NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Bast RC Jr, Kufe DW, Pollock RE, et al., editors. Holland-Frei Cancer Medicine. 5th edition. Hamilton (ON): BC Decker; 2000.

Cover of Holland-Frei Cancer Medicine

Holland-Frei Cancer Medicine. 5th edition.

Show details

Chapter 81Neoplasms of the Eye

, MD, , MD, and , MD.

Ophthalmic oncology is foreign to most adult and pediatric oncologists. It is a subspecialty dominated by a small number of uniquely qualified ophthalmologists who frequently work out of Eye Institutes, publish their findings in the ophthalmic literature, diagnose and treat diseases that are unfamiliar to most oncologists, and use a vocabulary unlike most others involved in cancer medicine. The anatomy, specialized imaging techniques, and methods of examination are ophthalmic based. Most striking is the fact that the major intraocular malignancies, retinoblastoma in children and uveal melanomas in adults, are routinely diagnosed and treated without pathologic confirmation. Systemic chemotherapy, ocular irradiation, and removal of one or both eyes are routinely performed throughout the world without needle biopsies, incisional biopsies, or pretreatment cytologic study of any form. Furthermore, the evaluation of local control is based on ophthalmic techniques with which most oncologists are unfamiliar: indirect ophthalmoscopy under anesthesia aided by fundus photography, fluorescein angiography, and ocular ultrasonography, all of which are performed by the ophthalmologist.

When pathologic specimens are available, they are usually interpreted by ophthalmic pathologists. Even expert pathologists working in large cancer centers are usually unfamiliar with ocular anatomy, pathology, and artifacts. In some cases, such as the interpretation of ocular melanomas, ocular pathologists have evolved their own classification schemes, cell type terminology, and descriptions that at times are at odds with traditional oncologic pathology.

Finally, the eye is a common site for metastasis, and it may be the general ophthalmologist who first detects that a patient has a metastatic tumor and needs to be referred to an oncologist.

Because the appropriate management of many ophthalmic malignancies routinely requires surgery, chemotherapy, radiation (external beam and brachytherapy), lasers, and cryotherapy, and because retinoblastoma has such a strong genetic pattern, successful management of these patients requires a well-integrated team of ophthalmologists, radiation oncologists, and radiation physicists, and often a team of pediatric oncologists, pediatricians, nurses, ophthalmic pathologists, genetic counselors, social workers, and diagnostic radiologists.

This chapter reviews benign and malignant ocular, orbital, and lid tumors in both children and adults. The most common of these are listed in Table 81.1.

Table 81.1. Most Common Ophthalmic Neoplasms, Benign and Malignant.

Table 81.1

Most Common Ophthalmic Neoplasms, Benign and Malignant.

Ophthalmic Oncology in Children

Ocular Disease

Benign Disease

Benign tumors of children’s eyes are very rare. Choroidal nevi, which are present in more than 10% of the adult population, are rare before puberty and are never seen in the infant.

Conjunctival nevi are also extremely rare in prepubertal children, as are iris nevi. Iris nevi in children may represent Lisch nodules, a manifestation of neurofibromatosis type 2. Benign retinal tumors are also rare. When found, they are usually astrocytic hamartomas and frequently part of tuberous sclerosis syndrome. Astrocytic hamartomas usually look like ill-defined transparent plastic-wrap overlying the retina and obscuring retinal blood vessels. They may enlarge and calcify with time. They may be confused with myelinated nerve fibers; the latter are usually fan shaped, white, and may characteristically fan out to be broader the more distant from the optic nerve, follow the distribution of the nerve fiber layer, and often obscure retinal vessels. Hamartomas of the retinal pigment epithelium are rare in children. They are frequently near the optic disc and pigmented, with distortion of retinal vessels and a crinkled, slightly opaque appearance. They have no malignant potential.

Malignant Disease (Retinoblastoma)

The most common primary ocular malignancy of childhood is retinoblastoma.1 An excellent summary of the disease is available for families at the web site Even more common, however, is ocular involvement from leukemias. Retinoblastoma is a true neoplasm of the retina. Although relatively rare, it has been the subject of great interest because of its well-established genetic pattern and because of the well-studied molecular mechanisms that characterize this tumor.2

Retinoblastoma occurs in one in 18,000 to 30,000 live births worldwide. Surveys suggest a relatively constant occurrence in this century.3 The incidence in the United States is relatively low, at 3.58 cases for each million children under the age of 15 years, and is closely correlated with age. For ages 1 to 4 years, the incidence is 10.6 per million; for 5 to 9 years, 1.53 per million; and for 10 to 14 years, 0.27 per million. It is the seventh most common pediatric cancer, but in some countries (e.g., Mexico), it is the most common solid tumor in childhood.

There is no difference in incidence by sex or by right or left eye.4 Some data suggest clustering, but convincing evidence is lacking. Retinoblastoma appears to occur more commonly in poor patients worldwide.

Retinoblastoma occurs in two forms: germinal and non germinal. In germinal cases, both eyes are usually affected, and the mean number of tumors distributed between the two eyes is 5. Inheritance usually follows an autosomal dominant pattern, with 90% penetrance. All patients with bilateral retinoblastoma have a germinal mutation on chromosome 13, although only 8% have an antecedent family history of the disease. About 15% of patients with germinal retinoblastoma have only one eye involved. When patients with germinal retinoblastoma have unilateral disease, it is almost always multi-focal.

Nongerminal retinoblastoma is always unilateral and unifocal, although seeding of the tumor, as it breaks apart because of a lack of cohesiveness, may cause hundreds of tiny intraocular seeds to appear. Genetic counseling on the basis of these factors is schematically presented in Table 81.2

Table 81.2. Genetic Counseling for Retinoblastoma.

Table 81.2

Genetic Counseling for Retinoblastoma.

Molecular Biology of Retinoblastoma

Retinoblastoma is one of the prototypical models demonstrating the genetic etiology of cancer. Knudson proposed the now classic two-hit model in 1971 after noting that the timing of tumor development (earlier diagnosis of bilateral tumors than of unilateral tumors) suggested a mechanism in which at least two events would be responsible for the development of the tumor.5 Patients with multifocal bilateral disease are germline carriers for the first hit, with only one second hit necessary to develop retinoblastoma. Unilateral, unifocal patients usually have a normal germline genome but develop both hits in the progenitor tumor cell.

The nature of these hits was suggested by karyotypic abnormalities on the long arm of chromosome 13 in rare cases and tumors. In 1986, Friend and colleagues isolated the rb1 gene, located on the long arm of chromosome 13, band 14.2.6 Further characterization of the gene revealed that it spans 200 kb and is composed of 27 exons. The gene encodes a 4.7-kb transcript, which is expressed in all adult tissues. The 110-kD nuclear phosphoprotein consists of 928 amino acids.

The protein seems to be a regulator at the cell cycle check point between G1 and the entry into the S-phase. The phosphorylation pattern of p110 RB varies during the cell cycle and the current model (Fig. 81.1) suggests that the unphosphorylated normal RB1 protein binds transcriptional regulators that promote entry into the S-phase. When the normal RB1 protein is phosphorylated, it dissociates from E2F (one of these transcription factors), freeing itself to bind to DNA and stimulate transcription of as yet unknown downstream genes that promote progression through the cell cycle. Loss of normal RB1 function, as in the case of the tumors, presumably allows for the uncontrolled entry into the S-phase, more rapid cell cycling, and, therefore, rapid cell division.

Figure 81.1. Artistic rendition of molecular mechanism of retinoblastoma gene action.

Figure 81.1

Artistic rendition of molecular mechanism of retinoblastoma gene action.

Because it is the loss of function of the rb1 gene that provides the impetus toward malignant transformation, the gene falls into the category of tumor suppressor genes. Retinoblastoma research has provided much of the initial understanding of this class of oncogenes. Whether replacement of normal rb1 into a fully transformed tumor cell is sufficient to reverse the malignant state is still controversial,7 and failure to do so would be consistent with the concept that while rb1 gene mutations are a necessary first step in tumorigenesis, they are not sufficient, and that additional steps are necessary to cause retinoblastoma.

Additional studies designed to better understand the function of the normal rb1 protein have included the creation of knock-out mice, which have been genetically engineered to lack the normal rb1 gene in their germline.8 Heterozygotes are normal at birth but have a propensity to develop pituitary tumors. Notably, these animals do not develop retinoblastomas. Animals homozygously lacking rb1 do not survive to birth, dying by embryonic day 16 due to hematopoietic and nervous system abnormalities.

Several groups of investigators have begun to use single-stranded conformational polymorphism analysis and DNA sequencing to identify specific mutations found in the rb1 gene in retinoblastoma tumors and in the germline of patients with the hereditary form. In cells with a single mutation (germline cells that carry the mutated gene), transcription of the abnormal allele cannot be detected, and it has been hypothesized that the normal allele somehow inhibits the expression of the abnormal mutant allele. This means that detection of the germline mutation in normal cells of potential carriers must be done at the DNA level, not on RNA.

While only limited amounts of data have been published to date, it seems that the majority of first hits are point mutations, with most of the remainder being small deletions, and only about 2 to 3% karyotypically visible larger deletions of material in the 13q14 band. Some of the children with this cytogenetic abnormality in their germline have a constitutional disease, with severe developmental delay and dysmorphic features, while others are phenotypically normal, other than their retinoblastoma. The second hit usually consists of loss of the normal allele (loss of heterozygosity), though a minority of patients may have a second independent point mutation, and an even smaller fraction a second independent small deletion.

Mutations in the rb1 gene seem to occur throughout the gene. The majority of the mutations are nonsense, introducing a premature stop codon and therefore producing a truncated and presumably nonfunctional protein.

While much work has focused on the rb1 gene mutations, it seems that retinoblastomas, like other cancers such as colon carcinomas and glial brain tumors, require other genetic abnormalities to occur prior to full transformation and development of the tumor. A rare benign clinical entity called a retinoma is thought to be the result of the loss of the rb1 gene without the subsequent acquisition of other mutations that allow progression to the full-blown malignant state. It is unclear what these subsequent steps in the development of the retinoblastoma tumor are, but cytogenetic studies have revealed some consistent abnormalities that provide clues. Squire noted gross aneuploidy of chromosome 1q(1q1) was present in 21 of 27 tumors studied.10 Additionally, a cytogenetic abnormality specific to the retinoblastoma tumor was noted, isochromosome 6p, or 6p1, in 15 of 27 cases. Which genes present at these loci are responsible for malignant transformation is at present unclear.

Other known oncogenes have been investigated in retinoblastoma tumors. Doz and colleagues were stimulated by the histologic similarity between retinoblastoma and neuroblastoma to investigate whether two genetic abnormalities common in stage 4 neuroblastoma (n-myc amplification and loss of material from the short arm of chromosome 1) are also common in retinoblastoma.11 They determined that only 1 of 45 retinoblastoma tumors had amplification of the n-myc proto-oncogene but did note that while only 9 of 43 primary retinoblastoma tumors had loss of heterozygosity of 1p, a higher proportion of distant metastases did, suggesting that a tumor suppressor gene on this arm may contribute to the metastatic potential of retinoblastomas.

While the amount of progress in our understanding of the molecular mechanisms underlying retinoblastoma development is impressive,12,13 it is important to put this into perspective and remember that the question that the patient’s family asks the physician is, why does my child have this disease? In the 90% of the cases without an antecedent family history, the reason for the rb1 gene mutation is still unclear, and therefore future discoveries that may explain the mechanisms responsible for the gene mutations are eagerly awaited.

