Correlation of CT Scanning and Pathologic Features of Ophthalmic Graves' Disease

Correlation of CT Scanning and Pathologic Features of Ophthalmic Graves' Disease STEPHEN L. TROKEL, MD, * FREDERICK A. JAKOBIEC, MDt Abstract: Correlating the CT scan features of patients with orbital Graves' disease with histopathologic observations allows one to focus more specifically on the distinguishing features of this disease with future research implications. Both CT scanning and pathologic studies have shown clearly that the extraocular muscles are the primary focus of the disease. Swelling of the extraocular muscles generally occurs within their bellys with sparing of the tendons. This contrasts with idiopathic inflammation of the muscles or myositis, which tends to involve the tendon as well. All of the associated findings in orbital Graves' disease probably flow from the enlarged volume of the extraocular muscles: proptosis, bowing of the medial lamina papyracea to accommodate the swollen belly of the medial rectus muscle, venous engorgement from stasis induced by direct compression of the orbital venous drainage, conjunctival and lid swelling, and lacrimal gland enlargement. Both radiographic and pathologic changes in the orbital fat are secondary and comparatively insignificant. While there appears to be no selective inflammation of the optic nerve meninges or the perineural connective tissues, enlargement of the extraocular muscle bellys where they converge at the crowded orbital apex brings about compression of the optic nerve, impairs its function, and causes visual decrease. Lymphocytic and plasmacytic infiltration along with edema within the endomysium of the extraocular muscles leads to the activation of fibroblasts with the production of acid mucopolysaccharides and progressive fibrosis. It is not known what attracts the lymphocytes to the extraocular muscles, why certain extraocular muscles are affected preferentially, why the disease may be asymmetrically unilateral, and whether a defect in T cell or B cell functions (or both) is immunologically at fault. [Key words: Graves' disease, CT scanning, pathology, B lymphocytes, T lymphocytes, extraocular muscles, orbital fat, proptosis.] Ophthalmology 88:553-564,

1981

From the Department of Ophthalmology, Columbia University College of Physicians and Surgeons, New York City*, the Departments of Ophthalmology and Pathology, The New York Hospital-Cornell Medical Center* and the Manhattan Eye, Ear and Throat Hospital, New York City and the Eye Bank for Sight Restoration, Inc., New York City.t Dr. Jakobiec is the recipient of the Research to Prevent Blindness Robert E. McCormick Scholars Award. This work was supported in part by NIH Grant EY03357. Reprint requests to Frederick A. Jakobiec, MD, Manhattan Eye, Ear and Throat Hospital, 210 East 64th Street, New York, NY 10021.

0161-6420/8110600/0553/$1.1 0 © American Academy of Ophthalmology

The pathophysiology of the orbital abnormalities in Graves' disease remains an enigma. Now that a large number of patients have been studied with high resolution computed tomography, it is an appropriate time to correlate the findings from these studies with what is known about the histopathologic changes that occur in the orbital tissues. Based upon our combined radiologic and pathologic experience with Graves' disease, we have concluded that the "shock organ" is almost exclusively the extraocular muscles. In this paper we would like to describe both the radiographic

553

OPHTHALMOLOGY • JUNE 1981 • VOLUME 88 • NUMBER 6

and the pathologic alterations within the extraocular muscles, and to highlight how they differ from other mimicking conditions, including idiopathic orbital myositis ("pseudotumor"), arteriovenous fistulas, and tumor metastatic to the muscles.

THE CLINICAL PROBLEM In the past, diagnosis of the eye signs of Graves' disease was difficult in the absence of a clear-cut history of abnormal thyroid function. In some patients without this history, clinical features accompanYlllg tl~e exophthalmos, especially the pattern of the motility disturbance, could be so characteristic that the diagnosis was suggested strongly and made presumptively. Only rarely would experienced clinicians suspect the diagnosis in the absence of the characteristic signs accompanying the exophthalmos. Laboratory support for these clinical impressions was obtained in some patients by the triiodothyronine suppression test and, more recently, by the TRH (thyrotropin releasing hormone) test. However, in a significant number of patients with exophthalmos, clinical abnormalities were present that could not be attributed definitely to Graves' disease and simulated the presence of an orbital neoplasm. Surgical exploration was required to define fully the pathologic processes in these patients and eliminate the possible presence of tumor. Computed tomography of continually increasing resolution has improved our ability to diagnose accurately these latter groups of patients. '-3 In addition to recognizing characteristic orbital changes in the aforementioned patients, we are recognizing identical morphologic changes in other patients in whom Graves' disease was unsuspected. Many of these patients had exophthalmos of unexplained origin that was attributed to inflammation or "pseudotumor," or remained obscure and undiagnosed.

