Knowledge of the implications of signal intensity and density changes in normal and abnormal tissue, together with a lesion’s morphology, location, and clinical presentation, enables accurate diagnosis. At no extra cost we have included tables of useful radiologic gamuts in Appendix A after this chapter. Detection of lesions begins with knowledge of normal anatomy and its variants. It is only through reviewing numerous cases day after day that one obtains a mental image of the normal anatomy from which deviations can be readily identified. This is the source of the speed of the readings made by the cagey older professors of neuroradiology—they have burned a CD template into their memory bank of the pattern recognition for normality. Thus, they can scan an image at light speed and still detect the subtlest abnormalities. The trainees who view the most cases will be served well in this regard—they will have the easiest time detecting lesions.
Let us start with the complicated modality—MR. MR lends itself to image analysis from several perspectives. These are based on intensity, morphology, and location. Liken yourself to a stock market technical analyst. Note the trends. You really do not have to know that much about the company (pathology) if you just follow the basics.
The first question is the intensity of the lesion on conventional pulse sequences: (1) T1WI, (2) T2WI, (3) FLAIR, (4) DWI, and (5) gradient-echo/susceptibility scans. Remember that the b0 images from a diffusion-weighted scan are a poor man’s gradient-echo scan in just 40 to 50 seconds; film them and use them. Box 18-1 provides useful information concerning the characterization of lesion types with intensity information. Unfortunately, all too often, MR lacks specificity because most lesions are dark on T1WI and bright on T2WI. There is presently no signal intensity pattern hallmark that clearly distinguishes, say, multiple sclerosis plaques from infarction or tumor. If a lesion decreases in intensity as the T2-weighting increases, you are more likely to be dealing with a lesion that has susceptibility effects. Gradient-echo scanning emphasizes susceptibility differences and flow. Therefore, lesions that are hypointense on pulse sequences emphasizing T2* information appear significantly more hypointense on gradient-echo scans if magnetic susceptibility is present. These are usually hemorrhagic lesions.
The morphology of the lesion is important in its categorization. Several criteria are critical here: mass effect, atrophy, texture, edema pattern, extent of lesion, nature (solid versus cystic), number (single or multiple), distribution (e.g., along vascular supplies, Virchow-Robin spaces, cranial nerves, meninges, white matter tracts), involvement of one or both hemispheres, and enhancement characteristics. The presence or absence of mass effect is usually obvious. Lesions possessing mass effect are usually “active,” whereas those that do not may be either old or very new. Examples of the former include a tumor or new stroke, whereas the latter would include an old stroke or old traumatic injury to the brain. Mass effect can be subtle, with slight effacement of sulci; however, such changes are highly significant with respect to arriving at the correct diagnosis. Careful observation is necessary. Absence of mass effect does not necessarily indicate benignity; rather, such lesions may be early in their evolution. The presence of focal atrophy (tissue loss) signifies past insult to the brain. Although the brain parenchyma decreases with age normally, focal loss of parenchyma is significant. Furthermore, global atrophic changes exceeding those for age suggest other processes, such as steroid use, neurodegenerative disorders, and human immunodeficiency virus (HIV) infection.
Certain lesions have rather characteristic textures. For instance, oligodendrogliomas have a rather heterogeneous texture, whereas lymphomas are more homogeneous. If edema is associated with a lesion, there is clearly an irritative element involved. The converse of this is not true; that is, if there is no edema, there is nothing harmful. Lesions without edema can be virulent; for example, cortical metastases, gliomatosis cerebri, Creutzfeldt-Jakob disease, and HIV infection. The extent of the lesion also gives some clues to the diagnosis. In general, a lesion spanning both hemispheres is most likely tumor, because vasogenic edema does not usually cross the connecting white matter tracts. Generally, lesions that are aggressive are poorly marginated and infiltrative. Again, the converse is not true. Many lesions have cystic components or are themselves cystic yet span the spectrum of aggressiveness. These include colloid cysts, craniopharyngioma, cystic astrocytoma, and necrotic glioblastomas. FLAIR images can distinguish structures containing cerebrospinal fluid (CSF), which are hypointense, from more complex cystic lesions, the latter being high signal intensity with contents that have complex constituents, including high levels of protein. Multiplicity of lesions also changes the radiologic diagnostic gamut (e.g., metastatic lesions, multiple strokes, multicentric tumor, neurofibromas, and multiple sclerosis).
Enhancement is very important because it establishes that the lesion has an abnormal blood-brain barrier. It does not, however, indicate whether a lesion is benign or malignant, nor does it always demarcate the border of the lesion. It is instrumental in increasing our sensitivity to detecting abnormalities, particularly extra-axial and cortical neoplasms. We err on the side of giving contrast because it improves our sensitivity and specificity and enables us to read faster with more conviction.
Finally, location (just as in real estate) is of critical importance in making the correct diagnosis. Is the lesion intra-axial, extra-axial, or both? Obviously, extra-axial lesions suggest a different differential diagnosis from those that are purely intra-axial. Multiplanar images help resolve this question. Certain lesions have a propensity for specific locations; for example, herpes simplex favors the temporal lobe, oligodendroglioma the frontotemporal lobe, and juvenile pilocytic astrocytomas the posterior fossa and suprasellar region. Does the lesion involve the cortex, white matter, or both? This provides the initial diagnostic algorithm. If the lesion is predominantly in the white matter we might consider multiple sclerosis, whereas if it affects both white and gray matter it may suggest a stroke. In the latter instance the lesion should follow a vascular distribution.
The interpretation of CT overlaps that of MR with respect to location, morphology, presentation, and enhancement features. The exception to this rule is that vessels enhance on CT, whereas, if there is fast flow, they do not enhance on MR. For your limbic pleasure, consider the density characteristics of lesions on CT in Box 18-2.
