Radiation and Chemotherapy Induced Injury
The spectrum of clinical and pathological effects of ionizing radiation on brain and spinal cord tissue is wide and multifactorial, depending on several factors including patient’s age, cumulative irradiation dose, type of radiotherapy, and concomitant chemotherapy or radiosensitizing agents. Radiation- and chemotherapy-induced neurotoxicity poses a challenge in clinical neuroradiology as they need to be promptly recognized, being aware of patient treatment histories to avoid futile discontinuation of an effective therapy. Furthermore, long-term complications of radiation therapy such as radiation necrosis, radiation-induced leukoencephalopathy, and secondary neoplasms may impact on the patient's management and clinical outcome. Radiological techniques such as conventional and advanced MRI play a pivotal role in the recognition of these entities in their acute, subacute, and late presentation. Ionizing radiation mainly affects glial and endothelial cells, the latter being involved both in the development of parenchymal lesions, such as radiation necrosis and radiation-induced leukoencephalopathy, and in vessel damage. This chapter evaluates the current knowledge in the diagnosis of acute and delayed sequelae of radiation therapy and concomitant or adjuvant chemotherapy on brain and spinal cord, with a particular focus on radiation injury, radiation-induced vasculopathy, and SMART (Stroke-like Migraine After Radiation Therapy) syndrome.
KeywordsRadiation-induced neurotoxicity Radiation necrosis Radiation-induced vasculopathy Postradiation imaging changes Treatment-induced leukoencephalopathy/neurotoxicity
List of Abbreviations
Apparent diffusion coefficient
Central nervous system
Dynamic Susceptibility Contrast
Diffusion tensor imaging
Hyper fractionated accelerated radiotherapy
Hypoxia-inducible factor 1α
Intensity-modulated radiation therapy
Volume transfer constant
Magnetic resonance angiography
MR proton spectroscopy
Positron emission tomography
Quality of life
Response Assessment in Neuro-Oncology
Relative cerebral blood flow
Relative cerebral blood volume
Region of interest
Stroke-like Migraine After Radiation Therapy
Vascular-endothelial growth factor
Fractional volume of the intravascular compartment (aka fractional plasma volume)
Whole-brain radiation therapy
Brain and Spinal Cord Radiation Injury
Definition of Entity
Radiation therapy (RT) is a major curative and palliative treatment of patients with primary CNS tumors as well as for brain and spinal cord metastases originating from extra-CNS neoplasms and other nonmalignant lesions such as arteriovenous malformations. Additionally, the radiation fields for upper digestive tract malignancies (i.e., pharynx and nasal cavities) often partially include normal brain and spinal cord tissue.
The effects of ionizing radiation on brain and spinal cord tissue vary from a mild edema to frank necrosis, and injured tissues can show transient focal lesions as well as diffuse white matter injury and grey matter atrophy. As such, the effects of RT on the CNS constitute a significant source of morbidity, which can be particularly relevant in patients already presenting with tumor-related clinical impairment. Three types of radiation injury to the brain and spinal cord can be identified on the basis of the time to development of signs and symptoms: acute toxicity, early delayed (or subacute) injury, and delayed (or late) injury.
Recent advancements in multimodal treatment protocols have led to improved survival rates in oncological patients. About half of the patients receiving RT survive more than 6 months, and in many of them long-term disease control is achieved. Consequently, the possibility to develop radiation-induced brain and spinal cord radiation injury has increased alongside longer survival, as well as the relevance of radiation effects on the nervous tissue and their impact on long-term morbidity, especially for cognitive functions and quality of life (QoL). Radiation effects on the brain present as late neurological sequelae with or without gross tissue necrosis, and 50–90% of long survivors exhibit cognitive deterioration which is often progressive and disabling. Affected cognitive domains include learning, memory, processing speed, attention, and executive function (Makale et al. 2017). Different patterns emerge according to the timing of onset with respect to RT.
Acute effects are rare and often asymptomatic with conventional dose fractionation schemes.
Early delayed radiation effects: Pseudoprogression (PsP) is the most common, with increasing lesions on imaging in high-grade glioma patients during or shortly after the end of concomitant chemo/radiotherapy, according to the Stupp protocol (Thust et al. 2018b). The clinical definitions of this phenomenon have been quite variable, which may explain some of the differences in reported incidences, which range from 9% to 30% (Radbruch et al. 2015) Patients with pseudoprogression tend to be significantly younger and less often clinically symptomatic than those with early progressive disease (34% vs. 57%), and their tumors were more often O6-methylguanine methyltransferase (MGMT) promotor methylated (Taal et al. 2008) (see also pseudoprogression and pseudoresponse in the “Intra-Axial Tumor” chapter).
