PURPOSE
To determine whether 18 F-Fluorodeoxyglucose ( 18 F-FDG) Positron Emission Tomography ( 18 F-FDG PET) can be utilized as a marker for extraocular muscle function, particularly in extraocular muscle paresis.
METHODS
A retrospective study was performed evaluating 18 F-FDG uptake in extraocular muscles of patients with normal ocular motility across a broad age range (range 1-87 years), to establish a normative reference database. Additionally, 18 F-FDG uptake was evaluated in paretic extraocular muscles and compared to the control cohort to assess whether 18 F-FDG PET could demonstrate changes in metabolic activity associated with functional impairment.
RESULTS
The normative database included a total of 857 rectus muscles from 111 patients. 18 F-FDG uptake declined progressively with age in all rectus muscles, with statistically significant associations for each muscle in the coronal plane. The medial and lateral recti demonstrated the steepest age-related decline in 18 F-FDG uptake ( P <.0005 and P <.0002, respectively). 18 F-FDG uptake in 23 paretic rectus muscles from 10 patients demonstrated markedly reduced uptake compared with controls for each rectus muscle ( P <.003). This difference remained highly significant when all paretic muscles were analyzed collectively using multilevel modeling ( P = 9.07e−18).
CONCLUSIONS
18 F-FDG PET shows promise as a surrogate marker of extraocular muscle function, particularly in hypometabolic conditions such as extraocular muscle paresis. The development of a muscle-specific normative reference database provides a foundation for broader applications of this modality, with the potential to better characterize a wide range of ocular motility disorders (NOTE: Publication of this article is sponsored by the American Ophthalmological Society).
INTRODUCTION
The extraocular muscles (EOMs) are uniquely specialized, performing a wide repertoire of precisely controlled movements that vary in speed, and are the most fatigue-resistant muscles in the human body. ,,, They are capable of rapid movement while also sustaining slow, steady motion and maintaining constant tonic activity in the absence of overt motion. Clinically, EOMs demonstrate selective vulnerability across neuromuscular diseases: they are prominently affected in conditions such as myasthenia gravis, mitochondrial myopathies and oculopharyngeal muscular dystrophy, yet remain relatively spared in others including Duchenne’s muscular dystrophy, limb-girdle muscular dystrophy and amyotrophic lateral sclerosis. ,,,
This broad functional spectrum of EOMs is reflected in the distinct anatomical and physiological differences compared with skeletal muscle. Skeletal muscle fibers are classified as slow-twitch (type I) and fast-twitch (type II) based on their contractile and metabolic properties. , The classification of EOMs is determined by their contractile features, innervation and anatomic location. ,, Each rectus EOM contains approximately 20 000 fibers organized into highly specialized and compartmentalized groupings. ,, They are composed of 2 principal layers: an inner global layer, which inserts directly onto the globe and lies adjacent to the optic nerve, and an outer orbital layer, which interfaces with the bony orbit. Individual EOMs are further subdivided into superior and inferior compartments, particularly within the horizonal rectus muscles. ,,,,
The innervation of muscle fibers within each EOM is markedly denser than that of skeletal muscle. Many fibers are singly innervated and exhibit a high mitochondrial density, while multiply innervated fibers demonstrate at least a tenfold higher nerve-to-muscle fiber ratio compared with skeletal muscle. Contractile behavior is largely determined by myosin isoform expression, and EOMs express approximately 10 distinct isoforms, including the persistence of neonatal isoforms. In contrast, adult limb muscles express no more than 4 isoforms. ,, Collectively, these structural, metabolic and molecular specializations distinguish EOMs from other skeletal muscles which underlie their unique responses to injury and disease.
Structural alterations of the EOMs and their associated connective tissues have been extensively characterized using computed tomography (CT) and high-resolution magnetic resonance imaging (MRI) in numerous disease processes. This includes cranial neuropathies and dysinnervation syndromes, chronic progressive external ophthalmoplegia, myasthenia gravis, thyroid eye disease and infiltrative disorders among others. ,, These imaging modalities have demonstrated muscle atrophy in paretic conditions, facilitating understanding of disease-specific structural segmentation. ,, They have also demonstrated muscle enlargement in inflammatory and infiltrative disorders such as thyroid eye disease and IgG4 related disease. ,, High-resolution MRI has also illustrated changes in inter-muscular relationships and alignment, highlighting the critical role of connective tissue pathology in disorders such as sagging eye syndrome and heavy eye syndrome, with important implications for surgical management. ,,,
Although these structural and anatomic imaging techniques provide insight into in vivo EOM morphology, muscle function and metabolism can only be inferred from such changes. Thus, advanced MRI techniques—including diffusion weighted imaging (DWI), T1- and T2-weighted sequences—have been employed to assess EOM fat and water content in various disease states. These approaches have demonstrated increased T2 signal in muscles affected by myasthenia gravis, hyperintense T1-weighted signal in mitochondrial myopathies and combined DWI and T2 abnormalities in thyroid eye disease. ,, Despite the advances, current MRI techniques remain limited in their ability to directly characterize the metabolic and functional status of EOMs.