Presenting Signs and Symptoms of Retinoblastoma

The presenting signs and symptoms of retinoblastoma vary depending on where in the world a child with retinoblastoma is seen. In the developing countries, children often present with extraocular disease: proptosis and orbital mass when the retinoblastoma has grown within the eye causing rupture of the globe and direct extension into the orbit (Fig. 81.2). Regional node metastasis may be found in the preauricular or submandibular nodes. These children are older (age 4–6 years), and few survive. In the United States, most children present with signs rather than symptoms.

Figure 81.2. Advanced orbital presentation of retinoblastoma (Courtesy of A.

Figure 81.2

Advanced orbital presentation of retinoblastoma (Courtesy of A. Wachtel, M.D., Lima, Peru)

The most common sign (60% of cases) is leukocoria, the term applied to a white pupillary reflex or cat’s eye reflex (Fig. 81.3).14 The reflex is caused by the tumor itself in the vitreous or by the retinal detachment caused by the underlying tumor.

Figure 81.3. Leukocoria (white pupillary reflex) caused by retinoblastoma.

Figure 81.3

Leukocoria (white pupillary reflex) caused by retinoblastoma. The tumor can be seen in the vitreous. There are seeds in the anterior chamber, anterior to the iris.

The second most common sign is strabismus or misalignment of the two eyes. Of the 22% of patients that so present, half have eyes crossed in (esotropia) and half out (exotropia). Esotropia in children, in general, is more common than exotropia so that an infant with exotropia must be feared to have retinoblastoma until proven otherwise. The crossed eyes are caused by tumor or retinal detachment in the area of central vision, the macula.

The next most common sign is painful glaucoma with inflammatory eye signs. These children may present to the pediatric emergency room with a picture-like orbital cellulitis; they appear systematically ill with irritability, failure to eat, and even low-grade fever. Although the clinical examination and diagnostic imaging may suggest extraocular disease, it is curious that these children usually have only intraocular disease, frequently with massive necrosis.

Other symptoms in the United States include anisocoria (different sized pupils), heterochromia (different colored irides), hyphema (blood in the anterior chamber), tumor hypopyon (tumor in the anterior chamber) and nystagmus. Less than 5% of our patients were detected on routine pediatric screening by a pediatrician. More than 90% of cases are detected by the mother.

When retinoblastoma is detected in the first 6 months of life, the patterns are different.15 Sixty percent of such patients have no signs or symptoms; they are examined because of a family history of retinoblastoma. Of those retinoblastomas we discovered in the first 3 months of life, none were referred with strabismus or inflammation, and only 30% had leukocoria.

The anatomic location within the retina and the age at which these tumors are found has been well studied(Figure 81.4). Retinoblastoma may be present anywhere in the retina at birth, but clear patterns were seen in time and location. By age 9 months, 50% of bilateral tumors were identified. By 2 years, 89% were diagnosed. There is a direct relationship between the age at tumor diagnosis and retinal topography. This relationship follows a central to peripheral distribution with macular tumors presenting earliest and peripheral, anterior tumors appearing later. While the average macular tumor was diagnosed at 5.6 months and none presented after 15.5 months, peripheral tumors were diagnosed at an average of 16.4 months and as late as 8 years.

Figure 81.4. Artisic rendition of relationship between age at diagnosis, surface area of involved retina and location of intraocular foci of retinoblastoma.

Figure 81.4

Artisic rendition of relationship between age at diagnosis, surface area of involved retina and location of intraocular foci of retinoblastoma.

The Reese-Ellsworth Classification scheme is the most commonly used for describing intraocular tumor.16 It is not a true staging scheme, for untreated patients do not progress from group I to higher groups, but it has served as an excellent ocular reference for comparison of different series and treatment schemes. Since that scheme does not deal with extraocular disease, we have created another extraocular classification. Both schemes are outlined in Table 81.3.

Table 81.3. Intraocular and Extraocular Classification of Retinoblastoma.

Table 81.3

Intraocular and Extraocular Classification of Retinoblastoma.

The diagnosis of retinoblastoma is usually made on examination of a child by indirect ophthalmoscopy with scleral indentation, under anesthesia, if necessary.17 The tumor begins as a glassy hemisphere in one or multiple sites, bulging from the retina. With time, the tumor becomes pink and vascularized and grows. As it grows, it may detach the retina and/or break apart, causing characteristic seeds within the vitreous and beneath the retina. Depending on the stage of disease progression, it may look similar to other ophthalmic conditions that are not malignant. Extensive lists of these simulating lesions, called pseudogliomas in ophthalmic texts, are presented elsewhere.18 The most common and difficult ones are presented in Table 81.4 and are discussed below.

Table 81.4. Lesions Simulating Retinoblastoma.

Table 81.4

Lesions Simulating Retinoblastoma.

Lesions Simulating Retinoblastoma

Astrocytic hamartomas may be seen in children with the syndrome of tuberous sclerosis or as incidental findings. If they have the syndrome, the children may have intracranial calcifications, seizures, delayed development, mental retardation, dermal adenoma sebaceum and characteristic ash-leaf skin findings. With time, astrocytic hamartomas become more defined, may increase in number, and calcify in part or completely. They require no ocular treatment.

The lesions of toxocara canis represent the retinal eosinophilic abscess(es) of presumably dead, migrated second-stage larvae of toxocara canis, the same round worm that causes the clinical syndrome of visceral larval migrans. Curiously, intraocular lesions are usually not found in the patients who have the full-blown syndrome. Children are diagnosed at an average age of 6 years (later than retinoblastoma) and get the disease by ingesting eggs in feces from puppies who are in the first year of life. It is a congenital infection of dogs and more prevalent in the warmer parts of the United States.

Coats disease is a purely ocular condition that mostly mimics retinoblastoma. It represents a unilateral (90%) retinal vascular anomaly of boys (80%), characterized in early stages by localized retinal vascular light bulb–like telangiectasia. With time these vessels leak an exudate that becomes rich in cholesterol crystals, macrophage laden, that detaches the retina simulating retinoblastoma. While it is not a malignancy it usually blinds the affected eye and may be very difficult to differentiate clinically from retinoblastoma.

Retinopathy of prematurity (ROP), formerly called retrolental fibroplasia (RLF), is a cicatricial disease of the vitreous and retina seen in low-birth-weight children who received oxygen at birth. It has also been seen in term babies who required no oxygen. It is usually bilateral, and both eyes are small (microphthalmic) and myopic.

Persistent hyperplastic primary vitreous (PHPV) is the term given to congenital findings of one eye that is smaller, contains a membrane that is vascularized behind the lens simulating a cataract, and usually has many other anomalies including abnormal iris blood vessels, a vascular stalk extending from the optic nerve to the back of the retrolental membrane, and anomalous formation of some of the layers of the retina. Though congenital, it is not heritable, and PHPV may stimulate retinoblastoma.

Diagnostic Tests

Needle biopsies are rarely, if ever, indicated in retinoblastoma. More than 50 years ago it was demonstrated that planned or unplanned puncturing of the eye allowed tumor cells to seep out of the eye and cause orbital invasion and death. As a result of this, many ancillary tests have been pursued to help the clinician make a correct diagnosis. While these tests may be helpful in some cases, in some institutions, they are not considered routine. As recently as 15 years ago, it was reported from some centers that as many as 20% of eyes enucleated for retinoblastoma did not contain this tumor. In our institution, no eyes that did not have retinoblastoma have been enucleated in more than 20 years.

Skull radiographs demonstrate intraocular calcification in 75% of cases of retinoblastoma. They have been supplanted by modern computed tomography (CT) scans in which more than 90% of retinoblastoma patients demonstrate intraocular calcification. Unfortunately, the following conditions in the pediatric age group also demonstrate CT calcifications: Coats disease, toxocara canis, retinopathy of prematurity, astrocytic hamartomas, PHPV, intraocular hemorrhage, and phthisis after trauma, infection or surgery. Fortunately, experienced radiologists can usually differentiate these lesions on the basis of topography, laterality, size of the globe, and associated findings. It should be emphasized that not all intraocular retinoblastomas demonstrate intraocular calcification. We have never seen an eye with complete replacement of the vitreous cavity that did not have tumor calcific action. The CT scan is unreliable for determining extension of the tumor into the choroid, optic nerve, or trans-sclerally.

Magnetic resonance imaging (MRI) for retinoblastoma has distinct disadvantages and advantages. Because it is difficult to detect calcification by MRI, the most diagnostic feature of retinoblastoma cannot be demonstrated. On the other hand, MRI has been the test that clinches the difficult clinical differentiation of retinoblastoma from Coats disease. Retinoblastoma is darker on T2-weighted images, whereas Coats disease is brighter, due to the proteinaceous exudate.

Ophthalmic ultrasonography, performed by the ophthalmologist at the same time as the examination without anesthesia, is extensively and routinely used. It demonstrates masses with high reflectivity that block sound, causing characteristic shadowing behind the tumor.

There has been great interest in aqueous taps for retinoblastoma.19 We no longer use this test because it is invasive and carries with it the risk of spreading the tumor. However many enzymes of the Embden-Myerhof pathways are in very high levels in the aqueous and vitreous of patients with retinoblastoma. Enzymes studied include lactic acid dehydrogenase, phosphoglucose-isomerase, and gamma-gamma enolase (or neuron-specific enolase). We have recently demonstrated high levels of other substances, including uric acid, Vanillylmandelic acid (VMA), homovanillic acid (HVA), and 3-methoxy, 4-hydroxyphenyl-glycol (MHPG).

Treatment of Retinoblastoma

The treatment of retinoblastoma can be divided into treatment of the intraocular disease20–22 and treatment of extraocular disease.21

Intraocular disease

Exenteration refers to the surgical removal of the eye and lids, orbital portion of the optic nerve, and all orbital tissue including extraocular muscles, fat, nerves, and muscles. It is rarely, if ever, used for retinoblastoma nowadays. Disease that is extensively within the orbit will not be cured by this technique, and excellent local control in cases like that can be obtained with external beam radiation and/or systemic chemotherapy. In cases of superimposed life-threatening orbital infection or bleeding, however, it may still be appropriate.


Enucleation refers to the surgical removal of the eye, leaving behind the lids, extraocular muscles, but removing as long a portion of the optic nerve within the orbit as possible. In children, the procedure is done under general anesthesia, though the children do not require overnight hospitalization. In adults, the procedure can be done under local anesthesia.

Clear consent must be obtained. Rather than having a distressed family sign permission for “enucleation OD” we write, “surgical removal of the complete right eye” and add “with permission to remove or biopsy any suspicious tissue behind the eye.” We always dilate both eyes of the patient on whom we are operating to inspect and confirm which eye is to be removed. Enucleation must be done with skill, getting a long stump of optic nerve and making sure not to perforate the globe.22 A ball is placed where the eye was (silicone, plastic), and 3 weeks later, a thin contact-like prosthesis is molded and painted on by an ocularist to match the fellow eye.