Risk factors for radiation injury
Type of RT and dose-volume predictors
IMRT (intensity-modulated radiation therapy)
• The total dose that leads to a 5% risk of significant toxicity (RN or cognitive decline) at 5 years for fractionated RT is 72 Gy
• The conventional upper limit of RT delivered to the brain with standard fractionation (i.e., 1.8–2 Gy per day) is 60 Gy
• Brain is particularly sensitive to fraction sizes >2.5 Gy and to fractions delivered more frequently than once daily
SRS (stereotaxic radio-surgery)
• For patient undergoing SRS (with or without WBRT) the risk of complications increases with the size of the target volume
• Toxicity increases rapidly once the volume of the brain exposed to >12 Gy is >5–10 cm3
Prior radiation exposure
• A meta-analysis of brain re-irradiation (interval between courses, 3–55 months) found no cases of RN when the total radiation dose was <100 Gy (Mayer and Sminia 2008)
• Prior WBRT or SRS and the time interval between re-irradiation influences the risk of RN
• The use of adjuvant chemotherapy increases the risk of RN by approximately fivefold
• Multiple transient enhancing cerebral lesions are frequently seen on MRI scans soon after high-dose chemotherapy and RT for childhood malignant brain tumors
• Concomitant administration of chemotherapy and RT has been associated with diffuse white matter injury in children (Rossi Espagnet et al. 2017)
Site of irradiation
• Hippocampus and frontal lobe sensitive; brainstem lowest risk
• For brain metastases treated with SRS, renal and lung carcinoma, HER2 amplification, and ALK/BRAF mutational status are predictive of RN (Miller et al. 2016)
• No differences between glioma histological types
Planning Target Volume (PTV) margin
• In SRS for brain metastases, larger GTV (gross tumor volume) to PTV margins gives a higher incidence of RN without improved local control
A peculiar manifestation of late delayed radiation injury can be observed in patients treated with SRS – the delivery of high doses of radiation administered to small targets in a single or in a limited number (3–5) of fractions, by using multiple, nonparallel radiation beams that converge on the target lesion (i.e., Gamma Knife, Linac, CyberKnife systems). The most frequent pathologies treated with SRS are metastatic tumors, meningiomas, vestibular schwannomas, arteriovenous malformations, pituitary adenomas, trigeminal neuralgia, and gliomas. In up to one half of these patients, a transient increase of lesion volume can be observed at follow-up imaging without any worsening of clinical symptoms, that usually resolves without any change in therapy.
Some patients are at greater risk of developing therapy-induced neurotoxicity than others, although predisposing factors remain poorly understood and therefore, it is difficult to accurately predict individual risks in a clinical setting. Nonetheless, it is known that the developing CNS of children is particularly vulnerable to the effects of radiation (Rossi Espagnet et al. 2017). Indeed, in the pediatric population, the incidence of RN in children with embryonal tumors treated with surgery, craniospinal irradiation, and chemotherapy is about 3.7% at 5 years in a large series of patients (Murphy et al. 2012), with the proportion of infratentorial brain volume receiving more than 50 Gy as main predictor of necrosis.
The risk of developing late delayed injuries increases when chemotherapy is associated with brain irradiation, often presenting with other patterns of neurotoxicity than RN. In children with primary malignant CNS tumors treated with RT and high-dose chemotherapy with myeloablative doses of thiotepa, transient white matter lesions (also referred to as transient focal enhancing lesions) can occur at follow-up imaging and usually resolve spontaneously within a variable period of time (Rossi Espagnet et al. 2017).
A delayed form of diffuse white matter injury after brain irradiation is commonly referred to as radiation leukoencephalopathy, the risk of which increases with concomitant administration of chemotherapy and whole brain irradiation. It is particularly common in patients treated for both solid and hematologic CNS malignancies. Increase risk of development of leukoencephalopathy is associated with higher radiation doses, methotrexate administration, and younger age of the patient.
Brain atrophy may appear as a side effect of RT and a progressive decrease in gray matter (GM) volume over time and a dependence on radiation dose. A regional vulnerability of the brain has been described, with evidence of atrophy and cortical thinning of hippocampal and higher-order association cortex regions, whose magnitude parallels 1-year atrophy rates seen in neurodegenerative diseases and may contribute in part to cognitive decline following brain radiation therapy.
Spinal cord radiation myelopathy, also known as radiation myelitis, is a rare complication in patients exposed to RT for malignancies not directly involving the spine such as lymphoma, gastric, head and neck, lung, or nasopharyngeal cancer. Multiple segments of the spinal cord included within the radiation field may develop myelitis, with higher risk when intrathecal chemotherapy has been administered, due to its radiosensitizing effect that reduces spinal cord tissue tolerance to RT. Furthermore, a higher incidence of radiation myelopathy has been shown in patients with recurrent malignancies due to the increasing role of stereotactic radiosurgery and re-irradiation.
Among the long-term sequelae of brain and spinal cord irradiation, the development of radiation-induced neoplasms is rare. The greatest risks have been associated with irradiation during childhood, leading to the development of meningiomas, malignant astrocytomas or medulloblastomas, and sarcomas typically presenting years to decades after therapy (Perry and Schmidt 2006).
Two main mechanism have been incriminated in the pathophysiology of radiation injury, with oligodendrocytes and endothelial cells as major targets, although several processes can be involved, including CNS inflammation, injury to neuronal lineages, glial cells, and their progenitors (Makale et al. 2017).
