Childhood osteomyelitis: imaging characteristics

Conventional radiography is the initial modality of choice to evaluate osseous changes. In the majority of cases it will be the only imaging technique used in the diagnosis and treatment of childhood osteomyelitis.

The ease of access and the relatively low radiation dose make it an ideal imaging tool for skeletal pathologies. Also it is possible to exclude other pathologies such as malignancies and fractures. However, conventional imaging should not be used to exclude osteomyelitis in the first 10 days of symptoms [4, 20].

Karmazyn et al. advise performing at least two orthogonal views of the body part of interest [19]. Comparison with radiographs of the opposite limb is seldom useful, and therefore they should not be routinely obtained.

Conventional imaging has a reported sensitivity of 20–75 % and a specificity of 75–83 %, but evidence is limited (level of evidence II–III) [2‐4, 9].

Ultrasonography, especially in children, is a very useful and versatile modality. It allows correlating the physical exam with US findings and comparing the affected side with the opposite side [3, 21, 22]. It is important to use a linear high-frequency transducer.

US has a high reported sensitivity for detection of sub-periostal abscesses and fluid collections [3, 22], also complications such as soft tissue abscess, sequesters, fistulas and sinus track formation can be evaluated by US. Furthermore, it is possible to perform direct intervention, such as aspiration or drainage of an abscess or joint effusion (Fig. 1). A drawback of US is the inability to evaluate bone marrow involvement and the operator dependency. The sensitivity of US in osteomyelitis is reported to be 46–74 %, with a high specificity of 63–100 % (level of evidence: III) [21, 23].

In the acute workup of osteomyelitis, computed imaging plays no role.

In chronic osteomyelitis, it is more useful and superior to MR in detecting cortical destruction, air and sequesters [3, 4, 19, 20, 24‐26]. The main advantage of CT is the high spatial resolution of bony structures and surrounding soft tissue, and can also be used in the diagnostic workup of CT for image-guided aspiration/intervention (Fig. 2).

A major disadvantage of CT is the relatively high amount of radiation; therefore, CT should be used according to the ALARA principle (As Low As Reasonably Achievable) [27].

No specific literature was found on the sensitivity and specificity of CT in detecting acute osteomyelitis and its complications. Based on a systematic review, the sensitivity and specificity of CT in chronic osteomyelitis were found to be respectively 67 % and 50 % (level of evidence: III) [28].

After conventional imaging MRI is the most important imaging modality for the evaluation of osteomyelitis. MRI not only allows excellent delineation of the osseous component but also soft tissue extension, joint effusion and complications of osteomyelitis such as abscess formation, especially in case of failure to respond to therapy [25].

In young children the main disadvantage of the use of MRI is the long scan time during which the patient has to remain immobile. This leads to the need for general anaesthesia (intubation) or sedation in young children. Other approaches, such as immobilisation with for example a vacuum mattress or a feed and swaddle protocol, have been advocated, thus obviating the need for general sedation [29]. For older children having enough time and patience to make the child feel comfortable and cooperative is mandatory. Adaptations to the environment, such as the use of an open MRI, the presence of parents or distraction during imaging, can also be helpful.

First of all, it’s important to choose the appropriate coil for the maximum zone of interest and adequate details. There always needs to be a balance between the signal-to-noise ratio and image resolution. In children, it is wise to start with the most important sequences, in case the examination has to be stopped before the end of the planned scan. Long sequences should be avoided, but if really needed, they can be separated in contiguous series. The most important sequences for the diagnosis and follow-up of osteomyelitis are summarised in Table 4 [30]. The use of gadolinium-enhanced MR is still under debate, especially when unenhanced series are normal. Nevertheless, research by Kan et al. shows that the sensitivity of contrast-enhanced MR is higher for abscesses than that of non-enhanced MR. Indications for using contrast are spinal osteomyelitis and cases where the initial studies are inconclusive and clinical symptoms persist [4, 31].

MRI has a high sensitivity for detection of osteomyelitis of 82–100 % and high specificity of 75–99 % (level of evidence: III) [3, 19, 24]. In cases of multifocal sites or if the localisation of the symptoms is doubtful, a whole-body MRI can be performed, although there is poor evidence (level of evidence: IV).

