Introduction
Emergency imaging and radiological workup in trauma care have been reported as a “perfect
storm” for radiological errors given the fact that radiologists and radiographers
have to cope with factors that impair proper scanning and image interpretation, such
as uncooperative patients, inadequate medical histories, time-critical decisions,
and concurrent tasks (especially during night shifts or weekend shifts) [1]. The majority of diagnoses missed in the emergency setting are fractures [1]. Thus the musculoskeletal region is the main area for missed pathologies in the
emergency radiology setting. For instance, the missed injury rate in a Danish casualty
department was between 0.5% and 2.2% when analyzing a cohort of n=15,806 [2]. In a Taiwanese emergency room department, the missed fracture rate in radiological
reports was 3.7% in a cohort of n=3,081 [3]. The authors identified the joints as risk areas and reported the prevalence of
missed fractures with the following locations: foot (8%), knee (6%), elbow (6%), hand
(5%), wrist (4%), hip (4%), and ankle (3%), while in a North American pediatric emergency
department the most frequently missed fractures were of the hand phalanges followed
by metatarsus, distal radius, tibia, and phalanges of the foot [4]. Other authors also reported that among all missed injuries, extremity fractures
make up between 14% and 60% [5]
[6]. Moreover, missing a fracture is the second most common error alleged in medical
malpractice suits against radiologists in the United States [7]. Of note, Wei et al. also reported that 70% of initially missed fractures could
be identified by a second review [3]. In addition, in this study, they described factors contributing to the missing
of a fracture ([Table 1]). The two predominant factors, which make up 70% of missed fractures, were first
‘subtle but still visible fractures’ and second ‘radiographically imperceptible fractures’.
Radiologists are able to lessen the factors contributing to missed fractures. For
example, implementing a ‘four-eyes principle’, i.e., two readers read the radiographs,
would help to overcome the missing of ‘subtle but still visible fractures’. The emerging
artificial intelligence (AI) applications may also be supportive in this respect.
Also, the additional use of (cross-sectional) imaging would help to overcome the missing
of ‘radiographically imperceptible fractures’. [Table 1] presents techniques to lessen the factors contributing to missed fractures.
Table 1 Factors contributing to missed fractures in a cohort of n=3,081 [3] and measures for lessening these factors.
Factors contributing to missing of a fracture
|
Measures for lessening the effect
|
Note: The percentage of the aforementioned factors contributing to the missing of a fracture
in the cohort of Wei et al. [3] is given in parentheses. *The ‘satisfaction of search’ error is a common error in
diagnostic radiology, and it typically occurs when the reporting radiologist fails
to continue to search for subsequent abnormalities after identifying an initial one,
because this initial detection of an abnormality satisfies the ‘search for meaning’
and the reporting of the case is prematurely ended. In emergency medicine, ‘satisfaction
of search’ errors are referred to as premature closure and this may entail identifying
two or more totally independent pathologies or two or more findings which together
take a case from a differential list to a definitive diagnosis [8].
|
Subtle but still visible fractures (37%)
|
Four-eyes principle, AI support
|
Radiographically imperceptible (33%)
|
Additional (cross-sectional) imaging
|
Block by splinting devices (7%)
|
Remove the splint before acquiring X-rays
|
Multiple fractures (9%)
|
Remember ‘satisfaction of search’ error*
|
Inappropriate or insufficient radiographs acquired (5%)
|
Teach radiology technologists, quality control
|
Lack of relevant clinical information (4%)
|
Talk to your clinical partners
|
Poorly positioned radiographs (2%)
|
Teach radiology technologists, quality control
|
Metal artifacts (2%)
|
Use metal artifact suppression techniques
|
Severe osteoporosis (2%)
|
Additional (cross-sectional) imaging
|
Pearls of wisdom to reduce the number of missed pathologies of the musculoskeletal
system in the emergency radiology setting
A review of the literature was performed by searching the PubMed and ScienceDirect
databases, using the keywords (‘missed injuries’ or ‘missed fractures’) and (‘emergency
radiology’ or ‘emergency room’) and (‘musculoskeletal’ or ‘bone’ or ‘skeleton’) for
the title and abstract query. The inclusion criteria were scientific papers presented
in the English and German languages. Among the 347 relevant hits between 1980 and
2024 as identified by the author of this review article, there were 114 relevant articles
from the years between 2018 and 2024. Based on this literature search and personal
experience from more than 24 years of radiological reporting and as a certified Fellow
of the European Society of Emergency Radiology since 2018, information to reduce the
number of missed pathologies of the musculoskeletal system in the emergency radiology
setting is presented in the following.
