Brain tumour post-treatment imaging and treatment-related complications
- 1.3k Downloads
The imaging of primary and metastatic brain tumours is very complex and relies heavily on advanced magnetic resonance imaging (MRI). Utilisation of these advanced imaging techniques is essential in helping clinicians determine tumour response after initiation of treatment. Many options are currently available to treat brain tumours, and each can significantly alter the brain tumour appearance on post-treatment imaging. In addition, there are several common and uncommon treatment-related complications that are important to identify on standard post-treatment imaging.
This article provides a review of the various post-treatment-related imaging appearances of brain neoplasms, including a discussion of advanced MR imaging techniques available and treatment response criteria most commonly used in clinical practice. This article also provides a review of the multitude of treatment-related complications that can be identified on routine post-treatment imaging, with an emphasis on radiation-induced, chemotherapy-induced, and post-surgical entities.
Although radiological evaluation of brain tumours after treatment can be quite challenging, knowledge of the various imaging techniques available can help the radiologist distinguish treatment response from tumour progression and has the potential to save patients from inappropriate alterations in treatment. In addition, knowledge of common post-treatment-related complications that can be identified on imaging can help the radiologist play a key role in preventing significant patient morbidity/mortality.
• Contrast enhancement does not reliably define tumour extent in many low-grade or infiltrative gliomas.
• Focal regions of elevated cerebral blood volume (rCBV) on dynamic susceptibility contrast (DSC) perfusion-weighted imaging are suggestive of tumour growth/recurrence.
• Brain tumour treatment response criteria rely on both imaging and clinical parameters.
• Chemotherapeutic agents can potentiate many forms of radiation-induced injury.
• Ipilimumab-induced hypophysitis results in transient diffuse enlargement of the pituitary gland.
KeywordsBrain neoplasms Glioma Neoplasm metastasis Radiotherapy Review
Primary and metastatic brain tumours are frequently encountered in the daily practice of neuroimaging. Many options are currently available to treat these neoplasms and revolve around a combination of surgery, radiation, and/or novel chemotherapeutic agents. These treatment choices are ever growing and each results in alterations in pathophysiology that can drastically change the imaging appearance of the tumour. The interpretation of post-treatment imaging has therefore become much more complex, most notably with high-grade gliomas where the combination of radiation and anti-neoangiogenesis drugs can result in either increasing or decreasing local enhancement, irrespective of tumour progression/regression. For this reason, it is imperative that radiologists have a thorough understanding of the advanced imaging techniques available to image these tumours as well as the possible common treatment-related complications. The purpose of this review is to briefly review brain tumour imaging techniques, discuss the most commonly used brain tumour treatment response criteria implemented in clinical practice, and illustrate the broad range of post-treatment-related complications that can be identified on routine post-treatment imaging.
Brain tumour imaging techniques
DWI provides a visual and quantitative representation of the diffusivity of water via creation of a DWI map (typical b value of 1000) and an apparent diffusion coefficient (ADC) map. Many pathophysiological processes result in diffusion restriction; however, in the context of brain tumour imaging the most common aetiologies include abnormally high cellularity, cellular injury, and peritumoural oedema. Highly cellular tumours, such as lymphoma and many high-grade gliomas, result in decreased water diffusivity through a relative reduction in the extracellular space for protons to move about . This results in low ADC signal and some studies have shown that glioma grade often inversely correlates with the minimum ADC value identified within the tumour . Knowledge of a tumour’s baseline cellularity is very helpful as any new area of low ADC signal on follow-up imaging should raise the suspicion for tumour recurrence/progression.
Peritumoural oedema refers to the area of abnormal signal surrounding the enhancing component of a brain tumour. The problem with this term is that it encompasses two separate pathophysiological processes (vasogenic and infiltrative oedema) that have very different implications for treatment. In vasogenic oedema, there is increased fluid in the extracellular space due to alterations in vascular permeability resulting in leakage of plasma fluid and protein. This commonly occurs with metastatic lesions and non-infiltrating primary brain tumours and is typically reversible. It classically does not result in diffusion restriction and the area of abnormal signal is not considered to be within the margins of the tumour. The major clinical implication of vasogenic oedema is in regard to any mass effect it creates on adjacent normal structures. In infiltrative oedema, there is both perivascular infiltration of tumour and leakage of fluid into the extracellular space. This commonly occurs with higher grade infiltrating gliomas and makes it difficult to accurately define tumour margins. Theoretically, perivascular infiltration of tumour should decrease water diffusivity producing low ADC signal. However, many studies have shown that the areas of tumour infiltration are quite small and DWI is not very reliable for differentiating vasogenic from infiltrative oedema . In these cases, DWI must be combined with other advanced imaging techniques so that accurate tumour margins can be identified and treatment can be planned accordingly.
