Rhino-orbital-cerebral mucormycosis is caused by the infection of fungi belonging to the order Mucorales, which are usually present in soil and vegetative matter. Inhalation of spores of these fungi followed by colonization of nasal mucosa is the primary route of infection. The aggressive extra-sinus spread through direct extension, bony destruction, or perineural/perivascular routes leads to high mortality in infected patients.
Immunocompromised state, hematological malignancies, uncontrolled diabetes mellitus, and steroid therapy predispose individuals for mucormycosis [8]. In our study, the majority of the patients were either diabetic or developed uncontrolled blood glucose levels during COVID-19 infection. COVID-19 infection itself causes impairment of cell-mediated immunity through immunomodulation which makes patients susceptible to fungal co-infections. History of steroid administration, oxygen therapy, and prolonged hospital stay was common in our studied population. Thus, both COVID-19 illness and its treatment make patients vulnerable to opportunistic fungal infections.
Patients with ROCM may present with nonspecific symptoms of sinusitis such as headache, facial pain, fever, nasal discharge, or congestion. Orbital symptoms such as swelling, proptosis, visual disturbances, and ophthalmoplegia or neurological symptoms such as confusion and altered sensorium should raise the suspicion of invasive fungal sinusitis, especially in the presence of predisposing factors. In the era of COVID-19 infection, all patients with sinusitis should be properly evaluated clinically, radiologically, and if required, histopathologically to rule out ROCM.
On imaging, there is variation in MR signals on T1- and T2-weighted imaging. However, the presence of low signals on T2-weighted imaging characteristically represents iron- and manganese-containing fungal elements. In contrast studies, ROCM shows multiple enhancement patterns. In the early stages, intense homogenous enhancement of mucosa is seen, which is similar to bacterial sinusitis. Presence of heterogenous enhancement or non-enhancing areas within the lesion points towards invasive fungal etiology. The disease is known to cause devitalization of sinonasal mucosa by vascular invasion and infarction which corresponds to non-enhancing tissues on contrast administration. The “black turbinate sign,” which is an early sign of ROCM, simply represents the non-enhancing turbinate on contrast MR imaging [9, 10]. The non-enhancement should persist on delayed images unlike the benign black turbinate, which shows a gradual increase in signal.
Periantral fat invasion which may appear as stranding of premaxillary/retroantral fat, increased signals on fat-saturated T2-weighted imaging, or enhancement on contrast sequences is also an early feature of ROCM. Fat stranding is also well visualized on CT images. In our study, 75.83% of the patients showed involvement of periantral fat. A subset of patients with periantral extension had no bony erosion suggesting spread through perivascular routes. Bony erosion or marrow edema was present in 60% of patients. Erosions were commonly present in walls of maxillary antrum and orbital walls, allowing spread in retromaxillary soft tissue and orbit, respectively. The erosion or rarefaction of bony walls was better visualized on CT images.
Extension of infection in masticator space and infratemporal fossa was common in our study. Muscles of mastication appeared bulky with heterogeneous enhancement in involved cases. Pterygopalatine fossa was involved in about one-third of the studied cases. It is a neurovascular crossroad that serves as a route to infection into the middle cranial fossa.
Extension of infection into the orbit can be seen in the form of stranding of intraconal or extraconal fat, the presence of abnormal T2 hypointense soft tissue with variable enhancement, and/or bulky heterogeneous or non-enhancing intraocular muscles. There may not be associated erosion of orbital walls as infection can spread via nasolacrimal duct and vascular channels. Direct spread of infection through ethmoid and maxillary sinuses causes lateral displacement of the medial rectus and upward displacement of the inferior rectus, respectively. In our study, 24 patients out of 84 with orbital involvement had no specific ocular complaints. Therefore, it is advisable to include imaging of the orbit along with sinuses, in all suspected cases of ROCM.
