, suggesting that elbow valgus laxity is induced by shoulder rotation with the elbow at 90° of flexion, which owing to bony constraints does not occur with the elbow extended. The effect of elbow valgus laxity on shoulder rotation was also described in a cadaveric study in which the pronator musculature, posterior band anterior bundle, and entire anterior bundle were progressively severed [9]. Elbow valgus laxity and shoulder ER increased to 4.1° after sectioning of the posterior band and anterior bundle and to 11° after total anterior bundle sectioning. Because the FEVER view is obtained in fixed near maximum ER with elevation of the elbow and application of weighted valgus stress, we expected to achieve substantial valgus stress across the elbow joint, thereby mitigating (if not eliminating) any counteracting effect of shoulder ER.
The moderate association between the ulnotrochlear joint space measurements in standard and FEVER views indicates the complementary information provided by these values, possibly owing to intrinsic UCL structural or functional characteristics. Extrinsic factors—such as the difference in elbow position between views, differences between views in the image angle or slice selection, or the valgus stress itself—could also contribute to variation in measurements. Increased interreader agreement on ulnotrochlear joint space measurements in the FEVER view compared with standard view may relate to the increase in ulnotrochlear joint space in the FEVER view, which may allow more precise cursor placement and decreased measurement error.
The lack of a significant overall difference in ulnotrochlear widening with valgus stress between symptomatic and asymptomatic throwers is consistent with prior reports of highly variable ulnotrochlear laxity in throwing athletes [1013]. In a study of asymptomatic pitchers, Shanley et al. [13] reported that an ulnohumeral joint space measurement of at least 5.6 mm increased by a factor of 6 the likelihood of the need for UCL reconstruction within 1 year. Ten percent of patients in the sample had presumed abnormal gapping; this finding suggests high baseline variability in these athletes. In a study of 736 pitchers, Ciccotti et al. [12] found that joint space widening increased a mean of 0.78 mm over time. In a study of baseball players undergoing stress radiography before UCL reconstruction, Bruce et al. [11] reported increased joint opening of 0.4 mm in patients with UCL injuries and 0.6 mm in those with complete tears. We observed that increase in the ulnotrochlear joint space in the FEVER view relative to the standard view was nonsignificantly greater in symptomatic than in asymptomatic patients. The importance of this finding is uncertain given the lack of prior studies of use of this stress technique in symptomatic patients and the lack of clinical and surgical corroboration and performance information. The impact is also uncertain because the MRI examinations were performed at various points during the baseball season, and we did not record how long each player had been active in throwing activities.
Six players who had undergone UCL reconstruction surgery reported no significant difference in pain or discomfort during FEVER compared with players with no history of reconstruction. Nonetheless, valgus stress was achieved by FEVER, as confirmed by ulnotrochlear joint space widening. This shows that patients who have undergone UCL reconstruction can be comfortably evaluated in the FEVER elbow position, which is important given the increasing prevalence of UCL reconstruction among MLB pitchers. Graft signal characteristics may be heterogeneous on anatomic imaging in both intact and insufficient reconstructions, potentially increasing the value of functional integrity assessment in these patients.
There was a small increase in the number of abnormal UCLs identified with the FEVER view. It is unknown from our study whether these cases represent increased detection of injured UCLs with the FEVER view or an increase in false-positive findings. Further studies with clinical, surgical, or performance follow-up are needed to determine diagnostic accuracy. Nonetheless, variability in elbow laxity creates challenges in the interpretation of all stress tests in the general and overhead athlete populations, not just for the FEVER view at MRI. In addition, although the difference was not significant, the readers reported fewer findings of periligamentous edema and retraction on FEVER view images. Periligamentous edema may have implications in determination of injury acuity, and ligament retraction may have implications for surgical management. The differences in these findings between views may represent traction across the UCL in the valgus stress position that in turn causes tightening and straightening of the ligament fibers, elimination or reduction of contour abnormalities, or flattening of fluid adjacent to the ligament. Because substantial abnormalities have been observed on imaging of asymptomatic patients [26], further investigation is needed to determine whether the differences for these findings represent fewer false-positive or more false-negative findings with the FEVER view.
This preliminary study had limitations. Most notably, it was conducted without other stress imaging (US or radiography), physical examination stress tests, performance data, or surgical correlation, preventing determination of sensitivity, specificity, and accuracy of the MRI findings. Imaging was not performed on patients' contralateral elbow or on control patients, precluding comparison with patients' nonthrowing elbow and precluding assessment of the normal variation in valgus laxity in the stressed and unstressed positions. Additionally, the extent of valgus stress was not quantified with a dynamometer or other standard bio-mechanical tools. Furthermore, the images were not evaluated at random, and readers evaluated FEVER view images jointly with the previously evaluated standard view images. We used this approach because the FEVER view was intended to achieve ulnotrochlear joint space widening between the standard and valgus stress elbow positions as an addition to rather than a replacement of standard elbow MRI sequences. Inclusion of more readers and blinding of readers to the patient's position (standard vs FEVER) would result in a more comprehensive assessment of the impact of the FEVER view on reader confidence and interpretation.
This pilot study assessed whether FEVER produces elbow valgus stress during MRI. Use of the FEVER view improved interreader agreement on ulnotrochlear joint space measurements and improved reader confidence for a range of abnormalities. The findings support use of the FEVER view as a practical addition to standard elbow MRI protocols for achieving elbow valgus stress in throwing athletes so that functional information can be collected to complement the high-resolution anatomic assessment with MRI. Further studies with larger sample sizes, comparison with the nonthrowing elbow, and clinical, performance, and surgical correlation are required to determine the accuracy and predictive value FEVER view MRI.