Acute radiation toxicity is thought to be secondary to radiation-induced cytokine release, vascular hyperpermeability, and vasodilation in which vascular endothelial growth factor (VEGF) signaling is upregulated, resulting in increased edema. These processes are likely to be enhanced by chemotherapy.
Early delayed (or subacute) injury is presumably due to damage to oligodendrocytes, which are thought to be the most radiation-sensitive glial cell type in the CNS. Damage to oligodendrocytes and their progenitors results in white matter injury such as impairment in myelin synthesis leading to transient, diffuse demyelination.
Late delayed injury is related to long-term damage to vascular endothelium which is considered to be a major mechanism for radiation-induced vasculopathy, whose features are fully detailed later in this chapter. Briefly, the exposure to ionizing radiation causes endothelial apoptosis, which in turn leads to increased oxygen-free radicals as well as release of pro-inflammatory mediators, HIF-1α and VEGF. This cascade leads to vessel narrowing and fibrinoid necrosis of small vessels, ultimately resulting in ischemia and cell death. The addition of chemotherapeutic agents may enhance multiple apoptotic pathways through several metabolic pathways, resulting in an increased risk of developing treatment necrosis.
Finally, the alteration of signaling microenvironment in progenitor cell niches in the brain and the hippocampus ultimately leads to a progressive neural loss which causes late cognitive impairment and memory loss.
Radiation-related histopathological changes commonly involve cerebral white matter with sparing of subcortical U-fibers. There is a predilection for periventricular involvement, particularly capping the ventricles or within the corpus callosum. This may be explained by the relatively poor blood supply of periventricular white matter from long medullary arteries, that lack collateral vessels being more vulnerable to ischemic effects produced by radiation-induced vasculopathy. Occasionally, lesions extend into the adjacent cortex or deep gray matter.
Radiation-related leukoencephalopathy refers to a more extensive white matter damage. Vascular injury of small to medium sized vessels, as well as some degree of superimposed radiation induced oligodendroglial and stem cell/progenitor cell toxicity, likely account for radiation leukoencephalopathy. The histopathology of the latter shows a spectrum of changes ranging from myelin pallor and gliosis to demyelination to coagulative necrosis, typically superimposed onto the classic radiation induced vascular changes already described for focal radiation damage.
Radiation myelopathy mainly involves myelinated fibers and blood vessels in the spinal cord, and the lateral and dorsal funiculi are most severely affected by demyelination. Vascular endothelium damage leading to hyalinosis of intramedullary vessel walls, necrosis, and local calcium deposits may also be seen.
Clinical Scenarios and Indications for Imaging
Acute radiation toxicity occurs during or shortly after RT and commonly manifests as signs of increased intracranial pressure with headache, nausea, and vomiting. A transitory worsening of neurological deficits related to the primary lesion may also be observed. These symptoms are typically mild and self-limiting with conventional dose fractionation schemes (1.8–2 Gy per fraction up to a total dose of 60 Gy), usually improving after steroid administration.
Early delayed toxicity occurs from a few weeks to a few months after RT. These changes are often asymptomatic or manifest as general neurologic deterioration, somnolence, and mild cognitive deficits. These effects often stabilize or diminish over time, even if they can progress to late effects following high-dose, large brain volume treatment, and concurrent chemotherapy. Pseudoprogression is a peculiar early delayed radiation effect, with a clinical deterioration during or shortly after RT (often within 3–6 months after the end of RT) with increasing lesions on imaging, and with subsequent improvement without a change in treatments.
Late delayed toxicity occurs from a few months to several years after RT. The clinical course of radiation necrosis is highly variable, ranging from asymptomatic presentation to worsening of neurological focal or diffuse signs and symptoms, including seizures, which may closely mimic relapse/progression of the treated lesions. Surgery may be required to reduce the mass effect and edema and allow to obtain histopathologic confirmation, although most of the cases show a mixture of predominantly radiation necrosis intermingled with limited residual and/or recurrent tumor can be found. Resolution may be obtained after corticosteroid therapy alone in some cases.
Radiation-related leukoencephalopathy often manifests with neurobehavioral symptoms and progressive cognitive impairment, which tend to progress over years and significantly alters QoL (Makale et al. 2017).
Delayed radiation myelopathy can occur from a few months to several years following radiation exposure. Clinically, it presents with a progressive onset of numbness and weakness with or without sphincter dysfunction. Radiation myelopathy is usually a progressive and permanent disease, even if some improvement in neurologic symptoms on follow-up has been reported.
MRI with gadolinium is the examination of choice in the neuroradiological follow-up after RT treatment, and thus to detect imaging changes related to brain and spinal cord radiation injury. Diffusion and perfusion MRI can be useful to determine necrosis and tumor viability, respectively.
Imaging patterns following radiation therapy usually refer to early and late delayed effects, as MRI in acute radiation injury generally does not show any change with respect to pre-treatment scan.