The first and, from my point of view, most important pearl of wisdom is to use additional
imaging in case of doubt. [Fig. 1] presents a case where the fracture is radiographically occult and is only visible
when using cross-sectional imaging. Radiologists know the best use of all available
radiological imaging modalities and should therefore advise patients and clinical
partners as to which imaging modality is appropriate for a certain clinical question.
For certain clinical conditions, image or “choosing wisely” concepts including emergency
imaging have been published [9]. Although radiographs are the mainstay for fracture assessment, a fracture may be
missed because it is radiologically invisible or equivocal [10]. In summary, when there is a high clinical index of suspicion for a fracture despite
initial negative radiographic findings, additional computed tomography (CT) or magnetic
resonance imaging (MRI) examinations are recommended, particularly if the results
would affect clinical management [10]
[11].
Fig. 1 57-year-old woman with fall on right knee and pain in the medial aspect. Radiographs
on two planes (a: anterior-posterior, b: lateral view) as well as CT (c: coronal reconstruction with 0.6 mm slice thickness, d: sagittal reconstruction with 0.6 mm slice thickness) did not demonstrate a fracture.
MRI (e: T1w-weighted sagittal sequence, f: proton-density fat-saturated sagittal sequence) shows the fracture of the tibial
head with substantial bone marrow edema (arrows) that, also on retrospect, was not
visible on both the CT and X-ray images.
The second pearl of wisdom is to perform proper imaging. This requires knowledge of
the patient’s clinical information ([Fig. 2]). The precise correlation of physical examination findings, such as the site of
maximum pain, with radiologic imaging is also helpful for proper reporting [10]
[11]. The benefit of sufficient clinical information to improve radiologists’ performance
with respect to selection of the best imaging protocol and reporting has been advocated
by both radiologists and trauma surgeons, thereby substantiating the value of close
interaction [10]
[12].
Fig. 2 57-year-old man with fall from a ladder. The radiographs ordered initially (a: calcaneus, b: ankle joint and second plane of calcaneus and ankle joint) were reported as uneventful.
Only, a minor plantar osseous spur was mentioned in the report as an auxiliary finding.
Because of the fact that pain was persisting for 2 months, an additional MRI was ordered
and performed 10 weeks after initial trauma. The MRI examination (c–e, short-tau inversion recovery (STIR) images) revealed a stress fracture of the fourth
metatarsal bone (arrow in c) and at the base of the second metatarsal bone (arrow in d) as well as a slight bone bruise within the calcaneus but without a fracture line
and partial tear of the plantar fascia near its insertion (open arrow in e).
The third pearl of wisdom is to know the indirect signs of fracture when interpreting
projection radiographs. [Fig. 3] shows as an example the ‘fat-pad sign’ resulting from hemarthrosis of the elbow
joint as an indication of a fracture. This may especially be important in pediatric
and adolescent patients. General radiologists who do not frequently report pediatric
musculoskeletal cases should pay careful attention to the following five pediatric
fracture types: 1. unstable fracture of the radial condyle, 2. luxation of the radial
head, 3. supracondylar humerus fracture with rotation, 4. proximal bowing fracture
of the tibia, 5. fracture of the medial malleolus. When one of these aforementioned
fracture types is overlooked, growth disturbance together with or without dysfunction
may result and missing one of these fractures is often a reason for liability claims.
