SMART syndrome: two cases highlighting a complex and rare complication of brain irradiation

Abstract

Stroke-like migraine attacks after radiation therapy (SMART) syndrome is a rare complication of radiotherapy with complex neurological impairment. Patients present with neurological symptoms and signs such as migraine, hemianopsia, hemiplegia, aphasia and/or seizures—without recurrence of neoplastic disease. In this report, we describe SMART syndrome in two adult patients 4 and 14 years following brain irradiation, respectively.

Background

Brain tumour survivors may develop stroke-like migraine attacks after radiation therapy (SMART) syndrome, a rare complication characterised by recurring episodes of complex neurological impairment without neoplastic recurrence.¹ The syndrome has been reported to develop between 1 and 37 years following brain irradiation.^{2 3} Its exact pathophysiology is poorly understood, and the pathogenesis is believed to be multifactorial.^{4 5} Onset is usually subacute, with neurological symptoms and signs such as migraine, hemianopsia, hemiplegia, aphasia and/or seizures.⁴ Radiation dose associated with SMART syndrome is usually ≥50 gray (Gy), but SMART cases in patients treated with less than 50 Gy have also been reported.^{3 6} The hallmark, although far from pathognomonic, MRI features are reversible, transient, unilateral cortical gadolinium enhancement with correlative abnormal T2 and fluid attenuated inversion recovery (FLAIR) signal without respect of vascular borders.^{4 7 8} Although SMART syndrome is usually reversible within weeks to months, there have been reports of permanent neurological sequelae and imaging abnormalities.^{4 9} There is currently no clear consensus regarding the optimal treatment approach. We here present two case reports fulfilling criteria of SMART syndrome.

Case presentation 1

A male patient in his 50s was admitted to his local hospital due to recurrent brief episodes of headache and paresis of the left hand (day 1). Fourteen years prior, he had been diagnosed with glioblastoma multiforme (GBM) WHO grade IV localised in the right parietal lobe. Isocitrate dehydrogenase mutation status and O6-methylguanine-DNA methyltransferase promoter methylation status were not tested for at that time, and hence unknown. Besides asthma, he was otherwise healthy prior to the GBM diagnosis. Following surgical resection of his GBM in 2006, neuro-oncological treatment consisting of fractionated partial brain radiotherapy (RT) to 60 Gy with concomitant and six cycles of adjuvant temozolomide was administered. The patient was thereafter followed for 14 years with no clinical or radiological signs of recurrence. Following his GBM diagnosis, he was successfully treated with valproic acid. He was later also diagnosed with diabetes mellitus type 2, a probable transient ischaemic attack and deep venous thrombosis.

Three weeks prior to hospitalisation for what was later diagnosed as SMART syndrome, a follow-up MRI showed no evidence of recurrent disease. The headaches and left-hand paresis causing hospital admission lead to a CT scan that did not reveal any new changes. Focal epileptic seizures were suspected, and valproic acid dosage was adjusted before the patient was discharged.

The patient was re-admitted the following day with focal status epilepticus (day 4). He suffered from headaches and had progressive attacks with dysphasia and left-sided hemiparesis. A repeated MRI the same day did not reveal any new changes. Several antiepileptic drugs (AEDs) were administered, unfortunately without sufficient antiepileptic effect, and due to intractable epileptic seizures, the patient was sedated to burst suppression at day 9 and remained sedated for 13 days. Electroencephalogram (EEG) on day 12 showed a rhythmic and partly sharp activity spreading from the right temporal lobe to both hemispheres when propofol was paused. Cerebrospinal fluid (CSF) examination was, except for a slightly elevated protein level (0.76 g/L), normal including negativity for neurotropic viruses.

SMART syndrome was suggested as a differential diagnosis at day 9, and steroid pulse therapy (methylprednisolone 1000 mg/day intravenously for 5 days) was initiated. In addition, AED was further optimised. A repeated MRI at day 12 revealed restricted diffusion and enhanced T2/FLAIR signal in right frontal, temporal and parietal areas. No cortical or subcortical contrast enhancement was seen, and the findings were interpreted as postictal ischaemia.

