The past few decades have witnessed major strides in the field of orbital surgery. Insights into the pathogenesis of orbital disease have refined surgical indications and objectives. Appreciation of anatomic relationships, aesthetically sensitive incisions and dissection paths, respectful tissue handling, and discrete hemostatic techniques have all enhanced effectiveness while reducing collateral damage. Arguably, the most significant development has been high-resolution imaging, which effectively transformed the orbitotomy from an exploratory adventure to a precisely planned, goal-directed mission.

Despite these advances, the management of small hemangiomas lodged deep in the orbital apex has remained a challenge. Some lesions, generally those found incidentally in asymptomatic patients scanned for unrelated reasons, remain unchanged in long-term follow-up. Others cause progressive functional deficits, and even minor incremental growth can have major impact in the densely packed region. In such cases, early intercession would seem to be clearly indicated. On the other hand, because of often tenacious attachments to fixed apical structures, even careful surgical removal can imperil vision. The surgeon and patient, therefore, are faced with difficult decisions of whether, when, and how to intervene. Although no single solution may be applicable to every case, consideration of a hemangioma’s pathogenesis and growth—particularly within the orbital apex—can inform our surgical judgment regarding indications, timing, and technique.

Cavernous hemangioma is the most common benign orbital tumor in adults. Quantitative data specific to the orbital apex are lacking, but hemangiomas likely represent a major proportion of well-circumscribed, round-to-oval masses in the region. Based on their signal intensity in conventional magnetic resonance imaging (MRI) scans (isointense to muscle in T1-weighted images; hyperintense in T2-weighted images), and their contrast-enhancement spread pattern in dynamic MRI scans (starting from one point or portion, rather than a wide area), imaging usually allows their differentiation from other apical tumors, such as schwannoma, neurofibroma, or meningioma. Based on their low-flow arterial input, they can generally be distinguished from other well-circumscribed vascular lesions, such as hemangiopericytoma, with multiphase dynamic computed tomography (CT) or MRI scanning.

A cavernous hemangioma is not a neoplasm in the usual sense of the term: it does not derive from a single, proliferating progenitor cell. Rather, it consists of interconnected, endothelially lined vascular channels with stromal septa of varied thickness ( Figure 1 , Top left and Top right). The observation of “budding” of the vascular spaces into the interstitium (a finding seen in arteriovenous fistulas) led Harris and Jakobiec to speculate that a low-grade, local hemodynamic shift could open new channels and promote their extension into surrounding tissue. We might take that pathogenetic concept a bit further, and relate it to the morbidity of both apical tumor growth and apical tumor removal.

Cavernous hemangiomas. (Top left) The lesions consist of intertwining plexuses of endothelially lined channels, which appear as juxtaposed blood-filled spaces in histologic section (hematoxylin-eosin, ×10). (Top right) Progressive branching of vascular channels into the neighboring stroma may be driven by an ongoing imbalance of local hemodynamic forces (hematoxylin-eosin, ×20). (Center) Slow enlargement induces a fibrous capsule at the perimeter, which is continually reconstituted as the process advances (Masson trichrome, ×2). (Bottom left and right) Hematoxylin-eosin (×4) and Movat elastin (×4) staining, respectively, demonstrate an apical cavernous hemangioma whose capsule has incorporated the adventitia of the ophthalmic artery.

As the neovascular network slowly advances into neighboring tissue, it induces a fibrous capsule at the interface ( Figure 1 , Center). Continued growth involves further budding of vascular channels into both the internal septa and the fibrous perimeter. The capsule, therefore, is continuously reconstituted just beyond the expanding plexus. In the process, it must displace, compress, or incorporate surrounding structures. In the anterior and middle orbit, the relatively mobile muscles, nerves, vessels, and loose fibrofatty matrix can generally accommodate that slow expansion by moving aside. Compression to the point of visual impairment tends to occur only with larger tumors, and substantial infiltration by the advancing capsule is uncommon. Because adhesions are tenuous, surgical extraction can often be easily accomplished with a cryoprobe.

In the orbital apex, in contrast, visually critical vessels and nerves are tightly compacted and directly apposed, with virtually no intervening fibrofatty matrix. Without room for displacement, tumor expansion takes a toll by compression and/or incorporation. Figure 1 (Bottom left and Bottom right) demonstrates the apical tumor of a patient with dense, diffuse visual field loss and acuity of 20/800 prior to removal. At one pole of the histopathologically confirmed cavernous hemangioma is the ophthalmic artery, whose adventitia was annexed by the lesion’s advancing capsule. The tumor’s fusion with the wall of the artery and compression of its lumen, rather than direct impact on the optic nerve, may have been the primary mechanism of the vision loss. That growth pattern also precluded surgical separation of the artery and tumor. The findings suggest that although cavernous hemangiomas are grossly well circumscribed and encapsulated, rather than diffuse and infiltrative, the distinction may be less clinically relevant in the crowded apex than elsewhere in the orbit. Indeed, the degree to which visually critical structures have been expropriated by tumor growth can be crucial in the surgical outcome. Unfortunately, such microanatomic details are not detectable preoperatively.

