Supratentorial intracranial mass in an adult (an approach)
The identification of a supratentorial intracranial mass in an adult is a fairly common clinical scenario, the appropriate management of which relies heavily on preoperative imaging. Often important clues will be present in the clinical history (e.g. immunosuppression, systemic malignancy, duration of symptoms) and it is difficult to over-estimate the importance in obtaining such information prior to interpreting images.
It is, however, not unusual for no specific information to be available and in such cases a systematic approach to CT and MRI interpretation will enable you to arrive at a firm preoperative diagnosis in most cases. This article focus on such cases, but deliberately only considers common entities, and important entities not to miss.
The primary aims of preoperative interpretation are to:
- localise the mass
- establish likely diagnosis
- identify time-critical findings
Additionally if you work in a center where surgery and treatment is going to take place, and thus multidisciplinary review of patients occurs, you will also want to establish framework within which histology results are to be interpreted.
Note: This article focuses primarily on supratentorial parenchymal and extra-axial masses in adults. For approaches to other scenarios please refer to:
- intraventricular lesions
- posterior fossa mass (coming soon)
- pineal region mass
- pituitary region mass
Before one can embark upon this task, appropriate imaging needs to be available. A good quality MRI with a combination of structural and advanced imaging is crucial (e.g. MRI brain tumour protocol).
A CT is not mandatory, however in practice it is usually available as it is often the first study obtained, and it can give important clues particularly in respect to bony changes or calcification.
Correct localisation of an intracranial mass is essential not only because it is essential in correctly narrowing your differential but also in aiding the treating surgeon in planning their approach.
Extra-axial vs intra-axial
The first step in correct localisation is determining if the mass is intra-axial or extra-axial. In many instances this is self evident, however, not infrequently the differentiation can be challenging. This is especially the case in the setting of a large peripheral heterogeneous mass (e.g. case 1 vs case 2).
Scrutiny of the images can usually confidently localise the mass into one of the two compartments. Occasionally this is not possible, and in such cases it is important to explicitly state that location is uncertain, as this often has important surgical implications.
The two most useful signs are identifying a CSF cleft (with interposition of pial vessels between the mass and the cortex) and white matter buckling sign. Presence of a dural tail is also quite helpful, although one must be mindful of placing too much emphasis on this well known sign as a number of intraparenchymal masses can also result in reactive dural thickening (e.g. PXA).
If intact cortex is visible between the mass and the adjacent white matter, then this is a very good sign that the mass is extra-axial. The converse is, however, not true as in many instances an extra-axial mass can result in marked thinning and/or oedema within the cortex resulting in it being invisible. This is not infrequently the case with meningiomas with adjacent oedema.
The most helpful signs of the mass being intra-axial are presence of thickened non-enhancing cortex with increased T2 signal - a common finding in diffuse glial tumours (e.g. diffuse astrocytomas), expansion of nearby gyri (this is the opposite of the white matter buckling sign) and thickening of the corpus callosum.
Next, after having established whether the mass is intra or extra-axial, the location of the lesion needs to be established. Essential, and too often incorrect is something as basic as "Which lobe is it in?" In many instances you should be able to narrow it down to a specific gyrus, particularly when the lesion is small and superficial.
Even more importantly, what is its relationship to nearby eloquent areas or major white matter tracts. Although a comprehensive overview of neuroanatomy is clearly beyond the scope of this article, as a minimum the following regions should be considered:
- central sulcus separating primary motor from primary sensory cortex
- internal capsule and corticospinal tract
- Wernicke's area and Broca's area for language
- calcarine fissure running through the primary visual cortex
- amygdala, hippocampus and fornix
- corpus callosum
Additionally a number of features are important for surgical planning:
- does the mass involve the cortex or is it entirely subcortical? (i.e. will an abnormality be visible on the surface of the brain)
- does the mass involve the subependymal brain? (i.e. is resection likely to result in communication with the ventricles)
- are there nearby vascular structures that are going to be at risk? (e.g. displaced or encased arteries; large overlying veins)
- are the dural venous sinuses compressed, invaded or occluded?
- is there evidence of bone invasion or extracranial extension?
Although the list of possible entities resulting in an intracranial mass (even excluding intraventricular, pineal and pituitary mass) is very long, limiting your attention to a handful of entities will allow you to make the correct diagnosis in the vast majority of cases. There are many situations where less common entities should be considered; this is beyond the scope of this article.
More importantly, you want to avoid mistaking a tumour-mimic for a neoplasm as management is going to be vastly different.
Therefore the main conditions to consider are:
- extra-axial masses
- intra-axial masses
From a purely statistical point of view, once you have established that the mass is extra-axial, by far the most likely diagnosis is that of a meningioma. Given the large number of histological meningioma variants (e.g. microcystic, chordoid, lipomatous etc..) each of which has variable imaging features, even if the appearances are not typical, meningioma in most instances remains the most likely diagnosis.