CT SCANNING: LOW RESOLUTION Initial CT scanners had resolution of the order of 110 cubic mm which allowed the imaging of gross orbital abnormalities. Using the EMI Mark I head scanner, it was possible to visualize enlarged extraocular muscles in those patients in whom exopthalmos was present. In the milder forms of Graves' disease, resolution was inadequate to detect pathology with certitude. Early scanners occasionally allowed differentiation of other specific causes of extraocular muscle enlargement in some patients. The large muscle of acute myositis could be distinguished because of the associated inflammatory signs, as could the muscle enlargement due to carotid cavernous fistula or apical compression. The rare metastasis to the extraocular muscles was recognized also. 554

CT SCANNING: HIGH RESOLUTION Scanners available to us in the past several years have spatial resolution which approaches 1.1 cubic mm, representing a hundredfold improvement as compared to the Mark I EMI head scanner. This improved resolution is obviously very important when we image extraocular muscles whose normal maximum thickness of 2.8 mm decreases to less than 1 mm at the tendinous insertion. With these new scanners, the pattern of muscle enlargement can be defined for all seven extraocular muscles. The higher resolution capability has extended already our understanding of the pathophysiology of vision loss in severe forms of Graves' disease by resolving the crowded structures in the orbital apex. In addition, resolution of the minimal extraocular muscle enlargement in Grade 1 and 2 Graves' disease has made this a sensitive as well as an accurate diagnostic technique. The improved resolution has, as hoped, enhanced our ability to distinguish between Graves' disease and those other clinical entities that can produce large extraocular muscles.

CT FINDINGS IN GRAVES' DISEASE EXTRAOCULAR MUSCLES

The brunt of the tomographic abnormalities in orbits of patients with Graves' disease is borne by the extraocular muscles, a conclusion based on over 200 orbits of patients with Graves' disease studied by high resolution computed tomography. Enlargement of these muscles is the most salient and constant finding. Abnormalities range from minimal enlargement of few muscles (Fig 1) in patients with mild Graves' disease showing only lid retraction, to enormous enlargement of mUltiple extraocular muscles (Figs 2-5) in patients with severe forms of the disease. The ability to resolve structures at the orbital apex shows that direct muscular compression of the optic nerve is always associated with the optic neuropathy of Graves' disease. Characteristically, the muscle enlargement spares (Fig 2) the tendinous insertion, a feature that helps to distinguish the enlargement of Graves' disease from that of inflammatory myositis. Functionally, the enlarged muscles (Fig 1) initially contract normally. After a variable period they maintain the ability to contract but become fibrotic and are unable to relax (Fig 3). These fibrotic extraocular muscles restrict motility by a tethering action and present as an underaction of the antagonist muscle. In severe and massive involvement of the muscle, the muscle may not only become tethered and fibrotic, but in addition, it may lose its ability to contract (Fig 4).

TROKEL AND JAKOBIEC • GRAVES'DISEASE

Fig 1. A coronal CT scan in a patient with lid lag, globe lag, and stare. The patient is euthyroid but has an abnormal TRH test. Mild enlargement of the inferior rectus in the right orbit is noted as well as a minimal enlargement of the inferior rectus on the left. The coronal section of the slightly enlarged muscle has a "hamburger patty" appearance rather than the flat or mildly concaved contour of the normal coronal section. This appearance is the most sensitive, specific sign of mild muscle enlargement.

BONE CHANGES

These may occur secondary to pressure of the massively enlarged extraocular muscles against the bones in the crowded orbital apex. The lamina papryacea of the ethmoid is a delicate bone plate with little structural rigidity. As shown (Figs 4-7), medial bowing of the lamina paprycea provides more room for the muscle and can be considered a form of spontaneous orbital decompression. Severe Graves' disease associated with vision loss is often associated with bilateral medial displacement of the medial walls of both orbits, forming a "wasp-waisted" or "Coca Cola bottle" contour (Fig 7). THE LACRIMAL GLAND

This structure is seen to be enlarged (Figs 2 and 5) with a degree of contrast enhancement in active phases of Graves' disease. The gland becomes larger in cases of progressively enlarging muscles. ORBITAL FAT

The radiodensity of the orbital fat appears normal in patients with all stages of Graves' disease. Careful search for abnormal radiodensity or asymmetric radiodensity in unilateral Graves' disease fails to re-

Fig 2. In this and subsequent apical CT scans, the orbits are being reviewed from below in terms of right and left orientation. Patient has esotropia with inability to abduct either eye. The CT scan shows enlargement of the medial rectus muscle bilaterally with mild enlargement of the lateral rectus muscle. Both eyes are held in the adducted position by the tethering effect of the fibrotic medial recti which continue to contract but no longer allow the globe to abduct. Notice the enlarged lacrimal gland in the right orbit and the mild medial bowing of the lamina papyracea in the left orbit.

