•
•
Computed Tomography
Ophthalmopathy 1
Computed Tomography in Graves'
Dieter Enzmann, M.D., William H. Marshall, Jr., M.D., A. Ralph Rosenthal, M.D., and Joseph P. Kriss, M.D. The CT scan with the 160 X 160 matrix demonstrated both the normal orbital anatomy and the abnormal orbital anatomy of Graves' ophthalmopathy in great detail. In Graves' ophthalmopathy, the cardinal pathologic feature of extraocular muscle enlargement was accurately reflected on the CT scan and was a distinctive, diagnostically reliable finding. Enlargement of the medial and lateral rectus muscles and of the apex of the muscle cone were the most consistent findings. The severity of the CT scan abnormalities correlated well with clinical severity. Because muscle cone abnormality was observed characteristically in those patients with sight loss, we suggest that pressure by the extraocular muscles on the optic nerve may contribute to visual acuity loss in this disease. INDEX TERMS:
Computed tomography, cranial. Eyes. Muscles, abnormalities. Nerves,
optic. Orbit Radiology 118:615-620, March 1976
• RAVES' DISEASE accounts for more cases of unilateral and bilateral exophthalmos in the adult population than any other single disease entity (5). The diagnosis of Graves' ophthalmopathy is usually not a clinical problem, but can be if proptosis and other eye signs precede thyrotoxic manifestations or goiter. Various serologic tests and diagnostic radionuclide procedures are helpful in documenting thyroid disease even in the euthyroid individual (6). However, evaluation of the precise location and extent of intraorbital changes leaves much to be desired. Computed tomography (CT) now makes it possible to obtain noninvasive cross-sectional images through the orbit with a resolution sufficient to assess the size of the extraocular muscles. Since gross enlargement of these muscles is one of the cardinal features of Graves' ophthalmopathy (8, 10), we decided to study the evaluative potential of CT scanning in patients with Graves' disease.
G
Table I: DIAGNOSIS
EMI EVALUATlDN
MEDIALRECTUS ENLARGEMENT
CLASSES INVDLVED
PAST TREATMENT
GRAVES'
B
6,4,3)
STERDIDS
H+
++
All 1+1
GRAVES'
T.5
6,5,4,3,2
RTX STERDIDS SURGERY
H+
H+
LATERAL RECTUS ENLARGEMENT
++
ALL 1+1
GRAVES'
T
6,4,3)
RTX
H+
(+)
GRAVES'
5
6,4,3,2
NONE
H+
GRAVES'
5
6,4,3,2
RTX
++
0
0
++
ALL
All (+) TWD 1+1 TWD I-I
GRAVES'
5
5,4,3,2
SURGERY
ALL (+)
GRAVES'
4
3,2
STEROIDS RTX
ALL 1+)
GRAVES'
3.5
3)
NONE
ALL 1+)
GRAVES'
3
4,3,2
STEROIDS SURGERY
++
SURGERY
++
4,3
MUSCLE CDNE ABNDRMALITY
0
0
ALL 1+1
GRAVES'
3
All (+l
GRAVES'
3
3,2
NDNE
ALL H
GRAVES'
1.5
3
STERDIDS RTX
0
ALLI)
GRAVES'
1
4
RTX
0
ALL I )
GRAVES'
4-UNILATERAl
4,3)
RTX
ALL H
GRAVES'
2·UNILATERAL
4,3
RTX SURGERY
0 0 0
0
0
ALL I-I
GRAVES'
0
0
NDNE
0
0
0
ALL ( I
GRAVES'
0
0
NDNE
0
0
ALL I ) ALL
NORMAL
0
0
NDNE
0
0
0
NDRMAL
0
0
NDNE
0
0
0
Table II: CLASS
The study group consisted of 24 patients with Graves' disease, all but 2 having ophthalmic complications. The diagnosis of Graves' disease was based on commonly accepted clinical and laboratory criteria. In those individuals without goiter or thyroid dysfunction, thyroid involvement was documented by the demonstration of elevated values of serum antimicrosomal and/or antithyroglobulin antibody and/or long-acting thyroid stimulator (4). These oatients were selected at random from the Nuclear Medicine thyroid clinic. Most, but not all patients had received prior treatment for thyrotoxicosis with radioiodine, antithyroid drugs and/or thyroidectomy. The ophthalmopathy in 19 cases had been treated with oral corticosteroids, orbital radiotherapy, and/or orbital decompression (TABLE I). The extent of ophthalmopathy was evaluated (J.P.K.) according to the classification
OPHTHALMOPATHY INDEX
All (+)
(4obsl
METHODS
Clinical and CT Scan Summary.
