BENIGN AND MALIGNANT TUMORS: GENERAL ISSUES, GROWTH PATTERNS, AND TREATMENT-RELATED CHANGES
KEY POINTS
- Growth patterns of benign and malignant neoplasms
- Perineural spread of malignancies
- Computed tomography, magnetic resonance imaging, and ultrasound appearance of benign and malignant neoplasms
- General patterns of postradiotherapy and postsurgical changes following treatment for malignancies
- Principles of posttreatment surveillance including fluorine-18 2-fluoro-2-deoxy-D-glucose positron emission tomography
The classification used here is intended to aid the reader in forming a general understanding of the benign and malignant neoplasms that affect the extracranial head and neck. Some neoplasms occur exclusively in the head and neck region, such as juvenile angiofibroma, but most others are seen elsewhere in the body. In some neoplastic processes, such as lymphoma, the head and neck involvement may represent only a small component of the disease. This chapter and those that follow in this section (Chapters 22–43) take a pathoanatomic and pathologic point of view. Each tumor type is considered in terms of its occurrence at various head and neck sites and in particular the effect of a neoplasm’s spread pattern, natural history, and cellular structure on its computed tomography (CT), magnetic resonance (MR), and ultrasound (US) appearance and its physiology as reflected on metabolic imaging studies such as fluorine-18 2-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET). Other sections and chapters emphasize the clinical and imaging implications of a particular type of tumor at a specific anatomic location.
GENERAL ISSUES
CT and MR and, to a lesser extent, US and radionuclide studies currently play a central role for any multidisciplinary team that is called upon to manage head and neck neoplasms. The value of an imaging study in helping to care for the patients of surgeons, radiation oncologists, medical oncologists, and others on the health care team is directly proportional to the skill and interest of the diagnostic imager. If the decision making is a shared responsibility performed in an open forum, such as a Head and Neck Tumor Board, patients can be fully informed of their treatment options and prognosis before a perhaps less optimal path is chosen. This ability to adequately inform is, in large part, due to the extraordinary amount of information concerning the extent of disease available with current medical imaging technology and the exceptional accuracy of pathologic diagnosis compared to that available just three decades ago.
Knowledge continues to evolve about the specific value and limitations of imaging studies in head and neck neoplasms. It is the task of the interdisciplinary team and especially diagnostic radiologists to refine their use in the safest and most efficient manner. However, newer imaging approaches must respect the more than 50 to 60 years of clinical experience in treating these disorders and augment those lessons learned rather than ignore or, worse, undo them.
The most basic contribution of imaging is its ability to find disease spreading beyond the limits of the physical examination of the head and neck, including more advanced endoscopic and physiologic testing procedures. This manifests itself in several ways. Spread beneath intact mucosa is not visible to the endoscopist, and imaging can lead to the initial diagnosis of a mucosally not apparent primary cancer (Figs. 21.1–21.4).1,2 Deep spread in a mucosally apparent tumor is often palpable but still usually difficult to assess objectively even with good bimanual palpation combined with inspection and history; however, experienced examiners can often anticipate involvement of many of the areas that are subsequently proven to be abnormal on imaging studies, principally CT and magnetic resonance imaging (MRI) (Figs. 21.5 and 21.6). Some locales are simply inaccessible for palpation, such as the skull base, pterygopalatine fossa, infratemporal fossa, orbits, and brain (Fig. 21.7). Critical management decisions frequently hinge on imaging findings alone (Fig. 21.7) and/or confirmation by imaging-guided biopsy (Fig. 21.8).
Perineural spread can also take malignancies far from the site of tumor origin and well beyond the limits of even the most skilled clinician’s physical examination (Figs. 21.9–21.11).3
In cancer patients, imaging can also provide an objective assessment of all cervical and retropharyngeal nodes proven for some time, although itself an imperfect method, to be far more accurate than the physical examination (Fig. 21.12).The extent of nodal disease is often crucial to both therapy and prognosis, and high-quality, detailed imaging can improve the accuracy of clinical staging of cervical and retropharyngeal nodal metastasis (Fig. 21.13).
Finally CT, MRI, and FDG-PET may be used to follow patients under treatment to monitor tumor response and try to detect recurrent or persistent disease before it becomes symptomatic and, hopefully, at a time when salvage therapy for cure is possible (Fig. 21.14). Each of these issues will now be considered in more detail.