Having said that, if the mass is large and destroys bone, and particularly if it has numerous microlobulations (bumpy surface) and prominent surface vascularity (rather than spoke wheel / sunburst appearances) then a haemangiopericytoma should be considered.
A dural metastasis is worth considering, particularly if there is an appropriate history of systemic malignancy (particularly breast cancer, lung cancer, prostate cancer). No reliable imaging features to distinguish them from somewhat unusual meningiomas exist, although usually they have less well defined margins and frequently have adjacent oedema, which unfortunately is not uncommonly encountered in meningiomas.
Most intra-axial masses that do not enhance are low grade diffuse gliomas (most commonly fibrillary astrocytoma). They typically involve the cortex which is thickened with increased T2 signal. When the majority of the mass is located within the cortex with relatively well circumscribed margins, oligodendroglial component is likely (oligodendroglioma or oligoastrocytoma). This should particularly be considered if calcification is present.
A crucial diagnosis to not be missed is HSV encephalitis. Usually presentation will be helpful, with patients having a short clinical decline. In some instances the presentation will be less obvious (often due to co-morbidities). The pattern of involvement in immunocompetent adults is usually bilateral but asymmetrical involvement of the limbic system, medial temporal lobes, insular cortices and inferolateral frontal lobes. The basal ganglia are typically spared. Some enhancement usually develops later in the course of disease.
When the abnormality is small, and mass effect is minor, a region of focal cortical dysplasia is the main differential and distinguishing between the two is difficult, but fortunately in most instances follow-up can be safely undertaken. Elevated CBV or a tumour trace on MR spectroscopy can be clues that a lesion is a tumour.
There are three main patterns of enhancement: peripheral, solid and irregular.
Peripheral enhancement, often referred to as ring enhancement, is the most common pattern. The most important diagnosis that needs to be considered and not missed whenever presented with an intra-parenchymal mass which demonstrates peripheral enhancement is that of a cerebral abscess. Surgical drainage in most cases is a matter of urgency; "never let the sun set on an undrained abscess". Having said that, being too timid in dismissing the possibility of a lesion representing an abscess is also a failing.
There are a number of features that are helpful in suggesting the diagnosis or cerebral abscess:
- restricted diffusion on DWI of the central fluid component
- thin regular peripheral ring enhancement, thinner towards the ventricle
- dual rim sign on SWI / complete low signal T2 / T2* ring
Of these, striking abnormal restricted diffusion is most helpful, with the one caveat being that care should be taken in interpreting diffusion signal in cavities that contain blood products.
Once you have satisfied yourself that a lesion is not an abscess, you want to exclude other tumour mimics, namely resolving cerebral haemorrhage and tumefactive demyelination. Both of these lesions usually require nothing more than medical management and follow-up. Surgical intervention in such cases is not only usually unwarranted but will likely lead to unnecessary morbidity.
In most cases of cerebral haemorrhage, an obvious clinical event will be evident and acute imaging with CT will have been obtained. In some cases though, particularly in the infirm elderly, this will not have occurred and it falls to the radiologist to make the diagnosis. Firstly, it is important to take the location of the lesion into account as most will either be a basal ganglia haemorrhages or lobar haemorrhages. Careful scrutiny of SWI images for appropriate distribution of microhemorrhages is a helpful corroborating piece of evidence.
Secondly, the lesion itself will demonstrate central signal of altered blood product (variable depending on timing; see ageing blood on MRI). The peripheral enhancement is usually relatively minor, without solid / nodular components. Surrounding the mass a variable amount of vasogenic oedema will be present, but no non-enhancing cortical expansion will be visible.
Tumefactive demyelination can be challenging to distinguish form tumours. The most helpful sign is that of incomplete open-ring enhancement, where the leading edge of demyelination is enhancing and diffusion restricting, but does not form a continuous sphere around the center of the lesion. This edge typically gets interrupted by grey matter.
Glioblastoma and cerebral metastases are the most common peripherally enhancing lesions and distinguishing between the two can be challenging. This is discussed in the next section on solidly enhancing lesions, as the clues to distinguishing between the two are mostly at the periphery of the lesion rather than in the center.
It should be noted however that if you eventually decide to favour a glioma, presence of definite central non-enhancement usually implies necrosis and thus the diagnosis of WHO IV glioblastoma. This should be distinguished from thin-walled cyst formation or areas of high T2 / FLAIR suppression as seen in some oligoastrocytomas / protoplasmic astroctyomas / gemistocytic astrocytomas.
Solid enhancement is seen predominantly in high grade astrocytomas (WHO III or IV), cerebral metastases or CNS lymphoma. Each of these has a few helpful clues.
High grade gliomas, even when the majority of the tumour enhances, very often have adjacent non-enhancing tumour, most easily identified as an area of thickened high T2 signal cortex. Unfortunately non-enhancing tumour confined to the white matter can be difficult to distinguish from oedema. In such cases examining ADC maps for areas of relatively lower ADC (signifying cellular tumour) and perfusion (demonstrating elevated CBV) can be helpful.