veal changes in the lucency of the orbital fat. This observation severely challenges the existing view that a portion of the exophthalmos is due to water drawn into the orbit fat by mucopolysaccharide deposition. This contrasts with a variety of orbital inflammations where edema fluid is seen infiltrating the fat, dramatically changing its radiodensity. The normal appearance of the orbital fat and the close correlation of muscle swelling with exophthalmos strongly suggests that the fat plays a secondary role in the eye changes of Graves' disease. In a single patient who was thyrotoxic and exophthalmic at age 16, an apparent increase in orbital fat (Fig 8) was seen when scanned at age 22. OPTIC NERVE

There is little change in the appearance of the optic nerve with the exception of those patients who show apical compression of this structure (Fig 6). Mild swelling and prominence of the nerve sheath can be seen in these cases. VASCULAR CHANGES

It is unclear whether in certain stages of Graves' disease prominent vascular channels can be consis-

555

OPHTHALMOLOGY • JUNE 1981 • VOLUME 88 • NUMBER 6

Fig 3. Axial CT scan through the lower third of an orbit in a patient with euthyroid Graves ' disease. The clinical presentation was a fixed right hypo tropia with inability to elevate the right globe above the midline. The CTscan shows an enlarged inferior rectus muscle in the right orbit. The oval radiographic appearance results from the oblique section through the enlarged muscle. This radiographic contour has been confused with an oval orbital mass resulting in unnecessary orbital and cranial surgical procedures. If there is any question as to the true nature of this density, coronal sections produce an unambiguous image. At this stage , the enlarged extraocular muscle will continue to contract. However, the muscle is no longer able to elongate normally and acts to tether the globe producing hypotropia and noncomitant vertical strabismus.

Fig 4. A massively enlarged medial rectus muscle neither adducts nor allows abduction in the left orbit. Note mild enlargement of the left lateral rectus, the enlarged right lacrimal gland and medial bowing of the medial wall of the left orbit. While the initial stages of muscle enlargement are accompanied by a tethering of the globe with inability to extend , massively enlarged muscles lose their ability to contract.

tently seen within the orbital fat. Certainly, this has not been a prominent feature of those patients who have been studied. DIFFERENTIAL DIAGNOSIS

It is important to consider all other clinical entities that can be associated with enlargement of the extraocular muscles. Three different processes exist: inflammatory myositis, vascular congestion, and metastatic tumor. Inflammatory myositis ("pseudotumor") is characterized (Fig 9) by enlargement of the extraocular muscles that extends anteriorly to involve the tendon. The associated edema leaks into the adjacent fat, perioptic connective tissues, Tenon's space, and other fascial planes. Systemic corticosteroids bring prompt resolution of the swollen muscles to normal size, which is demonstrable radiographically. A single muscle is usually involved in adults. Involvement of mUltiple muscles is common in children.

556

Fig S. Midaxial CT scan in a patient with grade VI Graves ' disease. Medical control was achieved in the left orbit with corticosteroids and radiotherapy. The right orbit shows greater exophthalmos, enlargement of the lacrimal gland and continued poor vision. Pressure was removed from the right optic nerve by removal of the medial wall of the orbit and the medial half of the floor . Rapid return of vision was noted even after cessation of all corticosteroids.

TROKEL AND JAKOBIEC • GRAVES' DISEASE

Figs 6A and B. Grade VI Graves' disease in the left orbit and grade V Graves' disease in the right orbit. Notice the greater distortion of the medial wall of the left orbit in the axial view. A, left , the axial section and B, right, the corneal section enlargement of the extraoc ular mu scles are seen. As the orbital apex is approached , all extraocular mu scles are seen. At the orbital apex all extraocular muscles impinge upon the optic nerve and compress it. Note the enlarged superior ophthalmic vein in the coronal sec tion at the right orbit. A component of venous stasis is suggested in association with compression at the orbital apex. In the coronal section an element of a fascial plane forming the muscle cone is seen. Note the predominance of swelling of the muscle fibers that spare the insertion of the muscles.