BASIC TUMOR MORPHOLOGY AND SPREAD PATTERNS: MAINLY COMPUTED TOMOGRAPHY AND MAGNETIC RESONANCE APPEARANCE
Frequently, the decision must be made to use either CT or MR as the primary tool for evaluating a particular tumor or site. Several factors contribute to making this choice. The contrast between tumor and normal tissue is one of the more important factors. On non–contrast-enhanced CT, most tumors appear to be roughly the same density as muscle. Vague areas of fluid density somewhat less than the “tissue density” of muscle may be visible within the mass. This trend is of little practical use since non–contrast-enhanced CT is hardly ever used to evaluate cancer. Contrast between fluid and solid elements of the tumor and the much lower density fat will be excellent on CT. Contrast between tumor and lymphoid tissue such as the tonsils and adenoids will generally be poor. The tendency for poor contrast between tumor, muscle, and lymphoid tissue severely limits the value of non–contrast-enhanced CT. If contrast-enhanced computed tomography (CECT) is not possible, then contrast-enhanced magnetic resonance (CEMR) is the preferred imaging tool.
On CECT, most tumors will accumulate iodine to a greater extent than surrounding muscle; thus, CECT improves tumor to muscle contrast (Figs. 21.14 and 21.15). Many tumors will invoke an inflammatory response at their interface with normal tissues. This enhancing margin further improves tumor to muscle contrast and may help in evaluating the aggressiveness of a lesion particularly at tumor–muscle and tumor–fat interfaces. Similar improvement in contrast between tumor and lymphoid tissue may occur, but lymphoid tissue, when inflamed, will also enhance, so there is no guarantee of improved tumor–lymphoid tissue contrast (Fig. 21.15). An exception is the proven value of CECT in demonstrating focal metastatic deposits in normal size, untreated lymph nodes; in these cases, the reactive node parenchyma typically enhances to a greater degree than metastatic foci (Fig. 21.16). Marked enhancement may suggest a more uncommon etiology than the ubiquitous epithelial malignancies that involve the head and neck regions (Fig. 21.17). The degree of enhancement as a marker of perfusion and surrogate for capillary and angiogenesis on CECT has more recently been offered, at least experimentally, as a factor that might help predict response to radiotherapy.
A main benefit of MRI is that image contrast is based on several intrinsic factors (T1, T2, proton density, flow) as opposed to one (electron density), as with CT as discussed in Chapters 1–3). Intravenous contrast may also be used with MR, but contrast between tumor and normal tissues can be varied by simply manipulating the pulse sequences. Non–contrast-enhanced T1-weighted (T1W) MR images generally provide suboptimal contrast between tumor and muscle or lymphoid tissue and excellent contrast between tumor and fat (Fig. 21.18 and Table 21.1). Short TI inversion recovery (STIR) images may improve tumor to muscle contrast while maintaining good tumor to fat contrast but are also limited in usefulness for technical reasons. T2-weighted (T2W) fast spin echo (FSE) images provide excellent tumor to muscle contrast; however, T2W FSE images inherently diminish the tumor to fat contrast unless they are fat suppressed using frequency selective fat suppression (Fig. 21.18). Exceptions to this are tumors with areas of necrosis, or micro- or macrocystic components, where these “cystic” components stay relatively hyperintense to fat even on T2W FSE images (Fig. 21.19). Most nonnecrotic solid neoplasms will show signal intensities that roughly parallel that of fat and lymphoid tissue on T2W sequences; however, some may be closer to muscle, especially if there is a relatively fibrous stromal component or the tumor is inherently fibrous (Fig. 21.9C). Contrast between tumor and lymphoid tissue may be poor on T2W MR (Fig. 21.18). The most obvious example of this is lymphoma, but this tendency is found with most other tumors (Fig. 21.20 and Table 21.1)
O, not predictable but helpful at times; +, predictably good contrast in most cases; FS, fat suppression; Gd, gadolinium; –, potentially misleading and must usually compare with one or two other sequences to be sure that the pathology is not going unnoticed; SD, spin density; FSE, fast spin echo.
a“Lesion” includes a benign/malignant tumor and inflammatory/infectious processes arising extracranially.
bFat suppression done with frequency selective saturation pulse.
cFSET2 represents fast spin echo T2 without fat saturation.
Gadolinium CEMR behaves in a manner similar to that described for CECT. Tumor to muscle contrast is improved on contrast-enhanced T1W images with the margins and internal morphology of the lesion becoming more apparent (Fig. 21.18). However, since fat is bright on T1W images, CEMR will diminish the contrast between tumor and surrounding fat. This may cause a significant problem in evaluating tumors in extracranial locations or those invading bones containing fatty marrow such as the mandible and skull base (Fig. 21.8).