PWI is a noninvasive imaging technique for measuring brain tumour vascularity. It indirectly provides information on tumour angiogenesis and altered capillary permeability, both of which are present in many types of brain tumours. This is of particular importance on post-treatment imaging, as areas of increased perfusion can be suggestive of tumour growth or recurrence. The most commonly used perfusion techniques include dynamic susceptibility contrast (DSC), dynamic contrast-enhanced (DCE), and arterial spin labelling (ASL) MR imaging.
DCE perfusion imaging involves performing repeated T1-weighted imaging after contrast administration to produce a contrast signal time intensity curve. These data can then be analysed to calculate multiple parameters reflective of tumour vascularity. The most commonly calculated parameter is that of the transfer coefficient ktrans, which represents the degree of permeability between the plasma and extravascular extracellular space . It is important to note that when the blood-brain barrier is intact, ktrans more truly reflects vascular permeability; however, when the blood-brain barrier is disrupted by tumour, ktrans becomes more representative of cerebral blood flow . Although still under investigation, many authors suggest that alterations in the ktrans of brain tumours likely occur because of the formation of immature hyperpermeable vessels, a common occurrence in growing/progression of brain tumours. Therefore, as with rCBV in DSC perfusion imaging, elevations in ktrans may also be a potential marker for tumour growth/recurrence .
ASL perfusion imaging is a non-contrast perfusion technique that uses inversion pulses to magnetically label water protons in inflowing arterial blood. By comparing the difference in signal between labelled images and non-labelled images, cerebral blood flow (CBF) can then be quantified. This technique has been widely established as a useful technique in assessing cerebral perfusion in stroke and dementia; however, more recent literature has demonstrated growing utility in brain tumour imaging. Although ASL imaging does not directly measure tumour blood volume, many studies have shown that blood volume and blood flow are strongly correlated, and therefore focal areas of elevated CBF on ASL imaging may be a marker for tumour growth/recurrence. Compared with DSC perfusion techniques, ASL has the added advantage of a higher signal-to-noise ratio (SNR) and no need for intravenous contrast. However, major limitations include longer scan time and lower spatial resolution compared with DSC perfusion imaging .
Future advanced imaging techniques
It is important to note that alternative brain tumour imaging techniques, such as chemical exchange saturation transfer (CEST) and sodium MR imaging, have been increasingly described in the literature. These techniques are primarily in the research stage but seem to show promise in better characterising tumours post treatment. For these reasons a discussion on brain tumour imaging cannot be complete without at least mentioning these techniques. CEST imaging is a novel form of MR imaging in which a radiofrequency pulse is applied resulting in saturation of a particular chemical species. Over time, this magnetic saturation is transferred to water molecules via an exchange of protons, resulting in a decrease in water signal. This decrease in water signal can then be detected and, from this, an indirect measurement of the originally saturated species can be obtained. Although many chemical species are currently being investigated, Park et al. have shown that CEST amide proton transfer (APT) imaging can reliably distinguish tumour progression from a treatment-related effect as well as improve the diagnostic accuracy when combined with conventional perfusion-weighted and contrast-enhanced T1-weighted imaging techniques . Sodium MR imaging is another novel MRI technique in which the magnetic moments of 23Na nuclei, instead of traditional 1H nuclei, are used to create image contrast. Sodium nuclei are extremely abundant throughout the human body and have been found to yield the second strongest nuclear magnetic resonance (NMR) signal after 1H nuclei . In post-treatment brain tumour imaging, this is particularly useful as elevations in the intracellular tissue sodium concentration (TSC) are thought to be directly related to an alteration in Na+/K+ pump exchange and cell death. Therefore, some authors suggest that sodium MR imaging performed during radiation treatment can be used to actively monitor the spatial distribution of tissue response and therefore allow for individualised changes in a given patient’s treatment algorithm .
Although these imaging techniques appear to demonstrate clinical utility, more research is needed to establish whether these techniques can be routinely used for post-treatment imaging.
Treatment response criteria
It is imperative that radiologists be familiar with the treatment response criteria that referring clinicians are using to assess patients with brain tumours. Many of these algorithms rely heavily on imaging and therefore radiologists play an important role in dictating clinical management. Although each brain tumour behaves differently and carries different histological grades/aggressiveness, it is helpful to focus on the two main categories of brain tumours: high-grade gliomas and metastases.
