The scaphoid bone plays a pivotal role in the wrist carpal complex as it preserves the normal alignment of the carpus creating normal wrist motion(1). Early diagnosis of a scaphoid fracture is critical to promoting healing and restoring alignment and hence conducive to a good clinical outcome. In the case of a missed diagnosis, inter-fragmentary motion will affect the fracture healing process and could lead to complications such as avascular necrosis, mal-union, delayed union and non-union. Furthermore, kinematic abnormalities resulting from malalignment could result in instability of the wrist carpal complex. Complications resulting from a missed diagnosis of a scaphoid fracture can lead to significant long-term disability(2–4). Early diagnosis and treatment improves clinical outcome and reduces complication rates. Imaging plays a vital role in diagnosis as it can be difficult to assess patients clinically with relatively innocuous symptoms.
Some authors have argued that subtle fractures which are not visible on the initial radiographs are likely to heal spontaneously without immobilisation; but this view is not widely shared as the scaphoid bone is affected by nearly every motion of the hand, wrist and forearm, subsequently causing movement and pressure on the fracture line(5). Routinely, scaphoid views would be ordered as a first line investigation and in the majority of cases these fractures can be diagnosed primarily if radiographs are of good quality. When radiographs are inconclusive and the patient is symptomatic a second line of imaging investigation should be considered. Bone scintigraphy (BS), computed tomography (CT) and magnetic resonance imaging (MRI) can be helpful when repeat radiographs are not diagnostic(6). Some authors have advocated the role of ultrasonography in diagnosing fractures based on many factors. Cortical bone can be reliably visualized on ultrasound (US) given its reflective acoustic characteristics; this renders ultrasound imaging of cortical bone a highly specific means of identifying fractures as small as 1 mm(7). Currently, ultrasound is widely utilized by emergency physicians to aid diagnosis of a variety of clinical entities with a high degree of reliability. The constant advancement of ultrasound imaging and its availability have led to its increased utilisation as a first line diagnostic modality amongst clinicians in general. This has subsequently led to more appropriate use of this diagnostic modality. Furthermore, ultrasound is non-invasive, radiation free and relatively inexpensive. On the other hand, ultrasound has limited penetration abilities and areas of deep bone or thick layers of soft tissues may not be well visualized(8). Ultrasound is also highly operator-dependent compared to other imaging modalities.
The purpose of this review is to systematically evaluate the literature concerning the role of ultrasound in the diagnosis of scaphoid fractures.
We searched the National Institute for Health and Care Excellence (NICE) database using the Healthcare Databases Advanced Search (HDAS) tool. In addition, we utilized the PubMed database to search the Medical Literature Analysis and Retrieval System Online (MEDLINE), Excerpta Medica database (EMBASE), Cumulative Index of Nursing and Allied Health (CINAHL) and Allied and Complimentary Medicine (AMED) databases.
The search of these databases was conducted from their year of inception to February 2018. Broad search keywords were used, rather than specific terms, to ensure no articles were missed. Abstracts of all studies identified by the search strategy were examined and, if relevant, the papers were then read in full. The reference lists of these papers were also searched. Study selection, assessment of quality, and data extraction were performed independently by two authors: Mohammed Ali (MA) and Mujtaba Ali (Muj A). The Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) methodology guidance was employed(9). Studies were included if they discuss the use of ultrasound to diagnose scaphoid fractures based on cortical interruption (CI), radio-carpal effusion (RCE) and scapho-trapezium-trapezoid effusion (STTE). We excluded single case reports, reviews, conference abstracts, studies on cadavers, technical notes and articles not published in English. Studies utilising tenderness induced by direct pressure with the ultrasound probe were also excluded.
One author (MA) extracted data using a standardized form for each of the following: author, year of publication, study design, sample size, demographics, operator specialty, confirmation test, sensitivity and specificity of reported values. This was verified by a second author (Muj A).
As all the included articles are non-randomized control trial studies, the quality assessment was performed using the methodological index for non-randomized studies (MINORS) score(10). It has 12 domains for which non-comparative studies use the first 8 domains. Each domain is scored out of 2 with ideal scores being at least 16 for non-comparative studies and 24 for comparative studies. The investigators discussed scores where more than two-point difference was recorded, until an agreement has been reached.
Systematic review registration: PROSPERO CRD42017082758.
The queries used in the NICE HDAS database were “scaphoid fracture” and “ultrasound scan”; these resulted in 416 articles. The search results are outlined in Figure 1.