External beam irradiation

Radiation therapy for retinoblastoma, in the unenucleated eye is designed and prescribed to encompass the entire tumor bearing portion of globe and at least 1 cm of optic nerve, while avoiding as much normal tissue as possible, including the very radiosensitive lens (and lacrimal gland, if possible). For children with bilateral disease, the conventional parallel opposing fields from left and right are designed with a “D”-shaped block.23 Using a CT simulator or plain films taken in the conventional radiation therapy simulator unit, and information about globe size and lens position derived from a head CT or MRI study, the field is designed with the straight edge of the “D” block located at the posterior pole of the lens. The curved portion of the “D” is shaped to encompass the eye as described above, with a margin of about 1 cm in all directions. For children with bilateral disease where one eye has been enucleated, we use a similar field arrangement involving a single field or bilateral fields, with the empty orbit receiving some exit dose of radiation. For patients with unilateral disease, we have designed a technique described in detail elsewhere, which employs two lateral oblique photon fields with wedges and one lateral electron field, to spare the contralateral globe.24

The dose prescribed to the retinal target volume ranges from 4,200 to 4,600 cGy, with the lower dose reserved for children under the age of 6 months, and the higher doses for children with advanced bilateral disease. Children are treated 5 days/week, with daily fractions of 180 cGy per day. Although occasionally sedation is necessary for immobilization of a child, a plaster of Paris cast of the child’s head and shoulders, bi-valved and molded to a restraining board which swaddles the baby’s or the child’s body, may actually induce a sense of security and sleep, and can serve as the immobilization technique.

Following radiation, frequent follow-up visits with the ophthalmologist are scheduled for examinations under anesthesia to assess tumor regression. Regression patterns have been described elsewhere in detail.

Of note, even after treatment, some scarring and residual calcification are the rule rather than the exception in these children’s eyes. After 10 years, 90% of treated tumors still have visible intraocular material, when viewed with the ophthalmoscope, ultrasound, or CT scan.


Since the late 1950s, thanks to the pioneering work of G. Meyer-Schwickerath,25 we have been able to photocoagulate retinoblastomas successfully. A comprehensive review of the subject is presented elsewhere.3 Traditionally, we focus light through the dilated pupil under anesthesia and burn the tiny blood vessels that supply the retinal tumors. When the blood supply is destroyed, the tumor involutes and is permanently removed. Photocoagulation has few side effects and causes complete remissions of treated tumors that are no larger than 3 mm in diameter. Most tumors require more than one session to be cured. In recent years, we have been able to treat tumors directly with lasers in the visible [argon] or invisible wave length [diode, infrared laser] with success.


In the late 1960s Lincoff demonstrated that cryotherapy could effectively destroy small retinoblastomas.26 There has since been extensive experience with this technique.27 Under anesthesia, a pencil-like blunted probe is precisely placed on the outside of the sclera directly behind an intraocular focus of retinoblastoma. Rapid freezing (faster than -90°C a minute) causes intracellular ice crystals to form. These crystals cause rupture of tumor cells in addition to vascular occlusion and are locally successful in more than 90% of tumors, when the tumor is < 3 mm in diameter.


Brachytherapy for retinoblastoma has been in use since the 1930s when it was devised and refined by the legendary British ophthalmologist Henry Stallard.28 Though the technique has remained similar, there has been an evolution of isotopes in recent years. Stallard used 60Co plaques, but in recent years,125I plaques have been employed. Use of this isotope allows for a custom-built plaque for size of the lesion, dose, and duration of application and is also advantageous to both patient and staff from the aspect of radiation safety. Plaques of radioactive ruthenium, gold, palladium and strontium have all been used with success. After localization of the tumor in the operating room, the plaque is sutured on the outside of the sclera overlying the intraocular tumor. The dose prescribed is 4,000 to 4,500 cGy at a rate of approximately 1,000 cGy per day. In a second operation, the plaque is removed. The response most commonly seen is a type 4 regression pattern.29 When plaques are used after failure of all other conservative measures, success rates of 50% are still attainable.30

With increasing awareness of the late side effects of external beam treatment in very young children, and the increasing use of systemic chemotherapy to shrink initial tumor volume, radioactive plaque therapy is used increasingly as primary local therapy, in place of the external beam treatment in appropriate patients. Dose and technique are as described above.

Systemic chemotherapy

Systemic chemotherapy for intraocular disease was introduced in the 1950s. Over the years, there have been waves of enthusiasm for different agents. It has been recognized that single or multiple agents cause dramatic reduction in the size of intraocular tumors but not permanent responses. As a result of this, chemotherapy is presently being investigated for chemoreduction of tumors; when the tumor shrinks, it is then treated with an additional modality, such as photocoagulation, hyperthermia, cryotherapy, or radioactive plaques, which appears to cause permanent inactivation of the tumor. This is an area of active clinical and research interest in the hope of replacing external beam radiation, and because of the hope that eyes that might previously have been enucleated could be locally cured with this newer approach.

Most recent work has explored the use of vincristine, carboplatin, and an epipodophyllotoxin, either etoposide or teniposide.31–33 Investigators from Toronto have added cyclosporine as a P-glycoprotein inhibitor and have suggested that they have achieved a better outcome.34 It is important to recognize several limitations to the recent literature. First, the use of chemotherapy as an alternative to radiation therapy is based largely on the hypothesis that the chemotherapy will have less long-term morbidity and mortality in retinoblastoma survivors. This cannot be known until many years of follow-up have been performed. However, it is important to recognize that alkylating agents have been demonstrated to increase the risk of secondary bone tumors in survivors of childhood cancer.35 Teniposide is well known to induce secondary leukemia, and platinum-based chemotherapy has recently been demonstrated to increase the risk of secondary leukemia in survivors of ovarian cancer.36 Therefore, the use of chemotherapy may prove to have significant long-term toxicities that will have to be introduced into the risk to benefit ratios that will guide the decision whether to use chemotherapy or radiation therapy for these patients. Secondly, while it is clear that retinoblastomas are chemotherapy-sensitive tumors, which patients should receive chemotherapy and what the optimal regimens are remain to be defined.

Extraocular disease

Historically, chemotherapy has been reserved for the treatment of extraocular disease, but precisely which patients deserve treatment is still controversial (Table 81.5), and the optimal agents and regimens to use are not well established. In the developed world, extraocular disease occurs only in a small minority of patients. It is seen in both unilateral and bilateral cases, but presents sooner after diagnosis in patients with unilateral disease (2.7 – 3.2 months) than in patients with bilateral disease (11.4 – 11.7 months).37

Table 81.5. Summary of Recent Reports of Chemotherapy for Extraocular Retinoblastoma.

Table 81.5

Summary of Recent Reports of Chemotherapy for Extraocular Retinoblastoma.

The literature regarding efficacy of various chemotherapeutic agents to treat retinoblastoma is relatively sparse. Alkylating agents are felt to be the most active agents.38 Nitrogen mustard and triethlyenemelanamine(TEM) were frequently used in the past, and cyclophosphamide and ifosfamide are currently thought to be the most active single agents. Other reportedly active single agents include vincristine, doxorubicin, and cytarabine: idarubicin has recently been determined to be highly effective in a phase II trial.39 Trials utilizing single-agent carboplatin have not been reported, but our unpublished experience clearly shows this to be an active agent in the context of intraocular disease. Most reports of multi-agent regimens have utilized cyclophosphamide and vincristine. Other agents used in multi-drug regimens include cisplatin, carboplatin, and etoposide.

Extraocular disease can be divided according to the sites of involvement (see Table 81.3). The natural course of untreated retinoblastoma is one of progressive, localized ocular involvement, with eventual extension into the brain along the optic nerve or extension into the overlying conjunctiva with spread to regional nodes. This pattern is still common in the developing countries, but treatment of the intraocular disease significantly alters this pattern. Since unilateral patients and bilateral patients, before and after intraocular treatments have different patterns and timing of spread, it has not been possible to develop a useful staging system for extraocular disease. We have, therefore, utilized a classification scheme as a guide for treatment (see Table 81.3). Retinoblastoma can spread via several routes. It may grow contiguously through the choroid and into the sclera (stage I) and then into the orbit (stage II) or may grow back through the optic nerve and then invade the brain (stage III). It may enter the subarachnoid space and then spread throughout the leptomeninges via the cerebrospinal fluid (stage IV). It may also spread hematogenously, causing metastatic disease in the bone marrow, bone, and organs such as the liver and spleen (stage V). Rarely, retinoblastoma may spread through the lymphatics to produce cervical node disease (stage IIb), but of note, the eye only has lymphatic drainage through the conjunctiva. An algorithm for the management of extraocular retinoblastoma is shown in Figure 81.5.

Figure 81.5. Algorithm for management of extraocular retinoblastoma.

Figure 81.5

Algorithm for management of extraocular retinoblastoma. *Controversial.

Optic nerve/choroid invasion

When eyes containing retinoblastoma are enucleated, the extent of disease present in the pathologic specimen has been used to determine the need for chemotherapy.40,41 There are a number of such analyses in the literature, with some lack of agreement about which criteria indicate a high risk of micrometastases and the presumptive need for chemotherapy. Two factors extensively discussed are whether optic nerve invasion is present and whether choroidal invasion is present. If optic nerve disease is present at the cut end (positive margin), there is little controversy that central nervous system (CNS) spread is likely, and that treatment is indicated. More commonly, however, invasion of the nerve is noted with a surgical margin free of tumor. An important landmark used to define the extent of optic nerve invasion is the lamina cribrosa, which is the extension of sclera at the site of the optic nerve. Of patients with optic nerve invasion beyond the lamina cribrosa but a negative surgical margin, 30 to 40% will later develop metastatic disease.40–42 Invasion of the choroid only, without associated optic nerve disease, is questionable as a prognostic factor for metastatic disease. Patients with both invasion of the optic nerve beyond the lamina cribrosa and massive choroidal involvement deserve to be treated with chemotherapy, and it has been suggested that they be investigated on a multi-institutional basis.

Whether a 30 to 40% risk of metastasis justifies treating the 60 to 70% of patients who would not develop metastases is a difficult question. The toxicity, expense, and success of treating needs to be considered in conjunction with the efficacy of therapy, once overt metastatic disease is detected. If a significant proportion of patients with recurrent or metastatic disease can be salvaged with contemporary therapies, it may be reasonable to follow up patients with optic nerve invasion beyond the lamina cribrosa, but with a negative surgical margin, and to reserve aggressive chemotherapy for the minority whose disease recurs.

Orbital disease

Until recently, disease extending into the orbit was uniformly fatal despite aggressive surgery and radiation therapy. In the past few years, however, reports of the successful use of aggressive systemic chemotherapy in conjunction with radiation therapy have appeared in the literature.43,44 Doz and colleagues reported 33 patients with histologically proven orbital disease treated between 1977 and 1991. Twenty patients had isolated orbital disease, while 7 also had CNS metastases and 6 had metastatic disease outside the CNS. Most patients received both intensive chemotherapy and orbital radiation. Several chemotherapeutic regimens were used, including cyclophosphamide, platinum compounds, etoposide, doxorubicin, and vincristine. An overall survival of 34% (± 68%) was noted, with most recurrences occurring within the first year following diagnosis of the orbital disease. Patients without CNS disease fared better, and there was a trend toward improvement in the outcome of the more recently treated patients (post-1985). While survival is clearly possible, orbital disease still carries a high risk of mortality, and aggressive multi-modality therapy is warranted. It is also important to note that several groups have reported good outcomes for patients with local nodal disease extension. These patients should not be lumped together with patients who have CNS or hematogenously spread distant metastases.

Central nervous system/distant metastases

Metastatic disease to the CNS or to distant sites was considered incurable until recently. Grabowski described 16 children with CNS or hematogenous metastases and showed that aggressive treatment with chemotherapy and radiation could cure a substantial proportion of these patients.37 Of the patients who received treatment at the New York Hospital, all 5 with intracranial disease were disease-free survivors, as were 4 of 6 patients with bone and bone marrow disease. Treatment included systemic chemotherapy with cyclophosphamide, doxorubicin, and vincristine, intrathecal methotrexate and cytarabine, and, in patients at least 1 year old, whole brain radiation (1,800 cGy) plus involved field boosts to a total of 4,500 to 5,500 cGy. In contrast, however, Schvartzman had poor results using a similar approach in children with distant metastases45 and Doz and colleagues reported that all 7 of their patients with CNS involvement in addition to orbital disease died despite aggressive therapy.

