Temporal lobe radiation necrosis

Clinical History

An asymptomatic patient, who had been treated for mucoepidermoid parotid carcinoma by surgery and radiotherapy 5 years previously, was found on routine follow-up Magnetic Resonance Imaging (MRI) to have a growing temporal lobe enhancing lesion with low vascularity.

Imaging Findings

An asymptomatic man, who had previously been treated for a left parotid tumour, was incidentally found to have a focal lesion in his left temporal lobe on routine follow-up Magnetic Resonance Imaging (MRI).

A left parotid mucoepidermoid carcinoma had been excised surgically 5 years previously, after which the patient had undergone local field radiotherapy (total dose 60Gy). He had remained symptom-free since then. The follow-up MRI revealed an irregular lesion of heterogeneous signal intensity showing contrast (Gadolinium DTPA) enhancement with perifocal white matter signal intensity changes in his left temporal lobe (Figs 1a and 1b). No residual mass was identified in the left parotid space.

Proton MRS using the point resolved spectroscopy technique (PRESS) with TR = 1500msecs, No. of acquisitions = 128, and TE = 135msecs (Fig. 1c) revealed raised choline:creatine ratios and markedly reduced NAA:creatine levels. In addition a lipid peak was also seen.

Perfusion sensitive MRI was performed using the contrast enhanced first pass bolus technique and the relative cerebral blood volume map (rCBV) (Fig. 1d) calculated revealed the lesion to have a low vascularity with a normalised CBV ratio of 0.81 (+/- 0.23).

In the subsequent 2 months, the patient developed slurring of speech with mild recent memory loss. MRI now revealed growth of the lesion to twice its previous size and an increase in the perifocal white matter changes (Fig. 2a). MRS using similar parameters revealed an increase in the choline:creatine ratios, loss of NAA and a lipid peak. Perfusion MRI still revealed the lesion to have similar or lower vascularity in comparison with the normal contralateral white matter (Fig. 2b). The growth in size, the associated mass effect and the MRS findings prompted surgery in spite of the perfusion MRI findings of a lesion of low vascularity.

Discussion

The portions of the brain vulnerable to radiation necrosis are the proliferative cells including the oligodendrocytes and endothelial cells, rather than the stable neurons. Since the former cells have a very low turnover, radiation injury typically is delayed rather than acute. Pathology reveals fibrinoid changes in blood vessels, demyelination and coagulative necrosis. Cavitation with liquefaction and mural gliosis can also occur.

Radiation necrosis can occur with a latent interval ranging from 1.5 to 13 years [1]. Imaging findings in radiation necrosis to the brain [2] have been divided into four categories: Normal (I), lesions with only oedema (II), lesions with contrast enhancement (III), and cystic encephalomalacia (IV). The patient in this case showed an enhancing focal lesion that demonstrated rapid growth over time, further confusing with the appearances of a tumour. The growth presumably represented increasing necrosis of the surrounding oedematous-appearing areas.

Proton MRS has been widely used in radiation injury [2,3,4]. The typical findings are loss of NAA (the earliest and most severe abnormality), elevated levels of choline (occurring later when possibly injury to oligodendrocytes leads to demyelination, or due to reactive astroglosis) and presence of lactate due to ischaemic compromise. The present case illustrates the overlapping features of MRS findings in radiation necrosis and tumour since elevated choline and NAA loss could occur with tumours. The presence of lactate/lipid can occur with any condition resulting in increasing glycolysis as occurs in high-grade gliomas.

Perfusion sensitive MRI has been used extensively [5] and soon could be established as a routine clinical tool in the evaluation of radiation necrosis/ tumour recurrence. In fact perfusion sensitive MRI is the only reliable and consistent modality to demonstrate low vascularity of a lesion. GHA SPECT in these cases can be misleading, since GHA freely enters areas of blood-brain barrier breakdown and accumulation of GHA does not necessarily define the true vascularity of the lesion.

In conclusion, the present case demonstrates the overlapping features of radiation injury and tumours as seen on conventional MRI. Emphasis has been laid on the fact that contrast-enhancing lesions can actually grow in size, years after irradiation. The case also further elucidates the utility of perfusion sensitive MRI and MRS in the diagnosis of radiation necrosis. Perfusion sensitive MRI should be considered as a strong tool in determining the vascularity of a lesion and hence suggest the diagnosis. A combination of these modalities together with consideration of the clinical background should enable a diagnosis to be confidently established in a large proportion of patients.

Differential Diagnosis List

Final Diagnosis

Temporal lobe delayed radiation necrosis

URL:	https://www.eurorad.org/case/1557
DOI:	10.1594/EURORAD/CASE.1557
ISSN:	1563-4086