In our 50-year collection of literature on orbital tumors (1948 to 1997), many clinicopathologic correlations, lists, and classifications pertaining to the frequency, prevalence, and incidence of these disorders have appeared. Some of these surveys from earlier years are now out of date because of the small size of the sample, the type of tumor selected for study, changes in nomenclature, and the like. Lists, however, can be considered “dynamic” and subject to change simply because of pathologic revisions and reclassifications.

We have reprinted the summary table from the third edition (see Table 3.1) and noted the variables that evolved in the tumor mix during the 40-year period 1948 to 1987. These variations from the expected norm may interest readers who want to compare and contrast their experiences with ours.

The survey tabulated in our third edition (1975 to 1987) added 612 consecutive cases to a cohort of 764 surveyed from 1948, bringing the total to 1,376 consecutive orbital tumors. The 612 cases seen over the 13-year period represent an average of 47 patients per year, 10 more per year than in the second series. A striking aspect of the second and third series is the consistency of percentages of primary, secondary, and metastatic lesions from 1948. The proportion of primary tumors declined from 45.0% to 41.7% with the addition of data from the third case series. In the secondary tumors, the change was a decline of only 0.9%, from 45.0% to 44.1%. The proportion of metastatic tumors increased from 7.0% to 8.1%. These minimal shifts in distribution underline the pathoanatomic stability of our tumor cohort.

A similar consistency prevailed among the major histologic tumor groups as cases were added to the series. Also, the five most common primary tumors remained the same from the first to the second series (i.e., hemangioma, non-Hodgkin lymphoma, inflammatory pseudotumor, meningioma, and optic nerve glioma).

In the third series, basal cell carcinoma replaced malignant melanoma as the fifth most frequent tumor among the secondary orbital tumors. The radiopaque therapy for ocular malignant melanoma probably is the contributing factor for this shift in incidence. On a further analysis of the third series, we found 1,036 neoplasms (1,036 out of 1,376), increasing the incidence to 75%. Of the 1,036 neoplasms, 659 were malignant and 377 were benign, a ratio of 1.7:1.

By 1987, several changes in terminology had occurred. Organizing hematoma, esthesioneuroblastoma, non-Hodgkin lymphoma, Langerhans histiocytosis, and angiomyoma replaced the terms hematic cyst, olfactory neuroblastoma, malignant lymphoma, histiocytosis X, and vascular leiomyoma, respectively. Acquired meningocele is a new addition, and we replaced the word generalized (second edition) with the more appropriate term multifocal.

About the time of the third edition, many physicians were interested in counting orbital tumors in pediatric patients. Four of the larger reviews of this type were by Eldrup-Jorgenson and Fledelius (1975), Iliff and Green (1978), Shields et al. (1986), and Bullock et al. (1989). Collectively they reported 809 cases. Our own tabulation totaled 212 cases, a 15% (212 out of 1,376) share of total patients.

Compilations of cases in the literature confirm the wide spectrum of tumor types in pediatric patients. Prominent in all these surveys is vascular hamartoma (capillary hemangioma), developmental cyst, rhabdomyosarcoma, optic nerve glioma, and plexiform neurofibroma. However, our series differs from other surveys because of the lower incidence of dermoid cyst. This lesion usually is said to be the most common orbital tumor of childhood. This

discrepancy probably is attributable to the inclusion of cysts in adnexal locations (eyebrow and eyelid) in many other series, whereas the cysts in our list were all retroseptal. The lower incidence of dermoid cyst in our series also may alter the ratio of benign to malignant lesions. In our pediatric cases, 70% (149 out of 212) were benign tumors and 30% (63 out of 212) were malignant. This is a considerably higher ratio of malignant tumors in children than that reported in the literature.

Also in the third edition, we referenced some 27 publications pertaining to the classification and frequency of orbital tumors. These were a heterogeneous collection of facts and figures. No two of these studies were identical in the coverage of the various parameters of an incidence survey. In an earlier year, 1979, after reviewing several surveys of such types, Jones and Jakobiec (1979) noted that there probably is no unbiased, parametrically complete tumor series, and efforts to blend and compare the data of one series with another still leave the goal of a true and perfect survey unrealized.

Since the tabulation of orbital tumors in the 1948 to 1987 series (third edition), an additional 419 consecutive orbital tumors were surgically confirmed between 1988 and 1997 (see Table 3.2). The composite total is now 1,795 cases (see Table 3.3).