Beyond the Globe: Molecular Diagnostics of Optic Nerve, Orbital, and Adnexal Neoplasms

Abstract

Recent advances in molecular pathology have transformed the diagnostic landscape and management of human cancer. Increasingly, integration of genomic and epigenomic data with conventional histopathology has improved tumor classification, refined prognostic assessment, and revealed previously unsuspected therapeutic targets. High-throughput techniques such as next-generation sequencing, gene fusion panels, and methylation arrays have expanded applicability to formalin-fixed tissue and enabled simultaneous evaluation of multiple cancer-defining alterations/genetic drivers. In optic nerve gliomas, MAPK pathway activation through BRAF or FGFR1 alterations, or NF1 inactivation, is a basic biological feature with diagnostic implications, while MEK inhibitors may be of clinical benefit in selected patients. Optic nerve sheath/orbital meningiomas demonstrate divergent molecular landscapes depending on location, with NF2-driven and non–NF2-driven pathways informing recurrence risk and behavior. In the orbit, characteristic genetic drivers facilitate diagnosis of mesenchymal tumors such as solitary fibrous tumor (NAB2::STAT6) and alveolar rhabdomyosarcoma (PAX3/7::FOXO1), while molecular profiling assists in distinguishing challenging peripheral nerve and melanocytic lesions. Similarly, lacrimal gland neoplasms parallel salivary gland counterparts, with recurrent fusions such as PLAG1, HMGA2, and MYB::NFIB of great diagnostic utility. These advances underscore the growing role of molecular diagnostics in improving accuracy, guiding prognostication, and refining the classification of rare ocular tumors. As high-throughput techniques continue to mature, integration with evolving spatial and single-cell–based approaches promises to expand our understanding and further personalize diagnostic and therapeutic strategies.

Introduction

Numerous advances in the diagnosis and classification of tumors of the eye and ocular adnexa have occurred, as with cancers at many other sites. A fifth edition of Eye and Orbit Tumours has been published¹ and, in the spirit of other World Health Organization (WHO) Classifications, started incorporating genetic information into diagnostic approaches.^2,3 Genetic drivers of many primary tumors have been increasingly clarified. These advances have been the consequence, in part, of The Cancer Genome Atlas and numerous high-throughput molecular studies undertaken in every major category of tumors. A variety of platforms and techniques with increasing applicability to formalin-fixed, paraffin-embedded material, including methylation-based arrays, next-generation sequencing (NGS), and gene expression profiling, have become important tools for integrating molecular data into traditional diagnostic approaches for a variety of tumor types. Methylation arrays, in particular, have found increased applications in the diagnostic approach to brain tumors, soft tissue, head and neck, and other tumors and have facilitated the identification and characterization of a variety of novel tumor types and subtypes.⁴ In this review, we focus on tumors involving the optic nerve, orbit (mesenchymal), and adnexa (predominantly lacrimal gland), where advances of genetics translating into molecular diagnostics have increasing practical applications (Fig. 1). Some tumors at ocular sites, in particular, have just recently begun to be characterized, and we may know more at the moment regarding corresponding tumor types that occur at other sites at a higher frequency (e.g., tumors of the optic nerve and brain, lacrimal gland tumors through salivary gland counterparts). Comprehensive coverage of tumors of the eye is covered in recent texts, including the WHO Classification.^1,5

Figure 1.

Special techniques used in detecting orbital tumors. A variety of high-throughput assays have found increased application to tumors at a variety of sites, including the eye and surrounding tissues. The integration of molecular assay results is used in conjunction with clinical and histopathologic assessment to refine and guide management of optic nerve, lacrimal, and orbital tumors.

Special Techniques

Conventionally, molecular pathology laboratories relied on targeted assays such as polymerase chain reaction and fluorescence in situ hybridization, which are useful in the detection of single-gene alterations, including mutations in putative oncogenes and tumor suppressors, copy number changes, and oncogenic fusion events. However, their performance is usually restricted by predefined genomic targets, which potentially miss unexpected and novel events, and the results do not provide a genome-wide context.

High-throughput techniques have found increasing applications at the present time, and the main advantage is that they allow for the simultaneous testing of multiple genes. This includes NGS (which is suitable to detect single-nucleotide variants, small indels, and copy number alterations) and RNA-based NGS, which is optimized to detect oncogenic gene fusions and outlier expression. In theory, these techniques can interrogate every encoding gene in a sample, but smaller panels are routinely designed to test for cancer-relevant genes in a practical manner.