Even if the promising results of Grabowski are replicated, the young age of most of these patients makes the use of craniospinal radiation extremely toxic to normal development, and severe developmental delay and intellectual impairment are frequent sequelae. This same problem has been addressed by pediatric oncologists for the treatment of young children with primary brain tumors. A strategy of utilizing high-dose chemotherapeutic agents that penetrate the blood-brain barrier has been employed to delay or avoid the use of radiation. We and others have used high-dose systemic chemotherapy in conjunction with autologous stem cell rescue for patients with CNS or metastatic retinoblastoma with promising results. Namouni treated 25 patients with high-risk retinoblastoma with high-dose carboplatin, etoposide, and cyclophosphamide, followed by autologous stem cell rescue. Five of 11 patients with metastatic disease not involving the CNS were event-free survivors.46 This strategy (without intrathecal medications) has been used to treat brain tumors that have a propensity to disseminate along the leptomeninges (such as medulloblastoma), and therefore the need for intrathecal therapy for retinoblastoma is unclear. Additionally, the role of intrathecal therapy must be questioned because of the in vitro data that suggest that retinoblastoma cell lines are invariably resistant to methotrexate, though some were sensitive to cytarabine.47

Second Malignancies

Since the first report of the successful treatment of bilateral retinoblastoma with combinations of surgery and irradiation,48 it has been recognized that some retinoblastoma patients develop second nonocular cancers years after the successful treatment for the eye neoplasm. Our center has repeatedly published on the largest such experience with the longest follow-up; we will emphasize the highlights of our findings here, but those interested in more details are encouraged to review the primary references.49–55

Only retinoblastoma patients with the germinal form of the disease are at increased risk for the development of second tumors. Unilateral retinoblastoma patients without the constitutional mutation of one allele are at no higher risk for developing cancer, even if treated exactly as a germinal retinoblastoma. In particular, unilateral patients without the mutation are at no higher risk than nonretinoblastoma patients, if they receive external beam irradiation, brachytherapy, photocoagulation, cryotherapy, or enucleation.

Patients with the germinal mutation are at risk for the development of second cancers. This represents any bilateral patient, whether or not that patient has a family history of the disease. It also represents virtually all the unilateral patients who have a family history of the disease. The risk also attends any unilateral retinoblastoma patient who has been shown by molecular techniques to have a germinal mutation. On clinical grounds we have found that unilateral patients who are diagnosed at a young age (under 6 months) and those who present with multi-focal unilateral tumors (not multiple tumors from seeding) are likely to have the germinal mutation and most of these are at risk for the development of second cancers. Since 100% of patients with germinal mutations have not yet developed second neoplasms, there may be other factors, as yet unknown, that also govern their development.

The cumulative incidence of second cancers is 1% per year, reaching 51% (+/-)6.2%, 50 years after the diagnosis of retinoblastoma. These cancers are more frequent in girls than boys. The cancers that are more common are (in descending order) sarcomas of bone, soft tissue sarcomas, pinealomas, cutaneous melanomas, brain tumors, Hodgkin’s disease and carcinoma of the female breast.50

Hereditary retinoblastoma survivors are also at significantly increased chances of developing lipomas (RR=8.2). Those who developed lipomas were at an even higher risk for developing second cancers.52

In addition to having the abnormnal gene and developing lipomas, two important studies have demonstrated the impact of radiation dose and timing. A dose response curve for the development of soft tissue sarcomas in bilateral retinoblastoma patients has been described. Odds ratios at 0 to 4.9 Gy were 1.0, at 5 to 9.9 Gy were 1.91, at 10 to 29.9 Gy were 4.6, at 30 to 59.9 Gy were 6.4, and >60 Gy were 11.7.54 The age at radiation also affected the incidence of second tumors. Children radiated in the first 12 months of life were twice as likely to develop second cancers than those radiated after the age of 1 year, suggesting that radiation was significantly safer, if administered after the age of 1 year.55 In the radiated patients, the second malignancies were in the radiation field two-thirds of the time and out of the field one-third of the time. In the nonradiated patients, the tumors were out of what would have been the field two-thirds of the time and in the field one-third of the time. Thus, some patients who had not been treated with radiation developed second malignancies in the field of radiation.

The cumulative probability of death from second neoplasms was 26% at 40 years; death was more frequent in females (RR-5.39) than in males (RR-5.22). The risk in germinal cases increased by irradiation was also apparent in unilateral retinoblastoma patients with the germinal mutation (53). Some of the lifelong risk is manifest in tissues during expansive growth. Osteosarcoma of the skull usually develops in the first 10 years of life, whereas osteogenic sarcoma of the long bones develops during teen-age years. The risk of acute leukemia, the most common pediatric cancer, is not increased.

An important and curious second cancer is now recognized. It has been referred to as trilateral retinoblastoma because it occurs in the primitive third eye, the pineal gland, in patients who have bilateral retinoblastoma. Typically, it has occurred in patients with bilateral retinoblastoma who have a family history of the disease, who have been diagnosed early in life (within the first 6 months), and who have been treated with radiation. All such children have died from their pinealblastoma with leptomeningeal spread, despite a variety of attempts to cure the tumor.

The most common cause of death of a retinoblastoma patient with the germinal mutation is not retinoblastoma but a second cancer. Pinealblastoma is the most common cause of death in the 5- to 10-year period (10%).56


Childhood leukemia, particularly acute lymphocytic leukemia (ALL) is the most common malignant tumor that involves the eyes of children. Leukemic infiltrates involve the eye in three distinct presentations of leukemia. In all cases, the site of involvement appears the same. Leukemic involvement primarily involves the uveal tract: the iris, ciliary body, and/or choroid. Leukemic infiltration is virtually impossible to detect ophthalmoscopically in the choroid or ciliary body. When present in the iris, the child may have iris infiltrates causing heterochromia (different color of the two irides), cells in the anterior chamber (mimicking and, in some cases, treated as idiopathic iritis), or bleeding in the anterior chamber (hyphema) that may be associated with glaucoma with a painful, photophobic, red, sensitive eye.

At autopsy, after death from leukemia, tumor has been identified in the choroid in 90% of eyes. There are no good clinical descriptions of its appearance, however. Tumor in the choroid is not seen because ALL diffusely invades the choroidal blood channels, and even with extravasation beneath these vessels, the only effect is that the choroid is thicker. B-scan ultrasound examination by an ophthalmologist, frequently performed at the bedside can reveal thickening of the choroid.57 Retinal involvement, which is rare, has been described: Roth’s spots and white patches are seen.

The three situations when leukemia is seen within the eye follow:


Simultaneous with the first presentation of leukemia: the majority of eyes demonstrate ultrasonic findings. Occasionally, the effect of thrombocytopenia or sludging is noted with hemorrhages, infarcts, and dilated vessels. When leukemia is treated, the choroidal involvement disappears within days.


As a recurrence following induction treatment and CNS radiation prophylaxis: in these children, the CNS has been treated with radiation, but the eye has been spared, creating a sanctuary site. Although rare, treatment of the eye alone in such cases may be justified.


As a sign of CNS recurrence: This is by far the most important presentation because ocular recurrence is a marker for CNS recurrence whether or not CNS tumor can be identified. These patients frequently develop cells near the posterior pole of the eye. At times, the cells appear to be coming directly from the optic nerve into the vitreous, clouding the appearance of the optic nerve itself. It was formerly thought that ocular recurrence indicated systemic hematogenous recurrence, but recent evidence58 suggests that the disease in itself sometimes recurs because it is seeded by the CNS directly through the optic nerve. This is important because if ocular recurrence occurs without apparent systemic or CNS recurrence, the brain should be treated. The traditional treatment for ocular recurrence of leukemia had been radiation with 800 to 1,800 cGy, but we have found that treating the eye via a CNS route (ventricular methotrexate, through an Ommaya reservoir) quickly and effectively eliminates ocular recurrence.

Orbital Tumors


Benign tumors of the orbit in children are frequently incidental problems detected on CT scan or because lid or orbital asymmetry is observed. Many require no treatment.

The most common benign orbital tumor of childhood is capillary hemangioma (unlike the most common benign orbital tumor of adults, which is cavernous hemangioma).59 Capillary hemangiomas are highly cellular, angioblastic, and rarely encapsulated. As many as 25% may not be apparent at birth, but they grow so rapidly and frequently bleed that they may be mistaken for a rhabdomyosarcoma. Biopsy of rhabdomyosarcoma is always done to differentiate it from capillary hemangioma. On CT scanning, capillary hemangiomas are usually associated with congenitally enlarged orbits while rhabdomyosarcomas are not. Despite growth that may appear explosive, their natural course is to stop growing after 6 to 12 months when they will slowly involute (over 2 to 5 years) and largely regress. They may be associated with strawberry hemangiomas of the nearby skin.

These tumors usually cause profound and permanent visual loss, if left untreated. The mechanism of visual loss is not simply ptosis, with occlusion of the pupil and consequent development of deprivation amblyopia. The mass displaces the eye and causes strabismus with strabismic amblyopia. The mass usually presses on the eye itself causing significant astigmatism, with the development of anisometropic amblyopia. Rarely, hemangiomas cause corneal damage by proptosis and exposure, and they may press on the optic nerve and cause optic atrophy. If they sequester platelets, they may cause thrombocytopenia. They may even cause high-output cardiac failure.

Treatment of these hemangiomas is difficult. Surgery, either conventional or with the KTP (532 nm potassium titanyl phosphate) laser, rarely removes enough of the lesion to prevent ocular complications. The tumors respond to low-dose radiation, and we have used fractionated doses up to 800 cGy with success. There have been reports of injecting sclerosing solutions into the lesions or even occluding them with interventional radiologic techniques, but experience is limited.

Fortunately, these tumors respond to local or systemic steroids. Hemangiomas have responded to systemic steroids, but all have recurred when the steroids were discontinued after a few weeks. Systemic treatment must continue for months, and with the doses recommended for the young children (20 mg/d) systemic toxicity is common and worrisome. Local injections of short- and long- acting steroids are probably the treatment of choice, when mandated by visual or overwhelming cosmetic reasons.

Dermoid Cysts

Dermoid cysts are among the most common orbital tumors of childhood. They are benign and represent congenital ectodermal rests. They are congenital, usually in the anterior portion of the superior orbit and cause a characteristic hollowed out appearance of bone. Occasionally, the dermoids have calcification and even aberrant tooth formation within the mass. Treatment is surgical but utmost care must be taken because many have dumbbell-shaped posterior orbital extensions, sometimes intracranially. When the cysts rupture, either from trauma or from incomplete surgical removal, a violent orbital inflammation may ensue.


Despite the fact that lymphatics do not exist in the orbit, benign lymphangiomas do.60 These tumors are thought to be congenital, have no malignant potential but in contrast to hemangiomas are rarely present at birth. They are most commonly seen at age 6 years with explosive proptosis, caused by bleeding of the tumor into the cystic spaces referred to as chocolate cysts. These lesions consist of primitive lymphatic channels with lymphoid hyperplasia that may enlarge during upper respiratory infections. Differentiation from rhabdomyosarcomas is important. As many as 20% of lymphangiomas have lymphangioma tissue on the palate, and many have anterior, visible orbital extensions. Their diffuseness on imaging techniques is unlike the typical more localized rhabdomyosarcoma.