Positron emission tomography (PET) is a relatively noninvasive imaging modality with the capacity to detect inflammatory, degenerative and neoplastic processes. A principal advantage of PET lies in its ability to detect metabolic alterations that precede structural changes, enabling detection of subtle or subclinical disease. Its widespread use in oncologic imaging to identify early metastatic disease has demonstrated significant clinical benefit. 18 F-Fluorodeoxyglucose ( 18 F-FDG) is a glucose analog radiotracer that reflects tissue glucose utilization and serves as a surrogate marker of metabolic activity. PET imaging with 18 F-FDG provides both qualitative and quantitative assessments, including the standardized units of value (SUV), which facilitates objective evaluation of tissue metabolism of the radiotracer.
More recently, the application of PET to characterize EOM functional activity has begun to emerge. Small case series, including work by our group, have described qualitative hypometabolism in paretic EOMs and increased FDG uptake in hyperactive or inflammatory conditions such as Graves’ ophthalmopathy. ,,,,, These findings highlight the potential of 18 F-FDG PET as a surrogate marker of EOM function in both normal physiology and disease states. However, these reports remain limited in scope and have not systematically evaluated the normal variability of 18 F-FDG uptake in unaffected EOMs. Consequently, a normative reference database is lacking, limiting meaningful comparison across disease states and impeding broader clinical application.
The primary objective of this study is to establish a normative database of 18 F-FDG uptake for each rectus muscle across the human lifespan. Defining age-related patterns of metabolic activity in normal EOMs will provide a critical reference standard for future investigations of EOM function in neuromuscular, metabolic and neurologic diseases. In addition, this framework will enable characterization of physiologic changes in rectus muscle metabolism associated with normal aging.
As a secondary objective, this study leverages a retrospective cohort of patients with extraocular muscle paresis to compare 18 F-FDG uptake in affected muscles against the established normative database. Through this comparison, we aim to determine whether 18 F-FDG PET can reliably serve as a surrogate marker of EOM metabolic activity and functional impairment in paretic disease.
METHODS
Design and study subjects
This study was approved by the Institutional Review Board of the University of California San Francisco (UCSF IRB #21-35465). We performed a retrospective analysis of all patients who underwent 18 F-FDG PET/CT at UCSF between 2015 and 2025. A database search was performed to identify patients who had undergone 18 F-FDG PET/CT and received a comprehensive ophthalmic examination at UCSF within 3 months of the imaging study. Medical records of eligible patients were reviewed to determine cohort assignment.
Control cohort
For inclusion in the control cohort, patients were required to have no evidence of extraocular motility dysfunction on clinical examination. Exclusion criteria included a history of neurodegenerative disease, neurologic dysfunction, strabismus, prior strabismus surgery or any condition known to affect extraocular muscle function. Ophthalmic examination records were reviewed to confirm that a complete extraocular motility examination had been performed and documented as normal.
Extraocular muscle paresis cohort
For the extraocular muscle paresis cohort, we performed a retrospective review of patients with clinically documented extraocular muscle dysfunction who had undergone 18 F-FDG PET/CT within 2 weeks of either an ophthalmic or neurologic examination demonstrating extraocular muscle paresis. Clinical documentation was reviewed to confirm persistence of the paresis following the PET/CT scan. Patients with a prior history of strabismus or strabismus surgery were excluded from this cohort.
FDG-PET and PET/CT imaging protocol
All FDG-PET/CT examinations were performed on a Biograph 16 (Hi-Rez) PET/CT Scanner (Siemens Medical Solutions) with an integrated PET and 16-MDCT scanner or a Discovery VCT PET/CT scanner (Siemens Medical Solutions) with an integrated PET and 64-MDCT scanner, a Philips Vereos PET/CT system, or a Siemens Biograph Vision 600 PET/CT system. Standard clinical protocol included the following: All patients fasted with hydration for at least 6 hours prior to PET/CT examinations. Patients had blood glucose levels <200 mg/dL prior to intravenous injection of 18 F-FDG (7-11 mCi ± 10% for adults, 0.1 mCi/kg (0.7 mCi minimum/10 mCi max for pediatric patients)) followed by a 10-mL normal saline flush. Patients rested for 60 minutes ± 15 minutes and voided before being positioned supine on the scanner table. CT and PET images were fused after correction and in contiguous 5-mm slices for the entire body. PET was performed immediately following CT, without patient repositioning. PET images were obtained at 5 to 10 bed positions per patient from the skull vertex through the mid-thigh with an acquisition time of 3 to 4 minutes per station.
