Purpose: Orbital cavernous hemangiomas (OCH) often present as incidentally found asymptomatic lesions, and it has been reported that such lesions do not grow. The authors' own experience has been different. They describe the natural history of OCH in a cohort with asymptomatic incidentally found lesions and compare them with patients with symptomatic OCH to identify any characteristics that might predict those which enlarge and produce symptoms. Methods: This was a retrospective comparative case series from 3 tertiary referral orbital centers in Australia. There were 104 patients with OCH, 31 of whom presented with incidentally found lesions. For patients with incidental lesions, only those with a minimum follow-up period of 12 months were included. Outcome measures included demographic features, clinical features at presentation and during follow up, anatomical location of the OCH, change in maximal linear dimensions of the OCH over time, follow-up period, and requirement for treatment in the follow-up period. Results: There were 104 patients included in the study. Thirty-one had an asymptomatic, incidental OCH on imaging performed for other indications. Seventy-nine patients had treatment, and 11 of these had presented with an incidental, asymptomatic OCH that enlarged and produced symptoms or new clinical findings. In the 20 other patients with incidentally found OCH, there was no or minimal change in the follow-up period of 1.2 to 20 years (mean 5.8 years, standard deviation [SD] 4.6 years). None of the factors analyzed in the study were able to predict those patients with an incidental OCH that enlarged and required treatment. Patients with incidental lesions that did not grow over several years did not change over much longer periods of time. When considering all patients with OCH, both incidentally diagnosed and symptomatic, there were 4 factors that predicted whether someone with an OCH would require treatment. These were maximal linear dimension, male gender, extraconal location (increased risk of requiring treatment), and being an incidental finding (decreased risk of requiring treatment). Conclusions: OCH may present as an incidental finding. A significant number will grow over time and require treatment. The authors could not identify any factor that would predict those lesions that enlarge over time and require treatment. If an incidental OCH does not change over several years, it is very unlikely to do so in longer periods of follow up.