AJR 2003; 181:1235-1238
© American Roentgen Ray Society
Should the Routine Wrist Examination for Trauma Be a Four-View Study, Including a Semisupinated Oblique View?
Lincoln D. Russin1,
Gabrielle Bergman1,
Lucy Miller1,
William M. Griffin1,
Marc Walter2,
Mythreyi Bhargavan3 and
Jonathan Sunshine3
1 Department of Radiology, Sutter Amador Hospital, 200 Mission Blvd., Jackson,
CA 95642.
2 Department of Emergency Medicine, Mark Twain St. Joseph's Hospital, 768
Mountain Branch Rd., San Andreas, CA 95249.
3 American College of Radiology, 1891 Preston White Dr., Reston, VA 20191.
Received April 1, 2001;
accepted after revision May 14, 2003.
Supported by the ACR Technology Assessment Studies Assistance Program.
Address correspondence to L. D. Russin.
Introduction
Fractures of the distal radius are more frequently missed on initial
radiographs than any other fracture except that of the navicular bone
[1]. Routine radiographic
examination of the injured wrist usually includes three views:
posteroanterior, posteroanterior oblique, and lateral. A semisupinated oblique
projection (Figs. 1A,
1B and
2A,
2B,
2C,
2D) is suggested as an
additional view when the routine study is inconclusive
[2]. This view shows the dorsal
cortex of the radius clearly, not obscured by superimposition on the ulna as
it appears on the lateral view. This view was described in 1926
[3] but has not been generally
accepted as an essential part of the routine examination. Indeed, whereas the
routine three views all have been rated 9 on a scale of 9 in the American
College of Radiology (ACR) Appropriateness Criteria for effectiveness
in imaging acute hand and wrist trauma
[2], the semisupinated oblique
view has a rating of only 2.
View larger version (149K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A. —Four radiographic images of right wrist of 32-year-old man
who fell with his hand extended. Posteroanterior radiograph shows mild
cortical irregularity of radius but no definite fracture.
|
|
View larger version (111K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2C. —Four radiographic images of right wrist of 32-year-old man
who fell with his hand extended. Lateral radiograph shows no definite
abnormality. Dorsal cortex of radius is obscured by superimposed ulna.
|
|
Subjective experience at a community hospital at which the semisupinated
oblique view is part of the routine examination suggests that this view is
particularly sensitive for detecting distal radius fractures. We wanted to
document the relative sensitivity of each of the four views for fracture
detection and to assess whether it is beneficial to make the routine
radiographic examination a four-view study.
Materials and Methods
After collecting 550 consecutive wrist examinations performed over a
7-month period at a community hospital at which the routine trauma study is
typically a four-view examination, we selected all cases in which an acute
fracture of any bone was diagnosed and four views were available for review.
Fifty-four examinations met these criteria and composed the study sample for
this retrospective analysis.
Three radiologists at another community hospital (a musculoskeletal
specialist with 15 years of experience and two general radiologists with 14
and 32 years of experience) and an emergency room physician with 20 years of
experience agreed to review the 54 examinations. These reviewers knew only
that the patients had wrist pain after trauma and that this study examined the
relative sensitivity of each view for detecting fractures.
To avoid the likelihood that a reviewer's perception of a fracture on one
view might influence interpretation of the other views in that examination, we
cut all radiographs to separate the views so that reviewers would evaluate
only one view at a time. A coded number identified each view. The 216 views (4
x 54 = 216) were mixed and placed in a box. In separate sessions, the
reviewers randomly selected individual views for interpretation and filled out
a data sheet for each view, noting the presence or absence of fractures of the
radius, ulna, navicular bone, other carpal bones, or metacarpals. They graded
each detected fracture as 1 (subtle), 2 (positive), or 3 (obvious). Reviewers
assigned a grade of 0 (negative) if they detected no fractures.
First, we analyzed the resulting data to find the number of fractures of
each bone detected on each view. In the subset of cases in which all reviewers
agreed that a radial fracture was present on at least one of the four views,
the authors determined the sensitivity of each view for detecting radial
fractures. Next, they identified how often each view was the only view showing
the fracture when the other three views were negative. Finally, they found how
often a view was negative when one or more of the other views were positive
for fracture. We used z tests to calculate the statistical
significance of differences between the average proportion of fractures
detected on the semisupinated oblique view and the average proportion of
fractures detected on each of the other three views, with significance defined
at the 5% level [4].
Results
The reviewers did not find fractures of any bone in two of the 54
examinations originally reported as positive for fracture. The true number of
radial fractures in the 54 examinations could not be determined because
follow-up examinations and clinical information were not available to confirm
the diagnoses. The number of radial fractures reported by the four reviewers
in the 54 examinations ranged from 39 to 45 (average, 40.8). However, there
were only 37 cases in which all four reviewers diagnosed a radial
fracture.
In this group of 37 unequivocally positive cases, one reviewer detected
radial fractures on the semisupinated oblique view in 36 examinations. In
contrast, the same reviewer found radial fractures on only 33 lateral views,
29 posteroanterior oblique views, and 27 posteroanterior views. Three
fractures were diagnosed only on the semisupinated oblique view. Conversely,
the semisupinated oblique view showed false-negative results only once,
whereas the lateral, posteroanterior oblique and posteroanterior views showed
false-negative results in four, eight, and 10 of these positive examinations,
respectively.
The data for all four reviewers were similar and are presented in
Table 1. Each reviewer detected
the highest number of radial fractures on the semisupinated oblique view. The
semisupinated oblique view was most frequently the only view positive. Also,
for each reviewer, the semisupinated oblique view was falsely negative least
often. Analysis of sensitivity of fracture detection (average number of radial
fractures detected on each view in this group of 37 unequivocally positive
cases) indicates that the higher sensitivity of the semisupinated oblique view
is statistically significant, as is the lower percentage of false-negative
semisupinated oblique views. The higher percentage of fractures detected only
on the semisupinated oblique view could not be tested for statistical
significance.