Methylation profiling array is another technique that has proven to be of great value in the classification of tumors of the nervous system and other body sites.⁴ It is typically successful when tumor content and DNA quality are adequate but requires more tumor purity than NGS-based assays. It also is helpful in identifying clinically relevant molecular subtypes with variable biology and prognosis. The two main tumors that involve the optic nerve, pilocytic astrocytoma and meningioma, are readily classifiable by this technique. However, their accuracy declines with low tumor cellularity or degraded DNA, and in some cases, borderline classifier scores may require histologic correlation and orthogonal confirmation. This prompts the exploration of higher-resolution modalities.

Newer techniques, including single-cell sequencing and spatial transcriptomics, offer high-resolution insights into the tumor microenvironment and genotype–phenotype relationships but are not yet standardized for routine clinical reporting. They typically require specialized analytics and can be costly compared with other molecular assays. As of now, they have enormous potential for robust research applications, which will continue to inform molecular diagnostics.

Tumors of the Optic Nerve

Optic nerve gliomas are slow-growing neoplasms with molecular genetic alterations different from the more common adult diffuse glioma counterparts. Most of these tumors represent, at the pathologic level, pilocytic astrocytomas (PAs). The molecular pathogenesis of PAs is driven by singular genetic alterations that lead to MAPK pathway activation (Fig. 2). Commonly affected genes within this pathway include BRAF, FGFR1, and members of the NTRK family.⁶ The most frequent genetic event in sporadic PAs (i.e., unassociated with a genetic syndrome) is a focal tandem duplication of chromosome 7q34, which results in a fusion of the N-terminal region of KIAA1549 with the kinase domain of BRAF. This fusion drives constitutive MAPK pathway activation and is particularly prevalent in cerebellar PAs.⁷ Although it is still common, it is observed less frequently in tumors arising in the optic pathway and hypothalamus, which together account for approximately 9% of PA cases.⁶ In one study including 59 biopsy specimens of optic pathway gliomas (OPGs), 73% harbored BRAF duplications, whereas generally they constitute 90% of cerebellar cases.^6,8

Figure 2.

Signaling pathways relevant to optic nerve tumors. NF2 loss of function (Merlin deficiency) promotes meningioma growth by deregulating adhesion and Hippo signaling, which leads to FAK/Src/AKT/mTORC1 and RAC1/PAK1 activity, reduced LATS1/2 restraint, and nuclear activation of YAP/TAZ, together enhancing proliferation and the survival of meningioma cells, as well as facilitating the invasion into surrounding orbital tissues. NF1 loss of function causes constitutive MAPK signaling, underlying optic nerve glioma (pilocytic astrocytoma). Mutually exclusive MAPK drivers commonly observed in these tumors are also listed, such as KIAA1549::BRAF fusion, BRAF p.V600E, FGFR alterations, and NTRK fusions.

Although BRAF can fuse with various gene partners, the resulting proteins consistently lack the N-terminal regulatory domain, leading to enhanced kinase activity.⁶ Other alterations include single amino acid changes at position 600 of BRAF from valine to glutamic acid, which may be of clinical importance given the available data and experience as a therapeutic target in other tumor types. Of interest, BRAF p.V600E mutant pediatric diencephalic gliomas have been found to be more aggressive than BRAF-rearranged ones in at least one retrospective study.⁹ Other rare MAPK-activating fusions, such as SRGAP3:RAF1, have also been reported and function through similar mechanisms.⁶ PTEN and CDKN2A deletions, markers of more aggressive gliomas, are very rare in the optic pathways.¹⁰ BRAF p.V600E point mutations and NTRK fusions are more frequently identified in supratentorial tumors, while FGFR1 alterations are commonly restricted to midline structures and may be detected in rare optic nerve examples.⁶

The NF1 tumor suppressor gene regulates the RAS/MAPK signaling pathway, and loss of function of this gene in the germline results in neurofibromatosis type 1 (NF1), which leads to 15% to 20% of children with NF1 to develop OPGs.¹¹ In NF1-associated tumors, there is a biallelic NF1 gene inactivation mechanism.^12,13 Interestingly, several studies of PAs have demonstrated that NF1 clinical status and other MAPK pathway alterations (e.g., BRAF) are mutually exclusive in most cases.^8,10,14