Data collection
Medical records were reviewed to determine age, sex and relevant medical and ophthalmic history. Detailed ophthalmic examination findings were reviewed with emphasis on extraocular motility including the presence or absence of extraocular muscle paresis. All PET/CT studies were reviewed using Visage 7 Client imaging software (Pro Medicus Limited). PET and corresponding CT images were displayed in orthogonal planes and regions of interest (ROI) were manually delineated to measure the maximum standardized uptake value (SUVmax) of each extraocular rectus muscle (superior, inferior, medial and lateral rectus). Superior and inferior oblique muscles were excluded due to limitations in resolution. Measurements were obtained in the coronal plane for all rectus muscles and additionally in the axial plane for the medial and lateral rectus muscles ( Figure 1 ).
Region of interest delineation on 18 F-FDG PET/CT images. There are 3 primary sequences that are used to determine extraocular muscle regions of interest and the corresponding SUVmax in the axial plane. A. Illustrates the noncontrast head CT used to visualize the orbital structures and EOMs. B. Is the 18 F-FDG PET scan through the corresponding region highlighting 18 F-FDG uptake in the medial and lateral recti in the axial plane. C. Is the merged PET/CT image with color overlay indicating areas of high 18 F-FDG uptake. A region of interest (circle) is created within the left medial rectus within which the SUVmax for the muscle will be obtained.
SUVmax represents the highest concentration of the 18 F-FDG radiotracer within the defined ROI and is calculated using the standard formula: SUVmax = (maximum radioactivity concentration in the ROI (Bq/mL))/ (injected dose (MBq)/patient weight (g)). In cases where adequate delineation of the EOM was not possible due to poor spatial resolution or adjacent cerebral 18 F-FDG avidity, ROIs were not created, and the affected muscles were excluded from analysis.
Data analysis
Patient age ranges and sex distribution were summarized. To determine the correlation between age and 18 F-FDG uptake in each rectus muscle, linear regression analysis was performed using SUVmax values from both the axial and coronal imaging planes. To evaluate the effect of extraocular muscle paresis on metabolic activity, 18 F-FDG uptake in paretic muscles were compared with uptake in corresponding muscles from the control cohort using the Wilcoxon Rank-Sum analysis. All statistical analysis and figure generation were performed using R statistical software package (R Foundation for Statistical Computing, Vienna, Austria). A P <.05 was considered statistically significant.
RESULTS
Control patient cohort and characteristics
111 patients who underwent 18 F-FDG PET/CT imaging and a comprehensive ophthalmic examination were identified for inclusion in the control cohort. There were 60 females and 51 males with a mean age of 58.4 years (SD 20.6, range 1-87 years). All patients demonstrated full extraocular motility and had no history of ocular motility dysfunction, neurologic disease, strabismus or prior strabismus surgery. The majority of patients underwent 18 F-FDG PET/CT imaging for their underlying malignancy (95/111). The remainder underwent imaging for malignancy screening (4/111), autoimmune conditions (6/111), seizure foci evaluation (3/111) and monitoring of nonmalignant lymphomatous conditions (3/111). A total of 857 extraocular muscles were included in the analysis (222 medial recti, 222 lateral recti, 222 inferior recti and 191 superior recti).
Age-related changes in 18 F-FDG uptake in normal extraocular muscles
Linear regression models were fit separately for each muscle, and imaging plane. In the control cohort, increasing age was associated with reduced 18 F-FDG uptake across all rectus muscles. 18 F-FDG SUVmax values declined progressively with age in every rectus muscle examined, with all associations being statistically significant in the coronal plane (all P <.005, Figure 2 ). The magnitude of age-related decline varied by muscle, with the steepest slopes observed in the medial and lateral rectus muscles. Specifically, the rate of decline in SUVmax in coronal sections ranged from −0.021 to −0.25 SUV/year. The medial rectus demonstrated the steepest decline (−0.0252 SUV/year), closely followed by the lateral rectus (−0.0250 SUV/year). The inferior rectus showed a decline of −0.0226 SUV/year while the superior rectus exhibited the smallest decline (−0.0215 SUV/year). Sex did not have a significant effect upon 18 F-FDG uptake in any rectus muscle.