Despite these efforts, the patient’s neurological impairment and treatment refractory epilepsy persisted. At day 26, nearly 4 weeks following the first hospital admission, a repeated MRI revealed gadolinium contrast enhancement, partly patchy and partly gyral located in the same areas which 2 weeks earlier had restricted diffusion (figure 1A). Some cortical areas had contrast enhancement and oedema, but most of the cortex was spared. T2/FLAIR signal intensity was unchanged compared with MRI at day 12 (figure 1B).

Figure 1.

Coronal MRI of the brain at day 26 and MRI superimposed on the previous radiation therapy plan (case 1). (A) T1-weighted sequence with gadolinium shows gadolinium contrast enhancement, partly patchy and partly gyral, located in right temporal and parietal areas. Some cortical areas had contrast enhancement and oedema, but most of the cortex was spared. (B) T2/FLAIR sequence shows enhanced signal in right temporal and parietal lobes. (C) T1-weighted sequence with gadolinium and dose distribution of previous radiation therapy plan. Green and red colours represent 50% and 95% isodose levels, respectively. (D) T2/FLAIR sequence and dose distribution of previous radiation therapy plan. Green and red colours represent 50% and 95% isodose levels, respectively. FLAIR, fluid attenuated inversion recovery.

The tertiary referral hospital where the patient had previously received radiation therapy was asked to do a second opinion of the cerebral MRI. Although most of the cortex was spared, SMART syndrome was believed to be the most probable diagnosis based on previous brain irradiation, clinical presentation, MRI findings and later also the autopsy report. When the latest MRI was imported into the radiation therapy plan administered 14 years earlier, all areas with MRI changes matched areas which had received a minimum of 45 Gy (figure 1C, D).

Outcome and follow-up

During hospitalisation, the patient developed pneumonia, which was initially clinically and biochemically successfully treated with broad-spectrum antibiotics. At day 30, 4 weeks following his first admission to hospital, the patient developed septic shock with multiorgan failure, and he deceased the following day. An autopsy was performed and showed pronounced radiation damage in the right cerebral hemisphere. Findings included pronounced radiation damage in blood vessels with thickening, sclerosis, partly calcified walls and severe stenosis of lumina, thinning of white matter with spongiosis and haemosiderin-laden macrophages, as well as gliosis with reactive astrocytes. There were also findings suggestive of multiple acute and subacute cerebral infarctions in the right cerebral hemisphere, mainly located subcortically. There were no signs of tumour recurrence. The only abnormal findings in other organ systems were congested pulmonary vessels with a number of granulocytes without signs of acute or established pneumonia.

Case presentation 2

A female patient in her late 60s was diagnosed with a left parietal and occipital oligodendroglioma. WHO grade was inconclusive but closer to grade III than grade II. The patient’s medical history included surgery for a herniated intervertebral lumbosacral disc and depression. A subtotal resection of what turned out to be an oligodendroglioma was carried out. Subsequent neuro-oncological treatment consisting of fractionated partial brain RT to 54 Gy followed by adjuvant chemotherapy with procarbazine, lomustine and vincristine (PCV) was administered. Due to haematological toxicity, she only received two cycles of PCV. Two and a half years after completion of antineoplastic therapy, the patient experienced recurrent episodes of progressive dysphasia, right-sided hemianopsia and right-sided central facial paresis believed to be simple focal epileptic seizures. Initially, treatment with the AED levetiracetam was successful. Clinical history at this time also included progressive cognitive impairment, and the patient was shortly after diagnosed with dementia.