Recent studies have demonstrated that MEK inhibitors, which target the hyperactivated RAS/MAPK pathway downstream of NF1 loss, can effectively reduce optic glioma volume and astrocyte proliferation in preclinical models. Clinically, MEK inhibition has also shown promise, with 26% of patients experiencing improved visual fields and the majority demonstrating disease stability.¹⁵

Meningiomas of the Optic Nerve Sheath

Meningiomas are the most frequent primary intracranial tumors, yet the molecular pathology of orbital and optic nerve meningiomas is not well studied. When considering meningiomas overall, integrating molecular markers such as copy number variations and methylation patterns is increasingly of value in the prediction of meningioma recurrence compared to WHO grading alone, particularly in borderline grade 1 and 2 tumors.¹⁶ This approach may also be applicable to the molecular pathology of optic nerve sheath meningiomas, as these markers can enhance both diagnostic accuracy and clinical decision-making.

The most frequently observed genetic event in meningiomas is the loss of chromosome 22/inactivation of the NF2 tumor suppressor gene (Fig. 2). This alteration is particularly common in fibrous and transitional meningioma subtypes, as well as tumors located in the cerebral hemispheres. However, tumors arising near the skull base, including those along the optic nerve, are often associated with non-NF2 mutations, suggesting that different molecular mechanisms drive their development.¹⁷ Of interest, in one cytogenetic study of 19 samples tested by single-nucleotide polymorphism microarrays, most spheno-orbital and primary intraorbital ectopic meningiomas contained monosomy 22/22q loss, but this alteration was found in only one (of five) meningiomas of the optic nerve sheath proper.¹⁸ Higher-resolution sequencing studies looking at pathogenic gene variants should clarify the true frequency of NF2 alterations in these subgroups.

Several other genes have emerged as important in non-NF2 meningiomas, many of which are relevant to optic nerve sheath meningiomas (ONSMs) and orbital meningiomas due to their frequent basal skull location. Among these are mutations in TRAF7, which encodes a proapoptotic E3 ubiquitin ligase; AKT1, which encodes a key effector of PI3K signaling; KLF4, which encodes three C2H2 zinc finger motifs; and SMO, which encodes a negative regulator of the hedgehog pathway. These alterations were all mutually exclusive of NF2 mutations. TRAF7 mutations very often occur concomitantly with AKT1 and KLF4 mutations. In two large-scale genome-wide genotyping and exome sequencing studies, most non-NF2 meningiomas were found in the medial skull base. All meningiomas with the SMO mutation were located near the midline anterior skull base, which is highly relevant to ONSMs. Conversely, the lateral and posterior skull base meningiomas had NF2 mutations or a loss of chromosome 22.¹⁷

Other molecular markers have also been linked to tumor grade and behavior. For example, chromosomal losses at regions such as 1p, 6q, 10q, and 14q have been associated with more aggressive meningiomas. Deletion of chromosomes 1p, 6q, and 19p were frequently found in spheno-orbital meningiomas compared to tumors involving the optic nerve proper.¹⁸ Notably, primary intraorbital ectopic meningiomas seem to have fewer chromosomal alterations than spheno-orbital and optic nerve tumors.

In intracranial meningiomas, additional alterations such as TERT promoter mutation and homozygous CDKN2A/B deletions suffice for a grade 3 diagnosis in meningioma,¹⁹ although grade 3 tumors are not typically encountered in the orbit and optic nerve.

Tumors of the Orbit

Despite being a relatively small compartment by volume, the orbit is rich in anatomic structures relevant to proper ocular function, and therefore, a variety of tumors may occur at this site, many with known molecular drivers (Table 1). Molecular diagnostics is of particular help in the evaluation of a variety of soft tissue tumors that may occur at this site. High-throughput sequencing and gene fusion panels are particularly suited for this purpose. A variety of soft tissue tumors and hematolymphoid malignancies may develop at this site, and therefore, the reader is directed to more authoritative textbooks for comprehensive coverage of individual tumor types.²⁰

Table 1.