Fat suppression techniques can be used with T1 or T2W pulse sequences to overcome some of the difficulties with tissue contrast that were just discussed. Fat suppression techniques are not a cure-all, and the ideal single pulse sequence for evaluating all head and neck neoplasms does not yet exist. Also fat suppression techniques all suffer from problems with susceptibility artifacts.
In summary, the contrast between various tumors and surrounding tissue will vary with the tumor histology, size, growth pattern, and natural history of the lesion. Some specific factors that will influence the appearance of particular tumors to be discussed in subsequent sections of this chapter include cellular elements, stromal elements, an acinar or microcystic morphology, macrocystic components or necrosis and cyst contents, hemorrhage, naturally occurring paramagnetic substances, calcification or ossification, growth pattern and peritumoral reactive changes, prior treatment, and the use of intravenous contrast enhancement. The physical basis of some of these factors is considered in the technical section and other discussions in Chapters 1 through 5 and 10 through 12.
While FDG-PET is useful for decision making in some head and neck cancer patients, it is not on its own sufficient for showing the extent of the primary tumor, including bone destruction and perineural spread. The physiologic data from FDG-PET can be fused to CT and MRI data sets or acquired simultaneously to help refine the gross anatomic limits of metabolically active tumor.
US is only of limited use in evaluating primary malignant and benign tumors. It is used primarily in the thyroid and major salivary glands. Some practices find it useful in the routine evaluation of cervical nodes. Its use in exploring the elastic properties of tumors as a measure of their invasiveness or simply for differential diagnosis is in the early stage of clinical investigation.
INTERNAL MORPHOLOGY OF NEOPLASMS
Most tumors grow as a discrete mass. They are generally shaped like a sphere or solid ellipse, even when malignant, but may be modified by surrounding anatomy; the growth patterns that result are discussed subsequently (Fig. 21.21). The mass will have an internal appearance on CT or MR based on the major factors discussed previously and technical factors related to image acquisition.
Homogeneously solid tumors on gross appearance may show tremendous differences histologically. Such differences in microscopic structure will usually affect the MR appearance of the lesion to a far greater degree than the CT appearance. Tumors composed of tightly packed cells, with relatively little water in the cytoplasm and little stroma, will usually be isodense to muscle on non–contrast-enhanced CT and isointense to slightly hyperintense to muscle on non–contrast-enhanced T1W images (Figs. 21.1A and 21.4A). T2W FSE images will show such a lesion to be roughly isointense to fat, slightly hyperintense to muscle, and likely slightly less intense than lymphoid tissue (Figs. 21.1, 21.2, 21.5, 21.17, and 21.18). These trends are those most commonly present in squamous cell carcinoma (Table 21.1). The tumor cells may also have a large amount of watery cytoplasm. A good example is the heavily mucoid-laden cell population of a chordoma. The more voluminous or watery cytoplasm cause such tumors to have higher signal intensity on heavily T2W images than the more common variety of tumor cell population (Fig. 21.22). Densely cellular tumors with limited stromal volume tend to show restricted diffusion.
Tumor cells may be mixed with a neoplastic or reactive stroma. The appearance of CT and T1W images may due to differences in stromal elements as well as the neoplastic cell population. For instance, a water-rich stroma may cause the lesion to appear somewhat hypointense to muscle on T1W images. Variations in stromal structure can greatly alter the appearance on T2W images. At the extremes of this are the very watery mucoid matrix of chordoma that causes the tumor to appear bright (Fig. 21.22) and often partially calcified fibrous matrix of fibro-osseous lesions that make them appear sometimes muscle equivalent on heavily T2W images (Fig. 21.23). Reactive fibroproliferative processes that may mimic tumor can have even a more densely collagenous makeup and profoundly low signal intensity on T2W images—a characteristic that is distinctly uncommon in all but a very few benign and malignant neoplasms (Fig. 21.24). One can extrapolate from these extreme examples the varied appearances that result from lesser amounts of fluid, cells, or fibrosis in a tumor matrix or stroma.
Tumors may be nearly entirely or partially microcystic. A classic example is the proliferative hemangioma (Fig. 21.25). These true neoplasms may appear slightly hypointense to muscle on T1W images. The effect of the increased fluid in myriad microscopic vascular spaces is most noticeable as markedly increased signal intensity on heavily T2W images. This reasoning can be extrapolated to thyroid adenomas and papillary-follicular carcinoma and epithelial salivary gland tumors (Chapter 22) among others. These tumors may contain colloid, microcysts, acini, or vascular spaces filled with fluid-rich contents that will have high signal intensity on T2W images with the actual signal intensity seen dependent on the relative amount of macromolecules and bound and unbound water present in the respective “cystic spaces.”