For instance, in Northern Germany physicians are found at fault in 60% of these liability
cases [13]. Of note, three of these fractures that are prone to be overlooked are situated
next to the elbow joint and misinterpretation of fractures of the elbow may result
in delays of consolidation with subsequent growth disturbances, joint dysfunction,
and malposition. Missed fractures are common in pediatric trauma patients because
of the substantial normal variation in the contour of developing bones and growth
plates as well as because of the subtlety of findings or the radiographical invisibility
[14]. Moreover, fractures have been reported to be the third most common diagnosis included
in medical malpractice lawsuits involving children in United States emergency departments
and urgent care centers [15]. Hence, knowledge of typical fractures for different ages, the individual bone nuclei
of the growth plate, and epiphyseal injuries is important for the correct radiological
diagnosis.
Fig. 3 17-year-old male adolescent who fell while horseback riding. The externally performed
X-ray in lateral projection a does not show a fracture (of note the radial head is superimposed by the coronoid process) but positive ‘fat-pad sign’ as the result of hemarthrosis (arrows). The Greenspan
view performed in-house of the radial head b demonstrates the intraarticular, non-displaced fracture of the radial head (Mason
type I, arrow) that is also shown at the sagittal reconstruction of the additional
CT scan (c, arrow). MRI (d: proton-density fat-saturated sagittal sequence) shows bone marrow edema of the radial
head and the distal posterior humerus (asterisks), the non-displaced intraarticular
Mason type I fracture (arrow) as well as the joint effusion (open arrow) causing the
‘fat-pad sign’ of the initial radiograph.
The fourth pearl of wisdom is to always acquire radiographs on two planes when trauma
sequelae are suspected and to know Aunt Minnie’s atlas [16], i.e. the typical (pathognomonic) imaging signs of the musculoskeletal pathologies.
[Fig. 4] demonstrates the typical ‘pooping sign’ of triquetrum fractures. Triquetrum fractures
are the second most common carpal fracture after the scaphoid and make up approximately
15% of all wrist fractures. They are usually the result of forced hyperflexion. Dorsal
triquetrum fractures are most common and they are most often caused by avulsion from
the attachments of dorsal radiocarpal ligaments, whereas transverse or sagittal fractures
of the triquetral body are far less common. They have been reported to occur in association
with a variety of different mechanisms, including crush injuries and perilunate fracture
dislocations. As a rule of thumb, the routine wrist series is usually sufficient for
fracture identification. On X-ray, dorsal triquetral fractures are seen only on the
lateral projection since the pisiform usually overlies and obscures the triquetrum
on the anterior-posterior projection of the wrist. In summary, many fractures are
visible on only a single view. Therefore, there is a risk of a false-negative interpretation
of a radiographic examination if not all necessary views are obtained, thus substantiating
the role of standardized protocols with two or more planes [11].
Fig. 4 41-year-old man with pain in the left wrist after a fall on flexed hand while ice
skating. The anterior-posterior radiograph a does not show a fracture. The lateral radiograph b depicts a tiny osseous fragment on the dorsal side of the carpal bones (arrow). The
“pooping duck sign” (arrows, c) indicates a triquetrum fracture (drawing courtesy of Henning Maschke, Hamburg). The CT scan (d: sagittal reconstruction with 0.625 mm slice thickness) better demonstrates the typical
“pooping duck sign” of the triquetral fracture but in addition reveals a fracture
of the tubercule of the trapezium bone (arrows; e: sagittal and f: paratransverse reconstruction of the CT dataset with a 0.625 mm slice thickness),
as well as cortical flakes of the hamate bone (open arrow) and capitate bone (not
shown). The fractures of the trapezium, hamate bone, and capitate bone were not visible
on the radiographs.
The fifth pearl of wisdom is to know the limitations of the selected imaging modality.