The number of ulnar fractures reported by the four reviewers in the 54
examinations ranged from 15 to 20 (average, 18.3, Tables
2 and
3). The number of navicular
fractures ranged from 1 to 6 (average, 3.8). The reviewers found fractures of
the ulna and navicular bone more readily on the posteroanterior and
posteroanterior oblique views than on the lateral and semisupinated oblique
views (Tables 2 and
3). They identified few
fractures of the other carpal bones and metacarpals in the 54 examinations. We
did not determine the sensitivity of each view for fracture detection in these
bones.
View this table:
[in this window]
[in a new window]
|
TABLE 2 All Fractures Detected Retrospectively by Four Reviewers in 54
Examinations Initially Diagnosed at Another Institution as Positive for at
Least One Fracture
|
|
Discussion
The semisupinated oblique view of the wrist appears to be the most
sensitive view for detecting radial fractures. Explanation for this finding
may be that the semisupinated oblique view shows the dorsal cortex of the
distal radius most clearly (Fig.
2D). It is this cortex that is primarily subjected to compressive
stress associated with a fall on an outstretched hand; the resulting fracture
of this cortex is one of the most commonly encountered fractures
[5]. Review of the
false-negative semisupinated oblique views revealed that these fractures were
usually of the anterior (volar) aspect of the distal radial metaphysis, which
is better visualized on the posteroanterior oblique and lateral views.
The lateral view is not as sensitive as the semisupinated oblique view
because the dorsal cortex of the radius is sometimes superimposed over the
ulna, obscuring visualization of the fracture site. However, the lateral view
has value for orthopedists in estimating the degree of impaction and
angulation of the fractures.
The posteroanterior and posteroanterior oblique views are the least
sensitive for detecting radial fractures, but the most sensitive for detecting
ulnar and navicular fractures (Tables
2 and
3). These bones are better
visualized on these views and are obscured by superimposed radius or other
carpal bones on the lateral and semisupinated oblique views. The value of
including a posteroanterior oblique view in trauma radiography of extremities
has recently been discussed
[6].
The reviewers' evaluations of the four separate views may not reflect the
way they would have evaluated the examinations normally, with all views
included on one or two sheets of film. Comparison of views would enable
reviewers to detect fractures that they might have missed on the
posteroanterior or posteroanterior oblique views if they had known that
fractures were present on the semisupinated oblique view or lateral view.
Our study has methodologic limitations that make our findings suggestive
but not definitive. Ours was not a prospective study showing added benefit
from including the semisupinated oblique view into the routine wrist trauma
examination. Negative findings were not included in the study sample;
therefore, specificity of fracture detection could not be measured.
Furthermore, the possibility that false-positive examinations might be
increased by the addition of the semisupinated oblique view was not
evaluated.
Ideally, a diagnostic test should be evaluated with reference to an
independent criterion (e.g., comparison of the sensitivity of biopsy methods
in patients with known malignancy). In our study, the diagnosis of
"fracture" depended on unanimous agreement among the reviewers
evaluating the radiographic examinations, thereby weakening the statistical
analysis of differences between the radiographic projections. Independent
confirmation of fracture diagnosis using bone scanning, MRI, or other
assessment was not available. Nevertheless, the data clearly demonstrate that
the semisupinated oblique view allows the detection of radial fractures more
frequently than the other three views routinely obtained.
It is often feasible to take four views of the wrist without using more
film than is generally exposed for a three-view study. We did not calculate
the cost of adding a semisupinated oblique view to the routine examination. It
would probably be more costly to bring patients back for additional views if
symptoms suggest that a fracture was missed on an initial three-view
examination.
In conclusion, our study suggests that the semisupinated oblique view
provides increased sensitivity for detecting radial fractures, which are
missed more frequently than any fracture except navicular fractures
[1]. Thus, we believe that the
routine wrist radiographic examination for trauma should be a four-view study,
including the semisupinated oblique view.
Acknowledgments
We thank Tushar P. Patel for work on an earlier version of this research,
Leah Russin for statistical analysis and encouragement in completion of this
work, and F. Frank Zboralske for editing the manuscript and providing
invaluable advice on presenting the data. We also thank Rebecca Lewis for
statistical analysis.
References
- Hyland-McGuire P, Guly HR, Hughes PM. Double take: fracture fishing
in accident and emergency practice. J Accid Emerg Med1997; 14:84
–87[Abstract]
- American College of Radiology. Appropriateness criteria,
2001. Reston, VA: American College of Radiology,2001
- McBride E. Wrist joint injuries: a plea for greater accuracy in
treatment. J Okla State Med Assoc1926; 19:67
–70
- Fisher LD, van Belle G. Biostatistics: a methodology for
the health sciences. New York: Wiley, 1993:185
–189
- Rogers L, ed. Radiology of skeletal trauma,
2nd ed. New York: Churchill Livingstone, 1992:841
–842
- De Smet AA, Doherty MP, Norris MA, et al. Are oblique views needed
for trauma radiography of the extremities? AJR1999; 172:1561
–1565[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
|
|
 
R. D. Welling, J. A. Jacobson, D. A. Jamadar, S. Chong, E. M. Caoili, and P. J. L. Jebson
MDCT and Radiography of Wrist Fractures: Radiographic Sensitivity and Fracture Patterns
Am. J. Roentgenol.,
January 1, 2008;
190(1):
10 - 16.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Introduction
Materials and Methods
