Myths in the Diagnosis and Management of Orbital Tumors.

Orbital tumors constitute a group of diverse lesions with a low incidence in the population. Tumors affecting the eye and ocular adnexa may also secondarily invade the orbit. Lack of accumulation of a sufficient number of cases with a specific diagnosis at various orbital centers, the paucity of prospective randomized studies, animal model studies, tissue bank, and genetic studies led to the development of various myths regarding the diagnosis and treatment of orbital lesions in the past. These myths continue to influence the diagnosis and treatment of orbital lesions by orbital specialists. This manuscript discusses some of the more common myths through case summaries and a review of the literature. Detailed genotypic analysis and genetic classification will provide further insight into the pathogenesis of many orbital diseases in the future. This will enable targeted treatments even for diseases with the same histopathologic diagnosis. Phenotypic variability within the same disease will be addressed using targeted treatments.

INTRODUCTION

Orbital tumors constitute a group of diverse lesions with a low incidence in the population. Tumors affecting the eye and ocular adnexa may also secondarily invade the orbit. The low number of patients with a specific diagnosis at many centers limits prospective randomized clinical trials. Therefore, all the available recommendations on diagnosis and management are from small prospective case series or retrospective studies. The lack of animal models and tissue bank material limit the extent of scientific studies to be conducted regarding the pathogenesis and treatment of orbital lesions. Genotypic analysis of orbital tumor cases is rare. However, recently more data has been gathered. Due to all these factors, various myths regarding the diagnosis and management of orbital tumors have developed over time and continue to influence the decision making process of orbital specialists.

The use of orbital imaging studies, including computed tomography (CT) and magnetic resonance imaging (MRI), allows localization of the mass in the orbit. This localization allows the correct surgical approach to biopsy the lesion. As a result, anterior orbitotomy techniques using a transconjunctival or skin approach are used more often now than the traditional lateral orbitotomy (Kronlein, Berke) that required a larger incision and osteotomy. MRI allows evaluation of the internal tissue features better through T1- and T2-weighted (T1W and T2W) imaging sequences and patterns of contrast enhancement.

Seven common orbital myths are listed below. These myths are discussed using case summaries and literature data.

MYTH #1: “CAN AN ORBITAL LESION BE DIAGNOSED USING IMAGING STUDIES WITHOUT THE NEED FOR AN ORBITAL BIOPSY?”

Although MRI and CT can provide clues about the presumptive diagnosis in some cases, it is best to do an orbital biopsy, either excisional or incisional, if the lesion is in a surgically accessible location. However, if the lesion is in a location that is difficult to reach surgically and morbidity is expected, then a clinical diagnosis based on imaging findings can be cautiously pursued.

Case report

A 43-year-old female presented with pain, proptosis, and decreased vision in the right eye. Orbital MRI demonstrated an ill-defined mass located in the right orbital apex isointense to the extraocular muscles on T1W images [Figure 1a], hyopintense on T2W images [Figure 1b] with moderate contrast enhancement [Figure 1c]. At this point, the MRI findings were consistent with either an inflammatory mass or a nonepithelial tumor such as lymphoma.1 Although some studies suggested that diffusion weighted images can be used to differentiate between these two disease conditions, there is still a considerable overlap between them.

Figure 1

A 42-year-old female presenting with painful proptosis. (a) T1-weighted axial magnetic resonance imaging shows ill-defined orbital lesion occupying the apex. The lesion is isointense to the cerebral gray matter and extraocular muscle. (b) The apical lesion shows marked contrast enhancement after gadolinium injection. (c) T2-weighted axial magnetic resonance imaging demonstrates that lesion is hypointense to the cerebral gray matter and extraocular muscle

The tumor had a location that was difficult to reach surgically with the potential of a number of surgical complications. Therefore, based on the MRI findings, a presumptive diagnosis of idiopathic orbital inflammation was made and the patient was treated with a tapering dose of 1 mg/kg/day oral prednisolone. The symptoms resided after corticosteroid treatment, and the patient continued to do well at 2 years follow-up without any recurrence. This case illustrates an example where the diagnosis was established based on the imaging studies without histopathologic evidence. If the patient had not responded to oral corticosteroid treatment or she had recurrent symptoms and signs after discontinuation of the drug, an incisional biopsy via orbitotomy surgery would have been necessary to establish the histopathologic diagnosis.