Vascular congestion of the muscle is most often due to increased pressure in the orbital veins associated with carotid-cavernous sinus fistulas or anteriovenous shunting. These cause enlargement of the superior and inferior ophthalmic veins and may be associated with visible vascular episcleral venous congestion. In some cases this involves the globe and causes dilation of the

central retinal veins and glaucoma. On CT scanning the enlarged extraocular muscles are seen (Fig lOA) in conjunction with enlargement of the superior ophthalmic vein (Fig. lOB) as well as prominent vascular channels within the orbit. The muscles appear to be more uniformly involved than seen in Graves ' disease. Apical compression by orbital neoplasms may also

Figs 7A and B. CT scans of two patients with grade VI Graves' disease. Contrast enhancement of the enlarged medial rectus and lateral rectus emphasize the extraocular muscle vis-a-vis the optic nerve . Patient's vision is reduced (A, left) to light perception on the right and count fingers' vision in the temporal field of the left eye. Bilateral medial orbital decompre ssions were performed. Note the relative sparing of the insertions of the extraocular mu sc les. Vision loss in (B , right) was greater in the left orbit. Note the medial bone deformation producing a "wasp-waist" or "Coca Cola bottle" contour.

557

OPHTHALMOLOGY. JUNE 1981 • VOLUME 88 • NUMBER 6

Fig 8. This is a midorbit CT scan of a 22-year-old patient who had Graves' disease since the age of 16. Thyrotoxicosis was resistant to drug treatment and required three doses of radioactive iodine to control. The exophthalmos had been stable for two years before the CT scans were made. This is the only patient of several hundred cases studied in whom there is an apparent increase in the volume of the orbital fat. The fat has normal radiodensity and does not appear to be infiltrated by water or fibrous tissue. Other sections show a mild enlargement of the inferior rectus muscle and increase in size of the lacrimal gland. We have never seen this apparent increase in orbital fat in older patients.

cause some degree of venous congestion. They are recognized because of the presence of the nodular tumor. Tumor may metastasize directly to the extraocular muscles. We have seen three patients with a metastatic carcinoma of the breast who have presented with impaired motility and an enlarged extraocular muscle. In the early stages a very slight enlargement may be nonspecific. However, as tumor growth proceeds, a characteristic (Fig 11) nodularity, focal contrast enhancement, and mass effect develop which distinguish this condition from other forms of muscle enlargement. UNEXPLAINED CLINICAL AND RADIOGRAPHIC SYNDROMES

A number of patients have appeared in whom the CT orbital findings lead to the presumptive diagnosis of CT signs of Graves' disease. The clinical presentation is so atypical in these patients that it was unsuspected. No endocrinologic evidence for Graves' disease can be shown to exist in most, but not all of these patients. One patient with 14 years of exophthalmos had normal motility. She was shown to have four enlarged extraocular muscles (Fig 12). Another patient with a complaint of periocular puffiness, mild conjunctival chemosis, and severe ocular irritation was initially thought to have a mild allergic blepharoconjunctivitis. CT scan showed similar enlargement of the extraocular muscles although no endocrinologic evidence to support the diagnosis was found. Clinical remission was induced with low doses of exogenous thyroid. 558

Fig 9. Midorbital scan of a 12-year-old boy with bilateral acute orbital myositis. This patient had rapid development of exophthalmos with mild conjunctival chemosis, injection and cutaneous erythema over several days. Motility was impaired. Note the edema surrounding the globes bilaterally. In myositis the inflammation extends quite anteriorly and usually involves the muscle tendon and periocular tissues. The lacrimal gland is not seen to be enlarged in these patients nor are bone changes seen. Correlation with clinical features may be necessary to evaluate properly these findings because on radiographic grounds alone they have much in common with those findings seen in early Graves' disease.

HISTOPATHOLOGIC FINDINGS The pathologic features of Graves' disease have received remarkably little attention of late. 4-8 One of us (F AJ) has evaluated approximately 50 biopsy specimens of fat and muscle in this disease. Additionally, through the courtesy ofthe Department of Ophthalmic Pathology at the Armed Forces Institute of Pathology, an exenterated orbit from a patient with Graves' disease (AFIP Acc. No. 692 463) has been reviewed to test the validity of the findings from the isolated biopsy specimens, and the data are consistent. The radiographic observation that the extraocular muscles are the predominant focus of the alterations in ophthalmic Graves' disease is substantiated on histopathologic evaluation (Figs. 13A and B). The extraocular muscles show some mixture of edema, mononuclear cell infiltration, mucopolysaccharides, and fibrosis. The lymphocytic infiltration is light and usually not hyperplastic. The fibroplasia will be mild in early stages of the disease, but will be marked in late stages. Biopsies of orbital fat obtained at the time of orbital decompression from patients with Graves' disease have consistently failed to show either significant

TROKEL AND JAKOBIEC • GRAVES'DISEASE

Figs lOA and B. Enlarged extraocular muscles in the midorbital scan of a patient with a traumatic carotid cavernous sinus fistula. The right orbit (A, left) is anophthalmic from a bullet injury, a piece of which is seen lodged in the floor of the sella. The enlarged extraocular muscles are seen in the left orbit. In the high orbital section (B, right) enlargement of the superior ophthalmic vein is noted, which identifies this as a carotid cavernous fistula. Similar findings are seen in arteriovenous malformations.