The radiologist’s role is to perform imaging wisely and radiologists know the pros
and cons of each available imaging modality better than their clinical partners (also
see the first pearl of wisdom). A good example of emergency radiology of the musculoskeletal
system where knowledge of the limitations of each imaging modality is mandatory is
when determining whether a scaphoid fracture is present. Scaphoid fractures account
for almost 80% of all carpal fractures and fractures in the middle third of the scaphoid
are the most common at around 60%. The typical trauma mechanism is a fall on the outstretched
hand [17]. The proper role of radiological imaging is not only to detect or rule out a scaphoid
fracture, but also to answer the question of fragment stability. The basis is conventional
radiography with three projections: dorsopalmar, lateral, and Stecher image (the scaphoid
is aligned parallel to the detector plane by closing the fist and ulnar deviation).
However, radiography only has a sensitivity of ≤ 70%. Hence, the early use of cross-sectional
imaging is crucial to avoid missing a scaphoid fracture. In German-speaking countries,
high-resolution CT is initially recommended using thin slices from 0.5 to 0.75 mm
and oblique-sagittal and oblique-coronal image reconstructions parallel to the longitudinal
axis of the scaphoid [18]. The advantage of CT is the superior, therapy-relevant representation of the fine
bony structures including the exact fragment dislocation (specificity 100%). The sensitivity
of CT for detecting fractures is only about 95%, i.e., a non-displaced fracture can
escape detection. If a fracture is clinically suspected and the X-ray and CT results
are negative, additional MR imaging must be carried out ([Fig. 5]). In the Anglo-American region, MRI is usually used immediately after the X-ray
procedure with thin slices of 1.5 or 2.0 mm and at least one slice plane parallel
to the scaphoid. MRI is advantageous because it detects all scaphoid fractures based
on trauma-induced bone marrow edema (sensitivity 100%). The disadvantage of MRI is
that it is often difficult to differentiate between a bone contusion (“bone bruise”)
and a fracture (specificity around 85%), which is why CT imaging must be supplemented
[17]. In summary, the early use of high-quality CT and/or MRI avoids delayed diagnosis
of a scaphoid fracture given the fact that up to 30% of acute scaphoid fractures are
missed on conventional radiography. Another often overlooked fracture of the wrist
is the hook of hamate fracture [19], since standard radiographs often fail to diagnose hamate fractures [10]. For an early diagnosis, when there is a high index of suspicion, the key is to
perform carpal tunnel projections or CT scans early [10]
[19]. The question as to whether there is a Lisfranc fracture is an example of often
missed trauma sequelae in the lower extremities. The lack of weight-bearing views
can lead to false-negative radiographic findings [11]. After midtarsal trauma, initial radiographs are typically non-weight-bearing anterior-posterior,
lateral and internal oblique views [10]. A high index of suspicion is key since 20% of Lisfranc fractures are missed at
first presentation, especially when no additional weight-bearing radiographs have
been performed [10]
[20]. It is important to look for misalignment on radiographs and to use CT scans early
[21], since CT’s superiority to conventional radiography in the diagnosis of bony disorders
of the Lisfranc joint has been demonstrated [22]. Hence, on radiographs, the congruence between the cuneiform bones and the cuboid
bone with the metatarsal bones should always be inspected closely and since findings
are usually very subtle, additional cross-sectional imaging may be necessary for safe
exclusion ([Fig. 6]). In addition, osseous injuries and especially avulsion fractures in growing adolescents
may be overlooked when using solely projection radiography as the imaging modality.
Avulsion fractures are typical between the ages of 12 and 22 years and they represent
overworking of the immature skeleton. Knowing the risk areas and checking for asymmetry
will help to avoid overlooking osseous avulsion injuries [20]. Besides avulsion injuries, other osseous injuries in growing adolescents that may
be underestimated on projection radiography are transitional fractures. Transitional
fractures are special forms of epiphyseal injuries in adolescents in whom the growth
plate is already partially closed. Due to the partial ossification, specific stereotypical
fracture patterns can develop and can be differentiated into biplane, triplane I,
and triplane II fractures, depending on the involvement of the metaphysis and the
number of fragments. At the beginning of the diagnostic cascade, conventional projection
radiography on two planes is used. However, due to the complex fracture patterns,
CT is often indicated and necessary for preoperative planning. Prognostically relevant
is above all the reconstruction of the joint surfaces, as early arthrosis can be a
risk if the incongruity remains [23]. Therefore, although the use of CT should be restricted in children, CT plays an
important role in proper preoperative assessment in the case of transitional fractures.