MRI findings can vary from transient white matter alterations to an increase in size of the contrast enhancing lesion with surrounding edema occurring early after the end of RT within the previous irradiating field, a phenomenon termed pseudoprogression, which is discussed below.
Perfusion-weighted imaging (PWI) has high reported diagnostic accuracy in recent meta-analyses, with sensitivities and specificities predominantly in the 80–90% range (Patel et al. 2017). Reduced rCBV values (both mean and maximum) on DSC perfusion have been described in PsP with respect to true tumor progression (Thust et al. 2018b), even if the methodological heterogeneity across studies makes it impossible to find clinically meaningful pooled rCBV thresholds to reliably diagnose PsP (Patel et al. 2017). DCE perfusion has also been used in the diagnosis of PsP, even if this technique is still scarcely employed in routine clinical examinations. Compared with true tumor progression, PsP demonstrated lower Vp and Ktrans values, but proposed thresholds are highly variable and depend on the pharmacokinetic models used and quantification analysis methods (Patel et al. 2017).
On diffusion-weighted imaging (DWI), PsP usually shows higher ADC values (both minimum and mean) compared with true tumor progression, with moderate diagnostic accuracy. This is likely related to the limited specificity of ADC measurements and the heterogeneity of diffusivity values within PsP lesions, ranging from low values of coexistent residual tumor tissue to high values consistent with vasogenic edema due to vascular hyperpermeability.
MR spectroscopy (MRS) in PsP may show a variable decrease in N-Acetyl-Aspartate levels with respect to baseline scan, which is likely related to transient neuroaxonal dysfunction, as well as lack of pronounced elevation of Choline levels, as observed in true tumor progression. However, differentiation of tumor progression from PsP using MRS has not gained widespread clinical use due to technical and quantification issues in the routine neuroradiological practice.
The issue of pseudoprogression led to a change in therapy response evaluation criteria (the Response evaluation in NeuroOncology [RANO] criteria) and to the recommendation not to enroll patients relapsing within 3 months from the end of radiotherapy in trials on recurrent glioblastoma, unless the recurrence is histologically proven or the progressive abnormalities lie outside the radiation field.
Late Delayed Effects
Radiation necrosis, which is detailed below, is considered the most common and severe form of late radiation injury, but other imaging patterns have been described, depending on different treatment approach and underlying pathophysiological mechanisms, such as treatment effects after SRS, transient focal enhancing lesions, diffuse white matter injury, and ultimately brain atrophy.
PWI (DSC) should be part of the neuroradiological follow-up after RT and usually demonstrates lower perfusion values in areas of RN with respect to recurrent tumor, possibly related to vascular impairment and ischemia-related changes. Reduced rCBV values (both mean and maximum) on DSC perfusion have been described in RN compared to recurrent tumor, with high diagnostic accuracy in individual studies (Patel et al. 2017). However, standardized rCBV cut-off values to reliably distinguish these two entities are lacking due to the methodological heterogeneity across scanners. This distinction is further hampered when mixed histological findings for necrosis and residual/recurrent tumors coexist within the same lesion, as is often the case. In some cases of RN, the enhancing component may also have a higher rCBV due to an increase of the vascular lumen volume secondary to aneurysmal or telangiectatic changes.
DCE perfusion has also been used in the diagnosis of RN, with lower Vp and Ktrans perfusion values in areas of RN with respect to recurrent tumor, with no widely established thresholds. Analysis of DCE T1 steady-state signal intensity curves from RN areas demonstrates a very slow increase in signal compatible with a leaky blood-brain barrier, which is different from residual/recurrent hypervascular tumor tissue, showing a vascular phase with rapid initial signal increase and washout.
DWI shows a broad range of ADC values within necrotic lesions, reflecting the coexistence of different pathological components. Increased ADC values with low signal on DWI images could be a consequence of an increase in water in the interstitial space following cell necrosis. However, low ADC values have been also reported in early stage necrosis, possibly as a consequence of coagulative necrosis with hemorrhagic components or early inflammatory infiltrate. Surrounding vasogenic edema shows elevated ADC values. Due to this variability of findings, DWI does not significantly improve the diagnostic accuracy in distinguishing RN from recurrent tumor.
Metabolic imaging using PET may provide relevant additional information in suspected radiation-induced lesions, particularly when radiolabeled amino acids are used. 11C-methyl-L-methionine (11C-MET), O-(2-[18F]fluoroethyl)-L-tyrosine (18F-FET), and 3,4-dihydroxy-6-[18F]-fluoro-L-phenylalanine (18F-DOPA) have been explored as potential tracers for differentiating between treatment necrosis and tumor recurrence, demonstrating sensitivities and specificities in the 75–95% range (Galldiks and Langen 2016). RN lesions usually show low radioactive tracer uptake, differently from progressing tumors that exhibit increased amino acid transport. However, these techniques still do not have a widespread clinical acceptance.