Treatment is difficult. They do not respond to steroids or radiation, and they are impossible to remove completely by surgery. Surgery with electrocautery, which may have to be repeated many times, with or without the KTP laser, is the treatment of choice. Satisfactory cosmetic results are difficult to attain.


The most common primary malignant orbital tumor of childhood is rhabdomyosarcoma. The average age at diagnosis is 6 years, and the sexes are equally affected, as are the two orbits. Although the hallmark of an orbital rhabdomyosarcoma is rapid, progressive, painless proptosis, the next most common finding is ptosis followed by a subconjunctival fleshy mass.61 The most common location in the orbit is a mass superonasally, and the most common histologic type is an embryonal rhabdomyosarcoma. When rhabdomyosarcomas present in the inferior orbit, they are usually histologically alveolar type. Interestingly, these malignancies do not originate from extraocular muscles but from skeletal rests within the orbit. CT scans may reveal extension into the sinuses. Rhabdomyosarcomas can originate from the sinuses and extend into the orbits or they may extend into the sinuses from the orbit. In both cases, if there is sinus involvement, prognosis for survival is worse.

Early biopsy is mandatory. All attempts should be made to biopsy the lesion directly without going through the sinuses or skull because of the possibility of tracking tumor.

Local excision of these tumors is rarely effective. Jones reported that exenteration offered a 3-year cure rate of 47%.62 External beam irradiation was started in the 1960s, and it was subsequently shown that irradiation combined with chemotherapy not only produced excellent local cures but better than 90% long-term survival since late metastases in orbital rhabdomyosarcoma do not occur. Radiation doses of 6,000 cGy were employed. The chemotherapy that was originally used—vincristine, cyclophosphamide and doxorubicin—has changed, as results of various intergroup rhabdomyosarcoma studies have demonstrated that the most effective combination for disease localized to the orbit is vincristine, actinomycin-D, and radiation.63 In giving radiation, the eye is not spared or shielded. Within 5 years, one-third of such eyes have been enucleated. Only 19% of such patients retain useful vision after 10 years, but the majority do keep their minimally sighted eyes.64

We recently reviewed the long-term experience in patients treated in the 1960s and 1970s with external beam irradiation alone.65 Their survival was similar to those patients treated with irradiation and chemotherapy. Interestingly, second nonocular tumors have occurred in rhabdomyosarcoma patients who received adjunctive chemotherapy but not in those patients who received irradiation alone.

Recurrent rhabdomyosarcoma following irradiation and chemotherapy occurs (though infrequently) 1 year after completion of therapy, with the same local signs (proptosis, ptosis, mass) and symptoms (diplopia, decreased vision) and poses a special problem for treatment. No long-term survivors had been reported after recurrence, but since 1988 we have used a system that is promising. To date, each of the five patients treated is alive and free of disease following exenteration and specialized brachytherapy. A mold of the orbit66 is made from an algae-based dental material which is then exactly replicated in plastic.125 I seeds are placed in holes drilled in the mold, which is left in situ until a total dose to the remaining microscopic tumor of 6,000 cGy is attained without significant re-irradiation of the brain.

While distant metastases from rhabdomyosarcoma are usually to the lungs, it is the absence of local spread to the sinuses that predicts success of treatment. If no sinus involvement is found at diagnosis, overall success is more than 90%; with sinus involvement (primary or secondary), success is almost halved. Metastasis from orbital rhabdomyosarcoma always occurs within 5 years from diagnosis.

The diagnosis of rhabdomyosarcoma should be considered whenever rapid progressive proptosis occurs in the first 20 years of life. Other conditions must be considered in the differential diagnosis, and they are listed in Table 81.6.

Table 81.6. Causes of Rapid Proptosis in Childhood.

Table 81.6

Causes of Rapid Proptosis in Childhood.

Adult Ocular Tumors


Benign tumors of the lid, conjunctiva, iris, and choroid are common, while those of the retina and cornea are rare. Benign tumors of the lens and vitreous do not exist. The most important benign tumors of the eye are the choroidal nevi.

Choroidal Nevi

Choroidal nevi are never present at birth. Pigmentation that looks like choroidal nevi in infancy is usually from choroidal neurofibroma, seen as a part of systemic neurofibromatosis. Nevi of the choroid can be seen before puberty, but they are unusual. With puberty, they become visible. In the United States, 10 to 13% of the adult population have choroidal nevi. They are racially related; choroidal nevi in African Americans are very rare.

Choroidal nevi are flat, pigmented, benign lesions with edges that can be feathered and irregular or rounded (Fig. 81.6). They are usually slate-gray to light-chocolate in color. With time (months to years), there may be associated findings with the nevi. Many demonstrate changes on the surface, such as drusen or subretinal fluid, and may cause overlying visual field defects; overlying neovascular membranes can appear. Since 10% of the adult population have choroidal nevi and there are only 1,500 choroidal melanomas in the United States yearly, it is assumed that the chance of a choroidal nevus becoming a melanoma is less than 1 in a 1,000. A number of studies have now demonstrated which nevi change into melanomas. The predictive factors are shown in Table 81.7.

Figure 81.6

Figure 81.6

Choroidal Nevus

Table 81.7. Predictive Factors: Nevus to Melanoma Transformation.

Table 81.7

Predictive Factors: Nevus to Melanoma Transformation.

The diagnosis of a nevus can be made with ophthalmoscopy alone. The development of retinal detachment over and around the tumor can be treated medically or with lasers. Many ophthalmologists treat symptomatic fluid with systemic acetazolamide. There are no controlled studies proving that the treatment works.

Photocoagulation with lasers of the leaking areas over the nevi has been done. While the fluid usually resorbs and leakage stops within days to weeks, it has been suggested that photocoagulation itself somehow weakens the layer between the retina and choroid (Bruch’s membrane) and, in some way, may stimulate transformation to malignancy.

A special type of nevus that is not flat has been described in the choroid. It is a melanocytoma, or pathologically a magnocellular nevus, with jet-black pigmentation. This consists of large cells in darker-skinned Caucasians, who are frequently of Mediterranean origin. These lesions originate in cells within the optic nerve itself and may obscure a view of the nerve. The lesions may be several millimeters high, grow slowly, and can affect the visual field or visual acuity. While these lesions are benign, rare cases of transformation to malignancy have been recorded.

The differential diagnosis of a nevus is straightforward. Other flat lesions in the choroid that have been confused with nevi and melanoma include hyperplasia of the retinal pigment epithelium (that lesion, also flat, is darker, almost black in coloration, frequently with bare spots devoid of pigment within the lesion that has very sharp edges and usually is circular in shape), hamartomas of the retinal pigment epithelium, and hemorrhages within the retina, especially hemorrhages beneath the retinal pigment epithelium (as part of macular degeneration). The differentiation of these lesions from melanoma is based on size, thickness, fluorescein patterns, and ultrasonography.

Iris Nevi

Iris nevi are, by definition, pigmented and flat. They are common, may be multiple, and occur more often in blue-eyed individuals. Iris nevi are also rarely present at birth, and like all other ocular nevi, become apparent around puberty. They are always benign and of no real consequence for the eye or for life. Confusion exists, however, in the use of the term iris melanoma. Ophthalmologists have traditionally described elevated pigmented lesions of the iris as iris melanomas to differentiate them from the flat nevi. This has been further confused by pathologic interpretation of elevated pigmented iris masses that were excised (or enucleated!) which were described as malignant. Elevated iris-pigmented masses may grow, may shed cells into the angle, clogging the trabecular meshwork and causing a severe secondary glaucoma that can blind the eye; but iris melanomas metastasize extremely rarely, if ever. Lack of metastasis may be a function of size. An iris melanoma that filled the anterior chamber would, if in the back of the eye (in the choroid) be small enough never to cause metastasis. Management of iris melanoma is based on what is best for the eye and the glaucoma and is not a decision about life. Ciliary body melanomas that present as iris lesions do metastasize, however.


The most common primary malignant tumor of the eye of adults is malignant melanoma, previously referred to as melanosarcoma of the choroid.67 There are about 1,500 new cases of choroidal malignant melanoma yearly in the United States. The average age at diagnosis is 55 to 65 years, with men and women equally affected and the two eyes equally susceptible. In the United States, 99% of choroidal melanomas originate in Caucasians. The most common reason for detection of the tumor is a routine examination (41%).68 Men more often present with symptoms, and when there are symptoms, the right eye is more often found to have the tumor. The most common symptom is a decrease in the peripheral visual field, followed by decreased vision. The lesion is not painful, unlike metastatic tumors to the eye, where pain is not unusual.

The visual field defect is characteristic. There is an absolute scotoma overlying the tumor, associated with a surrounding relative field defect that does not obey the horizontal meridian (as most ocular defects do) and does not observe the vertical meridian (as many CNS defects do).69

Melanomas of the choroid originate in melanocytes that normally lie within the choroid. The choroid, the layer between the sclera and retina, is a rich, high-flow syncytium of vascular lobules that not only supply blood to the photoreceptors (rods and cones) of the retina but also serve as a heat sink to dissipate heat energy liberated by absorbed visible light.

Whether melanomas of the choroid originate in nevi alone is not known, but patients with flat pigmented, untreated nevi followed up for more than 20 years have developed melanomas arising from the previously dormant lesion. The cause of choroidal melanomas is unknown but patients with melanosis oculi, nevus of Ota, plastic factory workers, and possibly HIV patients appear to be at higher risk for developing these tumors. We have also seen a high frequency in World War II holocaust survivors.

The diagnosis of choroidal melanoma can usually be made on ophthalmoscopic grounds alone. With the direct ophthalmoscope, it may be difficult to appreciate the three-dimensional shape, but with dilated pupils and the indirect ophthalmoscope, the tumor is easily identified. The tumor can have many shapes. A flat, diffuse type may be difficult to detect ophthalmoscopically, but most tumors are elevated, frequently dome shaped (the height of the tumor being half the base diameter), and occasionally multi-lobed. The tumor is held back by Bruch’s membrane, but when it ruptures this taut, transparent plastic layer it develops a rounded top on the surface of a dome-shaped mass; this is referred to as being mushroom shaped.

Pigmentation of choroidal melanomas varies from patient to patient and frequently from area to area within the tumor. As many as 40% of the tumors have no pigment clinically; they are frequently a dusky gray to charcoal in color, but occasionally they are deep brown. Black lesions within the eye are rarely melanomas.

All choroidal melanomas have associated retinal detachments. At times, it may be difficult to detect them ophthalmoscopically, while at other times, the retinal detachment may be so extensive that the melanoma is not seen (or suspected) clinically.

Since there are no lymphatics in the eye or within the orbit, melanomas of the choroid metastasize through the vascular channels. More than 75% of such metastases are first identified within the liver. Treatment strategies for metastatic melanoma are covered elsewhere. Fewer than 1% of patients with metastasis survive 5 years.

The diagnosis of melanoma is aided by fundus photography (Fig. 81.7) and fluorescein angiography, but the most commonly performed test is ocular ultrasonography. Typically the B-scan demonstrates an elevated solid tumor,70 and the A-scan demonstrates medium to low reflectivity.71 As recently as 1974, 20% of eyes enucleated for melanoma did not have a melanoma. In recent years, with standardization of criteria and careful attention to ocular ultrasonography the diagnosis has been shown to be accurate, without biopsy, in more than 99.5% of cases.