In formulating a management scheme, we should also consider what is known about the clinical behavior of these tumors. Orcutt and associates followed 11 visually asymptomatic patients with presumed cavernous hemangiomas that were found incidentally in scans for unrelated reasons (eg, headache, vertigo). All lesions were intraconal; those that were apical were not specified. All 11 patients were observed without treatment for an average of 3 years (range, 8 months to 10 years), and no tumor enlarged. The series, which did not include symptomatic patients who underwent surgery, suggests that at least some cavernous hemangiomas will cease growing spontaneously. In our pathogenetic construct, that might imply that local hemodynamic forces have equilibrated. Other hemangiomas, however, clearly do not lose their impetus for growth—even through decades of progressive expansion ( Figure 2 ).

Serial computed tomography (CT) scans documenting the slow progression of a cavernous hemangioma. (Left) CT in 1984 of a 56-year-old man with a 6-year history of progressive left proptosis and with visual acuity of 20/30 OS. He declined surgery. (Right) The patient returned in 1996, requesting tumor removal. Vision improved from hand motions to 20/40 after lateral orbitotomy and excision of a cavernous hemangioma. From Harris and Logani.

Reprinted with permission from Wolters Kluwer Health.

Harris and Jakobiec reviewed 66 cases of cavernous hemangioma that had been surgically removed, on average, 10 years earlier. In that archival series, almost every case antedated CT scanning, and rather than being incidental findings, the lesions had actively grown to cause proptosis or impair vision. Although several tumors among the 66 had been incompletely excised (according to the operating surgeon or the reviewing pathologist), none recurred. In those cases, partial resection and/or surrounding postoperative changes might have rebalanced hemodynamic factors. However, as with the natural history, the long-term course after incomplete resection is not uniform or predictable. Figure 3 demonstrates the slow, continued growth of a subtotally excised tumor (not detected in a scan 5 years after surgery, but clearly visible 15 years later). In yet another variation, Henderson and associates followed a patient with an incompletely excised cavernous hemangioma for 18 years. The residual tumor enlarged during the first 11 years, remained stable for 2 years, and then involuted.

Serial magnetic resonance imaging (MRI) scans showing slow growth of an incompletely excised cavernous hemangioma. (Left) MRI in 1989 of a 39-year-old woman with right-sided proptosis, acquired hyperopia, a relative afferent pupillary defect (RAPD), and disc swelling. Best-corrected vision was 20/25 ⁻² OD (+2.5 D). A lateral orbitotomy was performed, and the anterior lobe of a bilobed cavernous hemangioma was excised. Minimal manipulation of the fixed posterior lobe caused pupillary dilation, and a secondary craniotomy was planned for its removal. However, vision improved to 20/15 ⁻² OD (plano), and the disc swelling and RAPD resolved after the orbitotomy. (Center) MRI in 1994 showed negligible change in the posterior lesion. Vision was stable. (Right) MRI in 2009 demonstrated tumor progression. Vision remained 20/15 ⁻² OD (plano).

Although the functional outcome may be more related to a tumor’s microanatomic relationships and remaining growth potential than to the operative approach, the various surgical options should be considered in planning management. Anterior and lateral orbitotomy approaches have been commonly used to remove cavernous hemangiomas, and an increased risk to vision associated with apical tumors in contact with the optic nerve has been duly noted. Scheuerle and associates described 14 “cavernomas” of the apex that were completely resected using a craniotomy approach. Eleven of 14 patients functionally improved or remained stable; 2 of 14 experienced severe, permanent postoperative decreases in vision; and 1 of 14 suffered a subdural hematoma and permanent frontal lobe syndrome. Tsirbas and associates reported visual improvement in a patient with an inferomedial, extraconal hemangioma that was removed through a combination of lower fornix and endoscopic transantral approaches. Almond and associates described endoscopic transethmoidal decompression of the medial orbital wall—without tumor resection—in 4 patients with presumed cavernous hemangiomas. In 3 of 4 cases, visual fields improved or did not worsen (in relatively brief postoperative intervals up to 33 months). In 1 of 4 cases, preoperative vision of 20/30 continued to deteriorate after decompression (to light perception before, and to no light perception after, craniotomy with tumor excision).