Treatment Effects After Radiosurgery
Transient Focal Enhancing Lesions
Transient focal enhancing lesions have been recently described in patients treated with radiotherapy (especially hyper-fractionated accelerated radiotherapy, HART) and high-dose chemotherapy for the treatment of primary malignant CNS tumors in children, occurring at a median of 8 months after RT (range, 2–39 months) either in supra- or infratentorial regions with frequent involvement of the pons, cerebellum, and posterior cerebral hemispheres.
Diffuse White Matter Injury
Brain atrophy may appear as a side effect of RT and a progressive decrease in gray matter (GM) volume over time, which is often associated with cognitive impairment. The reduction of GM volume is partly mediated by incidental irradiation of normal brain tissue. RT planning consider the optic pathway, brainstem, and cranial nerves as organs at risk, whereas the brain parenchyma is treated as essentially homogeneous in terms of RT exposure risk, with only broad dose constraints to avoid overt radiation necrosis. As such, partial irradiation of normal brain tissue leads to a progressive cortical atrophy and thinning, which can be highlighted by MRI. Changes are already visible after 1-year post-RT and are largely dose-dependent. Quantitative MRI volumetry and measurements of cortical thickness can be used to assess the vulnerability of specific brain regions to radiation dose-dependent atrophy, with temporal and limbic cortex exhibiting the largest change in cortical thickness per Gy compared to other regions. These findings critically contribute to the development of cognitive disability in patients receiving RT.
Furthermore, hippocampal volume loss after chemoradiation has been also reported, which is thought to contribute to memory impairment after radiation therapy to the brain. Hippocampal atrophy at 1 year after treatment has been significantly associated with mean hippocampus radiation dose. Memory impairment may be partially mediated by a depletion of the neurogenic stem cell population residing in the hippocampal dentate gyrus. As a consequence, there have been increased efforts to spare the hippocampus during RT planning to reduce early cognitive decline.
Spinal Cord Imaging in Radiation Injury
Acute phase MRI findings such as cord swelling and enhancement tend to resolve on follow-up, although persistent intramedullary T2 abnormalities have been reported in half of the patients (Khan et al. 2018). Enhancing lesion becomes site of focal atrophy in late follow-up scans.
Diagnosis of DRM is difficult and often a diagnosis of exclusion. Differential diagnoses include the more common causes of transverse myelitis (partial and/or longitudinally extensive), including demyelinating diseases (multiple sclerosis, NMOSD), rheumatoid diseases (lupus, Sjögren syndrome), and infectious etiologies (viral, bacterial). Primary and metastatic tumors should also be considered. Spinal cord infarction can be clinically differentiated by the acute onset of symptoms.
Long-Term Sequelae of Brain and Spinal Cord Irradiation
Long-term sequelae of brain and spinal cord irradiation include radiation-induced vasculopathy and the development of radiation-induced neoplasms, the former being separately described in the next paragraph. As for radiation-induced tumors, the most common secondary tumors are meningiomas, nerve sheath tumors, pituitary adenomas, gliomas, sarcomas, and embryonal neoplasms, with latency periods varying greatly, but typically presenting years to decades after therapy. Irradiation during childhood hold the greatest risks of developing secondary neoplasms, with greater dose exposures generally associated with shorter latency periods and higher risks of malignancy. Radiation-induced neoplasms are usually clinically, radiologically, and histologically similar to their sporadic counterparts.
Definition of Entity
Among long-term sequelae of RT, radiation-induced vasculopathy shows a wide range of clinical and pathological manifestations. Vascular endothelium is among the first and most affected tissues by ionizing radiation; arteries and capillaries are both affected, while veins are involved to a lesser extent. Radiation injury may affect large vessels as carotid arteries at the neck as well as small vessels in the brain. Nevertheless, injuries to small vessels are more common than injuries to large arteries, usually manifest earlier, and cause the development of brain lesions that may be ischemic or inflammatory and mostly present as late effects. The late effects of this damage include cerebrovascular accident (CVA), lacunes, small vessels occlusive diseases such as Moya-Moya syndrome, cerebral microbleeds, vascular malformations, and rarely mineralizing microangiopathy.
The incidence of these late effects is widely variable among series and differs between adults and children exposed to RT (Murphy et al. 2015). The incidence of CVAs in the adult population exposed to radiation therapy is not definitely known, as most of the data come from old series of patients treated with old RT schemes. Nonetheless, it seems that RT itself is not an independent risk factor for CVAs. In pediatric cancer survivors, a cumulative incidence of late-occurring strokes of 0.73% in leukemia patients treated with WBRT and of 5.6% in brain tumor patients after RT was observed at 25 years. Furthermore, in brain tumor children, the risk of late CVAs is dose-dependent and increased by associated chemotherapy and endocrinopathy, younger age at treatment (<5 years), and specific tumor locations, being optic pathway gliomas and craniopharyngioma the highest risk tumors for developing vasculopathy due to their close proximity to arterial vessels.
Postradiation small-vessel injuries in children also often present as Moya-Moya syndrome, microangiopathy, and small vascular malformations as cavernous angioma and telangiectasia. From over 1800 patients all receiving prophylactic cranial RT, the incidence of Moya-Moya was 0.46% at 8 years. This incidence usually increases over time from RT can develop faster in younger patients. Furthermore, NF1 has been reported to be a risk factors, being patients more sensitive to the effect of RT.