High-grade glioma treatment response criteria
Historically, multiple different criteria were used to the assess treatment response of high-grade gliomas. The Levin criteria relied on qualitative changes in perilesional mass effect to ascribe treatment response . The World Health Organization (WHO) Oncology Response Criteria relied on quantitative changes in tumour size to ascribe treatment response . However, many soon realised that both of these criteria were only characterising the visual extent of the tumour and not accounting for biological features that were being seen clinically. In 1990, the Macdonald criteria were introduced to address this by adding two clinical parameters that were found to correlate with treatment response. These included: (1) decreased need for steroids and (2) stable or improved clinical status . Although better than prior criteria, the imaging component of the Macdonald criteria had a major weakness in that it relied solely on measuring changes in the enhancing portions of the tumour. As previously discussed, using post-contrast sequences alone is not an accurate way to measure the extent of many high-grade gliomas because of their propensity for infiltrative disease, which may not have enhancement. Therefore, in 2010 the Response Assessment in Neuro-Oncology (RANO) criteria were released emphasising the importance of T2 FLAIR sequences in assessing non-enhancing components of infiltrative lesions. In fact, any significant increase in T2 FLAIR non-enhancing lesions while on a stable or increasing dose of corticosteroids now met criteria for disease progression . Although other criteria have been published over the last several years (i.e., AVAglio and RTOG 0825), the RANO criteria are currently the most widely used in clinical practice .
In addition to the RANO criteria, two radiological phenomena have been well described in the literature and require a further in-depth discussion. These include pseudoprogression and pseudoresponse.
Metastasis treatment response criteria
Radiation-induced brain injury encompasses a wide variety of both clinical and radiological findings that result from fractionated or whole-brain radiation. These injury patterns are often described as an expression of time from the initiation of radiation under the following categories: acute (days to weeks), early delayed (weeks to months), and late delayed (months to years).
Acute and early delayed radiation injury is thought to result from alterations in vascular permeability and disruption of the blood-brain barrier resulting in varying degrees of brain oedema/swelling. In contrast to late delayed injury, this alteration in physiology is usually reversible and resolves spontaneously without any histopathological correlate . In the acute setting, patients typically present with vague signs related to increased brain oedema, most notably headache and drowsiness. However, it is important to note that on imaging, acute radiation injury may be difficult to identify as the radiation-induced brain oedema is often indistinguishable from the baseline vasogenic oedema incited by the brain tumour [22, 29]. In the early delayed setting, brain swelling may be accompanied by transient demyelination that often results in patients presenting with somnolence or attention/memory deficits. On imaging, this may appear as new areas of oedema or enhancement, typically within the vicinity of the irradiated tumour .
Late delayed radiation injury, on the other hand, is a much more serious form of injury due to a combination of vascular and glial injury resulting in irreversible and progressive white matter necrosis. The exact mechanism through which this occurs is not well understood but thought to be due to either vascular endothelial injury or direct parenchymal injury . The vascular hypothesis of late delayed radiation injury supposes that radiation induces structural changes in the vasculature (vessel wall thickening, wall dilatation, and endothelial cell nuclear enlargement), which in turn leads to ischaemia and white matter necrosis. In addition, animal models have shown that there is a quantitative decrease in vessel density after radiation , which in turn increases the risk for white matter ischaemia. The parenchymal hypothesis of late delayed radiation injury supposes that radiation induces direct injury to various parenchymal cell lines, in particular oligodendrocytes, astrocytes, microglia, and neurons. Shinohara et al. demonstrated that radiation induces loss of the oligodendrocyte type-2 astrocyte (O-2A) progenitor cells, preventing the formation of mature oligodendrocytes and ultimately resulting in demyelination and white matter necrosis . Hwang et al. demonstrated that radiation induces microglial activation, which in turns induces astrocyte expression of prostaglandin E2 (PGE2), stimulating gliosis, and brain oedema . Rosi et al. demonstrated that hippocampal neurons in mice exposed to radiation significantly decreased expression of activity-regulated cytoskeleton-associated (Arc) protein resulting in neuronal dysfunction and suspected cognitive impairment . Although not well understood, it is reasonable to surmise that late delayed radiation injury actually results from a complex interaction between both vascular and parenchymal dysfunction, although more study is needed to further define this dynamic process.
Although it is helpful to understand the temporal relationship of radiation and the pathophysiology of brain injury, many find it more helpful to classify radiation-induced brain injury by the pattern visualised on imaging. What follows is an imaging-based discussion of radiation-induced complications organised by which anatomic structures are affected.
Radiation-induced vascular injury
As previously stated, radiation can have profound effects on the intracranial vasculature through irreversible endothelial injury. This is typically a late delayed injury manifesting months to years after radiation exposure. On imaging, radiation-induced vascular injury produces three main appearances: radiation-induced vasculopathy, radiation-induced vascular proliferative lesions, and radiation-induced mineralising microangiopathy.
Radiation-induced mineralising microangiopathy refers to the development of dystrophic microcalcifications in the brain parenchyma. Histopathologically, this process consists of calcium deposition within damaged vessel walls as well as surrounding necrotic brain tissue . On imaging, calcification is typically seen in the basal ganglia and subcortical white matter, likely reflecting the inherit vulnerability that small perforating and peripheral vessels have to radiation injury (Fig. 11).