MYTH #2: “ORBITAL LESIONS WITH THE SAME HISTOPATHOLOGIC DIAGNOSIS REPRESENT THE SAME DISEASE AND RESPOND TO TREATMENT IN THE SAME WAY”

In some diseases such as idiopathic orbital inflammation (IOI), there is phenotypic variability from case to case with the same histopathologic diagnosis. It appears that IOI is a heterogeneous collection of diseases. Variables such as age, location, and histopathologic findings all affect the response to treatment.2 Recently, gene expression profiling methods have been used to demonstrate this heterogeneity among cases labeled as IOI. Genomic upregulation involving immunoglobulin, CXCR4, YKL-40, CXCL9, SLAM Family 7, and IL-7 receptor have been observed in some patients with IOI. Alternately, genomic downregulation involving alcohol dehydrogenase 1B, perilipin 1, adiponectin, leptin receptor, and C1Q have also been documented in some patients with IOI by gene expression array methods.2 Gene expression findings may help in prognostic and therapeutic decision making for patients with IOI. While some cases diagnosed histopathologically as IOI respond to a single course of tapering corticosteroid therapy, others demonstrate recurrence upon cessation of corticosteroid treatment and require other treatments such as external radiotherapy or surgical excision. Recently, a subgroup of IOI cases has been linked to IgG4 disease.3 Genotypic analysis will definitely shed more light on the phenotypic variability observed among IOI cases. Consequently, targeted subset treatment of IOI cases may become possible in the future.

MYTH #3: “WELL-CIRCUMSCRIBED ORBITAL TUMORS ARE COMMONLY CAVERNOUS HEMANGIOMAS. IS IT APPROPRIATE TO OBSERVE WELL-CIRCUMSCRIBED ORBITAL LESIONS RADIOLOGICALLY DIAGNOSED AS CAVERNOUS HEMANGIOMAS UNTIL THEY ENCROACH ON THE OPTIC NERVE AND CAUSE VISION LOSS?”

Cavernous hemangioma is the most frequent benign orbital tumor in adults. Cavernous hemangiomas are usually well-circumscribed lesions. Differential diagnosis of well-circumscribed orbital lesions includes schwannoma, solitary fibrous tumor, and some metastatic tumors in addition to cavernous hemangioma. However, cavernous hemangioma is by far the most common well-circumscribed orbital lesion. Figure 2a and b illustrate two cases of presumed orbital cavernous hemangioma on MRI. The case presented in Figure 2a had compressive optic neuropathy and vision decrease, and the orbital tumor was discovered with MRI. The patient presented in Figure 2b had a headache and the tumor was discovered once MRI was performed to investigate the cause of the headache.

Figure 2

(a) T1-weighted axial magnetic resonance imaging showing a well-circumscribed orbital lesion that compresses the optic nerve on the left side. (b) T1-weighted coronal magnetic resonance imaging is showing a well-circumscribed orbital lesion that does not encroach on the optic nerve on the right side. (c) Gross photograph of the tumor depicted in Figure 2a which proved to be a cavernous hemangioma on histopathologic examination after excision. (d) Gross photograph of the tumor depicted in Figure 2b which was also read as cavernous hemangioma

The treatment of cavernous hemangioma is surgical excision of the tumor.45 Sometimes an asymptomatic lesion located in the posterior orbit is difficult to access surgically, and the patient can be observed conservatively for signs of the globe, optic nerve compression, and visual loss.56 External beam radiation treatment has also been advocated in the treatment of orbital cavernous hemangiomas located in the orbital apex that cannot be resected.7 The tumor depicted in Figure 2a had a well-circumscribed orbital mass that was thought to be a cavernous hemangioma. Since the tumor produced optic nerve compression and vision loss, a decision was made to proceed with immediate surgery. An anterior orbitotomy was performed and the tumor was removed in toto with a cryoprobe without complications [Figure 2c].

The case depicted in Figure 2b also had an intraconal, well-circumscribed orbital mass. However, in this case, there was no optic nerve compression, vision loss, or disfiguring proptosis. Although immediate surgery is probably unnecessary in this case, the tumor may demonstrate growth in the future and lead to visual deterioration. Therefore, it is probably better to remove this tumor rather than wait for optic nerve complications to develop at which time surgery may be more difficult to perform. Additionally, permanent vision loss may result without timely treatment. We performed an anterior orbitotomy via a superolateral skin crease incision and removed the tumor in toto with a cryoprobe [Figure 2d]. Histopathologic examination findings were consistent with a cavernous hemangioma. For this case, some may recommend observation until tumor growth has been documented.5 We believe that it is best to remove these lesions before they encroach on the optic nerve and globe provided that they are in a surgically accessible location and the risk of complications is low.

MYTH #4: “SOME ORBITAL LESIONS PRODUCE CHARACTERISTIC FINDINGS ON IMAGING. IS IT CORRECT TO REACH DIAGNOSTIC DECISIONS BASED ON THESE IMAGING FINDINGS WITHOUT MAINTAINING A HIGH INDEX OF SUSPICION AND SKEPTICISM?”

