Dear Editor, Apically located orbital cavernous hemangiomas (OCHs) are notorious for their detrimental course and problematic management (McNab & Wright 1989). Of a series of 40 histologically proven orbital intraconal OCHs, operated between 1990 and 2010, by four different orbital surgeons in our institutions, 36 OCHs were completely removed within their capsula. All 36 excisions had a satisfying outcome. All these lesions appeared apple shaped and were located in the middle or anterior part of the orbit. Four apically located, pear-shaped OCHs, however, had a disappointing outcome. We here describe these cases and speculate about the causes of the surgical failure. Four otherwise healthy persons (three Caucasians, one Chinese; three women, one man), ranging in age from 37 to 60 years, with progressive visual loss (preoperative visual acuities were 0.5, 0.25, LP+ and 1.2) and/or extensive and progressive visual field defects (in all) underwent a lateral or anterior orbitotomy because of an apically located OCH. All tumours were partially removed after opening of the capsula in a piecemeal fashion. Three of them developed a retrobulbar bleeding during extubation, for which the orbits were reopened. The operation in the 4th patient, a 44-year-old woman, was complicated by profuse bleeding necessitating repeated bipolar coagulation. Final visual acuity was finger counting in 3 and no light perception in one. The incidence, clinical presentation and characteristics of orbital OCHs are nowadays well known (McNab & Wright 1989). Discussion remains on the origin of OCHs and how to deal with apically located representatives. Most OCHs present in the lateral part of the middle third of the orbit, which area is rich of vascular structures. In this region, OCHs can grow by easily pushing aside other structures (Harris 2010). These are easily accessible by an anterior or lateral orbitotomy and can be removed with subtle traction using, for instance, a cryocoagulation probe. It is this kind of OCH that has an apple shape on scans. In contrast, OCHs in the orbital apex, having a pear-shaped configuration, show a different growth pattern, in which the surrounding capsula fuses with adjacent vessels and nerves in an early stage of the disease. Total removal by orbitotomy is impeded by poor visualization, lack of manoeuvring space, excessive bleeding by traction and disruption of blood vessels, damaging of nerve vessels and the ciliary ganglion. Partial removal with or without bipolar cautery shrinkage has been suggested as an alternative procedure (McNab & Wright 1989; Harris 2010). Our experience does not support this option. Although histologic proof and size reduction was the goal of surgery in apical OCHs and these tumours were approached with minimal traction, significant bleeding during or shortly after the operation necessitated reopening of the wounds. Anaesthetic data showed no sudden increase in systemic blood pressure during or immediately after surgery. Unlike in more anteriorly located OCHs, in which the complete lesion can be removed within their capsula, the disruption of the lesion’s capsula in our four patients may be the explanation of the bleeding during or after surgery. Apparently, at least in our hands, a piecemeal OCH excision is not a viable option. Irradiation (Khan et al. 2009; Rootman et al. 2012) or a combined neurosurgical/endoscopic (Locatelli et al. 2011.) orbital approach (using neuronavigation) with microdissection and the use of an operation microscope may be better options to treat these lesions. Larger series, however, are needed to confirm this assumption.