Citation
Sande, M. A. J. van de. (2008, February 14). Rotator cuff degeneration in the rheumatoid shoulder : 'the issue is soft tissue'. Retrieved from https://hdl.handle.net/1887/12603
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/12603
Note: To cite this publication please use the final published version (if applicable).
Chapter 3.2
Proximal Migration can be measured accurately on standardized anteroposterior radiographs.
M.A.J. van de Sande, P.M. Rozing
Clin Orthop Relat Res. 2006 Feb;443:260-5
Summary
During the last decade, investigators of several studies have stressed the importance of early diagnosis and treatment of rotator cuff tears to improve outcome of surgical repair.
Proximal migration of the humeral head is thought to be indicative for rotator cuff disorders. We wanted to assess the accuracy of proximal migration measurement on anteroposterior radiographs. Computed tomography scans and anteroposterior radiographs of 43 shoulders in 26 patients were taken using the same protocol. Proximal migration was measured as the acromiohumeral interval and the upward migration index. The mean absolute difference for the upward migration index was only 0.06 (standard deviation, 0.07), this was less then 5% of the mean upward migration index measured on computed tomography reformations. The correlation coefficient showed a strong correlation between computed tomography and radiographic measurements, especially for the upward migration index (correlation coefficient, 0.82). Our results indicate that measurement of subacromial space on anteroposterior radiographs, controlled for positioning, scaling, and individual differences by using the upward migration index, presents an accurate measurement for proximal migration.
Introduction
In 1962, Golding suggested measurement of the subacromial space using plain radiographs was a sensitive diagnostic tool to evaluate rotator cuff disorders.33Since then, proximal migration of the humeral head was thought to be indicative of a rotator cuff tear, but not diagnostic.13; 36-39The prevalence of partial rotator cuff tears has been estimated between 20% and 30% and was correlated with age.132 Yamanaka and Matsumoto concluded that tears were likely to progress to full thickness with increasing age if left untreated.133 During the last decade, investigators of several studies have stressed the importance of early diagnosis and treatment of rotator cuff tears to improve the outcome of surgical repair.15; 134-136Goutallier et al stated that fatty degeneration is an important prognostic factor in rotator cuff surgery and presumed a relationship between fatty degeneration of the rotator cuff muscles and proximal migration of the humeral head.15 A subacromial space narrower than 5 mm has been considered pathologic and strongly indicative for Supraspinatus tendon rupture.33-39; 137; 138Numerous subsequent authors used the same principle to measure proximal migration of the humeral head.33-39; 137-139They measured the acromiohumeral interval as the smallest distance between the dense cortical bone marking the undersurface of the acromion and the most proximal articular cortex of the humeral head. Hirooka et al. described the upper migration index in 1996; they divided the distance from the center of the humeral
head to the undersurface of the acromion by the radius of the humeral head.13 Compared with the acromio-humeral interval, this ratio made data comparison between patients more reliable.13; 140To assess the influence of fatty degeneration and rotator cuff tears on proximal migration of the humeral head, we ascertained accuracy of measurement of the subacromial space on standardized radiographs. We hypothesized the upward migration index could provide an accurate measure for proximal migration of the humeral head.
Materials and methods
After the institutional review boards approved the cross-sectional study protocol, patients were selected prospectively based on the following criteria: (1) a clinical diagnosis of rheumatoid arthritis (RA) according to the American Rheumatism
Association criteria11 as having rotator cuff disorders (eg, rotator cuff tear and fatty degeneration) is seen commonly in rheumatoid shoulders; (2) patients older than 50 years. The age limit was chosen to impose the smallest risk from radiation exposure in a research setting (effective dose, 1.6 mSv) and still include enough patients with proximal migration of the humeral head because of rotator cuff pathology; and (3) no prior trauma or surgery to the shoulder. Between January 2003 and July 2004, 26 consecutive patients with rheumatoid arthritis (43 shoulders) were entered into our study. Patients were included after their treating physician ordered a bilateral anteroposterior (AP) radiograph to assess their shoulder complaints. All patients signed Informed consent forms. There were six men and 20 women with an average age of 66 years (range, 50–80 years). Thirty-one shoulders were symptomatic [pain and loss of range Figure 1. A radiograph shows proximal migration (A = undersurface acromion; C = center of the humeral head; R = radius of the humeral head; AH = acromiohumeral interval).
of motion (ROM)]. The mean interval between the diagnosis of RA and the computed tomography (CT) scan was 13 years (range, 1–40 years). A standard protocol AP radiograph was taken of all patients with the patient in the supine position, slightly turned to the image side (20°), and the arm in external rotation with the palm facing forward.37 Film-focus distance was measured at 115 cm and a 15° craniocaudal tilt was used to project the undersurface of the acromion perpendicular. This created a true AP projection 90° toward the glenohumeral joint. All radiographs were taken in a clinical setting in the presence of the principal investigator. Proximal migration was measured using the upward migration index where the smallest distance between the center of the humeral head to the undersurface of the acromion (CA) was divided by the radius of the humeral head (R) (Fig 1).13A circle fit was used to determine the center of the humeral head (Figure 1).