A 27-year-old male presented with a 5 years history of left proptosis, loss of visual acuity, and diplopia [Figure 3a]. The patient had been followed elsewhere with a presumptive diagnosis of orbital lymphangioma. Surgery was not performed for years due to concerns of intraoperative hemorrhage for years. Orbital MRI findings revealed a cavitary mass with fluid-fluid levels occupying almost the entire orbit [Figure 3b and c]. A detailed search demonstrated that tumors producing fluid-fluid levels include lymphangioma, hemangioma, aneurysmal bone cyst, chondroblastoma, osteoblastoma, intracranial schwannoma, peripheral nerve schwannoma, and extracranial neck and head schwannomas.8910111213 Nontumoral causes include fibrous dysplasia, tumoral calcinosis, and myositis ossificans.813 We thought that the lesion could be a lymphangioma as suspected, but it could also represent one of the rarer histopathologic entities listed above. We proceeded with an anterior orbitotomy and performed a piecemeal subtotal tumor excision [Figure 3d]. The lesion was yellow-white in color, had intrinsic vascularity, was unencapsulated and with a fragile structure during surgical dissection. Histopathology findings and immunohistopathology were consistent with schwannoma [Figure 3e and f]. This case illustrates that a high index of suspicion is required for imaging findings thought to represent the classic appearance of certain tumors.

Figure 3

A 31-year-old male with a 5 years history of proptosis in the left eye. (a) Facial photograph shows left proptosis. (b) T1-weighted magnetic resonance imaging demonstrates a cavitary mass in the left orbit. (c) T2-weighted magnetic resonance imaging shows fluid-fluid levels in the tumor. (d) Gross photo of the excised tumor shows yellowish hemorrhagic tumor fragments. (e) Histopathologic examination demonstrates schwannoma consisting mostly of spindle shaped tumor cells with elongated nuclei forming bundles and cellular areas without obvious mitosis (Antoni A pattern) (H and E, ×400). (f) Immunohistochemically, the tumor cells diffusely stain positive with S-100 (S-100, ×100)

MYTH #5: “LOCALLY ADVANCED EYELID AND PERIOCULAR CARCINOMAS WITH ORBITAL INVASION CAN ONLY BE TREATED USING RADICAL DISFIGURING SURGERY, HIGH DOSE RADIATION THERAPY WITH ITS OCULAR TOXICITY, AND CYTOTOXIC CHEMOTHERAPY WITH ITS SYSTEMIC COMPLICATIONS AND MORBIDITY”

Recently drugs that target sonic hedgehog and epidermal growth factor receptor (EGFR) pathways have been used in the treatment of unresectable (locally advanced or metastatic) eyelid carcinomas (basal and squamous cell, respectively) not amenable to surgical excision. Locally advanced eyelid and periocular carcinomas have traditionally been treated using wide surgical excision, exenteration, external beam radiotherapy, and systemic (intravenous) chemotherapy [Figure 4a-d].

Figure 4

A 65-year-old female with locally advanced squamous cell carcinoma of the eyelid with orbit and bone invasion. (a) Facial photograph shows ill-defined eyelid squamous cell carcinoma affecting the lateral canthal region. (b) Anterior segment photograph of the left eye demonstrates vascularized corneal opacification and conjunctival keratinization as secondary effects of locally advanced eyelid cancer. (c) Orbital computed tomography shows orbital and bone invasion from eyelid tumor. (d) Facial photograph 6 months after orbital exenteration, bone removal, external radiotherapy, and intravenous chemotherapy (cisplatin and doxorubicin)

Basal cell carcinomas (BCC) are known to be associated with abnormal Hedgehog pathway signaling that results in an uncontrolled proliferation of basal cells. This pathway is normally inhibited by patched homologue 1 (PTCH 1).14151617 However, when PTCH 1 is deficient, a transmembrane protein called smoothened (SMO) activates the Hedgehog pathway. Vismodegib selectively binds to the extracellular domain of SMO, thereby prevents Hedgehog signaling.14151617 Therefore, vismodegib is a treatment option for locally advanced BCC in the periocular region with extensive eyelid and orbital involvement. Patients who cannot undergo extensive surgery possibly exenteration because of medical reasons and comorbidities or patients who do not wish to undergo this type of disfiguring surgery are potential candidates for vismodegib. Vismodegib is prescribed orally and may be associated with serious side effects including muscle cramps, alopecia, and development of new squamous cell carcinomas.1417

Squamous cell carcinomas (SCC) have been shown to demonstrate overexpression of EGFR.1718 Patients with unresectable (locally advanced or metastatic) SSC of the skin were treated with cetuximab with a tumor response rate of 69%.18 Similarly, a study of 424 head and neck SCCs randomized to high dose radiation with or without cetuximab has shown superior survival with the addition of cetuximab.17 Advanced eyelid SCC with the orbital extension has been treated with intravenous cetuximab and oral erlotinib with good clinical and radiological response. Side effects of EGFR inhibitors include skin toxicity in the form of acne-like rash, papular and pustular eruptions, mucositis, esophagitis, and conjunctivitis.1718

Based on these experiences, patients with locally advanced or metastatic BCC and SCC with advanced age and comorbidities who are not amenable for surgery are potential candidates for targeted therapies. High cost, long-term treatment, and side effects are relative limitations of targeted therapies. Long-term results are currently unknown.