Figure 81.7

Figure 81.7

Melanoma of the choroid

Since clinical accuracy is so high, needle biopsy is rarely needed. Some centers have had large experience with needle biopsy and report high accuracy and few problems, but there has always been the concern that tumor cells might exit through the biopsy site, seed, and spread.

A number of clinical and pathologic features have been shown to correlate with patient survival.72


Size: The greater the size of the tumor, measured clinically using the height, and/or greatest base diameter to volume, the greater is the incidence of metastasis.


Location: Iris melanomas, as previously discussed, have the best prognosis. Ciliary body melanomas have a three-fold mortality, compared with choroidal melanomas.


Age: Patients younger than 60 years have better survival than those older than 60 years.


Extraocular extension: Patients with extraocular extension have mortality rates many times those who do not. The greater the amount of local extension into the orbit, the poorer is the prognosis.


Pathologic features: Many pathologic features (only available in cases where the eye has been removed) correlate with survival and are covered extensively elsewhere.72,73 The best known of these is cell type. Ocular melanomas that contain the plumper so-called epithelioid cells have a poorer prognosis.

The treatment of choroidal melanoma has received wide coverage in the ophthalmic literature.67,72,74–76 Most authors have classified choroidal melanomas as small, medium, and large. The actual definitions of these sizes vary despite the words, which makes it difficult to generalize about prior studies. The Collaborative Ocular Melanoma Study definition is seen in Table 81.8.

Table 81.8. Classification of Choroidal Melanomas.

Table 81.8

Classification of Choroidal Melanomas.

Meta-analysis reveals that patients with small melanomas have a 5-year survival of over 90%.75 Patients with medium-sized melanomas sustain a 5-year survival of 67%.75 Patients with large-sized melanomas have a 5-year survival of only 50%.75

A number of treatment options exist for choroidal melanomas.67,74,76 Although the best treatment is not known, the treatment options depend on the size of the tumor. Melanomas under 1.5 mm in height are not treated and probably have no metastasis nor mortality, if they stay that size. Melanomas between 1.5 mm and 10 mm in height have been treated with local excision, enucleation, or radiation with external photons, external protons, or brachytherapy. Melanomas larger than 10 mm in height are usually treated by enucleation. A nationwide prospective randomized trial of large melanomas showed similar survival in patients with large melanomas, when enucleation was compared with preoperative irradiation of 5 fractions of 200cGy external beam irradiation and enucleation within 72 hours.77

Radiation therapy for ocular melanomas has a long history. Brachytherapy for melanomas was first conceived 60 years ago by Henry Stallard.78 His original plaques utilized cobalt-60, and he reported that the majority of the tumors became flat. In recent years, cobalt has been largely abandoned because of safety issues, and several beta-emitting (strontium, ruthenium) or gamma-emitting (iodine, palladium, iridium) plaques have become available. In recent years, brachytherapy has been combined with concurrent hyperthermia. External photons have been used, and there has also been a large experience with helium ions and with protons.79 Some general conclusions about radiation therapy can now be made. About 70% of irradiated tumors decrease in size and there is a correlation between local control and survival. Tumors that continue to grow after treatment more commonly metastasize. Doses of 7,500 cGy to 10,000 cGy are used. The fractionation schemes vary markedly, without an apparent effect on either local control, metastasis, or complications. All radiation techniques have radiation complications.80 External irradiation is more commonly associated with anterior segment complications: dry eye, corneal complications, iritis, and cataract. Plaque therapy is more commonly associated with posterior complications, namely, radiation retinopathy and optic neuropathy.

There has been intense interest in finding out which of the therapeutic techniques is best for patient survival in the medium-sized tumor group.81 Recently, a 44-institution, prospective, randomized clinical trial (the Collaborative Ocular Melanoma Study) has closed to patient accruals in the United States and Canada. Medium-sized tumor patients were randomly assigned to enucleation or brachytherapy with125 I plaques at a tumor dose of 10,000 cGy. Patients are being followed up for recurrence and survival.

Metastatic Ocular Cancer (Adult)

The most common malignant neoplasm in the eye or orbit, in children or adults, is metastatic carcinoma to the choroid. For example, there are 350 cases of retinoblastoma yearly in the United States, 1,500 cases of choroidal melanoma, and over 30,000 cases of metastasis to the eye.

Metastasis to the eye most commonly occurs in adults between 55 and 65 years (the same distribution as ocular melanomas). The patterns in men and women are different. When men present with ocular metastasis, they are usually unaware of a primary neoplasm. The decreased vision caused by the metastasis is their first sign of cancer. In women, the diagnosis of cancer is usually well known and they may even be aware of other metastases besides that to the eye. To a large part, this has been because the metastases in men are usually from lung cancer, while in women, they have usually been from carcinoma of the breast. As lung cancer has now surpassed breast cancer as a cause of cancer deaths in women, it is anticipated that their pattern will become more similar to that of men. Many other cancers metastasize to the eye, including gastrointestinal, prostate, (though more commonly to orbital bones), thyroid, and ovarian cancers, cutaneous melanoma, and sweat gland cancers, to mention a few. Virtually all cancers have been found capable of metastasizing to the eye.

Most cancers metastasize to the uveal tract, but metastases to the lids, conjunctiva, optic nerve, orbit, extraocular muscles, and orbital bones are well known. Metastasis to the retina is rare. While metastases to the iris and ciliary body are not unusual, it is metastasis to the choroid that represents most of the metastases to the eye.

Choroidal metastases are usually amelanotic (whereas ocular melanomas are usually pigmented), multiple (whereas ocular melanomas are solitary in more than 99% of cases), bilateral (whereas bilateral ocular melanomas are extremely rare), minimally elevated (whereas most ocular melanomas are many millimeters high), and, when situated around the optic nerve or invading the sclera, painful (whereas ocular melanomas are painless in nearly all cases). Metastatic tumors, like melanomas always have an associated serous detachment, but the amount of detachment is proportionally greater in cases of metastasis. Ultrasonographically, most metastases have high reflectivity on ultrasound, in contrast to melanomas which usually have low to medium reflectivity.

In melanomas, most patients are detected as a result of routine examination; in metastases, the serous detachment causes decreased visual field and diminished acuity, so patients are usually seen because of symptoms. Metastases can be detected on routine examination in asymptomatic patients, however.

The most striking feature of metastatic ocular lesions is their association with concurrent CNS metastases. While the true concordance of these two is unknown, it has been our experience that more than 75% of cases of ocular metastasis have concurrent, CNS metastases, though frequently difficult to demonstrate. It has been speculated, therefore, that some ocular metastases do not arrive through blood-borne routes, but that CNS metastases may actually seed the choroid via the subarachnoid space as occurs in childhood leukemia.

Treatment for ocular metastasis is considered when symptoms of diminished vision, pain, or diplopia demand it. Treating the ocular lesion rarely has an impact on survival, except in carcinoid metastases but may significantly alter the quality of life. Many ocular metastases respond to chemotherapy the way other metastases respond. Chemotherapy and/or hormonal manipulation may cause rapid regression of the tumor and of subretinal fluid. Patients with undiagnosed ocular metastases occasionally demonstrate coarse brown pigmentation on the surface of a regressed choroidal metastasis, following successful chemotherapy. External beam irradiation is also used to palliate symptoms from ocular metastases. Choroidal metastases are treated with lateral photon fields, similar to the technique used for bilateral retinoblastoma described above, to a dose of 3,000 cGy in 10 fractions over 2 weeks, sparing the lens and anterior segment. Complications, other than transient skin erythema, are rare. When metastases involve the anterior segment, anterior irradiation is necessary, and if the patient survives for long enough, complications, such as dry eye, cataract, and red uncomfortable eyes, can appear. Except for carcinoid and breast cancers, however, median survival in patients with metastatic choroidal lesions is just over 6 months.

Cancer-Associated Retinopathies (CAR)

Paraneoplastic visual loss is an autoimmune disorder characterized by visual loss in patients with cancer. It has been shown that autoantibodies cross-react with retinal cell antigens, giving rise to loss of vision and optic atrophy. Lung and prostate cancers, endometrial sarcoma, and lymphoma have been associated with the syndrome. Steroids have been tried without much success. Plasmapheresis and intravenous immunoglobulin have been reported to work in a few patients.82

Ocular Lymphoid Tumors

Intraocular lymphomas are increasing in incidence because of their association with the acquired immunodeficiency syndrome (AIDS). Both benign and malignant types occur. Benign reactive hyperplasia may diffusely involve the uvea and be difficult to diagnose without biopsy. It is not associated with systemic findings.

Primary malignant lymphomas of the eye, also called reticulum cell sarcoma or microgliomatosis, usually present with cells in the vitreous and may have associated retinal and optic nerve involvement. These patients frequently have CNS disease but rarely systemic disease. Diagnosis is made by vitrectomy. Treatment is accomplished with systemic steroids and radiation therapy of 2,400 cGy to the affected eye and/or chemotherapy as for intermediate-grade systemic lymphoma. There is controversy about whether to treat the brain in these cases when diagnostic spinal tap and MRI studies show no cancer. Eventual CNS involvement is very common, as is bilateral ocular disease. Median survival is 3.5 years and usually determined by the brain involvement. Cure is rare, as is systemic spread.

Malignant lymphomas of the uvea present as diffuse uveal involvement and are usually associated with systemic disease, with involvement of lymph nodes and viscera but rarely with CNS involvement.

Orbital Tumors83

The diagnosis of orbital tumors has undergone a revolution in the past 20 years as a result of the widespread use of ultrasonography, CT and MRI. Prior to that, virtually all cases of proptosis required biopsy, and it was not unusual to be unable to find a tumor. As a result of noninvasive imaging, the number of orbits that require biopsy has decreased, and the chance of finding the diseased area has become much higher. Fortunately, malignant tumors of the orbit are unusual. In most cases, they come from adjacent sinuses or from the overlying skin. Biopsy is rarely needed for definitive diagnosis, even in cases of metastatic disease, as in metastasis of breast cancer to the orbit. Malignant primary cancers of the orbit that do require biopsy and surgical management arise almost exclusively from the lacrimal gland. Finally, lymphomatous lesions of the orbit are common; definitive biopsy may be necessary and, at times, is the easiest way to establish pathologically the type of lymphoma. The clinical findings suggestive of an orbital tumor are listed in Table 81.9.

Table 81.9. Findings Suspicious of Orbital Tumor.

Table 81.9

Findings Suspicious of Orbital Tumor.

All cases of suspected orbital tumor should have imaging: ophthalmic ultrasonography, CT with or without contrast material, and MRI with or without contrast. Any of these techniques alone, or two of them in combination, will lead to the anatomic location of the tumor. An algorithm for the differential diagnosis and treatment of orbital tumors is shown in Fig. 81.8

Figure 81.8. Algorithm for diagnosis and treatment of orbital tumors.

Figure 81.8

Algorithm for diagnosis and treatment of orbital tumors.

Bone Lesions

CT scanning best delineates bony lesions causing proptosis. Primary bony lesions include osteomas, osteogenic sarcomas, and fibrous dysplasia. Secondary lesions include metastases, especially from prostate, thyroid, lung, breast, and kidney cancers.

Lacrimal Gland Tumors

Lacrimal gland tumors can be easily found with ophthalmic B-scan at ultrasonography. The complete extent, especially bony involvement, is best demonstrated with CT scans. When lacrimal gland tumors are bilateral patients either have inflammatory lesions (sarcoid, pseudotumor) or lymphomas. Inflammatory lesions tend to be somewhat tender and represent the overwhelming majority of such lesions. Many clinicians treat painful bilateral lacrimal gland tumors without biopsy with high-dose aspirin, nonsteroidal anti-inflammatory agents, or systemic prednisone.