inflammation or marked fibrosis. A small amount of inflammation may be found in the fat immediately adjacent to the epimysial sheath (Fig 14). These findings correlate with the retained lucency of the orbital fat in virtually every case of orbital Graves' disease that we have studied by CT scanning. The tendons in the exenteration specimen were uninflamed (Fig 15), although a light inflammation in the neighboring fat was occasionally found. While one may rarely see a picture suggesting hypertrophy of normal density orbital fat on CT scanning, changes in its ground substance have not been noted by us. Dr. Carl Meyer at Columbia University studied two specimens of orbital fat obtained from patients with Graves' disease at the time of orbital decompression, and found that the level of the glycosaminoglycans in this compartment was undetectable. Similar findings were observed in controls from patients with tumors of the orbit. Daicker8 has reviewed recently the histopathologic features of ophthalmic Graves' disease in exenterated orbits obtained from five patients postmortem. Our experience is quite similar to his. The earliest changes consist of a lymphocytic and plasmacytic light infiltration within the connective tissue spaces (endomysium) of the extraocular muscles. This phase is followed by activation of the endomysial fibroblasts, which will produce a mixture of collagen and ground substance. The latter may be brought out with stains for acid mucopolysaccharides, such as the Alcian blue stain. As the disease progresses, the connective tissue

separating the individual myofibers becomes more prominent (Figs 16A and B). In specimens sectioned longitudinally, the endomysial edema and inflammation are well demonstrated and surround the surviving myofibers (Figs 17A and B). The perimysium surrounding bundles of myofibers undergoes progressive fibroplasia. In cross sections these areas of fibrosis may appear as hyalinized nodules surrounded by inflammation (Fig 18). Fibroplasia with collagen deposition is noted in the epimysial sheath of the extraocular muscles. The latter may, of course, extend from the outside of the muscles into the connective tissue septa that subdivides the orbital fat into lobules. In late stages, Daicker observed "lipomatosis" or the presence of mature fat cells even within the extraocular muscles. The mononuclear inflammation in both our biopsy specimens and the AFIP extenteration specimen was light. This may be the consequence of the fact that such tissues are obtained late in the disease (six months to a year after onset), but our opinion is that the lymphocytic infiltration at the onset of the disease is probably light and not hyperplastic. If one sees a prominent sheet-like lymphocytic infiltration of the orbital muscles, the diagnosis of lymphoid hyperplasia or .. pseudotumor," rather than Graves' disease, should be considered. Follicles are extremely rare in Gra ves' disease. "Pseudofollicles" are created by foci of degenerating muscle cells cuffed by lymphocytes (Fig 19). The mononuclear inflammation in Graves' disease is focal and diffuse (Fig 20). The cells are for 559

OPHTHALMOLOGY • JUNE 1981 • VOLUME 88 • NUMBER 6

Figs lIA and B. Adjacent axial CT scans through orbits of a 42-year-old woman with an initial superior oblique palsy. Progressive involvement of the medial rectus. a sensory deficit in the trigeminal distribution and finally loss of vision occurred. Note enlargement of the extraocular muscles (A, right) which are associated (B, leji) with focal nodul arity and high contrast definition from the surrounding structures. This patient had had a mastectomy 1Y2 years before the onset of her symptoms. Biopsy showed metastatic tumor consistent with origin from the breast. Three patients have been studied with focal met astatic disease to the extraocular muscles from the breast.

the most part lymphocytes and plasma cells, with a small number of associated mast cells (Fig 21). The only other tissue in the orbit that we have found to be consistently involved in Graves' disease is the lacrimal gland. Increased numbers of lymphocytes and plasma cells and accompanying edema will be noted within the acini of the lacrimal tissue, as well as in the interlobular connective tissue septa. Widespread fibrosis and obliteration of the lacrimal tissue, however, are not characteristic, in contrast to idiopathic dacryoadenitis ("pseudotumor"). The optic nerve was uninflamed in the exenteration specimen provided by the AFIP.

