Imaging the scaphoid problem : a diagnostic strategy for suspected scaphoid fractures
Beeres, F.J.P.
Citation
Beeres, F. J. P. (2008, May 14). Imaging the scaphoid problem : a diagnostic strategy for suspected scaphoid fractures. Retrieved from https://hdl.handle.net/1887/12857
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/1887/12857
Note: To cite this publication please use the final published version (if applicable).
Chapter
6
F.J.P. Beeres1 M. Hogervorst2 S.J. Rhemrev1 S. Le Cessie3 J.W. Arndt4 M.P.M. Stokkel5 K.A. Bartlema6 J.F. Hamming6
1Department of Surgery, Medical Centre Haaglanden, The Hague, the Netherlands
2Department of Surgery, Gelre Hospitals, Apeldoorn, the Netherlands
3Department of Medical Statistics and Bioinformatics, Leiden University Medical Centre, Leiden, the Netherlands
4Department of Nuclear Medicine, Medical Centre Haaglanden, The Hague, the Netherlands
5Department of Nuclear Medicine, Leiden University Medical Centre, Leiden, the Netherlands
6Department of Surgery, Leiden University Medical Centre, Leiden, the Netherlands
Reliability of bone scintigraphy for suspected scaphoid fractures
Based on:
Reliability of bone scintigraphy for suspected scaphoid fractures Clinical Nuclear Medicine 2007; 32:835-838
Abstract
The objective of the present study was to examine the observer variation for bone scintigraphy in the detection of occult scaphoid fractures in daily practice, using only the early bone scanning images.
Methods: One hundred consecutive bone scans of patients with a suspected scaphoid fracture but negative initial radiographs were prospectively included to calculate the inter- and intra-observer variation. Three nuclear medicine physicians independently evaluated all bone scans at 2 different points in time with a 3 months interval. The observers filled out a blinded scoring sheet for each patient. They scored if a scaphoid fracture was present or not. In addition, they scored the presence or absence of another fracture. The inter- and intra-observer variation was analysed using the kappa statistic.
Results: The inter-observer variation showed substantial agreement for a scaphoid fracture and almost perfect agreement for another fracture. The intra-observer variation showed almost perfect agreement for both a scaphoid fracture and another fracture.
Conclusion: In the present study, early static images of bone scintigraphy for suspected scaphoid fractures showed very little inter- and intra-observer variation.
In addition, expertise does not seem to have an influence on the results. This enhances the possibility of using bone scintigraphy in daily practice.
58
59
Chapter 6Observer variation of bone scintigraphy
Materials and methods
Patients
One hundred consecutive patients, who visited the Emergency Department between 2003 and 2006 with clinical signs, but without radiological evidence of a scaphoid fracture, were enrolled in a prospective study. All patients had a tender anatomic snuffbox, pain when applying axial pressure (first or second digit) and a recent history of acute trauma.
All patients were older than 18 years. The 100 patients included 56 men and 44 women of mean age 40 (range 18-84) years. In 3 patients bi-lateral scaphoid fractures were suspected.
Measurements
All bone scans were blind coded, i.e. without any patient identifying factors. Three independent observers scored all bone scans. They were randomly chosen, from 2 different institutions with different experience levels and scopes of interest. Observer 1 and 3 were both experienced consultant nuclear medicine physicians with 13 and more than 25 years of experience, respectively. Observer 2 was a final year resident, with 4 years of experience.
All bone scans were evaluated at 2 different points in time, 2 months apart, by all 3 observers. They filled out a standard scoring sheet, blinded to each other and blinded to any other data (i.e. scaphoid radiographs, other diagnostic modalities, clinical outcome, etc.). For every bone scan, each observer scored the following 2 items:
1. Scaphoid fracture (yes / no);
2. Another fracture (other fracture in the carpal region including carpal, metacarpal or ulnar fracture / distal radius fracture / distal radius and other fracture in the carpal region / no traumatic injury).
Statistical analysis
For the 2 items, both the inter- and intra-observer variations were calculated for all 100 patients. As there was a total of 200 scans, for both the inter- and intra-observer variations, all the 200 scans were used to calculate the kappa statistic. Of these, 103 scans were from patients with a suspected scaphoid fracture and 97 from the contra-lateral side of the patient. We analysed all 3 groups (200 scans: 103 suspected; 97 contra-lateral, unsuspected). The inter-observer variation was calculated for both the first and second points in time and was calculated for all 3 observers. The kappa statistic was calculated and its interpretation was based on the guidelines proposed by Landis and Koch.58These guidelines suggest that kappa values of 0 to 0.2 represent slight agreement, 0.21 to 0.40 fair agreement, 0.41 to 0.60 moderate agreement, and 0.61 to 0.80 substantial agreement.
A value above 0.80 is considered as almost perfect agreement. For a scaphoid fracture,
the simple kappa coefficient was calculated because the data were nominal (yes or no scaphoid fracture). For another fracture, the weighted kappa coefficient was calculated because the data were ordinal.
