Teaching Case

A 30-year-old man consulted for a right-sided pretibial swelling. He noticed this swelling since 2 years after being kicked in this region during a soccer game. There was severe pain at palpation, but no spontaneous or nightly pain. Personal history includes ulcerative colitis. No biochemical abnormalities were found.

Imaging work-up started with standard radiographs and ultrasound.
On the radiographs there was an approximately 10 millimeter sharply delineated translucency in the anterior cortex of the proximal half of the tibia (Fig. 1), without soft-tissue swelling.
The ultrasound showed an irregular anterior cortex of the tibia with a groove-like defect and a hypo-echogenic lesion of 15x16 millimeters in the cortical defect (Fig. 2a). There was no blood flow detectable inside the lesion and mild hyperaemia around the lesion (Fig. 2b).
For further characterisation, magnetic resonance imaging (MRI) with intravenous gadolinium administration was performed.
The MRI images demonstrated a fusiform thickening of the anterior cortex of the tibia.
The lesion which was located centrally in this cortical thickening was iso-intense to muscle on T1-images, hyperintense on T2-images and showed mild uniform enhancement (Fig. 3).
There was no oedema in the surrounding soft tissue or bone.

After this imaging work-up there was a preference for the diagnosis of adamantinoma and subperiosteal chondroma. On imaging it was not possible to differentiate between these two lesions, especially as the lesion was quite small (16 mm). The more extensive radiological differential diagnosis is listed below.

Because the lesion was very small the decision was made to remove the entire lesion as this would not impact the stability of the tibia and if biopsy was performed there might not be enough tissue to correctly diagnose the lesion. It was only afterwards that the definite diagnosis was made because of histopathological and immunohistochemical examination. This determined the lesion to be a chordoma.
Histopathologically, the tumour was composed of large cells with clear to eosinophilic cytoplasm in extracellular myxoid/chondroid matrix. Some tumour cells had a vacuolated ‘bubbly’ cytoplasm (‘physaliphorous cells’), compatible with the diagnosis of chordoma (Fig. 4). Immunohistochemistry was performed and showed strong and diffuse expression for pancytokeratin AE1/AE3, EMA and patchy expression for S100. The tumour cells were strong nuclear positive for the brachyury immunohistochemical staining, confirming the diagnosis of chordoma.

Background
Chordomas are rare malignant tumours that arise from notochordal remnants that are normally only found in the axial skeleton. They can demonstrate great variability in morphological appearance [1,2]. Chordomas are usually in the differential diagnosis when there is a lesion in the axial skeleton, they are however among the tumours with a lower incidence and therefore not the first diagnosis.
Chordomas are usually not appreciated on plain radiography. If they are detected they present as a sharply delineated lytic lesion exhibiting a very wide differential diagnosis. On CT there is better appreciation of the bony involvement (especially in the skull base).  On MRI these tumours will demonstrate intermediate to low-signal intensity on T1-weighted imaging with possible small hyper intense foci (mucus or haemorrhage). The most important sequences are T2-weighted images on which they will almost always demonstrate high-signal and contrast-enhanced T1-weighted images where they show inhomogeneous enhancement [3].
For the definite diagnosis there is always need of pathological and immunohistochemical analysis. On pathological examination there can be a specific image, but this is not always the case. However, immunohistochemical analysis have shown that brachyury is essential for notochordal differentiation and it has been shown that this molecule is specific for chordomas [2]. Extra-axial location of this tumour is extremely rare with only 28 cases (16 in soft tissue and 12 in bone) reported in the literature to date [1]. In these publications the appendicular chordomas usually occurred in a long bone, and presented as sharply delineated (sometimes multilocular) lytic lesions on plain radiography without sclerotic margin and without preference for a location in the bone (cortical, medullary, parosteal, epiphyseal). These lesions had a low to intermediate signal-intensity compared to muscle on T1-images, a high-signal intensity on T2-images and (usually) inhomogeneous contrast enhancement [4].
Clinically it is important to make the correct diagnosis as this affects the staging, treatment and suspected outcome as well as follow-up and possible therapy after local resection. The first choice of therapy for chordomas is local resection, but this is often not possible because of their location in the axial skeleton and their size at diagnosis. The most frequent location of metastasis for axial and extra-axial chordomas is the lungs.
In our patient, the diagnosis was only made after complete resection of the lesion. To our opinion, even in retrospect it was not possible to make the diagnosis of extra-axial chordoma on imaging because of the non-specific imaging features. Staging was performed afterwards and was negative for distant metastasis or chordomas in the axial skeleton. To detect local recurrence the patient is monitored by follow-up MRI.
Take home message/teaching points
Chordomas are rare malignant tumours that usually arise in the axial skeleton, however in rare cases they can manifest extra-axial.
Imaging features are not specific, especially in extra-axial cases, and pathological and immunohistochemical analysis, they are necessary to make the definite diagnosis [4].
Written informed patient consent for publication has been obtained.