References

1.
Regan WD, Korinek SL, Morrey BF, An KN. Biomechanical study of ligaments around the elbow joint. Clin Orthop Relat Res 1991; 271:170–179
2.
Dugas JR, Ostrander RV, Cain EL, Kingsley D, Andrews JR. Anatomy of the anterior bundle of the ulnar collateral ligament. J Shoulder Elbow Surg 2007; 16:657–660
3.
Morrey BF, An KN. Functional anatomy of the ligaments of the elbow. Clin Orthop Relat Res 1985; 201:84–90
4.
Fortenbaugh D, Fleisig GS, Andrews JR. Baseball pitching biomechanics in relation to injury risk and performance. Sports Health 2009; 1:314–320
5.
Wong TT, Lin DJ, Ayyala RS, Kazam JK. Elbow injuries in adult overhead athletes. AJR 2017; 208:[web]W110–W120
6.
Hibberd EE, Brown JR, Hoffer JT. Optimal management of ulnar collateral ligament injury in baseball pitchers. Open Access J Sports Med 2015; 6:343–352
7.
Camp CL, Conte S, D'Angelo J, Fealy S. Epidemiology of ulnar collateral ligament reconstruction in major and minor league baseball pitchers: comprehensive report on 1,313 cases. Orthop J Sports Med 2017; 5:2325967117S2325900369
8.
Conte SA, Fleisig GS, Dines JS, et al. Prevalence of ulnar collateral ligament surgery in professional baseball players. Am J Sports Med 2015; 43:1764–1769
9.
Mihata T, Safran MR, McGarry MH, Abe M, Lee TQ. Elbow valgus laxity may result in an overestimation of apparent shoulder external rotation during physical examination. Am J Sports Med 2008; 36:978–982
10.
Ellenbecker TS, Mattalino AJ, Elam EA, Caplinger RA. Medial elbow joint laxity in professional baseball pitchers: a bilateral comparison using stress radiography. Am J Sports Med 1998; 26:420–424
11.
Bruce JR, Hess R, Joyner P, Andrews JR. How much valgus instability can be expected with ulnar collateral ligament (UCL) injuries? A review of 273 baseball players with UCL injuries. J Shoulder Elbow Surg 2014; 23:1521–1526
12.
Ciccotti MG, Atanda A Jr, Nazarian LN, Dodson CC, Holmes L, Cohen SB. Stress sonography of the ulnar collateral ligament of the elbow in professional baseball pitchers: a 10-year study. Am J Sports Med 2014; 42:544–551
13.
Shanley E, Smith M, Mayer BK, et al. Using stress ultrasonography to understand the risk of UCL injury among professional baseball pitchers based on ligament morphology and dynamic abnormalities. Orthop J Sports Med 2018; 6:2325967118788847
14.
Sasaki J, Takahara M, Ogino T, Kashiwa H, Ishigaki D, Kanauchi Y. Ultrasonographic assessment of the ulnar collateral ligament and medial elbow laxity in college baseball players. J Bone Joint Surg Am 2002; 84:525–531
15.
Garcia GH, Gowd AK, Cabarcas BC, et al. Magnetic resonance imaging findings of the asymptomatic elbow predict injuries and surgery in Major League Baseball pitchers. Orthop J Sports Med 2019; 7:2325967118818413
16.
Gutierrez NM, Granville C, Kaplan L, Baraga M, Jose J. Elbow MRI findings do not correlate with future placement on the disabled list in asymptomatic professional baseball pitchers. Sports Health 2017; 9:222–229
17.
Roedl JB, Gonzalez FM, Zoga AC, et al. Potential utility of a combined approach with US and MR arthrography to image medial elbow pain in baseball players. Radiology 2016; 279:827–837
18.
Kim NR, Moon SG, Park JY, Choi JW, Oh KS. Stress ultrasound in baseball players with ulnar collateral ligament injuries: additional value for predicting rehabilitation outcome. J Shoulder Elbow Surg 2017; 26:815–823
19.
Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull 1979; 86:420–428
20.
Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit. Psychol Bull 1968; 70:213–220
21.
Fleiss JL, Cohen J, Everitt BS. Large sample standard errors of kappa and weighted kappa. Psychol Bull 1969; 72:323–327
22.
R Core Team website. R: a language and environment for statistical computing. R Foundation for Statistical Computing. 2018. www.R-project.org/. Accessed July 6, 2021
23.
Revelle WR. psych: procedures for personality and psychological research. R Project website. CRAN.R-project.org/package=psych. 2017. Accessed July 6, 2021
24.
O'Driscoll SW, Lawton RL, Smith AM. The “moving valgus stress test” for medial collateral ligament tears of the elbow. Am J Sports Med 2005; 33:231–239
25.
Yasui K, Mihata T, Takeda A, Watanabe C, Kinoshita M. A new manual method for assessing elbow valgus laxity. Sports Med Arthrosc Rehabil Ther Technol 2012; 4:11
26.
Bucknor MD, Stevens KJ, Steinbach LS. Elbow imaging in sport: sports imaging series. Radiology 2016; 279:12–28