Search results
Duplicates, non-clinical and studies unrelated to US were excluded. In addition, we excluded reviews, operative technique articles, case reports and ultrasound studies not utilising CI, RCE and STTE (Fig. 2). 6 articles(11–16) met the inclusion criteria for this review.
Study selection process
In total, 6 non-RCT studies were included for analysis(11–16). These articles included 236 patients, with sample sizes ranging from 15 to 63 patients per study and mean age ranging from 18 to 41.2 years. The quality of these articles ranged between moderate and high based on the MINORS score (Tab. 1).
Description of included articles
Author, year | RCT/CS | Number of patients | Mean age | Male/female | Minors |
---|---|---|---|---|---|
|
CS | 15 | 23.5 | 7/8 | 14 |
|
CS | 54 | 26 | 35/19 | 10 |
|
CS | 18 | 35 | not available | 10 |
|
CS | 24 | 42 | 13/11 | 12 |
|
CS | 62 | 41,2 | 29/33 | 14 |
|
CS | 63 | > 18 | not available | 12 |
All authors agreed on tenderness and swelling of the anatomical snuff box alongside tenderness on axial loading as criteria for clinical inclusion(11–16). Some authors added restriction of wrist movement(11,15) and wrist pain(15,16). All these studies focused on detecting cortical interruptions and joint effusions on ultrasonography.
Herneth(11) and Yildirim(16) used MRI to confirm the ultrasound findings, Fusetti(14) and Platon(15) CT scan, Senall(13) repeated the plain films while in Hauger’s(12) study there was no uniformed confirmatory test (Tab. 2).
Ultrasound characteristics and confirmatory tests
Author, year | Baseline test | Exam included | Waiting time | US device | Confirmation test |
---|---|---|---|---|---|
|
scaph views | CI, SPE | 72 hours | ATL, HDI 3000, 10–5 MHz | MRI |
|
scaph views | CI, SPE | within 7 days | HDI 3000&5000, 12 MHz | XR, CT, MR, B-scan |
|
scaph views | CI, SPE | 4.7 days | ATL, HDI 3000, 10–5 MHz | XR |
|
scaph views | CI, RCE, STTE | not mentioned | ATL-Philips 5000, 7–15 MHz | CT |
|
scaph views | CI, RCE, STTE | 3 days | Prosound SSD-5000, 5–13 MHz | CT |
|
Clinical exam | CI, RCE, STTE | 24 hours | DC3, Mindray, 5–13 MHz | MRI |
Herneth and Senall used the same ultrasound machine with similar low frequency probes (5 MHz) and both had the same sensitivity (78%) which was the lowest amongst the other studies. Fusetti and Hauger used high frequency probes (15 MHz and 12 MHz respectively) and they both achieved 100% sensitivity (Tab. 3 and Tab. 4) (Fig. 3).
Sensitivity and specificity
Author, year | US operator | Sensitivity | Specificity | Positive predictive value | Negative predictive value |
---|---|---|---|---|---|
|
Rad | 78% | 100% | 100% | 75% |
|
MSK Rad | 100% | 98% | 83% | 100% |
|
MSK Rad | 78% | 89% | 88% | 80% |
|
MSK Rad | 100% | 79% | 56% | 100% |
|
Rad | 92% | 71% | 46% | 97% |
|
AE Doctor | 85,70% | 100% | 100% | 100% |
Descriptive statistics
|
Minimum | Maximum | Mean | Std. deviation | |
---|---|---|---|---|---|
|
6 | 78.00% | 100.00% | 88.9500% | 10.03868% |
|
6 | 71.00% | 100.00% | 89.5000% | 12.21065% |
|
6 | 46.00% | 100.00% | 78.8333% | 22.78962% |
|
6 | 75.00% | 100.00% | 92.0000% | 11.40175% |
Descriptive statistics
The use of bedside musculoskeletal ultrasound has been supported by many authors based on the availability and practicality of this technique. Backhaus
In 2001, Herneth
Platon
Fractures were put into two groups according to their potential for complication: high (proximal or waist) and low (distal or tubercle). CT scans showed scaphoid fracture in 13 (21%) patients: 8 high potential (3 proximal pole and 5 waist) and 5 low potential (3 distal pole and 2 at the tubercle). US was 92% sensitive (12/13) in demonstrating a scaphoid fracture overall. It was 100% sensitive (8/8) in demonstrating a fracture with a high potential for complication. Platon considered these results to be related to the anatomical position of the scaphoid. As the scaphoid waist was the easiest to view on US, signs of fracture were reliably identified at this site. On the other hand, scanning the distal pole or the tubercle was more technically challenging. In a recent study, Yildirim(16) evaluated 63 patients with suspected scaphoid fractures using bedside ultrasonography and MRI. A scaphoid fracture was diagnosed by MRI in 14 patients. Ultrasound revealed signs indicative of scaphoid fracture in 12 of the 14 patients. Of these, 12 were corroborated by MRI and none found to be negative. In 45 cases, ultrasound did not reveal scaphoid fracture, of which two were deemed to have a scaphoid fracture following the MRI scan. The limitation of this study was a lack of blinding in the examining emergency physicians.