This ratio corrects for a possible magnification error, introduced during image acquisition, producing a comparable measure for proximal migration.13; 105; 141The acromiohumeral interval was measured as the smallest distance between the dense cortical bone marking the undersurface of the acromion and the most proximal articular cortex of the humeral head. All radiographs also were scored for progression of rheumatoid disease using the score described by Larsen et al.142All patients were scanned using a 16-slice CT scanner (Toshiba Aquilion, Tokyo, Japan) using the same protocol and calibration technique for each patient. The scanning parameters for this protocol were 120 kVp, 125 mAs, field of view 250 mm, and detector pitch of 15. A reconstruction filter (FC) 12 and Raster Artefact Suppression (RASP) were used, producing a 512 × 512 matrix slice thickness of 1 mm (slice overlap, 0.5 mm). Multiplanar reconstruction CT images were constructed in the parasagittal and coronal planes, parallel and perpendicular to the glenohumeral joint space (Figure 2). The smallest distance between the undersurface of the acromion and the most proximal part of the humeral head was pinpointed in these multiplanar reconstructions and measured in the AP plane (Figure 2). A circle fit was used to measure the largest radius of the humeral head in the AP plane controlled for the diameter in the sagittal plane throughout all multiplanar reconstructions (Figure 2). Computed tomography based measurement of the distance between the undersurface of the Table 1. The Mean and Standard Deviations for the Acromiohumeral Interval, the Radius of the humeral head, and the Upward Migration Index Measured on AP Radiographs and Parasagittal CT Slices
SD = standard deviation; CC = correlation coefficients between measurement techniques;
* = p < 0.001
Radiograph (humeral head radius) Computed Tomography Slices Patients Acromio- Humeral Upward Acromio- Humeral Upward humeral Head Migration humeral Head Migration (n=43) Interval (mm) Radius (mm) Index Interval (mm) Radius (mm) Index
Mean 8.4 27.6 1.3 6 24.1 1.2
SD 3.5 2.4 0.11 2.4 2.6 0.10
CC 0.62* 0.68* 0.82* 0.62* 0.68* 0.82*
acromion and center of the humeral head and its radius were considered the gold standard. The principal investigator was blinded to clinical information during all measurements. The parasagittal images also were used to measure fatty degeneration using mean muscle density. A histogram was constructed from all pixels in an outlined region of interest around the individual muscles borders. The mean muscle density was defined as the mean CT number (Hounsfield units) in one outlined rotator cuff muscle.143 Statistical analysis was done using SPSS statistical software (Version 11.5, SPSS Inc., Chicago, IL). To assess accuracy, the absolute difference and the standard deviation (SD) of the differences between radiographic and CT measurements were calculated. Pearson’s correlation coefficient and a Bland-Altman plot (a graphic method in which the differences between the two techniques are plotted against the averages of the two techniques) were used to evaluate the concordance and agreement between the upward migration index measured on the AP radiograph and the upward migration index measured on CT images. A p value less than 0.05 (two-sided) was considered significant.
Figure 2A–D. Computed tomography images show the radius of the humeral head and acromiohumeral interval measurements in the scapular plane parallel to the glenoid surface. (A) The oblique sagittal reformatted image parallel to the glenohumeral joint is used to pinpoint the largest radius of the humeral head. (B) The oblique coronal reformatted image perpendicular to the glenohumeral joint is used to measure the largest radius of the humeral head using a circle fit. (C) The horizontal reformatted image parallel to the glenohumeral joint is used to create the oblique coronal reformatted image perpendicular to the glenohumeral joint. (D) The oblique coronal reformatted image perpendicular to the glenohumeral joint also is used to measure the distance from the center of the humeral head to the undersurface of the acromion.
Results
The upward migration index presented the most accurate measurement for proximal migration when compared with the gold standard (CT reformations).
The mean absolute difference between the upward migration index measured on AP radiographs and CT images was only 0.06 (SD, 0.07). This was less than 5%
of the mean of the upward migration index measured on CT reconstructions (Fig 3). The mean difference between the acromiohumeral interval was 2.6 mm (SD, 2.1), 45% of the mean acromiohumeral interval measured on CT reconstructions.
Using the radius of the humeral head to scale the radiographs to the dimensions of the CT reformations, we still found a mean difference of 1.5 mm (SD, 1.6;
25%) in measurement of the subacromial space using the acromiohumeral interval (Table 1). We found concordance between both image modalities, but the upward migration index showed the strongest correlation between both techniques (Table 1) (Figure 4). An increased upward migration index did not influence this correlation as shown in the Bland-Altman plot (Figure 5). Age, gender, progression of rheumatoid disease (mean Larson classification, 2.2 (SD, 0.9)), or rotator cuff disease (Mean Muscle Density, 1.5 (SD, 0.7)) also did not influence this correlation. The partial correlation coefficient for the upward migration index in CT and radiographic images controlled for these parameters was 0.74 (p < 0.001).