Orbital Tumors With Specific Genetic Drivers

Tumor Type	Genetic Alteration
Mesenchymal tumors
Benign
Nodular fasciitis/proliferative myositis	USP6 rearrangements
Epithelioid hemangioma	FOS/FOSB rearrangements
Round cell sarcomas
Ewing sarcoma	EWSR1 or FUS rearrangements (EWSR1::FLI1, EWSR1::ERG)
Alveolar rhabdomyosarcoma	PAX3::FOXO1 and PAX7::FOXO1
Spindle cell sarcomas
Embryonal rhabdomyosarcoma	Mutations in the RAS pathway (NRAS, KRAS, HRAS, NF1, FGFR4), PI3K pathway (PTEN, PIK3CA), and other (FBXW7, CTNNB1)
Spindle cell/sclerosing rhabdomyosarcoma	MYOD1 p.L122R, EWSR1/FUS::TFCP2, or MEIS::NCOA2
Synovial sarcoma	SS18::SSX1 or SS18::SSX2 fusions
Solitary fibrous tumor	NAB2::STAT6 fusions
Inflammatory myofibroblastic tumor	ALK rearrangements
Other sarcomas
Mesenchymal chondrosarcoma	HEY1::NCOA2
Epithelioid sarcoma	SMARCB1/INI1 loss
Sclerosing epithelioid fibrosarcoma	EWSR1 or FUS rearrangements
Angiomatoid fibrous histiocytoma	EWSR1 and CREB (or ATF) family member fusions
Epithelioid hemangioendothelioma	WWTR1::CAMTA1, YAP1::TFE3
Extraskeletal myxoid chondrosarcoma	EWSR1::NR4A3, TAF15::NR4A3 fusions
Alveolar soft part sarcoma	ASPSCR1::TFE3
Peripheral nerve tumors
Neurofibroma	NF1 mutations
Schwannoma	NF2 mutations, SH3PXD2A-HTRA1, SOX10 indels
Malignant melanotic nerve sheath tumor	PRKAR1A mutations
MPNST	NF1, SUZ12/EED mutations
Melanocytic tumors
Melanocytoma	GNAQ, GNA11
Primary orbital melanoma	GNAQ, GNA11, EIF1AX, and SF3B1 mutations
Metastatic melanoma	BRAF p.V600E, NF1, TERT promoter mutations
Hematolymphoid neoplasms
Diffuse large B-cell lymphoma	MYD88 and CD79B mutations in a subset
Extranodal marginal zone lymphoma	t(14;18)(q32;q21)/IGH::MALT1; TNFAIP3 mutations
Extramedullary myeloid sarcoma	t(8;21)/RUNX1::RUNX1T1, inv(16)/CBFB::MYH11
Histiocytoses (e.g., Erdheim-Chester)	Alterations resulting in MAPK pathway activation (e.g., BRAF p.V600E)

Soft Tissue Tumors

Solitary fibrous tumor is a distinctive spindle cell neoplasm characterized by variable amounts of collagen deposits and staghorn vessels that may involve the orbit. Morphologic variants include those composed of large pleomorphic cells, previously termed giant cell angiofibroma, and with a predilection for this anatomic site.²¹ The tumor is characterized by the distinct oncogenic gene fusion NAB2::STAT6,^22,23 which results in nuclear localization of STAT6, detectable in most instances by immunohistochemistry, a practical surrogate marker. These tumors have a high risk of local recurrence and may metastasize, even late in their clinical course.

Rhabdomyosarcomas (RMSs) are the most common primary orbital malignancy of childhood,^24,25 with approximately 10% of all rhabdomyosarcomas occurring at this site.²⁶ Most of these tumors are embryonal rhabdomyosarcomas (ERMSs), followed by alveolar rhabdomyosarcoma (ARMS). Pleomorphic and spindle cell/sclerosing rhabdomyosarcomas are exceedingly rare at this site.¹

One-third of ERMSs are diagnosed before the age of 5, but they can develop at any age. Half of all ERMS cases originate in the head and neck region, including the orbit.²⁷ Compared to ARMSs, ERMSs possess wider genetic heterogeneity and greater genomic instability. Common alterations in ERMSs include gains on chromosomes 2, 8, and 13 and loss of heterozygosity of the short arm of chromosome 11. Pathogenic gene variants have been reported in TP53, FGFR4, FBXW7, CTNNB1 (β-catenin), NF1, PDGFRA, PIK3CA, and RAS family genes (with NRAS or KRAS point mutations present in ∼35% of cases).^25,28–30