Radiation-induced parenchymal injury
Stroke-like migraine attacks after radiation therapy (SMART) syndrome
Chemotherapeutic agents can result in direct toxicity to various structures of the central nervous system. The particular CNS structure involved and the degree to which they are affected vary depending on the drug administered and the dose that was given. Given the ever-growing list of new chemotherapeutic agents, each with its own toxic profile, a thorough discussion of CNS drug toxicity is beyond the scope of this article. However, that being said, it has been shown that most drugs tend to affect similar structures and produce similar patterns of injury on imaging. In addition, certain drugs can result in classic pathological imaging findings (i.e., ipilimumab-induced hypophysitis) that are easy to diagnose once familiar with their imaging appearance. What follows is a discussion of these common chemotherapy-related injury patterns.
It has been widely described in the literature that of all the CNS structures, the white matter is particularly vulnerable to drug injury. Although the mechanism through which this occurs is poorly understood, it has been suggested that leukotoxic agents result in disruption of neural transmission, particularly in the neurobehavioural pathways, often resulting in the classic clinical presentation of “altered mental status”. This injury results in a drug-related toxic leukoencephalopathy, producing a clinical and radiographic picture similar to radiation-induced leukoencephalopathy. In fact, many chemotherapeutic agents can actually potentiate many of the radiation injuries previously discussed. The most common leukotoxic agent used in clinical practice is methotrexate; however, many other agents including carmustine, cisplatin, cytaribine, fluorouracil, and interleukin-2 have also been implicated [50, 51]. The incidence with which neurotoxicity occurs is quite variable; however, the route of administration has been shown to be important. Filley et al. stated that toxic leukoencephalopathy may occur in less than 10% treated with intravenous methotrexate, but in up to 40% treated with intrathecal methotrexate . On imaging, chemotherapy-related leukoencephalopathy typically produces non-enhancing T2 hyperintensity in the frontoparietal white matter, but can sometimes demonstrate diffuse involvement of the periventricular and deep white matter, similar to radiation-induced leukoencephalopathy . On DWI, focal or diffuse areas of reversible restricted diffusion may be seen in acute cases and typically show improvement over time after cessation of the offending drug .
Primary and metastatic brain tumours are frequently encountered in the daily practice of neuroimaging. The wide array of treatment options currently available to treat these tumours has made interpretation of post-treatment imaging quite complex. Knowledge of post-treatment imaging techniques, treatment response criteria, and commonly encountered treatment-related complications will simplify the approach to this challenging topic.
We would like to thank Margaret Kowaluk from the graphics department at our institution for her help in preparing the figures for publication.
- 5.Chan AA, Nelson SJ (2004) Simplified gamma-variate fitting of perfusion curves. 2nd IEEE International Symposium on Biomedical Imaging: Nano to Macro 2(2):1067-1070Google Scholar
- 6.Kassner A, Annesley DJ, Zhu XP et al (2000) Abnormalities of the contrast re-circulation phase in cerebral tumors demonstrated using dynamic susceptibility contrast-enhanced imaging: a possible marker of vascular tortuosity. J Magn Reson Imaging 11:103–113Google Scholar
- 21.Chinot O, Macdonald D, Abrey L, Zahlmann G, Kerlogëguen Y, Cloughesy T (2013) Response assessment criteria for glioblastoma: practical adaptation and implementation in clinical trials of antiangiogenic therapy. Curr Neurol Neurosci Rep 13(5):347Google Scholar
- 22.Hygino da Cruz LC Jr, Rodrigues I, Domingues RC, Gasparetto EL, Sorensen AG (2011) Pseudoprogression and pseudoresponse: imaging challenges in the assessment of posttreatment glioma. AJNR Am J Neuroradiol 32(11):1978–1985Google Scholar
- 25.Balaña C, Capellades J, Pineda E et al (2017) Pseudoprogression as an adverse event of glioblastoma therapy. Cancer Med 6(12):2858–2866Google Scholar
- 28.Lin NU, Lee EQ, Aoyama H et al (2015) Response assessment criteria for brain metastases: proposal from the RANO group. Lancet Oncol 16:270–278Google Scholar
- 42.Shah R, Vattoth S, Jacob R et al (2012) Radiation necrosis in the brain: imaging features and differentiation from tumor recurrence. Radiographics 32(5):1343–1359Google Scholar
- 54.Rodrigues BT, Otty Z, Sangla K, Shenoy VV (2014) Ipilimumab-induced autoimmune hypophysitis: a differential for sellar mass lesions. Endocrinol Diabetes Metab Case Rep 12. https://doi.org/10.1530/EDM-14-0098
- 56.Gempt J, Förschler A, Buchmann N et al (2013) Postoperative ischemic changes following resection of newly diagnosed and recurrent gliomas and their clinical relevance. J Neurosurg 118:801–808Google Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.