D
0
0
0
0
Method for Computing Ophthalmopathy Index.
INVOLVEMENT CATEGORY
0
NO EVECHANGES
1
'PROPTOSIS, NO OTHER SIGNS
2
SOFT TISSUE
EACH CATEGORY GRAOEO MINIMAL MOOERATE MARKED
=1 =2 =3
3
• PROPTOSIS
OPHTHALMOPATHY INOEXIEVE)
SUMOF CATEGORIES
4
EXTRA·OCULAR MUSCLE
OPHTHALMOPATHY INDEX(PATIENT)
AVE OF BOTH EYES (EXCEPT UNILAT)
5
CORNEA
6
SIGHT LOSS ·KRAHN EXOPHTHALMOMETER READINGS SCOREO MINIMAL (20-23mm) = 1 MODERATE ( 23-27mml = 2 MARKED ( 27mm) = 3
accepted by the American Thyroid Association (TABLE II) (10), and an ophthalmic index was calculated (4) (TABLE II) which is a numerical representation of the various eye signs present (TABLE I). For 2 patients with unilateral involvement, the index of the involved eye only represented the degree of involvement. Two patients, one with a suspected acoustic neuroma and the other with prostatic carcinoma, served as normal orbit contro!s.
1 From the Department of Radiology and Division of Ophthalmology, Stanford University School of Medicine, Stanford, Calif. 94305. Accepted for publication in September 1975. shan
615
616
DIETER ENZMANN AND OTHERS
March 1976
Fig. 1. Comparison of anatomic sections and comparable CT scan cuts of a normal orbit at an inferior orbital plane A, midorbital plane B, and superior orbital plane, C. Note the accurate anatomic detail of the globes, retrobulbar fat, optic nerve, and extraocular muscles on the CT scan in comparison to the corresponding anatomic section. (See text for further description.)
Several CT scans of patients with one normal orbit but obvious unilateral orbital tumors were available for comparison but not included in this series. The commercially available EMI (Eleotrical and Musical Industries Ltd.) scanner was used with scans performed at 0 degrees to Reid's baseline. The first 7 patients were studied with the 80 X 80 matrix with 8 mm collimators. The remaining 19 patients were studied with the 160 X 160 matrix, 12 with 8mm and 7 with separated 5mm collimators (7). A series of Polaroid prints of the 26 orbital CT scans were independently interpreted by four radiologists (three neuroradiologists and one resident). All prints were photographed at a setting of -20 EMI units and a window width of 75-100 EMI units (9). The following orbital structures were evaluated: optic nerve; medial, lateral, superior, and inferior rectus muscles; oblique muscles; apex of the muscle cone;
density of intraorbital fat; and globe displacement. A scale ranging from 0 to +4 was used for grading enlargement or displacement. A category of "not visualized" was used if the scan was judged to be technically inadequate for evaluating a particular structure. The reliability of the radiologic interpretations was evaluated by tabulating the number of disagreements between radiologists in evaluating each structure in each eye. A disagreement was defined as a grading disparity of two Or more points (i.e., + 1 and +3); or any grade V5. a reading of "nonvisualized." Clinical correlation was evaluated by comparing the radiologic assessment of muscle size to the calculated ophthalmic index of the patient. The EMI densities for an area of definite retrobulbar fat, medial and lateral to the optic nerve, were calculated from the 160 X 160 numerical printouts. None of these patients was evaluated with orbital ultrasonography.
Vol. 118
COMPUTED TOMOGRAPHY IN GRAVES' OPHTHALMOPATHY
617
Computed Tomography
Fig. 2. CT scan (inferior orbit A,B; midorbit C; superior orbit D) of Graves' patient with ophthalmic index of 3 but previously treated with bilateral-lateral orbital decompression. Bilaterally symmetric medial (+2) and lateral rectus (+2) enlargement is seen with herniation of the right lateral rectus into the operative defect of the lateral bony orbit. Optic nerve is of normal size.