ARMS has a preference to develop in the extremities but may affect other sites, including the head and neck.²⁷ In ∼80% of ARMS cases, there are chromosomal translocations t(2;13)(q35;q14) resulting in PAX3::FOXO1 fusions and t(1;13)(p36;q14) resulting in PAX7::FOXO1 fusions.^24,28 Additionally, an upregulated expression of anaplastic lymphoma kinase protein is often found in ARMS and progressive embryonal RMS.²⁵

Spindle cell/sclerosing rhabdomyosarcoma favors the paratesticular and head and neck regions, and although rare in the orbit, it has been documented.³¹ Genetic alterations reported in this subtype include a MYOD1 p.L122R missense mutation or VGLL, COTED2, NCOA2, MEIS1, ESWR1, and TFCP2 rearrangements.²⁰ Intraosseous spindle cell rhabdomyosarcoma, a recently described subgroup, shows a predilection for craniofacial bones and may affect the orbit.³² It is associated with EWSR1/FUS::TFCP2 or MEIS::NCOA2 fusions.^14,33

Germline mutations also play a role in RMS predisposition. TP53 mutations have been implicated in both somatic and hereditary RMS, with approximately half of germline TP53-mutant RMS cases occurring in the head and neck region.²⁸

Peripheral Nerve Tumors

Most peripheral nerve sheath tumors that develop in the orbit are benign and include neurofibroma, schwannoma, and, rarely, hybrid nerve sheath tumors.³⁴ Neurofibroma may develop as several clinicopathologic types, including small skin growths (cutaneous neurofibroma); they may also involve the skin in a more diffuse manner, affect soft tissue or viscera, or invade multiple peripheral nerve fascicles (plexiform neurofibroma). Genetic inactivation of the NF1 gene is a consistent feature of all neurofibroma subtypes and is frequently the sole driver of these tumors.^35,36 Orbitofacial neurofibromas/NF have been recognized as a clinicopathologic variant that may be associated with more aggressive biological behavior but little potential for malignant transformation. Distinct epigenetic and gene expression profiles have been detected in these tumors.³⁷

The term atypical neurofibromatous neoplasms of uncertain biologic potential (ANNUBP) was proposed for neurofibromas exhibiting two or more worrisome histologic features: atypia, loss of neurofibroma architecture, hypercellularity, and increased mitotic activity, defined as >1/50 but <3/10 high-power fields.³⁸ Since CDKN2A deletions are enriched in ANNUBP (termed atypical neurofibromas in some studies),³⁹ in updated consensus recommendations, a diagnosis of ANNUBP can also be reached by the detection of homozygous CDKN2A deletions in a neurofibroma without histologic ANNUBP features or with heterozygous CDKN2A deletions in a tumor with at least one ANNUBP feature. The term low-grade MPNST has been replaced with ANNUBP with increased proliferation.⁴⁰

Schwannomas form well-circumscribed encapsulated masses, often developing in association with a single nerve fascicle, and may compress vital structures while growing, particularly in tight compartments such as the orbit. NF2 gene inactivation is frequent in schwannoma, occurring in most tumors (>75%). Additional alterations uncovered by genomic sequencing include mutations of ARID1 family members and a SH3PXD2A-HTRA1 fusion in approximately 10% of tumors.⁴¹ A subset of the latter may involve intracranial and head and neck sites, including the eyelid.⁴² Additionally, germline mutations in SMARCB1 or LZTR1 occur in a subset of patients with schwannomatosis.^43–45 More recently, SOX10 indel mutations have been found in a subset of schwannomas lacking NF2 mutations, which are enriched in tumors involving nonvestibular cranial nerves.⁴⁶ A curious case of SOX10-mutant plexiform schwannoma involving the orbit has been reported.⁴⁷

Hybrid nerve sheath tumors. Clear identification of the different benign nerve sheath tumor subtypes is straightforward in most instances, using a morphologic and immunophenotypic approach. However, there is increased recognition of hybrid features in these tumors.⁴⁸ Hybrid neurofibroma–schwannoma tumors have a predilection to develop in patients with neurofibromatosis 1 and schwannomatosis.⁴⁹ Hybrid schwannoma–perineuriomas have been increasingly recognized in soft tissue and viscera through immunohistochemistry, which may highlight 100+ Schwann cells and EMA⁺ perineurial cells.^50–52 Rare examples have been reported in the orbit.³⁴ Hybrid neurofibroma–perineuriomas are more uncommon overall.⁵³ VGLL3 gene rearrangements were detected in hybrid schwannoma–perineuriomas,^54,55 while ERBB2 mutations have been detected in hybrid neurofibroma–schwannomas.⁵⁶