Unilateral lacrimal gland masses almost always require biopsy, although their clinical presentations may be distinct. Benign mixed tumors are painless, slowly (over years) enlarging masses of the gland in patients 30 to 40 years old, causing infradisplacement of the globe. They present a rounded or globular soft tissue mass in the lacrimal gland that may even have formed a fossa within bone. Treatment is surgical.

Adenoid cystic carcinoma, the most common malignant epithelial tumor of the lacrimal gland, may also present at age 30 to 40 years with infradisplacement of the globe. Frequently, there is pain, numbness, diplopia, and visual disturbance. Symptoms are less than 1 year in duration, and CT scanning shows a circumscribed lacrimal gland mass, frequently with ragged, infiltrating edges into bone. Treatment is surgical. Recurrences may take many years to appear. Nonetheless, 90% of patients develop metastases and die within 15 years.

Pleomorphic adenocarcinomas (malignant mixed tumors) occur in older patients (50–60 years) in three distinct clinical patterns: (1) Patients may present with a painless mass de novo; (2) they may have a history of years of a painless lacrimal gland mass (probably representing a pleomorphic adenoma) with sudden (3–6 months) increase in size; and (3) they may have a history of a prior biopsy of a benign mixed tumor that was incompletely excised many years before. Treatment is surgical. More than three-quarters of patients die of metastases within 5 years.

Benign mixed tumors and carcinomas represent only 25% of lacrimal gland tumors. Nonepithelial lesions account for 75% of lacrimal gland tumors. Of these, 80% are inflammatory; previously called orbital pseudotumor, now they are lumped together as benign reactive lymphoid hyperplasia. These lesions can be slow growing or sudden, painful or painless, and may be associated with any of the earlier-mentioned signs and symptoms. They are treated with systemic high-dose aspirin, steroids, or in rare cases with low-dose irradiation. Their cause is unknown.

The remainder of the lacrimal gland tumors are lymphomas. As with all lymphomas, careful pathologic analysis, aided by fresh tissue with appropriate marker studies and systemic staging, help to define the disease and guide treatment with chemotherapy alone, local radiation alone, or a combination of both. There is much controversy about the clinical course and prognosis of periocular lymphoid lesions, but it appears that histology, stage at diagnosis, and anatomic site of involvement are predictive.84 One-third to two-thirds of patients develop systemic disease, if it is not already present at the time of orbital biopsy, and almost all of these do so within 2.5 years of orbital biopsy. The higher the grade of lymphoma, the greater is the chance of widespread disease (which is the opposite of nodal disease). Of patients with localized, extranodal disease, 86% were disease free at a median of 51 months after radiation treatment; 32% of patients who had disseminated disease had died.85

Some studies suggest that patients with bilateral orbital involvement have a poorer prognosis. Conjunctival lesions have the lowest chance (20%) of developing systemic disease. Eyelid lymphomas have the highest chance (67%), with orbital lesions in between (35%).69

Thyroid Ophthalmopathy

Thyroid-related orbital disease represents the most common cause of both bilateral and unilateral proptosis. The eye findings may be associated with clinical or laboratory evidence of hyperthyroidism, hypothyroidism, or euthyroidism. The disease is four to five times more common in women, usually those in middle age. While the eye findings in this disease may be subtle and varied, in cases of mistaken diagnosis of orbital tumor, proptosis is always present. A straightforward clinical examination of the patient with thyroid eye disease usually differentiates it from a true orbital tumor. In both cases, there may be proptosis. In the case of a tumor, the eye is displaced forward, which causes stretching of the upper lid, and the distance between the eyelid and the lid fold is increased. In cases of thyroid disease, because of “retraction” attributed to overactivity of the levator muscle, the displaced eye is associated with a shorter distance between the lid margin and the lid fold.

The signs and symptoms of thyroid eye disease include lid retraction, lid lag, lid edema, inability to fully close the eye, with corneal drying, infection, and ulcers. Because of the muscle involvement there may be limitation of ocular movement with diplopia. Elevated intraorbital pressure, especially on looking up, may cause a severe, blinding glaucoma. CT and/or MRI may demonstrate the proptosis, enlargement of one (usually the inferior rectus) or any combination of extraocular muscles, with increased fat in the orbit, but no clear masses.

Laboratory testing for thyroid ophthalmopathy is frequently confusing. Among the tests used are T4 (thyroxine) and TSH (thyroid-stimulating hormone). T3 (triiodothyronine), T3 uptake (a measure of the relative saturation of thyroid-binding globulin), and radioactive iodine uptake are sometimes used. In cases of autoimmune thyroid disease, microsomal antibodies and thyroid stimulating immunoglobulin are also measured. The thyroid-releasing hormone (TRH) stimulation test may be useful in cases of so-called euthyroid Graves’ disease.

Treatment is directed at cosmetics, vision, double vision, and problems related to corneal exposure. Glaucoma may have to be treated in one eye only. Treatment includes normalization of thyroid function and systemic steroids, irradiation (2,000 cGy), or surgical decompression. Systemic diuretics, cyclophosphamide, azathioprine, and cyclosporin have also been used with some success.

Well-Defined Orbital Masses

The most common benign orbital tumor of adults is the cavernous hemangioma (in contrast to the capillary hemangioma of children). Patients have slowly progressive painless proptosis, with a mass indenting the globe, showing striae in the retina and a flattened globe on imaging studies. Treatment is surgical, and complete removal is possible.

A mucocele is a cystic, encapsulated mass originating in a paranasal sinus (usually the frontal sinus) that follows repeated bouts of sinusitis, often leading to recurrent orbital cellulitis. The bony wall is not intact on imaging studies. Treatment involves excision of the mucocele and surgical attention to the involved sinus.

Diffuse Orbital Mass

Diffuse orbital masses usually require a biopsy and include lymphoma, benign reactive lymphoid hyperplasia (orbital pseudotumor), orbital cellulitis, fibrous histiocytoma (benign and malignant), neurofibromas, and sarcomas. Their management is presented in Figure 81.8.

Ophthalmic complications of radiation and chemotherapy

We have designed a teaching card (Figs. 81.9A and Figs. 81.9B) outlining the different toxicities of various chemotherapeutic agents and therapeutic radiation on the eye and orbit Abramson, Servodidio 1996,1997®)

Figure 81.9A. Graphic representation of the ophthalmic side effects of radiation and chemotherapy.

Figure 81.9A

Graphic representation of the ophthalmic side effects of radiation and chemotherapy.

Figure 81.9B. teaching card outlining the different toxicities of various chemotherapeutic agents and therapeutic radiation on the eye and orbit.

Figure 81.9B

teaching card outlining the different toxicities of various chemotherapeutic agents and therapeutic radiation on the eye and orbit.