The role of CT scans is to recognize the complexity of transitional fractures and
epiphyseal injuries of partially closed growth plates, because injuries of growth
plates may be hard to detect on radiographs [23]. In summary, especially in an emergency setting, radiologists should recommend supplementary
imaging procedures, such as cross-sectional imaging, to detect radiographically imperceptible
injuries or fractures that are difficult to classify on radiographs. Radiologists
best understand the limitations of radiography for certain diagnoses and thus can
best recommend more advanced imaging to establish a correct diagnosis in a reasonable
time frame [10].
Fig. 5 Non-displaced scaphoid fracture in a 34-year-old man after a fall on the left hand.
The radiologist reporting the initial radiographs (a: anterior-posterior, b: lateral view) presumed a non-displaced scaphoid fracture and recommended an additional
CT examination. The CT scan performed the same day did not reveal a fracture and thus
an additional MRI examination was recommended when pain would persist (c: coronal and d: oblique-sagittal image reconstruction parallel to the longitudinal axis of the scaphoid).
In the MRI examination performed 13 days after initial radiographs because of persisting
wrist pain (e: proton-density fat-saturated and f: T1-weighted coronal sequences), the fissure (arrow) and the bone marrow edema within
the scaphoid bone could be depicted.
Fig. 6 18-year-old man with radiographs of the left foot after crush injury. The initial
radiographs (non-weight-bearing anterior-posterior (not shown), internal oblique views
a, and lateral b) have been reported as uneventful. The patient received an MRI examination seven
days later showing injury of the Lisfranc joint (arrows, c: T1-weighted and d: proton-density weighted fat-saturated sequences) and CT performed 9 days after the
initial radiographs (e: transverse and f: sagittal reconstruction) best demonstrates the several small osseous flakes around
the tarsometatarsal joints (arrows) and the avulsion fractures of the lateral and
intermediate cuneiform, navicular, and cuboid bones as well as the bases of the second
and third metatarsal bones.
The sixth pearl of wisdom is to look for subtle findings. The Segond fracture is a
very good example of a subtle finding on projection radiography that typically indicates
a severe trauma and major damage that only cross-sectional imaging can identify ([Fig. 7]). A Segond fracture is a vertically oriented avulsion fracture at the lateral tibial
plateau, predominantly caused by a varus force on the knee when the foot is firmly
planted [20]. It is common in running athletes. MRI may reveal a bone contusion of the medial
femoral condyle and the posteromedial tibial plateau. Segond fractures are associated
with anterior cruciate ligament tears in 75–100% of cases and lateral meniscal tears
in 33% of all cases [20]. This fact explains why it is so important not to overlook the often tiny osseous
avulsion injuries on projection radiography. In summary, small avulsion fractures
may be easy to overlook and, as in the Segond fracture, they indicate major injuries
[11]. Fractures of the coccyx are another example of subtle findings and coccydynia is
one of the most overlooked symptoms in clinical practice ([Fig. 8]). The most common cause of coccydynia is trauma, e.g., from water slides and falls,
and radiologists should be familiar with the coccyx's morphologic appearance and with
morphologic parameters and hypermobility causing coccydynia [24].
Fig. 7 Segond fracture in a 56-year-old man after knee distortion. In the radiographs in
anterior-posterior a and lateral projection b, the small osseous flake is only visible on the anterior-posterior projection (arrows
in a and the zoomed inset c). The MRI examination demonstrates rupture of the anterior cruciate ligament (arrow,
proton-density fat-saturated sagittal sequence d), partial rupture of the medial collateral ligament (arrow), avulsion of the anterior
lateral capsule (open arrow), and bone marrow edema within the lateral femoral condyle
(asterisk; proton-density fat-saturated coronal sequence e), as well as a fracture of the dorsal tibial plateau (arrow, T1-weighted sagittal
sequence f).