The reported incidence of cavernomas after RT has been quite variable in different series, ranging from 5 to over 40% (Murphy et al. 2015). Cavernomas develop at a range of 3–9 years after RT, with shorter time in children less than 10 years of age.
Cerebral microbleeds (CMBs), a specific form of small-vessel vasculopathy, are a common injury after RT, with a cumulative incidence of 48% at 5 years in pediatric brain tumor survivors. Craniospinal or WBRT was the main risk factor for the development of CMBs. The incidence is even higher in adults treated for gliomas with RT, and approximately all the patients develop CMBs at approximately 2 years after the completion of RT, with an increasing number of lesions over time and a higher density in in the area of relatively high-dose radiation (>30 Gy) (Wahl et al. 2017).
Pathophysiology and Pathological Features
The process begins with progressive loss of endothelial cells which can last for months after the end of RT, involves apoptotic processes, and is dose dependent. The consequent disruption of the BBB leads to vasogenic edema and neural tissue hypoxia that can manifest acutely or lately, depending on the damage entity that is also related to radiation dose. Endothelial loss initially causes hemorrhage and thrombi that may bring to vessel wall necrosis and significantly increase risk of rupture when large radiation doses are administered. Weakness of the vessel wall is also the basis for the formation of fistulae between damaged vessels and normal adjacent ones (Murphy et al. 2015). Another relevant aspect is the inflammatory response to radiation damage to the vessel wall, which promotes basal membrane thickening, vessel dilatation, and edema. Furthermore, several pro-inflammatory genes are activated by radiation exposure and contribute to elevation of VEGF that, apart from promoting endothelial proliferation, also contributes to the increase of BBB permeability, BBB disruption, tissue edema, and consequent necrosis.
All the aforementioned pathophysiological processes contribute over time to the macroscopic development of abnormal stenotic and dilated vessels, with leaky, necrotic, and later fibrotic walls with hyalinization and fibrinoid necrosis, sclerosis, and thrombosis, affect both large and small vessels.
Small vessel sequelae of endothelial degeneration can occur months to years after initial damage and includes microvascular glomeruloid proliferation with telangiectasia, microvascular dilatation, as well as thickening and hyalinization of the vessel wall. Mineralizing microangiopathy is a very rare long-term complication, mainly seen in in pediatric patients treated for leukemia with WBRT and chemotherapy. Large vessels such as carotid arteries may develop wall thickening and/or atherosclerosis with subsequent irregular stenosis and thromboembolism. Radiation-induced atheroma tends to be macrophage-rich, lipid filled and shows intra-plaque hemorrhage more frequently than classic atherosclerosis.
Clinical Scenario and Imaging Features
Topography of radiation-induced vasculopathy depends on the location of the treated lesion. The relative location of tumors to vascular structures impacts the likelihood of long term effects, as in the aforementioned cases of RT for tumors of the middle cranial fossa which are more prone to develop vasculopathy of the circle of Willis and major cerebral arteries. As described above, postradiation vascular changes are essentially of two types: stenotic as a consequence of fibrosis and hyalinization/mineralization and malformative as a consequence of endothelial loss and thrombosis. Stenosis involves more frequently large vessels while malformations are more related to damage to medium and small intracranial vessels.
The most frequent sites of stenotic vasculopathy in adults population are carotid arteries at the neck in patients with head and neck cancer and intracranial carotid siphon in patients treated for pituitary adenomas. On CT and MR angiography, the involved vessel appears irregularly stenotic; at the neck level the length of the stenotic tract may be longer than the more common atheromatous stenosis and is usually included within the radiation field.
Radiation induced cerebral aneurysms may be fusiform or saccular, multiple and irregular in shape, and of variable size; moreover, since their site is related to radiation field, they are usually located at the level of distal branches of the Willis circle (see case report 2, Figs. 15 and 16).
Microbleeds, telangiectasias, and cavernomas are evident as markedly hypointense areas on T2∗ Gradient-Echo and SWI sequences due to the “blooming” effect of the underlying hemosiderin deposits. On SWI scans, microbleeds are usually identified as small hypointense foci that did not correspond to linear vessels on consecutive slices.
Mineralizing microangiopathy is well recognizable at nonenhanced brain CT and is characterized by calcifications in the basal ganglia and in the subcortical white matter.
As an effect of radiation-induced vasculopathy on small vessels, lacunar infarcts may develop in deep brain locations, having the same imaging features of common lacunar strokes.
Definition of Entity and Clinical Highlights
Stroke-like Migraine After Radiation Therapy (SMART) syndrome is a rare condition that occurs as a late delayed consequence of brain irradiation. It is clinically characterized by transient “stroke-like” episodes and acute/subacute migraine attacks with or without seizures in patients previously exposed to brain RT, both focal and WBRT or cranio-spinal irradiation. This syndrome usually manifests 10–20 years after brain irradiation especially for posterior fossa or posterior cerebral hemisphere tumors, although shorter delays have also been described.