The stage at which ultrasonography can be incorporated into the diagnostic process of suspected scaphoid fracture is a matter for debate. Current practice in the United Kingdom encountered by the authors in multiple centres entails initial conventional radiography in the acute setting followed by repeat radiographs after a set interval if the patient continues to be symptomatic. If the second set of radiographs remains inconclusive, and after confirming suspicion by orthopaedic surgeons, a more superior imaging modality is considered. Commonly this would involve MRI although CT is also widely utilized for this purpose. During this period, which can often be lengthy, the affected wrist and hand will commonly be immobilized. Ultrasonography can potentially be introduced at different stages within this pathway. The first consideration is for ultrasonography to be utilized as a primary imaging investigation by emergency care physicians. The superiority of ultrasonography over conventional radiography concluded in this review would be the main advantage of this approach as it can potentially facilitate either earlier confirmation or exclusion of the diagnosis and hence either facilitate earlier management or avoid prolonged periods of immobilisation respectively. The main limitation of this would be the need for adequate training required for emergency physicians, particularly that ultrasonography is highly user-dependent. The second consideration would be the introduction of ultrasonography as a secondary confirmatory test after failure of conventional radiography. The main advantage to the second consideration would be reducing the number of costly investigations such as MRI while also reducing waiting times as ultrasonography is more accessible. Limitations to the second consideration would be a lack of large volume clinical trials that support the reliability of ultrasonography for this purpose. An increasing practice in recent years has been the introduction of one stop clinics run collaboratively by orthopaedic surgeons and with musculoskeletal radiologists available in the clinic. The incorporation of ultrasonography in the setting of a one stop clinic would be our third consideration. In the authors’ opinion, this would be the most ideal setting for the introduction of ultrasonography into the diagnostic process of suspected scaphoid fractures given the instant availability of an experienced musculoskeletal radiologist to perform the examination. Furthermore, a combined decision can be made at the time as to the necessity of further imaging modalities such as MRI. This would also be an ideal setting to perform large volume prospective clinical trials.
The choice of secondary confirmatory test is an important consideration in future clinical trials looking into ultrasonography in the diagnosis of radiographically occult scaphoid fractures. The confirmatory test varied in the studies reviewed in this article. Bone scintigraphy, CT and MRI have been widely used and studied for this purpose. Nonetheless, uncertainty remains regarding the most appropriate(20). In addition to sensitivity and specificity, other factors should be taken into account when deciding upon the confirmatory imaging test. These include invasiveness and exposure to ionising radiation. Clinicians should be aware of advantages and limitations of each modality. It is the authors’ view that MRI would constitute a suitable confirmatory test in future studies given its superior sensitivity and lack of ionising radiation exposure.
Overall, not many authors have studied the role of ultrasound in the diagnosis of scaphoid fractures. The current literature has revealed that ultrasound has high sensitivity and specificity. Multiple factors including technical differences in ultrasound machines and probes, small sample sizes and variability of subsequent confirmatory tests have created a challenge in determining the ultimate reliability of ultrasonography in the diagnosis of occult scaphoid fractures. Considering these factors and limitations, largesample and high-quality clinical trials are needed to adequately assess its reliability for this purpose. The stage at which ultrasonography can be incorporated into the diagnostic process of suspected scaphoid fractures is a debatable matter and would largely depend on local procedures and facilities available. Concluding an approach that could be generalized would be difficult and therefore, a selective approach by individual centres would seem appropriate at present. In the authors’ opinion, combined one stop clinics would constitute the ideal setting for the introduction of ultrasound diagnosis of scaphoid fractures if these are locally available. Furthermore, this would be an ideal setting to perform large volume prospective unbiased clinical trials.