PATHOPHYSIOLOGIC MECHANISMS If the results of the CT scan and pathologic studies on orbital Graves' disease incriminating the extraocular muscles as the primary locus of the disease are accepted, then the proptosis can probablY be ascribed almost entirely to enlargement of the extraocular muscles without any concomitant changes in the orbital fat. As the mass of the extraocular muscle increases, the intraorbital pressure will become elevated, leading to a "tight orbit." Both the edema attendant upon the inflammation of the extraocular muscles and the increased orbital pressure probablY result in a slowdown of the normally precarious removal of the fluids in the interstitial space (since there are no orbital lymphatics), producing swelling of the lacrimal gland and some puffiness of the conjunctiva. Undoubtedly, venous stasis resulting from compression of the apical veins

560

by the enlarged muscles also contributes to this. The lid swelling is probably the result of both prolapsed orbital fat and some stasis of the interstitial fluids. All of the malfunctions of the extraocular muscles are postulated to be the result of the progressive deposition of collagen. In early stages, the edema in the endomysial space and the light mononuclear inflammation are probably susceptible to both radiotherapy and corticosteroid administration , but as the fibrosis becomes more severe, few modalities will be able to reverse the changes. As the muscle bellys of the swollen extraocular muscles converge at the annulus of Zinn at the orbital apex, the enclosed optic nerve will become compressed, causing a decrease in visual acuity. While the tendons of insertion onfo the globe of the extraocular muscles are spared in Graves ' disease, at the orbital apex the tendons of origin are of negligible size, so that the swollen muscles in this crowded precinct are in tight juxtaposition to the optic nerve. This compressive effect rather than inflammation of the meninges may be reduced by orbital radiotherapy or by surgical openings in the orbital walls. The injury to the extraocular muscles is caused by the lymphocytes and plasma cells which infiltrate initially within the endomysial compartment. Kroll and Kuwabara5 have demonstrated that the muscle fibers themselves on the ultrastructural level do not show early signs of degeneration, failing to support a primary muscle fiber injury. Electromyographic studies also point to an essentially normal muscle function. 9 The injury provoked by the immune system is mediated by activated fibroblasts, capable of producing both collagen and mucopolysaccharides. The production of mucopolysaccharides, therefore, is not a specific and

TROKEL AND JAKOBIEC •

Fig 12. Midorbital CT scan of a patient with a 14-year history of stable exophthalmos of the right eye. There was mild injection over the insertion of the lateral rectus muscle . No lid lag, stare or motility deficit was apparent. Multiple enlarged extraocular muscles resemble most the orbital findings in Graves' disease . These apparently stable enlarged extraocular muscles presumably were present for many years and may represent a highly atypical form of Graves' di sease. The absence of inflammatory sign s or venous congestion mitigate s against fistula or orbital inflammatory myositis. The appearance of the muscles most resemble s that of Graves' disease, although certainly cannot be distingui shed from a postulated chronic orbital myositis . Similar patients with highly atypical yet suggestive clinical finding s have been studied which show only enlarged extraocular muscles. We have tentatively identified this group of patient s as " radiologically presumed Graves ' disease ." They are distingui shed from those patients with euthyroid Graves ' disease, normal TRH te sts, and more typical patt ern s of motility disturbance th at are considered " clinically presumed Graves' disease with radiologic confirmation."

unique feature of the mYOSItIs of Graves' disease . Earlier work on Graves' disease led to the proposal that excess production of mucopolysaccharides ("ground substance") with attendant water binding was responsible for the ocular findings of Graves' disease. Pretibial myxedema and certain animal models lent further credence to this idea. Mast cells once thought to produce the mucopolysaccharides are present in small numbers in Graves ' disease, as well as in most inflammations, but are probably merely fellowtravelers in the inflammatory process. In our view, the mucopolysaccharide production is not dramatic or excessive in Graves' disease . All immature connective tissue is comprised of variable mixtures of immature collagen and mucopolysaccharides elaborated by activated fibroblasts. In Graves' disease mucopolysaccharides are produced early in the process of intramuscular fibroblast activation, followed by col-