Fig. 8 54-year-old woman after fall on buttocks 2 weeks ago with persisting pain. The coccyx
fracture of Co1 is visible on the radiograph (arrow, a), but could be easily overlooked. MRI (b: sagittal T1-weighted and c: sagittal STIR sequence) demonstrates the fracture-related bone marrow edema (arrows)
of Co1 and Co2.
The seventh pearl of wisdom is to know the bone nuclei when reading pediatric radiographs.
Common misinterpretations of pediatric radiographs are due to the numerous epiphyses
and apophyseal nuclei as well as accessory bone nuclei, because these are partly interindividual
and, depending on the level of development of the child, can imitate bony fractures
or fragmentations in the case of ignorance of age-specific X-ray anatomy. One typical
example is the lateral apophysis at the base of the fifth metatarsal bone. This lateral
apophysis, depending on the projection, is relatively distant from the base to the
representation. It should be noted that fractures of the fifth metatarsal base typically
proceed horizontally and a pure rupture of the apophysis is extreme rare [25] ([Fig. 9]). It is also helpful in this context to look at the relevant atlases illustrating
the numerous epiphyses and apophyseal nuclei as well as the accessory bone nuclei
of the growing skeleton.
Fig. 9 Fracture of the base of the fifth metatarsal in a child (courtesy of Christina Hauenstein, Rostock). The arrow points at the vertically oriented apophyseal nucleus of the fifth metatarsal
bone, while the open arrow indicates the more horizontal course of the fracture line.
The eighth pearl of wisdom is a really straightforward and obvious one: always look
at all available images. [Fig. 10] presents a case within the field of spine fractures illustrating the importance
of always also reviewing the CT scout views. Spine fractures have been reported to
be difficult to diagnose, especially on radiographs due to superimposing structures,
and missing them can be associated with increased neurologic injury and resulting
morbidity [26]. Reports from the literature underline the fact that the scout views should always
be read. Otherwise, the missing of 2–5% of pathologies in CT examinations has been
reported with obvious negative medicolegal and ethical considerations [1]
[27]
[28]
[29]. In summary, scout images are an integral part of any CT examination and thus should
be carefully reviewed for findings that may or may not be included in the field-of-view
of the study.
Fig. 10 83-year-old man with dementia who fell on his head in the nursing home. He received
anticoagulation because of known atrial fibrillation. The initial CT scan performed
at the emergency department was reported as no traumatic brain injury and no fracture
a. The patient was transferred from trauma surgery to the neurological department of
the tertiary care hospital and the suspected diagnosis was worsening of the dementia.
A follow-up cranial CT scan was performed the next day because of distinctive behavioral
changes of the patient at the neurological ward. Also, the follow-up CT scan the next
day was reported as no traumatic brain injury and no fracture (not shown). After another
4 days, the patient presented with further clinical worsening, respiratory insufficiency
and paresis of his left arm. Another cranial CT scan was ordered again showing no
brain ischemia or bleeding (not shown) but suspicion of a fracture of the odontoid
process of the second cervical vertebra was raised. The patient was then transferred
back from neurology to trauma surgery and a CT scan of the cervical spine six days
after the first presentation demonstrated the dislocated fracture (arrow) of the odontoid
process of the second cervical vertebra b. Of note, when thereafter reviewing again all radiologic images, the scout views
of the first c and second (not shown) CT scan both show the fracture of the odontoid process (arrow)
and it became evident that some radiologists of the team unintentionally had the routine
presentation of scout views deactivated in their PACS settings. The latter was then
fixed for the entire team and the case was presented in an interdisciplinary morbidity
and mortality conference.
