AJNR: 19, September 1998 EDITORIALS 1385 head and neck coil has considerable appeal, and such viding the “payoff” of very rapid coronal slab se- “combo coils” will soon be widely used. At the very quences obtained sequentially, while maintaining sig- least, the current “conventional” cervical-cranial nificantly high intravascular signal to limit flow- MRA techniques can be supplemented with the tech- related artifacts. Thus, LeClerc et al take another step nique described by LeClerc et al for evaluation of the up the stairway, the top of which is the complete aortic arch and proximal cervical vasculature. Second, replacement of conventional angiography with MRA with further technical development, it is likely that for evaluation of cerebrovascular disease. coverage can be extended to allow visualization of the ostia and the carotid siphons in one acquisition. Newer contrast agents that do not exit the intravas- MICHAEL BRANT-ZAWADZKI,MD cular space will optimize the ability to do repeated Hoag Memorial Hospital Presbyterian MR sequences after only a single injection, thus pro- Newport Beach, California

Detection of Perineural Spread: Fat Is a Friend

One could easily argue that the search for perineu- affected foramen, enhancement of the nerve, mass ral tumor spread is the most important task of the effect in the Meckel’s cave region, and obliteration of radiologist examining a patient with head and neck the fat in the pterygopalatine fossa. I would like to carcinoma. emphasize the importance of this last finding: oblit- Certainly the description of a primary tumor site is eration of normal fat just external to the neural fora- important. The relationship of tumor to bone has men. definite implications in surgical and radiation treat- Each of the major head and neck neural pathways ment planning. Many imaging findings alter the sur- transit some amount of fat immediately external to gical plan, but adjustments, though important, tend to the skull base. The first division of the trigeminal be relatively minor. Nodal metastases are a definite (ophthalmic) neural pathway, which potentially car- determinant of prognosis, but the changes in therapy ries tumor from the lacrimal gland, passes through effected by imaging definition of nodal metastasis are the fat of the superior orbital fissure. The second relatively few. The discovery of a tumor that follows a division trigeminal (maxillary), with connections to nerve to or through the skull base, however, has an the face, palate, and maxillary sinus, must traverse the immediate and profound effect on the perception of a fat-filled pterygopalatine fossa just external to the patient’s disease. The chance for surgical cure plum- foramen rotundum. The third division (mandibular) mets, long-term prognosis is significantly changed, can carry tumor from connections of the submandib- and alternative therapies are considered. Detection of ular gland, parotid gland, or potentially, the oral cav- perineural spread is crucial. The findings can be very ity through the foramen ovale. Immediately below the subtle, and the radiologist must seek out any help foramen ovale, a fat pad sits just medial to the lateral available. pterygoid muscle. Tumor following the path of this In any discussion of perineural spread, terminology nerve must traverse this fat before entering the skull is important. Perineural tumor spread must be distin- base. The facial nerve passes though the fat of the guished from the designation perineural tumor that stylomastoid foramen before penetrating the tempo- can be found in the histopathologic report of a pri- ral bone. The hypoglossal and glossopharyngeal mary lesion. The report of perineural tumor in a nerves pass through a small amount of fat as they pathologic report indicates the relationship of tumor follow the carotid and jugular just below the skull to a nerve. This designation may have an effect on base. prognosis, but does not necessarily imply that the Tumor has an appearance very different from fat tumor has left the primary area. In perineural tumor on both CT and on T1-weighted MR images. In our spread, the tumor actually appears to use the nerve as experience this fat is, therefore, very sensitive to tu- a conduit, selectively following the nerve away from mor spread. An enlarging nerve obliterates that fat so the primary site. The tumor moves through the skull tumor can be detected. Conversely, demonstration of base along the same path as the affected nerve. This an intact fat pad is a reassuring finding, indicating route allows apparent distant tumor spread because strongly that the tumor has not passed the fat pad. tissues in between remain relatively undistorted. Thus, the radiologist should be aware not only of the Ginsberg and DeMonte in this issue of the Ameri- important neural pathways and the foramen that each can Journal of Neuroradiology (page 1417) give an traverses but also of the fat pad or fossa sitting im- excellent demonstration of one of the most important mediately external to the particular foramen. Oblit- of these perineural pathways: the second division of eration of this fat is often the first, and perhaps the the trigeminal nerve from palate to pterygopalatine easiest, finding for the radiologist to appreciate. fossa through foramen rotundum to Meckel’s cave. Which fat planes or pads are most crucial? This Their superb images show the key findings of peri- certainly depends on the site of the primary tumor in neural tumor spread. There is enlargement of the question. Although many tumors such as lymphoma, 1386 EDITORIALS AJNR: 19, September 1998 melanoma, and squamous cell carcinoma spread The anatomy of the head and neck is indeed com- along nerves, in our experience and in the current plex. Following neural pathways through the region study, adenoid cystic carcinoma is the most common can be an arduous task. Demonstration of an en- traveler along the perineural pathway. These tumors larged foramen, an enhancing nerve, or an enlarged arise in salivary glands both major and minor as well cavernous sinus is indeed ominous. The fat planes as in the lacrimal glands. Neural connections from adjacent to the skull base are equally important and these areas are of obvious importance. The foramen deserve the same understanding and scrutiny. Indeed, ovale and the stylomastoid foramen are the key tran- when searching for perineural tumor spread, as in sit points from the parotid. The highest concentration many situations in head and neck imaging, fat is truly of minor salivary glands in the body is in the posterior a friend. roof of the mouth, and adenoid cystic carcinoma frequently originates in this location. The pterygopal- HUGH D. CURTIN, M.D. atine fossa becomes an extremely important land- Massachusetts Eye and Ear Infirmary, mark as shown in the current article. Boston, MA