Figure 3. The correlation between the Upward migration Index (UI) measured using radiographs and CT scans are presented in a scatter plot projected over the line of equality (—). The correlation coefficient was 0.82.
Discussion
Measurement of the acromiohumeral interval has long been used to assess rotator cuff disease.37However, conflicting results have been presented as several measurements of the subacromial space were related to rotator cuff disorders (Table 2).13; 135 These conflicting results may be caused by no systemic errors in measurement of the sub- acromial space (eg, angulations, rotation, scaling, patient characteristics). This error has been described, but to our knowledge, accuracy of radiographic measurement of the subacromial space has not been analyzed.105To assess if proximal migration can be a reliable indicator for rotator cuff disorders, we first needed to assess the accuracy of this radiographic measurement. There are several issues in interpreting the data. A standard protocol should be used in acquiring the AP radiographs and caution should be taken in measuring upward migration when suboptimal AP radiographs are used. Patient positioning (eg, upright-supine) might influence the subacromial space, and therefore Table 2. Mean outcome of subacromial space measurements in literature (Acromiohumeral Interval in AP Radiographs, MRI, CT and arthroscopy)
Distance measured:
a The smallest distance between the dense cortical bone marking the under surface of the acromion and the most proximal articular cortex of the humeral head.
b The smallest distance between the dense cortical bone marking the under surface of the acromion and the most proximal articular cartilage of the humeral head.
c The smallest distance between the dense cortical bone marking the under surface of the acromion and the most proximal part of the Supraspinatus tendon.
Author / Year n Modality Diagnosis Mean
Distance Distance
Golding 1962145 150 AP Radiographa Normal distance 7-13 Petterson et al. 1984145 175 AP Radiographa Normal 10
Chen et al. 1996146 29 MRI b Impingement 6
Tillander et al. 2002147 15 Arthroscopyc Normal 16 Tillander et al. 2002147 30 Arthroscopyc Impingement 8 Azzoni et al .2004144 70 AP Radiographa Normal cuff 11 Azzoni et al .2004144 54 AP Radiographa Cuff tendinopathy 13.9 Azzoni et al .2004144 20 AP Radiographa Partial cuff tear 10.1 Azzoni et al .2004144 26 AP Radiographa Complete cuff tear 9.0 Van de Sande et al. 2006148 43 AP Radiograha Rheumatoid
arthritis 8.4
Van de Sande et al. 2006148 43 CT scana Rheumatoid
arthritis 6
should be kept constant. We could not find any literature in which supine and upright positioning in radiograph measurements were compared. All measurements were performed by the principal investigator, and therefore, they could not be controlled for interobserver or intraobserver variance. Yet investigators in earlier reports state that these measurements on AP radiographs are highly reproducible within and between investigators.13; 37; 105 The 15° caudal to cranial tilt in the radiographic view was not implemented in the reformatted CT images because of software restrictions. The 15°
caudal to cranial tilt presents the largest subacromial interval on film but does not evenly influence the radius of the humeral head. This might explain the relatively small difference between the upward migration index in CT scans and AP radiographs and the larger difference between the acromiohumeral interval measured on CT or radiographs. The magnification factor introduced by the film-focus distance on radiographs also could explain this relatively large difference. Throughout the literature, the difference in Figure 4. A Bland-Altman plot is a graphic method that shows the differences between the two imaging techniques plotted against the averages of the two techniques.
Figure 5. The absolute outcome for the upper migration index (UI) measurements using radiographs and CT scans are plotted in a bar graph. The mean difference was 0.06, less then 5% of the mean upward migration index measured on CT reformations.
subacromial space was not controlled for patient characteristics or magnification factors.
Using a ratio (upward migration index) to assess the subacromial space, we think this error is minimized and studies are made more comparable. Our results for measurement of the subacromial space concur with results of earlier reports (Table 2). Only Azzoni et al have compared sonographic evaluation of the subacromial space with radiographic measurement.144They concluded both measurements correlate but did not discuss the accuracy. To our knowledge, our study is the first in which radiographic measurement of the subacromial space was compared with a gold standard (CT). Measurement of the subacromial space using the upward migration index in standardized radiographs can be very accurate as shown by the close agreement of the two measures, the statistical correlation, and the small standard deviation of the differences between both measurements. We think direct measurement of the acromiohumeral interval on standardized AP radiographs corrected for positioning, scaling, and individual differences by using the upward migration index is a reliable and accurate indicator for proximal migration of the humeral head. The relevance and implications of measuring proximal migration needs to be studied to analyze the underlying mechanism and assess its clinical implications.