Abramson D H. Retinoblastoma. Pediatr Emerg Casebook. 1985;3:3–15.
Abramson D H. Retinoblastoma: diagnosis and management. CA Cancer J Clin. 1982;32:130–140. [PubMed: 6804033]
Abramson D H. The focal treatment of retinoblastoma with emphasis on xenon arc photocoagulation. Acta Ophthalmol Suppl. 1989;194:3–63. [PubMed: 2559573]
Abramson D H, Ellsworth R M, Grumbach N, Kitchin F D. Retinoblastoma: survival, age at detection and comparison 1914–1958, 1958–1983. J Pediatr Ophthalmol Strabismus. 1985;22:246–250. [PubMed: 4078667]
Knudson A G. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci -USA. 1971;68:820–823. [PMC free article: PMC389051] [PubMed: 5279523]
Friend S H, Bernards R, Rogelj S. et al. A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature. 1986;323:643–646. [PubMed: 2877398]
Muncaster M M, Cohen B L, Philips R A, Gallie B L. Failure of RB1 to reverse the malignant phenotype of human tumor cell lines. Cancer Res. 1992;52:654–661. [PubMed: 1732054]
Harlow E. Retinoblastoma: for our eyes only. Nature. 1992;359:270–271. [PubMed: 1406928]
Yandell DW. Long-Term Complications of Treatment of Children and Adolescents for Cancer. Third International Conference, Jun 1994, Niagara Falls, NY.
Squire J, Gallie B L, Philips R A. A detailed analysis of chromosomal changes in heritable and non-heritable retinoblastoma. Hum Genet. 1985;70:291–301. [PubMed: 4018796]
Doz F, Peter M, Schliermacher G. et al. N-mycamplification, loss of heterozygosity on the short arm of chromosome 1 and DNA ploidy in retinoblastoma. Eur J Cancer. 1996;32A:645–649. [PubMed: 8695269]
Bookstein R, Lee W H. Molecular genetics of the retinoblastoma suppressor gene. Crit Rev Oncog. 1991;2:211–227. [PubMed: 1888791]
Schubert E L, Hansen M F, String L C. The retinoblastoma gene and its significance. Ann Med. 1994;26:177–184. [PubMed: 8074836]
Abramson D H, Frank C M, Susman M. et al. Presenting signs of retinoblastoma. J Pediatr. 1998;132:505–508. [PubMed: 9544909]
Abramson DH, Servodidio CA. Retinoblastoma in the first year of life. In: The eye in infancy, 2nd ed. Eisenberg SJ, editor. St. Louis: Mosby; 1994. p. 426–436.
Reese AB. Tumors of the eye, 3rd ed. New York: Harper & Row; 1976.
Abramson D H. The diagnosis of retinoblastoma. Bull N Y Acad Med. 1988;64:283–317. [PMC free article: PMC1629340] [PubMed: 3069163]
Shields JA. Diagnosis and management of intraocular tumors. St. Louis: Mosby; 1983.
Abramson DH. Lactate dehydrogenase and retinoblastoma. In: Ocular and adnexal tumors. Jakobiec FA, editor. Birmingham AL: Aesculaoius; 1978. p. 454–459.
Abramson D H. The surgical management of retinoblastoma. Ophthalmic Surg. 1980;11:596–598. [PubMed: 7422244]
Grabowski E, Abramson D H. Intraocular and extraocular retinoblastoma. Hematol Oncol Clin North Am. 1987;1:721–735. [PubMed: 3323180]
Abramson DH. Treatment of retinoblastoma. In: Retinoblastoma (Contemporary issues in ophthalmology). Blodi FC, editor. New York: Churchill Livingstone; 1985. p. 63–93.
Abramson D H, Jereb B. External beam radiation for retinoblastoma. Bull NY Acad Med. 1981;57:787–803. [PMC free article: PMC1805280] [PubMed: 6948598]
McCormick B, Ellsworth R M, Abramson D H. et al. Results of external beam radiation for children with retinoblastoma: a comparison of two techniques. J Pediatr Ophthalmol Strabismus. 1989;26:239–243. [PubMed: 2795413]
Meyer-Schwickerath G, Drance SM. Light coagulation. St. Louis; Mosby, 1960.
Lincoff H, McLean J, Long R. The cryosurgical treatment of intraocular tumors. Am J Ophthalmol. 1967;63:389. [PubMed: 6019526]
Abramson DH. Cryotherapy in retinoblastoma. In: Advanced techniques in ocular surgery. Jakobiec F, editor. Philadelphia: W.B. Saunders; 1984. p. 433–437.
Stallard H B. The conservative treatment of retinoblastoma. Highlights Ophthalmol. 1963;6:129–130.
Buys R, Abramson D H, Ellsworth R M, Haik B G. Radiation regression patterns after cobalt plaque insertion for retinoblastoma. Arch Ophthalmol. 1983;101:1206–1208. [PubMed: 6882247]
Abramson D H, Ellsworth R M, Haik B G. Cobalt plaques in advanced retinoblastoma. Retina. 1983;3:12–15.
Shields C L, DePotter P, Himelstein B P. et al. Chemoreduction in the initial management of intraocular retinoblastoma. Arch Ophthalmol. 1996;114:1330–1338. [PubMed: 8906023]
Kingston J E, Hungerford J L, Madreperla S A, Plowman P N. Results of combined chemotherapy and radiotherapy for advanced intraocular retinoblastoma. Arch Ophthalmol. 1996;114:1339–1343. [PubMed: 8906024]
Murphree A L, Villablanca J G, Deegan W F. et al. Chemotherapy plus local treatment in the management of intraocular retinoblastoma. Arch Ophthalmol. 1996;114:1348–1356. [PubMed: 8906025]
Gallie B L, Budning A, DeBoer G. et al. Chemotherapy with focal therapy can cure intraocular retinoblastoma without radiotherapy. Arch Ophthalmol. 1996;114:1321–1328. [PubMed: 8906022]
Hawkins M M, Wilson L M K, Burton H S. et al. Radiotherapy, alkylating agents, and risk of bone cancer after childhood cancer. J Natl Cancer Inst. 1996;88:270–278. [PubMed: 8614005]
Travis L B, Holowaty E J, Bergfeldt K. et al. Risk of leukemia after platinum-based chemotherapy for ovarian cancer. N Engl J Med. 1999;340:351–357. [PubMed: 9929525]
Grabowski EF, Abramson DH. Retinoblastoma. In: Clinical pediatric oncology, 4th ed. Fernbach DJ, Vietti TJ, editors. St. Louis: Mosby; 1991. p. 427–435.
White L. Chemotherapy in retinoblastoma: current status and future directions. Am J Pediatr Hematol Oncol. 1991;13:189–201. [PubMed: 2069230]
Chantada G L, Fandino A, Mato G, Casak S. Phase II window of idarubicin in children with extraocular retinoblastoma. J Clin Oncol. 1999;17:1847–1850. [PubMed: 10561224]
Magramm I, Abramson D H, Ellsworth R M. Optic nerve involvement in retinoblastoma. Ophthalmology. 1989;96:217–222. [PubMed: 2704542]
Shields C L, Shields J A, Baez K. et al. Optic nerve invasion of retinoblastoma: metastatic potential and clinical risk factors. Cancer. 1994;73:692–698. [PubMed: 8299091]
Zelter M, Damel A, Gonzalez G, Schwartz L. A prospective study on the treatment of retinoblastoma in 72 patients. Cancer. 1991;68:1685–1690. [PubMed: 1913508]
Doz F, Khelfaoui F, Mosseri V. et al. The role of chemotherapy in orbital involvement of retinoblastoma. Cancer. 1994;74:722–732. [PubMed: 8033054]
Goble R R, McKenzie J, Kingston J E. et al. Orbital recurrence of retinoblastoma successfully treated by combined therapy. Br J Ophthalmol. 1990;74:97–98. [PMC free article: PMC1041999] [PubMed: 2310733]
Schvartzman E, Chantada G, Fandino A. et al. Results of a stage-based protocol for the treatment of retinoblastoma. J Clin Oncol. 1996;14:1532–1536. [PubMed: 8622068]
Namouni F, Doz F, Tanguy M L. et al. High-dose chemotherapy with carboplatin, etoposide and cyclophosphamide followed by a haematopoietic stem cell rescue in patients with high-risk retinoblastoma: a SFOP and SFGM study. Eur J Cancer. 1997;33:2368–2375. [PubMed: 9616283]
Chan H S L, Canton M D, Gallie B L. Chemosensitivity and multidrug resistance to antineoplastic drugs in retinoblastoma cell lines. Anticancer Res. 1989;9:469–474. [PubMed: 2751270]
Reese A B, Merriam G R, Martin H E. Treatment of bilateral retinoblastoma by irradiation and surgery. Report on 15-year results. Am J Ophthalmol. 1949;32:175–190. [PubMed: 18111333]
Abramson D H, Ellsworth R M, Zimmerman L E. Non-ocular cancer in retinoblastoma survivors. Trans Am Acad Ophthalmol Otolaryngol. 1976;81:454–458. [PubMed: 1066869]
Abramson D H. Second non-ocular cancers in retinoblastoma; a unified hypothesis. The Fraces Chetti Lecture. Ophthalmic Genetics. 1999;20:193–204. [PubMed: 10610188]
Abramson D H, Ellsworth R M, Kitchin F D, Teung G. Second nonocular tumors in retinoblastoma survivors: are they radiation-induced? Ophthalmology. 1984;91:1351–1355. [PubMed: 6595610]
Li F P, Abramson D H, Tarone R E. et al. Hereditary retinoblastoma, lipoma and second primary cancers. J Natl Cancer Inst. 1997;89:83–84. [PubMed: 8978411]
Eng C, Li F P, Abramson D H. et al. Mortality from second tumors among long-term survivors of retinoblastoma. J Natl Cancer Inst. 1993;85:1121–1128. [PubMed: 8320741]
Wong F L, Boice J D, Abramson D H. et al. Cancer incidence after retinoblastoma. Radiation dose and sarcoma risk. JAMA. 1997;278(15):1262–1267. [PubMed: 9333268]
Abramson D H, Frank C. Second nonocular tumors in survivors of bilateral retinoblastoma. A possible age effect on radiation-related risk. Ophthalmology. 1998;105:573–579. [PubMed: 9544627]
Blach L E, McCormick B, Abramson D H, Ellsworth R M. Trilateral retinoblastoma: incidence and outcome: a decade of experience. Int J Radiat Oncol Biol Phys. 1994;29:729–733. [PubMed: 8040018]
Abramson D H, Jereb B, Wollner N. et al. Leukemic ophthalmopathy detected by ultrasound. J Pediatr Ophthalmol Strabismus. 1983;20:92–97. [PubMed: 6864428]
Gomez D G, Manzo R P, Fenstermacher J D, Potts D G. Cerebrospinal fluid absorption in the rabbit. Optic pathways. Graefes Arch Clin Exp Ophthalmol. 1988;226:1–7. [PubMed: 3342970]
Haik BG. Vascular tumors of the orbit. In: Ophthalmic and orbital plastic reconstructive surgery. Hornblass A, editor. Baltimore: Williams & Wilkins; 1989. p. 509–517.
Jones I S. Lymphangiomas of the ocular adnexae: an analysis of 62 cases. Am J Ophthalmol. 1961;51:481–509. [PubMed: 13790538]
Abramson D H, Ellsworth R M, Tretter P. et al. The treatment of orbital rhabdomyosarcoma with irradiation and chemotherapy. Ophthalmology. 1979;86:1330–1335. [PubMed: 233865]
Jones I S, Reese A B, Kraut J. Orbital rhabdomyosarcoma: an analysis of 62 cases. Am J Ophthalmol. 1966;61:721–735. [PubMed: 5931269]
Wharam M, Beltangady M, Hays D. et al. Localized orbital rhabdomyosarcoma. An interim report of the Intergroup Rhabdomyosarcoma Study Committee. Ophthalmology. 1987;94:251–254. [PubMed: 3587902]
Abramson D H, Notis C. Visual acuity after radiation for orbital rhabdomyosarcoma. Am J Ophthalmol. 1994;118:808–809. [PubMed: 7977611]
Notis C M, Abramson D H, Sagerman R H, Ellsworth R M. Orbital rhabdomyosarcoma: treatment or overtreatment. Ophthalmic Genet. 1995;16:159–162. [PubMed: 8749052]
Abramson D H, Fass D, McCormick B. et al. Implant brachytherapy: a novel treatment for recurrent orbital rhabdomyosarcoma. J AAPOS. 1997;1:154–157. [PubMed: 10532778]
Johnson RN. Choroidal and ciliary body malignant melanoma: diagnosis and management. Korea: Retina Research Fund; 1993.
Servodidio C A, Abramson D H. Presenting signs and symptoms of choroidal melanoma: what do they mean? Ann Ophthalmol. 1992;24:190–194. [PubMed: 1637129]
Abramson D H. Computerized visual defects of choroidal melanomas. Glaucoma. 1988;10:39–44.
Coleman D J, Abramson D H, Jack R L, Franzen L A. Ultrasonic diagnosis of tumors of the choroid. Arch Ophthalmol. 1974;91:344–355. [PubMed: 4821788]
Byrne SF, Green WL. Ultrasound of the eye and orbit. St. Louis: Mosby; 1992.
Shields J, Shields C L, Donoso L A. Management of posterior uveal melanoma. Surv Ophthalmol. 1991;36:161–195. [PubMed: 1776122]
Shammas H, Blodi F C. Prognostic factors in choroidal and ciliary body melanomas. Arch Ophthalmol. 1977;95:63–69. [PubMed: 836204]
Shields JA. Diagnosis and management of intraocular tumors. St Louis: Mosby; 1983.
Diener-West M, Hawkins B, Markuwitz J A. et al. A review of mortality from choroidal melanoma. II. A meta-analysis of 5-year mortality rates following enucleation 1966 through 1988. Arch Ophthalmol. 1992;110:245–250. [PubMed: 1531290]
Char D. Clinical ocular oncology, New York: Churchill Livingstone; 1988.
Collaborative Ocular Melanoma Study Group. The Collaborative Ocular Melanoma Study (COMS) randomized trial of pre-enucleation radiation of large choroidal melanoma II. Initial Mortality Findings COMS report No. 10. Am J Ophthalmol 1998;125:779–796. [PubMed: 9645716]
Stallard H B. Radiotherapy of malignant intraocular neoplasms. Br J Ophthalmol. 1948;32:618–639. [PMC free article: PMC512144] [PubMed: 18170499]
Char D H, Kroll J M, Castro J. et al. Ten-year follow up of helium ion therapy for choroidal melanoma. Ophthalmology. 1988;91:1281–1289.
Abramson DH, Servodidio CA. Ocular complications due to cancer treatment. In: Survivors of childhood cancer: assessment and management. St. Louis: Mosby; 1994. p. 111–132.
Fine S. Do I take the eye out or leave it in? Arch Ophthalmol. 1986;104:653–654. [PubMed: 3518676]
Jakobiec F A, Neri A, Knowles D M. Genotype monoclonality in immunophenotypically polyclonal orbital lymphoid tumors. A model of tumor progression in the lymphoid system. The 1986 Wendell Hughes lecture. Ophthalmology. 1987;94:980–994. [PubMed: 3658376]
Henderson JH. Orbital tumors, 2nd ed. New York: Georg Thieme; 1980.
Jakobiec F A, Neri A, Knowles D M. Genotypic monoclonality in immunophenotypically polyclonal orbital lymphoid tumors. A model of tumor progression in the lymphoid system. The 1986 Wendell Hughes lecture. Ophthalmology. 1987;94:980–994. [PubMed: 3658376]
McNally L, Jakobiec F, Knowles D M. Clinical, morphologic, immunophenotypic, and molecular genetic analysis of bilateral ocular adnexal lymphoid neoplasms in 17 patients. Am J Ophthalmol. 1987;103:555–568. [PubMed: 3494404]
Reference not available .
© 2000, BC Decker Inc.
Bookshelf ID: NBK20865


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...