Usually, gliomas have a single area of enhancement, however multiple areas of enhancement is not uncommon. In such cases the regions of enhancement are usually connected by T2 signal abnormality (multifocal glioma) or along expected white matter tracts. Having said that, although gliomas are usually solitary, they tend to have complex non-spherical shapes with satellite nodules.
Gliomas also have a predilection for crossing white matter tracts, particularly the corpus callosum (something which metastases do only rarely) and for subependymal spread.
Cerebral metastases, on the other hand, have no non-enhancing tumour component (all tumour tissue usually enhances - other than central necrosis). They are surrounded by vasogenic oedema with facilitated diffusion and no elevation of CBV. Metastases have a predilection for the grey-white matter junction and rarely involve the periventricular brain or corpus callosum. Although solitary metastases are common, multiple lesions are frequent and when present are randomly distributed through the brain.
For a more in depth discussion on distinguishing metastases for GBM see glioblastoma vs cerebral metastasis.
CNS lymphoma is the poster-child of solid enhancement, typically demonstrating very homogeneous vivid contrast enhancement with slightly blurry margins. These tumours are very cellular (small round blue cell tumours) and thus demonstrate low ADC values and are hyperdense on non-contrast CT. Lymphoma has a predilection for periventricular white matter, including the corpus callosum, cortex and deep grey matter. It is important to note that distribution, enhancement and morphology are significantly altered with immunosuppression or administration of steroids.
Irregular enhancement is a relatively uncommon presentation. In addition to diffuse gliomas (e.g. anaplastic astrocytoma; gemistocytic astrocytoma) and tumefactive demyelination, the main differential is that of a subacute cerebral infarction.
In such cases there are two main clues. Firstly the T2 abnormality will be confined to a vascular territory. Secondly the enhancement is largely restricted to the cortex. This is not always obvious, depending on the orientation of the sulci relative to the imaging plane. Unfortunately ADC values are often pseudonormalised, and most of the DWI high signal is attributable to T2 shine through. It is also worth noting that subacute infarcts will change rapidly and thus short interval follow-up will often easily confirm the diagnosis.
Time critical findings
The threshold of calling the referring clinician with the findings will depend on many factors including where the patient is (emergency department, ward vs outpatient) and what the clinician knows about the patient already (is this further assessment of a known abnormality or is it an unexpected finding).
A number of conditions and findings are life-threatening and in most instances require direct communication with treating clinician at the time of reporting. These include:
- presumptive diagnosis of cerebral abscess or HSV encephalitis
- acute obstructive hydrocephalus with transependymal oedema
- herniation (particularly transtentorial and tonsillar)
One of the important functions of a preoperative neuroradiology report should be to establish the parameters within which histology is to be interpreted. This requires a multidisciplinary approach but can avoid serious mistakes.
For example, as gliomas are heterogeneous tumours, with many glioblastomas arising from preexisting lower grade tumours (so-called secondary glioblastoma), under-grading due to sampling bias is not uncommon. A mass has definite central non-enhancement, with a continuous rim of peripheral enhancement, and no features to suggest a tumour mimic, it almost certainly is going to be a WHO IV glioblastoma or a cerebral metastasis. If the histology purports to be a WHO II or III astrocytoma, then most likely there has been sampling error and a second look at other parts of the specimen should be undertaken to seek out confirmatory necrosis or epithelial hyperplasia (both WHO IV features). This is important as therapy for gliomas depends on grade, and as does eligibility for certain newer drugs or trials.
Another common example is a diffuse brain tumour with high CBV. If histology comes back as WHO II oligodendroglioma or oligoastrocytoma then one is happy to accept the lower grade, as oligodendrogliomas have a propensity for elevated CBV (chicken wire vascularity). In contrast if the histology suggests a WHO II diffuse fibrillary astrocytoma, then one should be suspicious that this is either a high grade tumour (either sampling bias, or early IDH-1 negative tumour) or an oligodendroglial component has been overlooked. In any case follow-up and management should take this discrepancy into account.
A preoperative MRI brain report for a patient with an intracranial mass needs to accomplish a number of tasks that should be reflected in the conclusion of the report.
- It should localise the mass as being intra- or extra-axial, anatomically and relative to eloquent areas and vascular structures.
- It should establish a likely diagnosis, giving a sense of how certain each differential is.
- Cerebral abscess should be considered and preferably avoiding equivocation.
- Tumour mimics (tumefactive demyelination, subacute haemorrhage and infarction) should be considered to avoid unnecessary resection.
- Treating clinicians should be contacted with time critical findings.
- A framework within which histology results are to be interpreted should be evident. In other words, which diagnoses are considered acceptable as histology results.
- A discordance between radiology report and histology report should prompt a multidisciplinary review of the case.