The human factor
Humans can get tired. This obvious statement is also true for all medical professionals
including radiologists. A Scandinavian study has assessed the total number of correctly
diagnosed and missed fractures per hour of a day and they observed that the rate of
missed fractures surpassed the number of correctly diagnosed fractures between 5 p.m.
and 7 a.m., i.e., during the night shifts [30]. Other studies have also reported that on-call duty is a factor contributing to
the missing of fractures and other injuries [31]
[32]. Other factors include a lower level of experience on the part of the reader (most
on-call duties are performed by radiologists in training) and image interpretation
under stressful conditions in the emergency room [10], lack of clinical information, absence of previous imaging studies, suboptimal reading
room atmosphere, multitasking and increased workload [10] – all factors that are often typical of on-call duties and night and weekend shifts.
Another human factor is perceptual errors when reporting radiological images [33]
[34]. The ‘satisfaction of search’ error [8]
[11] ([Table 1]), where the detection of an abnormality results in premature termination of the
search for further issues, is one of these. In other words, fractures are missed on
radiological images, for instance, because other fractures are found. Another perceptual
error and form of incomplete search pattern is visual isolation, where the search
pattern of the radiologist is truncated to the main areas of an image, while little
or no attention is given to peripheral areas. Inattentional blindness can also be
included as a perceptual error. It is defined as the failure to notice a fully visible,
but unexpected object because attention was otherwise engaged [34]. These errors are also the reason why delayed diagnoses were not recognized on subsequent
radiologic examinations in about one third of cases [35]. Perceptual errors combined with time pressure are especially challenging in the
reading of whole-body CT scans of polytrauma patients. For instance, although osseous
wrist and hand injuries are present on about 12% of whole-body CT scans after polytrauma,
about 93% of these injuries were missed primarily in a recent study on 506 polytrauma
CT scans resulting in a diagnostic accuracy of 6.8% for the primary reporting [36]. The authors additionally mentioned that motorcycle accidents predispose for these
injuries and often cause additional fractures of the extremities [36]. Furthermore, the presence of more than two injured body parts has been identified
as an independent predictive factor for missed injury [37]. Moreover, missed injuries have been reported to be more likely in severely injured
and intubated patients [38] and missed foot injuries have been especially reported to occur in patients having
been in car accidents or having fallen from great height [39]. Besides fractures, there are non-skeletal injuries (such as parenchymal injuries)
and non-traumatic incidental findings (such as neoplastic findings) that are common
and, given the aspects mentioned above, can also be easily overlooked on polytrauma
CT scans. For example, abdominal injuries are potentially life-threatening and occur
in 20–25% of all polytraumatized patients with the liver (40%) and spleen (32%) as
the most commonly injured parenchymal organs [40]. In addition, incidental imaging findings unrelated to trauma that require urgent
treatment or further clarification have been reported in 8.4% of all patients, most
frequently in the thorax and in the abdominal/pelvic region, in a study on 2,440 patients
with multiple trauma undergoing whole-body CT at admission [41]. Approximately 40% of these patients had incidental findings requiring either immediate
or delayed treatment/follow-up with the most frequent findings being lesions suspicious
for malignancy or definite malignancies as well as inflammation [41]. Disregarding severity, the most common incidental findings in the aforementioned
study were mucosal swelling/chronic sinusitis (20%), healed fractures (10%), renal
cysts (10%), hepatic steatosis (9%), and hepatic cysts (9%) [41]. To overcome these problems, the standardized use of the ‘four-eyes principle’,
clinical re-evaluation, a second review of radiological imaging, and/or artificial
intelligence (AI) support may be advantageous. For example, re-evaluation of existing
imaging as part of tertiary surveys, i.e. the re-examination of patients after emergency
care typically within 24 hours of admission [6]
[36]
[39], significantly reduces the number of missed fractures and other injuries [42]
[43].