Results
In the group of 103 scans suspected of having a scaphoid fracture, there were on average 27.7% occult scaphoid fractures scored. In addition, 40.5% of these patients had other occult fractures.
In the group of 97 scans of the contra-lateral, unsuspected side, most of the observers scored no fracture. Of these 97 scans, an average of 3 false positive scaphoid fractures was scored. Therefore, it is not useful to calculate the observer variation for this group.
Inter-observer variation
The inter-observer variation was calculated for 3 observers at the 2 different points in time. Concerning:
1. A scaphoid fracture.
For the 200 scans, on average, there was substantial agreement (kappa 0.77).
For the 103 scans with a suspected scaphoid fracture, on average, there was substantial agreement (kappa 0.74);
2. Another fracture.
For the 200 scans, on average, there was almost perfect agreement (kappa 0.88).
For the 103 scans with a suspected scaphoid fracture, on average, there was substantial agreement (kappa 0.74).
Intra-observer variation
The intra-observer variation was calculated for 3 observers. With respect to:
1. A scaphoid fracture.
For the 200 scans, on average, there was almost perfect agreement (kappa 0.82).
For the 103 scans with a suspected scaphoid fracture, on average, there was almost perfect agreement (kappa 0.82);
2. Another fracture.
For the 200 scans, on average, there was almost perfect agreement (kappa 0.92).
For the 103 scans with a suspected scaphoid fracture, on average, there was almost perfect agreement (kappa 0.93).
60
61
Chapter 6Observer variation of bone scintigraphy
Figure 6.1 gives an example of a bone scan scored by all 3 observers at both times as a scaphoid fracture. Figure 6.2, 6.3 and 6.4 are examples of bone scans with an inter- and / or intra-observer variation for scaphoid, other and distal radius fractures.
Figure 6.1
Example of an occult scaphoid fracture scored by all 3 observers at both moments.
Figure 6.2
Example of a patient with a suspected scaphoid fracture on both sides. The results revealed an inter- and intra- observer variation for a bi-lateral scaphoid fracture.
Figure 6.3
Example of an inter- and intra-observer variation for a distal radius fracture. In addition was there an inter- observer variation for another carpal fracture.
Discussion
The purpose of this study was to examine the observer variation in daily practice for bone scintigraphy, using only the early images, in detecting occult scaphoid fractures.
Concerning scaphoid fractures, the observer variation for scaphoid fractures showed substantial to almost perfect agreement. The observer variation was even better for another fracture. In addition, the results were not related to knowledge and expertise of the observers. Therefore, the early nuclear images can be seen as a consistent tool in the diagnostic management of scaphoid fractures with a high accuracy even in daily practice.
Our results underline the clinical problem of diagnosing a scaphoid fracture; nearly 30%
of all patients, without radiologic evidence of a fracture on scaphoid radiographs, had a scaphoid fracture on bone scintigraphy. In addition, there were approximately 40% of other occult fractures diagnosed by bone scanning. These results are in agreement with the literature.4-5 As 4 observers evaluated all radiographs, it is suggested that these fractures have been missed because of the insensitivity of the radiographs and not because of observer errors. Therefore, conventional radiography of the scaphoid bone is not accurate, which hampers decision making in the management of a suspected scaphoid fracture. As a delay in treatment of a scaphoid fracture increases the risk of non-union, there is the tendency to over-treat patients.8,10,14,30,55A false positive diagnosis is therefore considered less serious than a false negative diagnosis. Besides, differentiation between a scaphoid fracture and another fracture in the carpal region is important, as it may affect the management.59
Nuclear imaging is very accurate in the diagnosis of scaphoid fractures, but it is not widely used because it is invasive, carries a varying specificity, is time-consuming and is costly.
Figure 6.4
Example of a patient with a suspected scaphoid fracture on his right hand. He had no complaints of his left hand or wrist. There was an inter- and intra-observer variation between all observers for a scaphoid, distal radius and other fracture.
62
63
Chapter 6Observer variation of bone scintigraphy
On the other hand some studies indicate that bone scintigraphy could be cost-effective in the diagnosis of scaphoid fractures.56However, in comparison to some other diagnostic modalities, bone scintigraphy seems to be less cost-effective.60Therefore, further research is necessary to establish fast and reliable diagnostic tools in the diagnostic work-up of suspected scaphoid fractures and further studies for cost-effectiveness should be conducted. One study reported a moderate inter- and intra-observer agreement for early bone scanning.61An almost perfect agreement was found when both the dynamic and static phases were evaluated. In the present study, only the early static bone scans were used for practical reasons; it could make bone scintigraphy more useful in daily practice.
In the present study, early static bone scanning images for suspected scaphoid fractures showed very little inter- and intra-observer variation. In addition, expertise does not seem to have an influence on the results. This enhances the possibility of using bone scintigraphy in daily practice. Nevertheless, bone scintigraphy is costly and invasive and further studies towards cost-effectiveness are necessary.