Bone scintigraphy can play a role when localisation of osteomyelitis, based on clinical information or other radiological methods, is not possible [32]. However, in the last few years nuclear imaging has increasingly been replaced by whole-body STIR MRI, thus allowing for one-stop-shop imaging.

The use of FDG-PET CT in the diagnosis of osteomyelitis has mainly been reported in adults [33]. For paediatric use a drawback is the high effective radiation dose, which can be in the range of 5–18 mSv [34].

Although limited evidence is available on paediatric osteomyelitis, PET CT can effectively differentiate between an active and inactive focus of chronic osteomyelitis [19]. Literature on the use of FDG PET in adults shows a sensitivity of 94–100 % and a specificity of 75–99 % (level of evidence: IV) [19].

In the first 48 h, deep soft tissue swelling in the metaphyseal region and loss of fat planes can be visible on conventional imaging (Fig. 5).

Several days after the onset of symptoms, periostal reactions can be seen (Figs. 6 and 5), and after 7 to 21 days lytic lesions can be seen in affected bones (Figs. 6 and 4). Eventually, this can lead to marked destruction of the bone with extension into the cortex (“endostal scalloping”). Bone destruction can appear as lucency but also in a permeative pattern. Destruction can be limited to the metaphysis, but can also, in rare cases, if misdiagnosed or left untreated, extend into the physis and epiphysis (Fig. 7). This can be very difficult to diagnose on conventional radiographs because in children the epiphysis is not yet ossified [3, 4, 6, 19, 26]. The presence of a joint effusion can be a helpful clue, although plain radiographs have a low sensitivity for detecting joint effusion (Fig. 8).

US allows evaluation of soft tissue swelling due to oedema or fluid collection, and hyperaemia can be detected using Doppler techniques. Also periosteal thickening and sub-periostal collections are seen. These findings can precede abnormalities on conventional imaging in the first week (Figs. 5 and 7).

US has high sensitivity for detection of intra-articular fluid [3, 4, 21, 24, 25, 30] (Figs. 6 and 8).

In more extensive cases, cortical defects can also be detected (Fig. 8). It can be useful for detecting foreign bodies, especially non-radioopaque foreign bodies such as wood splinters as a causative agent. Sometimes it is possible to image formation of a sequesters, especially when surrounded by pus (Fig. 9).

CT is useful in the evaluation of bony destruction (Fig. 10) and in cases with complications such as abscess, fistula or sequester formation (Fig. 9). Intravenous contrast agents can be useful for the evaluation of soft tissue extension, although MRI has a higher sensitivity for the evaluation of soft tissue involvement (Fig. 10).

After 3–5 days of onset of infection, MRI can detect osteomyelitis [35]. MRI is the best technique available to evaluate changes in the bone marrow/water content of bone marrow. Alteration of bone marrow signal intensity can be visible 1 or 2 days after the onset of infection (Figs. 6 and 8) [3].

MRI criteria for diagnosis are low signal intensity on T1, high signal intensity on T2/STIR, and enhancement of (subperiosteal) bone and/or soft tissues and abscess/collections after contrast administration [36].

MRI is also useful for evaluation of complications such as (subperiosteal) abscesses, joint effusions and soft-tissue extension that would require surgical treatment (Figs. 8, 10 and 11) [19].

Acute osteomyelitis on a three-phase bone scintigraphy can show increased blood flow and blood pool activity, and positive uptake on 3 h images in areas of osteomyelitis. However, at this time blood pool flow and pool activity can be normal or photon deficient because of devascularisation of the periostium due to rapid raising of the periostium [4, 37].

PET-CT will show high uptake in areas of osteomyelitis, in bone marrow and in soft tissue.

Since defining a differential diagnosis solely based on imaging characteristics can be very difficult, clinical and laboratory parameters and communication with clinicians remain very important in the diagnosis of osteomyelitis. In cases of doubt, fine-needle aspiration or an image-guided biopsy will eventually be necessary to determine the definitive diagnosis.

In patients without any infectious parameters, the differential diagnosis is mainly based on disease of vascular origin, neoplasm or trauma. This group without infectious parameters can be the most challenging for the differential diagnosis.