RESULTS
The 80 X 80 matrix CT scans resulted in numerous disagreements between observers, and consequently this matrix was judged to yield images inadequate for the evaluation of orbital structures in Graves' patients. On these grounds, these seven scans were eliminated from further analysis. No significant difference in subjective image quality was discernable between the 8mm and separated 5mm cones using the 160 X 160 matrix. A comparison of a series of 3 anatomic sections and 160 X 160 matrix scans at comparable orbital levels in a normal orbit is shown in Figure 1. The lowest anatomic section (Fig. 1, A) demonstrates the globe and a portion of the inferior rectus muscle near the apex of the
Fig. 4. CT scan (inferior orbit A,B; midorbit C; superior orbit D) of Graves' patient with ophthalmic index of 8 and visual acuity loss. There is marked bilateral symmetric enlargement of the medial (+3), lateral (+2) and inferior rectus muscles; the superior rectus is also enlarged. The muscle cone is enlarged (+3) and its possible pressure effect on the optic nerve is shown anatomically. The optic nerve is of normal size. The retrobulbar fat is of normal appearance (and density) although accounting for less area on midorbital cuts (B,C) (see Fig. 3).
Fig. 3. CT scan (inferior orbit A; midorbit B,C; superior orbit D) of Graves' patient with ophthalmic index of 1.5. Bilaterally symmetric medial rectus enlargement (+ 1) is seen and superior rectus enlargement. The inferior rectus is identified (A) but cannot be evaluated for enlargement. The muscle cone is normal as is the optic nerve. Note normal appearance of the retrobulbar fat.
bony orbit. This muscle was often not well visualized on the CT scan (left orbit). Because of the oblique orientation of the inferior and superior rectus muscles to the x-ray beam, the entire muscle body was rarely visualized on one section. Imaging was also degraded by a partial thickness arti..
Fig. 5. CT scan (inferior orbit A; midorbit B; superior orbit C) of Graves' patient without ophthalmopathy. No extraocular muscle or muscle cone enlargement is seen.
618
DIETER ENZMANN AND OTHERS
CORRELATION OF MEDIAL RECTUS ENLARGEMENT AND CLINICAL SEVERITY
+4...----...----...----.-------, t-
~
•
:!: +3
w t:1
a:
•••
•
Computed Tomography
Ophthalmopathy 1
Computed Tomography in Graves'
Dieter Enzmann, M.D., William H. Marshall, Jr., M.D., A. Ralph Rosenthal, M.D., and Joseph P. Kriss, M.D. The CT scan with the 160 X 160 matrix demonstrated both the normal orbital anatomy and the abnormal orbital anatomy of Graves' ophthalmopathy in great detail. In Graves' ophthalmopathy, the cardinal pathologic feature of extraocular muscle enlargement was accurately reflected on the CT scan and was a distinctive, diagnostically reliable finding. Enlargement of the medial and lateral rectus muscles and of the apex of the muscle cone were the most consistent findings. The severity of the CT scan abnormalities correlated well with clinical severity. Because muscle cone abnormality was observed characteristically in those patients with sight loss, we suggest that pressure by the extraocular muscles on the optic nerve may contribute to visual acuity loss in this disease. INDEX TERMS:
Computed tomography, cranial. Eyes. Muscles, abnormalities. Nerves,
optic. Orbit Radiology 118:615-620, March 1976
• RAVES' DISEASE accounts for more cases of unilateral and bilateral exophthalmos in the adult population than any other single disease entity (5). The diagnosis of Graves' ophthalmopathy is usually not a clinical problem, but can be if proptosis and other eye signs precede thyrotoxic manifestations or goiter. Various serologic tests and diagnostic radionuclide procedures are helpful in documenting thyroid disease even in the euthyroid individual (6). However, evaluation of the precise location and extent of intraorbital changes leaves much to be desired. Computed tomography (CT) now makes it possible to obtain noninvasive cross-sectional images through the orbit with a resolution sufficient to assess the size of the extraocular muscles. Since gross enlargement of these muscles is one of the cardinal features of Graves' ophthalmopathy (8, 10), we decided to study the evaluative potential of CT scanning in patients with Graves' disease.