Information & Authors

Information

Published In

American Journal of Roentgenology
Pages: 1176 - 1183
PubMed: 34076462

History

Submitted: February 5, 2021
Revision requested: February 16, 2021
Revision received: March 19, 2021
Accepted: May 14, 2021
Version of record online: June 2, 2021

Keywords

  1. baseball
  2. retraction
  3. ulnar collateral ligament
  4. ulnotrochlear joint
  5. valgus stress

Authors

Affiliations

Pamela Lund, MD
SimonMed Imaging, Scottsdale, AZ
Gary L. Waslewski, MD
Arizona Sports Medicine Center, Scottsdale, AZ
Ken Crenshaw, BS
Arizona Sports Medicine Center, Scottsdale, AZ
Mari Schenk, MD
Arizona Sports Medicine Center, Scottsdale, AZ
Graham Munday, BS
SimonMed Imaging, Scottsdale, AZ
Thomas Knoblauch, MS
Imgen LLC, 3560 E Flamingo Rd, Ste 100, Las Vegas, NV 89121
University of Nevada, Las Vegas, Las Vegas, NV
Alan Rodriguez, MD
Imgen LLC, 3560 E Flamingo Rd, Ste 100, Las Vegas, NV 89121
Travis Snyder, DO
Imgen LLC, 3560 E Flamingo Rd, Ste 100, Las Vegas, NV 89121
University of Nevada, Las Vegas, Las Vegas, NV
Touro University Nevada, Henderson, NV
HCA Healthcare, Nashville, TN
Mountain View Hospital, Las Vegas, NV

Notes

Address correspondence to T. Knoblauch ([email protected]).
The authors declare that they have no disclosures relevant to the subject matter of this article.

Metrics & Citations

Metrics

Citations

Export Citations

To download the citation to this article, select your reference manager software.

Articles citing this article

View Options

View options

PDF

View PDF

PDF Download

Download PDF

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media