Patients with a history of use of medications such as corticosteriods or for example patients with sickle cell disease must raise the suspicion of vaso-occlusive disease. Imaging will show typical serpentinous sclerotic lesions on conventional imaging, whereas MR imaging will show linear hypointense T1- and T2-weighted changes in the meta- and epiphysis [26].

Ewing’s sarcoma produces a large soft tissue mass without calcifications or bone matrix in the tissue. Onion-skin periostitis on conventional imaging is typical.

Osteosarcoma shows a lesion usually in the diaphysis with signs of aggressive (cortical) bone destruction and periostic reactions such as Codman’s triangle and sunburst spiculated periostitis.

In a patient with a history of trauma, stress fractures will show linear, hypointense changes on T1-weighted imaging, without enhancement. Bone marrow oedema can be visible on STIR imaging.

When imaging shows an osteolytic lesion with a central nidus surrounded by a sclerotic margin in a patient with a typical history of night pain relieved by use of NSAIDs, one should consider the diagnosis of osteoid osteoma.

In patients with suspicion of haematologic malignancy such as ALL, diffuse infiltration of the bone marrow can be seen. Imaging will show diminished signal intensities on T1 imaging; T2 will show heterogenic signals. Metastastic neuroblastoma can show multiple lesions usually characterised only by high signal intensity on STIR imaging.

In young children with a red painful or asymptomatic lesion on the sternum with or without infectious parameters, one should consider a self-limiting sternal tumour of childhood (SELSTOC) [38]. Ultrasound in these patients will show a dumbbell-shaped lesion extending to the area behind the sternal bone. No additional imaging will be needed in these patients.

In patients with infectious parameters in combination with backache, one must consider a spondylodiscitis. Imaging shows low signal of the disc with surrounding fluid/abscess with destruction of the vertebrae and rim enhancement after gadolinium administration (Fig. 10); Pott’s disease in tuberculosis will be more pronounced in the facet joints.

After a period of septicaemia, such as a fulminant meningococcemia, septic emboli can occur. This can lead to partial premature closure of the physis leading to joint deformities (Fig. 12).

Conventional imaging will show the same features as in acute osteomyelitis, although more extensively (Fig. 7). Complications such as fractures or sequester formation can be seen.

CT plays an important role in the diagnosis and analysis of chronic osteomyelitis. Imaging of complications of osteomyelitis such intraosseous abscesses, sequesters, and fistulas or cloaca formation is important (Figs. 8 and 9).

Soft tissue extension can be evaluated, but to a lesser extent compared to MR. Imaging characteristics are comparable to those for acute osteomyelitis.

MR can show soft tissue extension and abscess. Also intraosseous abscesses and fistulas can be seen. Imaging characteristics are comparable to those of acute osteomyelitis. Use of total body STIR can exclude multifocal disease.

PET scans will show uptake in the area of osteomyelitis. It can exclude multifocality.

The main differential diagnosis of chronic osteomyelitis consists of neoplasms.

In Ewing’s sarcoma a large soft tissue mass without calcifications or bone matrix in the tissue can be seen. Onion-skin periostitis is typical.

LCH can be difficult to differentiate from multifocal osteomyelitis. LCH typically produces punched-out lesions on conventional imaging. Therefore, conventional imaging can provide a valuable clue.

Bone metastases are usually multifocal. In this diagnosis also the clinical parameters are important signs; in most cases, there are no signs of inflammation.

Chronic recurrent multifocal osteomyelitis (CRMO) can be part of SAPHO syndrome (synovitis, acne, pustulosis, hyperostosis, osteitis). It is a very uncommon condition, of unknown aetiology, with a peak incidence between the ages of 4–14 years. Characteristically, there are multiple locations, but only one is symptomatic [39]. The imaging strategy usually has a symptomatic focus, as in acute osteomyelitis, but multiple lesions will be shown. Especially MR and PET can be used to analyse multiple lesions. STIR and T2 series show multiple spots of high signal intensity and series after contrast show enhancement. Imaging characteristics are comparable to those of acute osteomyelitis. The focus of osteomyelitis and symptoms can change over time (Fig. 13).