Approximately 4 years postirradiation, the patient experienced repeated simple and complex focal epileptic seizures with neurological symptoms and signs as described above. Because of this, she was repeatedly admitted to her local hospital over a period of 3 months. Progressive cognitive impairment had occurred over the last few years without any MRI changes suggestive of progressive neoplastic disease or any other pathology. EEGs were repeatedly pathological with slow multifocal theta and delta activity but without epileptiform activity. Adjustment of levetiracetam dosage seemed to solve the epileptic problems. Nonetheless, at the end of this 3 month period, she was again admitted to her local hospital due to an episode of protracted complex focal epileptic seizures with dysphasia and right-sided central facial paresis. EEG now demonstrated generalised cerebral dysrhythmia most pronounced over the left cerebral hemisphere, partly interpreted as epileptogenic activity. An MRI taken during this admission revealed pronounced cortical changes with oedema, laminar contrast enhancement and partly increased diffusion without decreased apparent diffusion coefficient (figure 2A). Changes affected all of the left occipital lobe, the majority of the left temporal and parietal lobes, as well as cranial and posterior parts of the left frontal lobe. T2/FLAIR hyperintense signal distribution was at this time point located also in the cortex, in addition to all areas in the MRI 3 months earlier (figure 2B). Recurrent disease was suspected, but due to previous haematological toxicity and cognitive impairment, chemotherapy was not an option. Neither re-irradiation nor surgery was recommended. Perampanel was added to levetiracetam as AED treatment, and the patient was discharged.

Figure 2.

Coronal MRI of the brain and MRI superimposed on the previous radiation therapy plan (case 2). (A) T1-weighted sequence with gadolinium: pronounced cortical contrast enhancement in the left occipital lobe, the majority of left temporal and parietal lobes, as well as in cranial and posterior parts of the left frontal lobe. (B) T2/FLAIR sequence shows enhanced signal in cortical and subcortical areas of left temporal and parietal lobes. Signal enhancement in white matter areas was similar to previous MRI due to post-radiation gliosis. (C) T1-weighted sequence with gadolinium and dose distribution of previous radiation therapy plan. Green and red colours represent 50% and 95% isodose levels, respectively. (D) T2/FLAIR sequence and dose distribution of previous radiation therapy plan. Green and red colours represent 50% and 95% isodose levels, respectively. FLAIR, fluid attenuated inversion recovery.

The tertiary referral hospital where the patient had previously received radiation therapy was involved for a second opinion of the cerebral MRI. Because of the changes described above, SMART syndrome was suspected, and a follow-up MRI was recommended. When the MRI was superimposed onto the radiation therapy plan administered 4 years earlier, all areas with MRI changes matched the areas that had received a minimum of 45 Gy, further supporting the diagnosis of SMART syndrome (figure 2C, D).

Outcome and follow-up

Follow-up MRI 4 months later showed complete regression of previous contrast enhancement and pathological T2/FLAIR signal, further substantiating SMART syndrome as the correct diagnosis. However, the patient’s cognitive and neurological status worsened, and she died 9 months after being admitted with SMART syndrome. AED treatment with levetiracetam and perampanel was initially successful, but she later required two admissions to her local hospital due to epileptic activity; the most recent of which occurred 4 weeks before she deceased. Unfortunately, an autopsy was not performed.

Discussion

Increased awareness of the rare SMART syndrome as a complication to brain irradiation has led to several case reports and some reviews.⁹ We contribute with a report of two additional cases.

Shuper and colleagues were the first to publish case reports consistent with SMART syndrome when they, in 1995, reported of complicated migraine-like episodes in four children following brain irradiation.¹⁰ Approximately a decade later, in 2006, Black and colleagues described and defined SMART syndrome.¹ Four diagnostic criteria were established; a history of brain irradiation without evidence of recurrent disease, debut of prolonged and reversible unilateral cortical symptoms and signs years after irradiation, transient, diffuse and unilateral cortical gadolinium enhancement of cerebral gyri sparing the white matter within the previously irradiated area, and no other attributable cause.

Both cases presented here had previously received high-dose partial brain irradiation. Onset of SMART syndrome was 14 and 4 years after irradiation, respectively, fitting well with SMART syndrome in which debut is reported to range from 1 to 37 years postirradiation.^{2 3} Both our cases presented with a subacute symptom onset and signs consistent with those described in published cases.⁴ As the first patient deceased due to a septicaemia 4 weeks after the debut of SMART syndrome, we do not know if the neurological and radiological changes would have been reversible. The second case had reversible radiological changes but deteriorated neurologically and cognitively—probably due to progressive dementia. Complete recovery is reported in 60%–85% of cases.^{9 11} A retrospective study of 25 patients with SMART syndrome showed that prolonged recovery was associated with older age, temporal lobe involvement and diffusion restriction on MRI.¹¹ The first case presented here had temporal lobe involvement and diffusion restriction on MRI.