G
Table I: DIAGNOSIS
EMI EVALUATlDN
MEDIALRECTUS ENLARGEMENT
CLASSES INVDLVED
PAST TREATMENT
GRAVES'
B
6,4,3)
STERDIDS
H+
++
All 1+1
GRAVES'
T.5
6,5,4,3,2
RTX STERDIDS SURGERY
H+
H+
LATERAL RECTUS ENLARGEMENT
++
ALL 1+1
GRAVES'
T
6,4,3)
RTX
H+
(+)
GRAVES'
5
6,4,3,2
NONE
H+
GRAVES'
5
6,4,3,2
RTX
++
0
0
++
ALL
All (+) TWD 1+1 TWD I-I
GRAVES'
5
5,4,3,2
SURGERY
ALL (+)
GRAVES'
4
3,2
STEROIDS RTX
ALL 1+)
GRAVES'
3.5
3)
NONE
ALL 1+)
GRAVES'
3
4,3,2
STEROIDS SURGERY
++
SURGERY
++
4,3
MUSCLE CDNE ABNDRMALITY
0
0
ALL 1+1
GRAVES'
3
All (+l
GRAVES'
3
3,2
NDNE
ALL H
GRAVES'
1.5
3
STERDIDS RTX
0
ALLI)
GRAVES'
1
4
RTX
0
ALL I )
GRAVES'
4-UNILATERAl
4,3)
RTX
ALL H
GRAVES'
2·UNILATERAL
4,3
RTX SURGERY
0 0 0
0
0
ALL I-I
GRAVES'
0
0
NDNE
0
0
0
ALL ( I
GRAVES'
0
0
NDNE
0
0
ALL I ) ALL
NORMAL
0
0
NDNE
0
0
0
NDRMAL
0
0
NDNE
0
0
0
Table II: CLASS
The study group consisted of 24 patients with Graves' disease, all but 2 having ophthalmic complications. The diagnosis of Graves' disease was based on commonly accepted clinical and laboratory criteria. In those individuals without goiter or thyroid dysfunction, thyroid involvement was documented by the demonstration of elevated values of serum antimicrosomal and/or antithyroglobulin antibody and/or long-acting thyroid stimulator (4). These oatients were selected at random from the Nuclear Medicine thyroid clinic. Most, but not all patients had received prior treatment for thyrotoxicosis with radioiodine, antithyroid drugs and/or thyroidectomy. The ophthalmopathy in 19 cases had been treated with oral corticosteroids, orbital radiotherapy, and/or orbital decompression (TABLE I). The extent of ophthalmopathy was evaluated (J.P.K.) according to the classification
OPHTHALMOPATHY INDEX
All (+)
(4obsl
METHODS
Clinical and CT Scan Summary.
D
0
0
0
0
Method for Computing Ophthalmopathy Index.
INVOLVEMENT CATEGORY
0
NO EVECHANGES
1
'PROPTOSIS, NO OTHER SIGNS
2
SOFT TISSUE
EACH CATEGORY GRAOEO MINIMAL MOOERATE MARKED
=1 =2 =3
3
• PROPTOSIS
OPHTHALMOPATHY INOEXIEVE)
SUMOF CATEGORIES
4
EXTRA·OCULAR MUSCLE
OPHTHALMOPATHY INDEX(PATIENT)
AVE OF BOTH EYES (EXCEPT UNILAT)
5
CORNEA
6
SIGHT LOSS ·KRAHN EXOPHTHALMOMETER READINGS SCOREO MINIMAL (20-23mm) = 1 MODERATE ( 23-27mml = 2 MARKED ( 27mm) = 3
accepted by the American Thyroid Association (TABLE II) (10), and an ophthalmic index was calculated (4) (TABLE II) which is a numerical representation of the various eye signs present (TABLE I). For 2 patients with unilateral involvement, the index of the involved eye only represented the degree of involvement. Two patients, one with a suspected acoustic neuroma and the other with prostatic carcinoma, served as normal orbit contro!s.
1 From the Department of Radiology and Division of Ophthalmology, Stanford University School of Medicine, Stanford, Calif. 94305. Accepted for publication in September 1975. shan
615
616
DIETER ENZMANN AND OTHERS
March 1976
Fig. 1. Comparison of anatomic sections and comparable CT scan cuts of a normal orbit at an inferior orbital plane A, midorbital plane B, and superior orbital plane, C. Note the accurate anatomic detail of the globes, retrobulbar fat, optic nerve, and extraocular muscles on the CT scan in comparison to the corresponding anatomic section. (See text for further description.)