Epidemiology and Pathophysiology
The pathophysiology of SMART is largely unknown, and scarce histopathological reports failed to demonstrate specific patterns. Impaired vascular reactivity, endothelial damage, and injury to the trigemino-vascular system have been advocated as causative, along with a cerebral hyperexcitability with altered ion channel reactivity, that may lower the threshold for cortical spreading depression and results in migraine-like headaches and increased risk for seizures (Black et al. 2006).
Perfusion images in SMART show abnormally elevated CBF and CBV, consistent with an abnormal vascular reactivity as demonstrated by transcranial Doppler ultrasound. MRS may show a decrease in N-acetyl-aspartate and increased levels of Creatine and Choline, consistent with transient neuroaxonal dysfunction or neuronal cell loss with mild nonspecific gliosis. In some cases EEG may show the presence of slow waves ipsilateral to MRI alterations, mostly without epileptiform abnormalities (Black et al. 2006).
In most of cases, both symptoms and MRI findings typically resolve in 2–5 weeks and often benefit from treatment with high-dose steroids on empirical grounds. Nonetheless, cases with long lasting or permanent neurological sequelae and irreversible brain damage with cortical laminar necrosis have been described. Furthermore, relapses are common and have been reported in more than half of the patients.
Peri-ictal pseudoprogression (PIPG) belongs to the same spectrum of phenomena as SMART, demonstrating similar MRI findings but absence of headache, less significant neurologic impairment, and more rapid clinical recovery as differentiating characteristics.
Recommended Imaging Protocol for Suspected Radiation Injury
Serial contrast-enhanced MRI represents the current backbone for monitoring treatment response following RT. Knowledge of irradiation fields and treatment dosimetry is crucial to correctly interpret imaging changes after treatment. An immediate pre-RT scan should be obtained as the baseline for comparison with the post-RT scan. Imaging protocol should refer to the EORTC-NBTS consensus recommendations (Ellingson et al. 2015) that have been recently endorsed by the ESNR (Thust et al. 2018a). The standardized protocol includes anatomical T1-weighted images acquired pre- and postcontrast administration, T2-weighted and FLAIR sequences and also includes recommendations for DWI, with a maximum b-value of 1000 s/mm2. PWI should be included as well, with DSC-derived rCBV being the most validated parameter to distinguish therapy effects from tumor progression. DCE could be performed optionally, preferably as an adjunct (Thust et al. 2018a). Quantitative assessment of PWI is strongly advocated, being aware that threshold values are not simply transferable between institutions, as they very much depend on scan parameters and postprocessing methods (Thust et al. 2018a). MRS is recommended as an optional modality, to be considered on an individual case basis in combination with other advanced techniques such as PWI.
There are no specific vascular imaging protocol or guidelines to follow-up radiation-treated patients at risk for developing a vascular lesion, apart for the inclusion of T2∗ Gradient Echo and/or SWI sequences to increase the detection of microbleeds, telangiectasia, and cavernoma. We refer to the specific chapter for the most suitable MR imaging (including MRA) or CT angiography protocol to search for intra-extracranial vascular stenosis or intracranial malformations, the most suitable one being the less invasive, but with sufficiently high diagnostic accuracy.
Interpretation Checklist and Structured Reporting
Details of the underlying pathology (e.g., tumor histology, previous surgery) and clinical highlights.
Date of RT initiation and details for the applied treatment regimen such as irradiation field, number of fractions, total radiation dose, concomitant/adjuvant chemotherapy, previous RT.
Presence of lesion(s) suggestive for RT-related effect within the irradiation field.
Details of T2/FLAIR-weighted, T2∗/SWI- and post-contrast T1-weighted signal pattern of the lesion, including location, ‘blooming’ artifacts and hemorrhagic lesions, (in)homogeneous gadolinium enhancement, areas of necrosis.
Description of edematous changes and mass effect, if present.
DWI/ADC changes within the lesion described as increased, attenuated, restricted diffusion, and comparison with respect to baseline pre-RT scan. For diffusion evaluation, quantitative ADC comparison to normal brain is recommended, to overcome the potential pitfall of lesions surrounded by edema appearing dark on the ADC map, even in the absence of diffusion restriction.
PWI changes described both qualitatively (e.g., homogeneous/heterogeneous pattern within the lesion, location of perfusion “hotspots” with respect to conventional imaging findings such as enhancing areas) and quantitatively, by measuring and reporting normalized rCBV values (and, when available, DCE-derived Ktrans and Vp values) and their variations with respect to previous exams, if any.
When MRS has been performed, any decrease in N-Acetyl-Aspartate ratios with respect to previous scans should be reported along with changes in Choline and Creatine levels over time; the presence of pathological metabolites such as lipids and lactate should be annotated.
Clinical Case and Sample Reports
Possible treatment effects in patient with controlled systemic disease.
Purpose of MR Study
Rule out possible RT effects vs. tumor progression.