The potential of artificial intelligence
An important and intriguing aspect is the emerging role of AI. There are now numerous
AI-based solutions on the market designed to assist radiologists in reporting, akin
to a ‘four-eyes principle’. Applications of AI for image interpretation in the musculoskeletal
region consist of the determination of body composition measurements, bone age, identification
of fractures, screening for osteoporosis, evaluation of segmental spine pathology,
detection and temporal monitoring of osseous metastases, diagnosis of primary bone
and soft tissue tumors, and grading of osteoarthritis [44]
[45]. The number of publications per year in PubMed using the keywords (‘artificial intelligence’
and ‘fracture’ and ‘radiology’) has increased steadily from 30 publications in 2019
to 151 publications in 2023, and several AI algorithms, specifically deep learning
algorithms, have been applied to fracture detection and classification, which are
potentially helpful tools for radiologists and clinicians [46]
[47]
[48]. For instance, it has been reported that AI has the potential to automate and improve
the accuracy of scaphoid fracture detection on radiography, thereby aiding in early
diagnosis and reducing unnecessary clinical examinations, as well as reducing the
risk of missed fractures and complications and reducing reading time and observer
fatigue [49]
[50]
[51]. It should be noted that the combination of AI and the radiologist's analysis provided
the best results regarding wrist fracture detection in a recent study including 1,917
radiographs [52]. As another example, a recent multicenter study including 600 adult patients with
multi-view radiographs after a recent trauma demonstrated that the AI aid improved
the sensitivity of physicians by 8.7% and their specificity by 4.1% and reduced the
mean reading time by 15.0% [53]. Besides the assessment of fractures in the emergency radiology setting, deep learning
algorithms have been used to detect, for instance, free fluid on Focused Assessment
with Sonography for Trauma, to identify intracranial hemorrhage on head CT scans,
and to identify injuries to organs like the spleen, liver, and lungs on abdominal
and chest CT [54]. It should be noted that AI is not replacing radiologists but could be especially
helpful to optimize workflow and augment diagnostic performance [46]
[55]
[56], particularly in light of increasing workloads and staffing shortages, which require
radiologists to review more images, particularly cross-sectional ones.
Conclusion and take-home message
In conclusion, one take-home message regarding the lessening of the number of easily
missed pathologies of the musculoskeletal system in the emergency radiology setting
is that a systematic approach is necessary in emergency and trauma care [20]. Some pearls of wisdom and measures have been summarized for this purpose in this
review article. It is important to quickly evaluate the high-risk areas first (and
maybe to re-evaluate them in a second look) and to know what is commonly missed [20]
[33]. The categories of missed fractures that should be remembered are [33]: a) ‘common but challenging’ (e.g., scaphoid fracture), b) ‘out of mind out of sight’ (e.g., reviewing the scout view), and c) ‘satisfaction of search’ (e.g., polytrauma such as motorcycle accidents). Strategies for mitigating perceptual errors such as the ‘satisfaction of search’ are,
for example, the use of checklists, self-prompting routines, and structured reporting
within an institutional culture of safety and vigilance [34]
[57]. In addition, case conferences focusing on missed findings may enhance radiological
training, because it is easier to see what you know to look for and, of course, it
is easy to miss what you do not know [33]. Moreover, in the emergency radiology setting, the initial modality that is used
and the patient populations that are involved have to be considered. For example,
different strategies are necessary for polytrauma CT scans compared to conventional
hand X-rays for elderly patients after a fall. The pearls of wisdom mentioned in this
review article help to address this issue. For all these reasons, radiologists should
be and continue to be adequately trained in emergency radiology, thereby providing
an invaluable service to clinical colleagues by ensuring that patients do not suffer
from delayed diagnoses [10]. For this purpose, the Emergency Radiology working group of the German Radiological
Society was founded in Germany in 2023 [58]. On an European level, the European Society of Emergency Radiology (ESER) was founded
in 2011 with the goal of providing emergency radiology training and certification.
One particular aim of ESER is to advance and improve radiological aspects of emergent
patient care and to advance the quality of diagnosis and treatment of acutely ill
or injured patients using imaging, by among others providing polytrauma imaging guidelines
[59] and by offering a dedicated curriculum that certifies a subspecialty in emergency
radiology with a European diploma [60]. Last but not least, multidisciplinary meetings (e.g., morning meeting of radiologists
together with trauma surgeons to review the cases of the night shift) are effective
for detecting missed fractures.