Malignant nerve sheath tumors involving the orbit are much more uncommon than benign nerve sheath tumors, and they may be difficult to separate from desmoplastic/spindle cell melanomas, which are by far more frequently encountered in the head and neck region. Molecular diagnostics may be of great help in this regard, detecting pathogenic gene variants typical of melanoma (e.g., BRAF) or MPNST (SUZ12, EED) or a UV-mutational signature that favors melanoma.⁵⁷

Melanocytic Tumors

BRAF mutations, particularly BRAF p.V600E, are present in up to 60% to 70% of cutaneous melanomas but are largely absent in uveal melanoma, which instead harbors an increased frequency of pathogenic variants in genes such as GNAQ, GNA11, SF3B1, and EIF1AX.⁵⁸ However, primary orbital melanoma (POM) has been increasingly recognized and characterized as a rare primary site.⁵⁹ In a cohort of six POM cases, immunohistochemistry revealed all cases were positive for melanocytic markers S100, Melan-A, HMB45, and SOX10. Genomic profiling across these cases identified frequent chromosomal gains in 6p, 8q, 17q, 6q, and 20p and recurrent losses in 1p, 9p, 16q, and 17p. One case showed both monosomy 3 and 8q gain, accompanied by loss of BAP1 nuclear staining. Somatic mutations were identified in several known melanoma-related genes: GNAQ (one case), GNA11 (one case), SF3B1 (two cases), NRAS (two cases), and the TERT promoter (two cases).⁶⁰

Additional studies reinforce the involvement of these genetic alterations in POM. Sequencing analysis of 12 tumors with suspected orbital melanoma identified GNAQ p.Q209L mutations in two cases, although it was noted that these tumors may represent occult spread from uveal primaries. A recurrent SF3B1 p.R625H mutation appeared to be associated with indolent tumor behavior, while EIF1AX N4S was found in a case with a nonaggressive clinical course. The presence of SF3B1 and EIF1AX pathogenic variants may serve as markers of disease progression in orbital melanoma.⁵⁹

In one reported case of meningeal melanocytoma with orbital extension, a GNA11 p.Q209L pathogenic gene variant was identified in a patient with an ipsilateral nevus of Ota, adding further evidence to the involvement of G-protein signaling pathway mutations in orbital melanocytic tumors.⁶¹ Another case of intermediate-grade melanocytoma revealed no gene mutations or copy number alterations.⁶² These small series and rare case reports highlight the possibility that a subset of melanocytic tumors of the orbit, as in meningioma, may represent extension from intracranial primaries or involvement from uveal primaries.

Together, these findings demonstrate that while orbital melanocytic tumors are rare, they share several molecular features with uveal and meningeal melanocytic neoplasms. Recurrent mutations in GNAQ, GNA11, SF3B1, and EIF1AX, along with characteristic chromosomal alterations, may help refine diagnosis and guide future classification efforts in these still poorly understood neoplasms (Table 1).

Hematolymphoid Tumors

A variety of lymphomas may involve the orbit and ocular adnexa primarily or through secondary involvement from systemic disease. Almost any lineage (B-cell, T-cell, natural killer) may affect these sites, but B-cell lymphomas predominate. Extranodal marginal zone lymphoma (EMZL) is the most common primary lymphoma of the ocular adnexa. Although the diagnosis is made primarily through morphologic and immunophenotypic analysis, EMZLs have been reported to develop preferentially certain genetic alterations, such as t(14;18)(q32;q21)/IGH::MALT1, occurring in approximately 25% of cases.⁶³ Pathogenic variants in TNFAIP3, a negative regulator of the canonical NF-κB pathway, occur in EMZL at a variety of sites, including ocular adnexa.⁶⁴ Diffuse large B-cell lymphoma is a relatively common lymphoma at a variety of body sites and is the primary intraocular lymphoma subtype. When involving the ocular adnexa, it appears to be pathologically and molecularly heterogeneous. Pathogenic gene variants in MYD88 are frequently encountered at this site,⁶⁵ as well as CD79A, to a lesser extent.⁶⁶

Extramedullary myeloid sarcomas (aggregates of blasts outside of the bone marrow) may also affect the ocular adnexa in rare instances. Genetic alterations typical of acute myeloid leukemia can be detected (see Table 1).