The periantral soft tissue and orbits were the most common sites of disease extension in our study. The study conducted by Therakathu et al. [11] also showed orbits as the most common site of extra-sinus spread seen in 76% of patients, followed by facial soft tissue (57%). Metwally et al. [7] also found periantral soft tissue involvement in 74.6% of patients. However, pterygopalatine fossa invasion was more common in their study (77.8%). Yadav et al. [12] in their study on 50 patients with mucormycosis found periantral soft tissue and orbital involvement in 74% and 76% of patients, respectively. The close proximity to the commonly infected maxillary sinus, direct invasion through bony erosion, and tendency to spread along perineural/perivascular routes possibly explain the frequent involvement of periantral fat. The orbits are separated from sinonasal mucosa through thin lamina papyracea and communicate to the nasal cavity via the nasolacrimal duct. Thus, orbits are also anatomically prone to invasive fungal infection.
The presence of enhancing soft tissue at the orbital apex may cause involvement of optic nerve with a risk of extension of infection into the cavernous sinus. Optic nerve involvement leads to irreversible vision loss and is an ominous sign. The spread of infection in the cavernous sinus may lead to thrombophlebitis or thrombosis of the sinus. Various divisions of trigeminal nerves and other cranial nerves traversing the cavernous sinus may get involved leading to extension into the posterior cranial fossa [12]. The walls of the cavernous portion of the internal carotid artery show thickening and enhancement if involved, with or without the presence of a thrombus in the vessel lumen.
In our study, meninges were the most commonly involved intracranial structure, which appeared thickened with post-contrast enhancement. The involvement of meninges is the earliest sign of the intracranial spread of infection. Although intracranial spread through perineural or perivascular routes may not always show meningeal enhancement. Invasion of the brain can lead to the formation of fungal abscess which is common in the temporal lobe. It has a characteristic low signal on both T1- and T2-weighted sequences. The walls of these lesions show enhancement on contrast images and restricted diffusion on diffusion-weighted imaging (DWI). DWI is also useful in the identification of cerebral infarcts which occur due to the angioinvasion property of mucor. The pathology of mucormycosis is characterized by the proliferation of angioinvasion hyphae with the elastic lamina of large or medium-sized vessels [13], which necessitates the use of angiographic studies in suspected cases for meticulous assessment of vasculature.
On classifying the cases on the basis of MRI findings in our study, half of the patients were found in orbital stage (stage III disease), while only 7 patients had stage II disease. The majority of the patients with orbital invasion also had soft tissue lesions in periantral region and various neck spaces, suggesting there occurs rapid invasion of orbits post-maxillofacial soft tissue infection. Simultaneous invasion of orbits and periantral region is another possibility. If the latter is considered, then stage II and stage III disease can be clubbed together in a three-stage classification as proposed by Therakathu et al. and Yadav et al. However, despite a multitude of recent literature on rhino-orbital-cerebral mucormycosis, no consensus on imaging-based classification has been reached.
Nevertheless, imaging is indispensable for management in patients with rhino-orbital-cerebral mucormycosis (ROCM). Imaging is requisitioned not only for early diagnosis but for a variety of reasons that include pre-surgical mapping to detect early orbital involvement and demonstrate possible extension to the skull base, anterior, and middle cranial fossa, which may be clinically silent, and for guided biopsy to ensure maximum diagnostic yield.
Rapid and progressive intracranial spread of the mucor infection occurs either by direct extension across the neural foramina, cribriform plate/ethmoid, walls of the frontal and sphenoid sinuses, or angioinvasion of the walls of the arteries and veins, causing vascular thrombosis, occlusion, and infarction [6, 14, 15]. Therefore, the radiological diagnosis of complications is equally important to determine the type and intensity of treatment required. The modality of choice is contrast-enhanced magnetic resonance imaging in view of its superior contrast resolution for soft-tissue and marrow abnormalities. The ability to depict cross-sectional anatomy and pathology with better tissue characterization is a distinct advantage of MRI over CT scan. Contrast administration is useful to identify non-viable necrotic areas. CT scan is complementary to MRI to better demonstrate bone erosion and demineralization.
The present study is one of the largest studies conducted on imaging of COVID-19-associated ROCM. However, there are certain limitations to our study. First, we did not correlate the imaging findings with the clinical outcomes. Second, a comparison of the findings from our study with those found in ROCM not related to COVID-19 illness was not done.