MRI is the modality of choice to diagnose SMART syndrome, and the hallmark features are reversible, transient, unilateral cortical gadolinium enhancement and correlative abnormal T2 and FLAIR signal.^{4 8} In the first case presented here, cortical areas were relatively spared, but based on clinical history, the totality of MRI findings, distribution of RT high doses corresponding to MRI changes and autopsy findings, we believe SMART syndrome is the most probable diagnosis. It has also been suggested that there is a time window to capture the hallmark radiological features of SMART syndrome, possibly explaining why all MRI changes were not present.¹² Our second case had the typical radiological hallmarks of SMART syndrome, and all these had disappeared on follow-up MRI 4 months later.

Due to a broad spectrum of differential diagnoses, it may be difficult to exclude other likely causes with certainty. In both cases presented here, the most likely differential diagnosis was focal epileptic seizures. In addition, tumour recurrence was initially suspected in the second case, but for several reasons antineoplastic therapy was not an option. A challenge with tumour recurrence as a differential diagnosis to SMART syndrome is that in most cases antineoplastic therapy cannot be postponed for weeks to months, and biopsy of a SMART lesion is not recommended since the pathological characteristics of SMART syndrome have yet to be demonstrated.⁴ Other important differential diagnoses of SMART syndrome include second primary tumours, migraine syndromes such as hemiplegic migraine and pseudomigraine, circulatory disturbances such as ischaemic strokes, cerebral infection, and various encephalopathies and vasculopathies.¹³

If SMART syndrome is suspected, recommended diagnostic investigations are gadolinium-enhanced brain MRI, MR or CT angiography, CSF analysis, EEG, serum lactic acid level and blood pressure monitoring, aiming at diagnosing and excluding other causes.^{1 4 12} We believe that superimposing the SMART-suggesting MRI onto the radiation therapy plan to see if MRI-changes are located in the irradiated area is of great value. Both cases presented here underwent investigation with MRI and EEG; CSF analysis was performed in the first case, but angiography was not performed in either of them.

Currently, there is no consensus regarding the treatment of SMART syndrome, and no controlled study has been performed due to the rarity of the syndrome. However, treatment to control symptoms such as optimising AED is recommended, since it is challenging to completely distinguish SMART syndrome and epileptic seizures.² Glucocorticoid treatment, although not proven, is often administered as empirical acute-phase treatment to accelerate recovery.⁹ Furthermore, glucocorticoids can help distinguish SMART syndrome from tumour recurrence, possibly obviating unnecessary biopsies and antineoplastic therapy. Complete recovery was observed in four of five case reports treated with steroid pulse therapy (methylprednisolone 1000 mg/day, 5 days).¹⁴ Our first case received such steroid pulse therapy, but it was not possible to evaluate effect since he deceased due to a bacterial infection. Verapamil, a calcium channel blocker, has also been suggested as beneficial acute-phase treatment based on the hypothesis of vascular dysfunction as one pathophysiological mechanism of SMART syndrome.¹⁵ However, the same retrospective study as referred to above did not find verapamil to be associated with reduced risk of recurrent SMART syndrome or faster resolution of symptoms.¹¹

SMART syndrome after brain irradiation may potentially be underdiagnosed due to its complexity and rarity. With these case reports, we aimed to draw attention to SMART syndrome, its diagnostic work-up and treatment options, as well as to increase awareness and knowledge of the syndrome.

Learning points.

• Stroke-like migraine attacks after radiation therapy (SMART) syndrome after brain irradiation is a complex and rare diagnosis of which knowledge and awareness is important.

• Superimposing SMART syndrome suggesting MRI series and radiotherapy dose distributions can be helpful when considering SMART syndrome as a possible diagnosis.

• We encourage seeking a second opinion by a tertiary referral hospital or the hospital responsible for the radiotherapy administration whenever SMART syndrome is suspected.

Ethics statements

Patient consent for publication

Consent obtained from next of kin.