Histiocytic neoplasms. A group of histiocytic neoplasms that may represent localized proliferations or multisystem disorders has been recognized as involving a variety of organ sites, including the orbit. Past classifications regarded some of these lesions as reactive. However, these disorders represent true neoplasms, given their patterns of growth and clonal nature, being driven by activating mutations in the MAPK signaling pathway.

Erdheim-Chester disease (ECD) is an aggressive histiocytic neoplasm that is almost never localized and may involve any anatomic compartment, including body cavities and bone. However, bilateral orbital involvement is a dramatic and recognized manifestation of the disease.⁶⁷ Pathogenic gene variants involving components of the MAPK pathway occur in most cases, with BRAF p.V600E present in approximately 50% of all cases of ECD. Targeted therapies have resulted in dramatic responses in some patients.⁶⁸ Rosai-Dorfman disease (RDD) is more likely to be localized than ECD, with the eye and ocular adnexa being involved in approximately 10% of cases.⁶⁹ Clonal alterations in MAPK pathway components have been detected in a minority of RDD cases occurring in a variety of sites, but BRAF p.V600E appears to be rare.^70,71 Juvenile xanthogranuloma is another histiocytosis that may present as an isolated lesion or as part of a systemic process. Pathogenic gene variants involving ARAF, KRAS, and NRAS have been reported,⁷² while NF1 loss is expected in tumors developing in patients with NF1. Langerhans cell histiocytosis is a distinctive disorder composed of cells containing nuclei that are typically multilobated and/or clefted. A variety of sites may be affected, but circumscribed lesions in skull bones, including orbit, are typical. BRAF p.V600E is detected in approximately half of the cases, with alterations in other components of the MAPK pathway detected in many BRAF p.V600E–negative cases.

Tumors of the Lacrimal Gland

A variety of epithelial and nonepithelial tumors may arise in the lacrimal gland, with many of these tumors being more common and better characterized in the salivary glands. These tumors have distinctive genetic drivers, particularly oncogenic gene fusions (Table 2 and Fig. 3). The most common benign epithelial tumor at this site is pleomorphic adenoma, which is characterized by rearrangements involving PLAG1 or HMGA2 genes. Conversely, adenoid cystic carcinoma is the most common malignant neoplasm at this site and is characterized by MYB::NFIB fusions.

Table 2.

Primary Epithelial Lacrimal Gland Tumors

Tumor	Relative Frequency of All Lacrimal Gland Tumors (From Folberg et al.5), %	Genetic Driver
Benign	15–66
Pleomorphic adenoma	10–32	PLAG1 or HMGA2 rearrangements
Malignant	1–29
Adenoid cystic carcinoma	1–12	MYB::NFIB fusion
Carcinoma ex pleomorphic adenoma	<1–4	PLAG1 or HMGA2 rearrangements + other chromosomal alterations
Primary ductal adenocarcinoma	<1	ERBB2 amplification/overexpression
Secretory carcinoma	<1	ETV6::NTRK3
Mucoepidermoid carcinoma	<1	CTRC1::MAML2
Epithelial–myoepithelial carcinoma	<1	HRAS mutations
Acinic cell carcinoma	<1	NR4A subfamily rearrangements

Figure 3.

Primary epithelial lacrimal gland tumors and hallmark genetic drivers. Pleomorphic adenoma is characterized by PLAG1 or HMGA2 rearrangements. Adenoid cystic carcinoma (1%–12%) typically harbors MYB::NFIB fusions. Alterations in less common epithelial lacrimal tumors, of variable malignant potential, are also listed.

Conclusions

Molecular diagnostics currently have numerous applications for the evaluation of human tumors, including those involving the optic nerve, orbit, and ocular adnexa. This offers refined diagnostic accuracy and clinically meaningful prognostication with evolving markers. Current advances already influence management, while emerging multiomic technologies are poised to drive the next generation of precision diagnostics for patients with ocular neoplasms.