Brain MRI scan including T2/FLAIR-weighted, DWI, SWI, unenhanced, and contrast-enhanced T1-weighted scans and DSC and DCE perfusion sequences according to a standardized protocol (Anzalone et al. 2018).
Baseline pre-SRS MRI (Fig. 13) acquired 15 months earlier showed a right parietal nodule with inhomogeneous hyperintensity and scarce edema on T2-weighted images (A), ADC similar to normal brain tissue (B), and faint enhancement on postcontrast T1-weighted images (C).
Current MRI (Fig. 14) demonstrates a heterogeneous lesion on T2-weighted images (A) with ring-like enhancement on postcontrast T1-weighted images (B), associated with surrounding edema which shows elevated ADC (C) and mass effect. SWI images (D) show multiple, punctate hypointense foci, consistent with cerebral microbleeds.
Qualitative assessment of DSC (Fig. 14e) and DCE (G) PWI parametric maps shows relatively low rCBV along with moderately increased Ktrans in the site of the lesion, with a small central area of increased rCBV and Ktrans (F and H, DSC, and DCE maps superimposed on T1-weighted image). Quantitative analysis demonstrates heterogeneous values of rCBV and Ktrans within the enhancing lesion, with a central area of high rCBV and moderately increased Ktrans (ROI 1: rCBV = 4.4 and Ktrans = 0.077 min−1, ROI 2: rCBV = 2.4 and Ktrans = 0.033 min−1), consistent with vascular and permeable metastatic tumor vasculature, and relatively lower values of perfusion and moderately increased permeability at the enhancing margins (ROI 3: rCBV = 1.5 and Ktrans = 0.038 min−1), possibly related to radiation injury. Despite thresholds are largely variable in literature, the presence of such mixed PWI values suggests the coexistence of radiation necrosis and residual tumor.
Conventional MRI and PWI, together with the clinical history of patient, suggest a prevalence of treatment-related effects, possibly mixed with residual tumor. The patient underwent surgical resection of the mass, and histopathological analysis confirmed radiation-related injury with necrosis and a small central area of residual metastatic tumor tissue from lung carcinoma.
Clinical examination revealed a GCS score of 14.
Purpose of Imaging Studies
Rule out the cause of cerebrovascular accident in a patient previously treated with WBRT.
Plain CT study, digital subtraction angiography (DSA), comparison with previous contrast-enhanced MRI.
Plain CT scan (Fig. 15a) demonstrated the presence of a right frontotemporal hematoma associated to subarachnoid hemorrhage and presence of intraventricular hemorrhage; it revealed also the presence of two calcified sphenoetmoidal meningiomas. DSA (Fig. 15: b, oblique view; c, lateral view) performed soon after demonstrated the presence of a diffuse irregular aspect of right middle and anterior cerebral arteries, characterized by multiple fusiform aneurysms. The patient was then treated by endovascular embolization of a small M2 aneurysm that was considered to be responsible for the acute hemorrhage.
A previous contrast-enhanced MRI (Fig. 16: a, T2-weighted axial images; b, T1-weighted coronal images) performed in another center to follow up the treated meningiomas showed the presence of enhancement along the right middle cerebral artery (white arrowheads) within the radiation field.
This is an unusual case of postradiation arteritis with development of irregular aneurysms within the field of radiation performed to treat two right sphenoetmoidal meningiomas. The presence of vessel wall enhancement on MRI can be considered a sign of inflammation/neovascularization, possibly increasing the risk of bleeding.
- Kumar AJ, Leeds NE, Fuller GN, Van Tassel P, Maor MH, Sawaya RE, Levin VA. Malignant gliomas: MR imaging spectrum of radiation therapy- and chemotherapy-induced necrosis of the brain after treatment. Radiology. 2000;217:377–84. https://doi.org/10.1148/radiology.217.2.r00nv36377.CrossRefPubMedGoogle Scholar
- Patel P, Baradaran H, Delgado D, Askin G, Christos P, John Tsiouris A, Gupta A. MR perfusion-weighted imaging in the evaluation of high-grade gliomas after treatment: a systematic review and meta-analysis. Neuro Oncol. 2017;19:118–27. https://doi.org/10.1093/neuonc/now148.CrossRefPubMedGoogle Scholar
- Ruben JD, Dally M, Bailey M, Smith R, McLean CA, Fedele P. Cerebral radiation necrosis: incidence, outcomes, and risk factors with emphasis on radiation parameters and chemotherapy. Int J Radiat Oncol Biol Phys. 2006;65:499–508. https://doi.org/10.1016/j.ijrobp.2005.12.002.CrossRefPubMedGoogle Scholar
- van Dijken BRJ, van Laar PJ, Holtman GA, van der Hoorn A. Diagnostic accuracy of magnetic resonance imaging techniques for treatment response evaluation in patients with high-grade glioma, a systematic review and meta-analysis. Eur Radiol. 2017;27:4129–44. https://doi.org/10.1007/s00330-017-4789-9.CrossRefPubMedPubMedCentralGoogle Scholar