MYTH #6: “OPTIC NERVE GLIOMAS ESPECIALLY THOSE ASSOCIATED WITH NEUROFIBROMATOSIS TYPE 1 ARE GENERALLY ACCEPTED TO BE STABLE LESIONS WITH A LOW MITOTIC INDEX. HOWEVER, SOME OF THESE CASES CAN DEMONSTRATE AGGRESSIVE BEHAVIOR, GROW TO INVOLVE THE CHIASM AND THE OTHER EYE, AND INTRACRANIAL STRUCTURES. SURGERY/EXTERNAL BEAM RADIATION TREATMENT/CHEMOTHERAPY ARE STILL THE MAINSTAYS OF TREATMENT UNDER THESE CIRCUMSTANCES”

Optic nerve gliomas are generally accepted to be low-grade pilocytic astrocytomas. These tumors occur in children younger than 10 years of age. Optic nerve gliomas demonstrate an association with neurofibromatosis type I (NF1). As outlined above, BRAF duplicate mutations have been found in some patients with optic nerve gliomas and NF1 demonstrating tumor growth and local invasion on serial examinations.19 In another study, BRAF duplicate mutations were found in almost all patients with optic nerve/chiasmal gliomas.20 Therefore, inhibitors of BRAF and other pathway components such as sorafenib have been tested in clinical trials as a treatment for optic gliomas. However, sorafenib produced an unexpected and unprecedented acceleration of tumor growth in children with pilocytic astrocytoma, irrespective of NF1 or tumor BRAF status.21 In vitro studies with sorafenib indicate that this effect is likely related to paradoxical ERK activation.21 Therefore, novel agents that modulate signal transduction and are thought to inhibit tumor growth may paradoxically induce tumor progression. In this case of optic gliomas, the myth still holds true that radiotherapy and chemotherapy are the mainstays of treatment.2122

MYTH #7: “ADENOID CYSTIC CARCINOMA OF THE LACRIMAL GLAND IS A MALIGNANT NEOPLASM, AND LONG-TERM SURVIVAL IS POOR. RECENTLY, AGGRESSIVE TREATMENT EMPLOYING NEOADJUVANT INTRAARTERIAL CHEMOTHERAPY FOR SHRINKING THE TUMOR FOLLOWED BY TUMOR EXCISION AND POSSIBLY POSTOPERATIVE EXTERNAL BEAM RADIATION TREATMENT HAS SHOWN INCREASED SURVIVAL COMPARED TO TRADITIONALLY EMPLOYED TREATMENTS INCLUDING EXENTERATION AND EXTERNAL BEAM RADIATION TREATMENT. STILL MANY PATIENTS WITH THIS DIAGNOSIS DO POORLY IN THE LONG-TERM”

Recent evidence suggests that treatment with sorafenib, a tyrosine kinase inhibitor, may be beneficial in metastatic adenoid cystic carcinoma of the salivary glands unresponsive to conventional chemotherapy.23 In another report, inhibition of KIT tyrosine kinase by imatinib mesylate achieved a significant reduction of tumor burden making the adenoid cystic carcinoma amenable to surgical resection.24 Lacrimal gland adenoid cystic carcinomas are frequently positive for MYB, and this finding suggests that MYB and its downstream targets are potential therapeutic targets for these tumors. Therefore, targeted treatments for MYB and its downstream products holds promise for ACC of the lacrimal gland.2526

Aggressive locally recurrent and metastatic adenoid cystic carcinomas may express vascular endothelial growth factor (VEGF) and c-kit. In a phase II clinical trial of sunitinib, a novel agent which acts as a multi-targeted inhibitor of VEGF receptor, c-kit, FMS-like tyrosine kinase 3 in ACC of the salivary gland, no responses were observed but sunitinib was well tolerated and prolonged tumor stabilization of ≥6 months was achieved in 62% of assessable patients.27 This experience indicates the challenges in evaluating molecular targeting agents.

A Phase II clinical trial on dovitinib, an inhibitor of receptor tyrosine kinase, in metastatic salivary gland adenoid cystic carcinoma is already underway and initial data on tumor regression in encouraging.

In summary, our understanding about orbital diseases is expanding based on the genomic analysis and targeted therapies. Collaborative research bringing together a larger number of patients will improve therapeutic approaches to this diverse group of diseases.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.