GRAVES'D ISEASE

lagen production that becomes progressively hyalinized . The central mystery of Graves ' disease is why lymphocytes infiltrate the extraocular muscles in the first place . The suggestion that there are lymphatic connections between the thyroid gland and the orbit is highly speculative because the orbit has not been shown convincingly to possess endothelium-lined lymphatics. In many regards Graves' disease suggests an autoimmune disorder, wherein lymphocytes and/or immunoglobulins attack self-antigens, in this case those present on cells composing the extraocular muscles, whether they be fibroblasts or surface membrane antigens of the striated muscles themselves. Due to defects in immunoregulation, a "forbidden clone" of lymphocytes may be allowed to proliferate with a homing pattern for the extraocular muscles. The possibilities of a foreign antigen or antigen-antibody complexes localizing to the extraocular muscles and bringing about the infiltration of lymphocytes must also be entertained. Whether the latter mechanisms , or lymphocytic attack on unaltered surface antigens of the extraocular muscles is at fault , one is still perplexed by the preferential involvement of certain extraocular muscles over others. The lymphoid system is now divided into the B cell and the T cell compartments. lO B lymphocytes produce immunoglobulins , which may deposit on or between the extraocular muscles , whereas T lymphocytes are involved in delayed hypersensitivity, and their attack would be direct on the extraocular muscles by means of surface membrane receptors or of certain secretory products referred to as Iymphokines. Very few immunological studies have been performed directly on the extraocular muscles in Graves' disease. Immunoglobulins localized to the surfaces of the extraocular muscles would point to a B cell-mediated mechanism. More recent studies have indicated an abnormality in circulating T lymphocytesY T lymphocytes have regulatory functions with respect to B lymphocytes: T helper cells will facilitate B cells to produce immunoglobulins, whereas T suppressor cells will prevent certain B cells from proliferating.12 A model of this pathogenetic mechanism is afforded by multiple sclerosis, wherein exacerbations of the clinical disease occur during intervals where there is a relative depletion of circulating T suppressor cells. 13 The "released" B cells secrete increased amounts of immunoglobulin, which deposit within the myelin plaques in the diseased nervous system. It is conceivable that defects noted in circulating T lymphocytes in patients with Graves' disease may be ascribable to a depletion of T suppressor cells, resulting in uncontrolled B cell proliferation with the production of immunoglobulins that could be directed to the extraocular muscles. The presence of many plasma cells in the extraocular muscle infiltrates of Graves' disease would seem to implicate an immunoglobulin-mediated mechanism. Studies on the peripheral blood lymphocytic sUbpopulations and on excised orbital tissues may be helpful in elucidating the relative contributions of T and B cell abnormalities to the extraocular muscle injuries of Graves' disease. 561

OPHTHALMOLOGY. JUNE 1981 • VOLUME 88 • NUMBER 6

Figs 13A and B. A, top left, massively enlarged extraocular muscle with scattered lymphocytic aggregates dominates the right portion of the field, Note the absence of any inflammation in the orbital fat on the left. B, top right, another mildly inflamed and massively fibrotic extraocular muscle shown in juxtaposition to un inflamed orbital vessels and nerves on the left (hematoxylin and eosin, original magnification x60), Fig 14. Center left, a small amount of lymphocytic infiltration is present in the fat and in a fibrous septum of the adipose tissue immediately subadjacent to the involved muscle (hematoxylin and eosin, original magnification x60). Fig 15. Center right, an extraocular muscle tendon shown below is totally uninflamed. A small amount of mononuclear inflammation is identified in the fat next to the tendon (hematoxylin and eosin, original magnification x 120). Figs 16A and B. A, bottom left, a longitudinal section through an involved extraocular muscle discloses only scattered surviving striated muscle fibers which stain intensely eosinophilic. The muscle is for the most part almost totally replaced by fibrous tissue containing a light mononuclear inflammation. B, bottom right, in cross section, the degenerating muscle cells are circular, have central nuclei, or are multinucleated (A-hematoxylin and eosin, original magnification x 100; B-hematoxylin and eosin, original magnification x 220).

562

TROKEL AND JAKOBIEC • GRAVES'DISEASE

Figs 17A and B. A, top left, longitudinally sectioned orbital biopsy specimen shows pronounced perimysial fibrosis stained yellow in this modified trichrome stain. Scant inflammation is present, shown toward the top right. The surviving striated muscle fibers stain red, and a degenerating muscle fiber stains brilliantly red toward the top. B, top right, the endomysial compartment exhibits edema and a light mononuclear inflammation (A-hematoxylin, phloxin, saffron, original magnification x 100; B-hematoxylin, phloxin , saffron, original mag· nification x 100; B-hematoxylin, phloxin , saffron, original magnification x 180). Fig 18. Center left. hyalinized fibrotic nodules. cross sectioned within the extraocular muscles, are surrounded by mononuclear inflammation (hematoxylin and eosin. original magnification x 80). Fig 19. Center right, a "pseudofollicle" within the extraocular muscles consists of a cuff of lymphocytes surrounding the degenerating striated muscle cells and fibrosis . True follicles with germinal centers are rare in the infiltrates of Graves ' disease (hematoxylin and eosin , original magnification x 80). Fig 20. Bottom left . the focal patches of inflammation in the muscles are diffuse and not particularly hyperplastic ; they are hardly ever confluent (hematoxylin and eosin , original magnification x 80). Fig 21. Bottom rif?ht. the infiltrate in the extraocular muscles consists of lymphocytes and abundant plasma cells . Two mast cells shown toward the right have vivid eosinophilic granular cytoplasm and a centrally positioned dark nucleus (hema~oxylin and eosin . original magnification x 240).