No Drug Is Benign

Despite the excellent safety record of gadolinium- The results obtained by Ray et al. indicate that, based compounds for enhanced MR imaging, the although the effects of gadopentetate dimeglumine potential neurotoxic effects of these drugs are not and gadodiamide are similar, they produce pathologic completely understood. Normally, the blood-brain effects that vary in character and location. This vari- barrier protects cerebral tissue from foreign chemi- ance can, in part, be attributed to differences in cals and the overaccumulation of native ones. A va- chemical composition: one is ionic, the other non- riety of pathogenic processes are known to heighten ionic. Striking pathologic manifestations were re- the permeability of the blood-brain barrier, however, ported after high-dose intraventricular administration allowing a significant increase in the local concentra- of both gadolinium compounds, however; indicating tion of a given pharmacologic compound. Creating that the acute excitatory effects were not agent-spe- reproducible and consistent models for the study of cific. the compromised blood-brain barrier remains chal- The authors admit that when these drugs are in- lenging. Obviously, histologic analysis of cerebral tis- jected intravenously with the typical doses prescribed sue after administration of contrast medium must be for conventional MR imaging, it would not be likely performed in an experimental animal rather than a that tissue concentrations would produce changes of human subject. Examination of animal tissue with an the severity reported in their study. We should nev- intact blood-brain barrier does not provide insight ertheless be careful with these drugs. A broken blood- into gadolinium-induced toxicity, and the results of brain barrier may allow a drug to reach unusually high concentrations in the cerebral parenchyma; it would induced blood-brain barrier compromise are often be important to know the threshold of accumulation inconsistent and nonquantitative. that would allow lesser, but significant, pathologic Ray et al in this issue of the American Journal of changes. It is worthwhile to remember that there Neuroradiology (page 1455) describe the use of the is always individual variation among tissues and intraventricular injection model as a reproducible and subjects. consistent way of studying the effects of the local There is a tendency for physicians to maximize the accumulation of a compound that may diffuse into amount of, or even overuse, a drug: if a little is good, the cerebral parenchyma after injection. This type of a lot may be better. The recent protocols for “triple- testing has always been performed for drugs, such as dose gadolinium” are a good example. While these myelographic compounds, intended for intrathecal protocols have not been associated with any higher and intraventricular space injection. While the gado- rate of occurrence of overt neurologic complications, linium-based agents were initially developed for in- we do not really know about the potential pathologic travenous injection, other methods of administration changes that may be induced from such doses, espe- are well known. Ray et al do not cite the results of any cially when the blood-brain barrier is broken. Al- tests initiated by pharmaceutical developers that ex- though we radiologists always want to see more, we amined the effects of gadolinium-based compounds must remember that there is a risk-benefit ratio to after intrathecal or intraventricular injection in the any compound. experimental animal. These tests may have been car- There has been the question of injecting gadolini- ried out previously; the manufacturer of the contrast um-based compounds into the intrathecal space for material administered in the present study was inter- cisternography to diagnose, for example, subtle leak- ested enough in this line of investigation to fund the age of cerebrospinal fluid. Other indications for in- project. trathecal and intraventricular gadolinium injection