Age- and sex- specific linear regressions for standard uptake value max for 18 F-FDG uptake in extraocular muscles of the control cohort in the coronal and axial planes. A. 18 F-FDG uptake in extraocular muscles of the control cohort (n = 111), coronal sections. B. 18 F-FDG uptake in extraocular muscles of the control cohort (n = 111), axial sections. Linear regression models for each extraocular muscle (lateral rectus [LR] n = 222, medial rectus [MR] n = 222, inferior rectus [IR] n = 222, superior rectus [SR] n = 191) demonstrate SUVmax values with aging. Sex: F = female, M = male. The SUVmax values declined with age for each muscle. Lateral rectus (coronal): −0.0250 SUV/yr ( P <.0002), Lateral rectus (axial): −0.0160SUV/yr ( P <.0228), Medial rectus (coronal): −0.252 SUV/yr ( P <.0005), Medial rectus (axial): −0.0265 SUV/yr ( P <.0003), Inferior rectus (coronal): −0.0229 SUV/yr ( P <.0054), Superior rectus (coronal): −0.0252 SUV/yr ( P <.0005).
Pairwise comparisons of 18 F-FDG SUVmax between rectus muscles in the control cohort demonstrated similar metabolic activity across all muscles. Adjustment for age and sex did not change the pattern of results. However, using Tukey-adjusted pairwise comparisons derived from a linear mixed-effects model revealed that the medial rectus demonstrated a significantly higher 18 F-FDG uptake than the lateral rectus and this effect remained significant after adjustment for age and sex (difference = 2.02 SUV, SE 0.52, P =.0006). Differences between the other muscle pairs were not statistically significant.
Effect of EOM paresis upon 18 F-FDG uptake
10 patients with clinically documented EOM paresis who underwent 18 F-FDG PET/CT during the period of their motility deficit were identified. Among these patients, 23 paretic rectus muscles were included for analysis (10 lateral rectus, 4 medial rectus, 5 superior rectus, and 4 inferior rectus). Using Wilcoxon Rank-Sum Analysis, the 18 F-FDG SUVmax values in paretic muscles were compared with the corresponding muscles from the control cohort. Significant reductions in 18 F-FDG uptake were observed across all paretic rectus muscles in the coronal plane: inferior rectus ( P <.0049) medial rectus ( P <.0018), superior rectus ( P <.0029) and lateral rectus ( P <.0005). In the axial plane, the medial rectus continued to demonstrate a significant reduction in FDG uptake ( P <.00372), but this was most pronounced in the paretic lateral rectus muscles ( P <.00001) ( Figure 3 ). There was no significant difference of 18 F-FDG uptake in the nonparetic muscles in the “paretic cohort” compared to the control cohort for all nonparetic recti muscles ( P >.05).
Effect of extraocular muscle paresis upon 18 F-FDG uptake. There is reduction in uptake of 18 F-FDG within paretic extraocular muscles ( P ) compared to the control cohort (Control). Significant reductions in 18 F-FDG SUVmax were observed in all paretic muscles in both the axial and coronal planes. Inferior rectus (IR) ( P <.0049), superior rectus (SR) ( P <.0029), lateral rectus (LR) coronal plane ( P <.0005), lateral rectus axial plane ( P <.0001), medial rectus (MR) coronal plane ( P <.0018), medial rectus axial plane ( P <.00372).
When all paretic muscles were analyzed collectively, 18 F-FDG SUVmax values were markedly and consistently reduced compared with control EOM ( Figure 4 ). This global effect of paresis remained highly significant in multilevel modeling, indicating that the metabolic consequence of neuromuscular paresis is strong and generalizable across all rectus muscles. The magnitude of the difference corresponded to a very large effect size, reflecting minimal overlap between the distributions of paretic and nonparetic muscles, supporting 18 F-FDG uptake as a strong biomarker of EOM dysfunction.
Collective analysis of 18 F-FDG uptake of all paretic muscles. There is global reduction of 18 F-FDG uptake in paretic extraocular muscles ( P ) compared to controls (Control) when all paretic muscles were analyzed collectively using Wilcoxon Rank-Sum Analysis ( P = 9.07−e-18).
Longitudinal changes in 18 F-FDG uptake with muscle recovery
One patient underwent sequential 18 F-FDG PET/CT scans during EOM paresis and recovery, performed as part of surveillance for metastatic parotid gland carcinoma with involvement of the right cavernous sinus. The initial PET/CT scan was performed within 2 months of symptom onset and demonstrated severe right lateral rectus dysfunction (clinical examination demonstrated −3.5 limitation). Two follow-up PET/CT scans were performed 7 months and 2 years following the initial study, during which time progressive clinical improvement of the right lateral rectus was documented ( Figure 5 ). With clinical improvement of right lateral rectus function, there was a corresponding increase in 18 F-FDG uptake in the right lateral rectus over time. This longitudinal observation provides preliminary evidence that changes in 18 F-FDG uptake may reflect dynamic alterations in EOM functional status.