563

OPHTHALMOLOGY. JUNE 1981

DIFFERENTIATION FROM IDIOPATHIC ORBITAL INFLAMMATION ("PSEUDOTUMOR") The single most important differential diagnosis from Graves' disease is idiopathic orbital inflammation. This condition is often referred to as orbital inflammatory pseudotumor, but with the advent of high resolution computed tomography, it has become possible to identify the exact site of the idiopathic orbital inflammation, leading to more satisfying designations such as idiopathic sclerotenonitis, myositis, perineuritis, dacryoadenitis, etc. Regardless of the designation, this cluster of entities should be distinguished from Graves' disease, and this has not always been clearly accomplished in previous papers on Graves' disease and "pseudotumor." From the clinical point of view, idiopathic orbital inflammation normally has an explosive onset, in contrast to the slower, more insidious onset of Graves' disease. 14 In idiopathic orbital inflammation there is usually an abrupt onset of an extraocular motility problem, decrease in visual acuity, pain, and erythema and swelling around the eyes, often reaching a crescendo shortly after the onset of the disease. Inflammatory signs may be sufficiently prominent to suggest an orbital cellulitis. This situation contrasts with that of Graves' disease, where the onset is usually less explosive. Extraocular motility and visual disabilities occur later in the course of the disease compared with idiopathic inflammation. The CT scan will show involvement not only of the extraocular muscles in idiopathic inflammation, but also of the orbital fat, the perineural connective tissues, Tenon's space, and the sclera. A distinct "ring" sign outlines the globe in the presence of anterior orbital inflammation. The inflammatory signs are more steroid-sensitive in idiopathic inflammation than Graves' disease. From the pathologic point of view, 15 idiopathic orbital inflammation also has distinguishing features from Graves' disease, as initially highlighted by Jellinek. 6 In histopathologic specimens obtained from patients with idiopathic orbital inflammation, one generally finds more exuberant lymphocytic infiltration of the extraocular muscles and the fat. Lymphoid follicle formation is more common in idiopathic inflammation than in Graves' disease. Additionally, one may see a

564

• VOLUME 88 • NUMBER 6

more mixed polymorphous inflammation in orbital idiopathic inflammation, including eosinophils and polymorphonuclear lymphocytes. The histopathologic features of idiopathic orbital inflammation are described elsewhere in this issue in the article by Mottow-Lippa et alY

REFERENCES 1. Trokel SL, Hilal MK. Submillimeter resolution CT scanning of orbital diseases. Ophthalmol 1980; 87:412-7. 2. Trokel, SL, Hilal SK. Recognition and differential diagnosis of enlarged extraocular muscles in computed tomography. Am J Ophthalmol 1979; 87503-12. 3. Enzmann D, Marshall WH, Rosenthal AR, Kriss JP. Computed tomography in Graves' ophthalmology. Radiology 1976;

118:615. 4. Tengroth B. Histological studies of orbital tissues in a case of endocrine exophthalmos before and after remission. Acta Ophthal 1964; 42588-91. 5. Kroll A, Kuwabara T. Dysthyroid ocular myopathy: anatomy, histology, and electron microscopy. Arch Ophthalmol 1966;

76:244-57. 6. Jellinek EH. The orbital pseudotumor syndrome and its differentiation from endocrine exophthalmos. Brain 1969; 92:35- 58. 7. Riley FC. Orbital pathology of Graves' disease. Mayo Clinic Proc 1972; 47:975-9. 8. Daicker B. The histological substrate of the extraocular muscle thickening seen in dysthyroid orbitopathy. Klin Mbl Augenheilk

1979; 174843-7. 9. Jensen SF. Endocrine ophthalmoplegia. Is it due to myopathy or to mechanical immobilization. Acta Ophthalmologica 1971;

49:679-84. 10. Knowles DL II, Jakobiec FA, Halper J. Immunologic characterization of ocular adnexal lymphoid neoplasms. Am J Ophthalmol 1979; 87:603-19. 11. Sergott R, Felberg N, Savino P et al. E-rosette formation in Graves' ophthalmopathy. Invest Ophthalmol Vis Sci 1979;

18:1245-51 12. Reinherz EL, Schlossman SF. Regulation of the immune response: inducer and suppressor T-Iymphocyte subsets in human beings. N Engl J Med 1980; 303:370-3. 13. Reinherz EL, Weiner HL, Hauser SL, et al. Loss of suppressor T cells in active multiple sclerosis. N Engl J Med 1980;

303125-9. 14. Mottow L, Jakobiec FA. Idiopathic Inflammatory orbital pseudotumor in childhood I. Clinic~1 characteristics. Arch Ophthalmol 1978; 961410-7 15. Mottow-Lippa L, Jakobiec FA, Smith M. Idiopathic inflammatory orbital pseudotumor in childhood. II. Results of diagnostic tests and biopsies. Ophthalmol 1981; 566-75.