Several CT scans of patients with one normal orbit but obvious unilateral orbital tumors were available for comparison but not included in this series. The commercially available EMI (Eleotrical and Musical Industries Ltd.) scanner was used with scans performed at 0 degrees to Reid's baseline. The first 7 patients were studied with the 80 X 80 matrix with 8 mm collimators. The remaining 19 patients were studied with the 160 X 160 matrix, 12 with 8mm and 7 with separated 5mm collimators (7). A series of Polaroid prints of the 26 orbital CT scans were independently interpreted by four radiologists (three neuroradiologists and one resident). All prints were photographed at a setting of -20 EMI units and a window width of 75-100 EMI units (9). The following orbital structures were evaluated: optic nerve; medial, lateral, superior, and inferior rectus muscles; oblique muscles; apex of the muscle cone;
density of intraorbital fat; and globe displacement. A scale ranging from 0 to +4 was used for grading enlargement or displacement. A category of "not visualized" was used if the scan was judged to be technically inadequate for evaluating a particular structure. The reliability of the radiologic interpretations was evaluated by tabulating the number of disagreements between radiologists in evaluating each structure in each eye. A disagreement was defined as a grading disparity of two Or more points (i.e., + 1 and +3); or any grade V5. a reading of "nonvisualized." Clinical correlation was evaluated by comparing the radiologic assessment of muscle size to the calculated ophthalmic index of the patient. The EMI densities for an area of definite retrobulbar fat, medial and lateral to the optic nerve, were calculated from the 160 X 160 numerical printouts. None of these patients was evaluated with orbital ultrasonography.
Vol. 118
COMPUTED TOMOGRAPHY IN GRAVES' OPHTHALMOPATHY
617
Computed Tomography
Fig. 2. CT scan (inferior orbit A,B; midorbit C; superior orbit D) of Graves' patient with ophthalmic index of 3 but previously treated with bilateral-lateral orbital decompression. Bilaterally symmetric medial (+2) and lateral rectus (+2) enlargement is seen with herniation of the right lateral rectus into the operative defect of the lateral bony orbit. Optic nerve is of normal size.
RESULTS
The 80 X 80 matrix CT scans resulted in numerous disagreements between observers, and consequently this matrix was judged to yield images inadequate for the evaluation of orbital structures in Graves' patients. On these grounds, these seven scans were eliminated from further analysis. No significant difference in subjective image quality was discernable between the 8mm and separated 5mm cones using the 160 X 160 matrix. A comparison of a series of 3 anatomic sections and 160 X 160 matrix scans at comparable orbital levels in a normal orbit is shown in Figure 1. The lowest anatomic section (Fig. 1, A) demonstrates the globe and a portion of the inferior rectus muscle near the apex of the
Fig. 4. CT scan (inferior orbit A,B; midorbit C; superior orbit D) of Graves' patient with ophthalmic index of 8 and visual acuity loss. There is marked bilateral symmetric enlargement of the medial (+3), lateral (+2) and inferior rectus muscles; the superior rectus is also enlarged. The muscle cone is enlarged (+3) and its possible pressure effect on the optic nerve is shown anatomically. The optic nerve is of normal size. The retrobulbar fat is of normal appearance (and density) although accounting for less area on midorbital cuts (B,C) (see Fig. 3).
Fig. 3. CT scan (inferior orbit A; midorbit B,C; superior orbit D) of Graves' patient with ophthalmic index of 1.5. Bilaterally symmetric medial rectus enlargement (+ 1) is seen and superior rectus enlargement. The inferior rectus is identified (A) but cannot be evaluated for enlargement. The muscle cone is normal as is the optic nerve. Note normal appearance of the retrobulbar fat.
bony orbit. This muscle was often not well visualized on the CT scan (left orbit). Because of the oblique orientation of the inferior and superior rectus muscles to the x-ray beam, the entire muscle body was rarely visualized on one section. Imaging was also degraded by a partial thickness arti..
Fig. 5. CT scan (inferior orbit A; midorbit B; superior orbit C) of Graves' patient without ophthalmopathy. No extraocular muscle or muscle cone enlargement is seen.
618
DIETER ENZMANN AND OTHERS
CORRELATION OF MEDIAL RECTUS ENLARGEMENT AND CLINICAL SEVERITY
+4...----...----...----.-------, t-
~
•
:!: +3
w t:1
a:
•••
