Intraocular–Central Nervous System Lymphoma Clinical Features, Diagnosis, and Outcomes Esen Karamursel Akpek, MD,1 Irma Ahmed, MD,2 Frederick H. Hochberg, MD,3 Masoud Soheilian, MD,1 Thaddeus P. Dryja, MD,2 Frederick A. Jakobiec, MD,2 C. Stephen Foster, MD1 Objective: To analyze the clinical features, laboratory investigations, and diagnosis of intraocular– central nervous system (CNS) lymphoma in a cohort of patients who underwent diagnostic vitrectomy. Design: Retrospective case series. Method and Study Materials: Thirty-four vitreous biopsy specimens obtained from 26 patients with treatment-resistant or unusual uveitis were re-evaluated in a masked fashion. The specimens were classified into three groups: “negative,” “suspicious of malignancy,” and “positive” based on the cytologic features, immunomarkers, and flow cytometry. The medical records of the patients were reviewed retrospectively. Main Outcome Measures: The reliability of vitreous cytology in diagnosing intraocular–CNS lymphoma and the differences in clinical features of patients with intraocular–CNS lymphoma and uveitis. Results: The two ocular pathologists concurred in their criteria for interpretation of all specimens. There was 100% concordance between the cytologic reports read independently by the two ocular pathologists over the 5-year period and the read-out done in a masked fashion at the time of the study. Ten patients were diagnosed with intraocular–CNS lymphoma based on the vitreous cytology and clinical features. The time interval between the initial presentation and vitreous biopsy was 1 week to 2 years, with 80% of the patients diagnosed within the first year. Retinal involvement in the form of lymphomatous subretinal pigment epithelial infiltrates, vasculitis, and apparent retinochoroiditis was present in six cases. Initial neuroimaging studies revealed concomitant CNS involvement in three patients, and an additional three developed CNS lymphoma following diagnosis by vitreous biopsy. Patients were treated with radiotherapy, chemotherapy, or both. Two of the four patients with a follow-up of greater than 12 months died due to CNS involvement. Conclusions: Vitreous cytology is a sensitive, reliable, and reproducible method of diagnosing intraocular– CNS lymphoma. A high index of suspicion based on the clinical findings and course of the uveitis is critically important in decision-making for diagnostic vitrectomy. Central nervous system involvement is frequent and associated with a high mortality rate. Ophthalmology 1999;106:1805–1810 Intraocular– central nervous system (CNS) lymphoma, an infrequent neoplasm, typically represents a diagnostic challenge, masquerading as uveitis, vitritis, or chorioretinitis. Central nervous system involvement, usually appearing several years after ocular symptoms, affects 50% to 80% of these patients.1,2 Conversely, approximately 20% of patients with primary CNS lymphoma exhibit ocular involvement at the time of diagnoOriginally received: May 28, 1998. Revision accepted: May 20, 1999. Manuscript no. 98275. 1 Ocular Immunology Service, Massachusetts Eye and Ear Infirmary, Boston, Massachusetts. 2 Ocular Pathology Service, Massachusetts Eye and Ear Infirmary, Boston, Massachusetts. 3 Neuro-oncology Service, Neuro-oncology and Brain Tumor Center, Massachusetts General Hospital, Boston, Massachusetts. Presented in part at the American Uveitis Society meeting, San Francisco, California, October 1997. Address correspondence to C. Stephen Foster, MD, Immunology Service, Massachusetts Eye and Ear Infirmary, 243 Charles Street, Boston, MA 02114. E-mail: [email protected]
sis.3 More than 150 cases of ocular lymphoma have been reported in the ophthalmic literature,4,5 and most of these cases were initially misdiagnosed as idiopathic uveitis or vitritis.6 The vitritis is initially “responsive” to steroid therapy because not all of the vitreous cells are malignant but rather are “reactive” lymphocytes or histiocytes. The diagnosis of intraocular–CNS lymphoma is difficult and requires a high index of suspicion. The best, accepted current diagnostic approach for patients suspected of having intraocular–CNS lymphoma involves a complete neurologic investigation, cerebrospinal fluid (CSF) cytology, and vitreous biopsy.7,8 In the earlier stages of the disease, cytologic examination of vitreous biopsy specimens may disclose seemingly reactive lymphocytes because few of the lymphocytes in the vitreous are frankly neoplastic. Additionally, because CNS involvement lags ocular manifestations by months to years (1 month to 10 years), results of cytologic analysis of CSF, neurologic examinations, and imaging studies are frequently normal.2,4,7 The purpose of this study is to analyze the clinical features, laboratory investigations, and diagnosis of intraoc-
Ophthalmology Volume 106, Number 9, September 1999
Figure 1. Photomicrograph of a positive vitrectomy specimen showing a reactive lymphocyte and an atypical lymphocyte (arrow) with an elevated nucleus-cytoplasm ratio and a round nucleus with a coarse chromatin pattern (Papanicolaou stain; original magnification, ⫻320).
ular–CNS lymphoma in a cohort of patients who underwent diagnostic vitrectomy.
Materials and Methods Over a 5-year period (1990 –1995) at the Massachusetts Eye and Ear Infirmary, 34 vitreous cytology specimens were obtained from 26 patients who underwent diagnostic vitrectomy to rule out intraocular–CNS lymphoma. The cytologic features of the specimens were reviewed by two ocular pathologists (FA Jakobiec and
Figure 2. Case 1. Fundus photograph of the right eye shows characteristic subretinal pigment epithelium yellow infiltrates.
TP Dryja) in a masked fashion at the Cogan Eye Pathology Laboratory of the same hospital. All of the vitreous specimens were obtained through a standard three-port pars plana vitrectomy. Before instituting the infusion, 1 ml of undiluted vitreous was aspirated into a syringe attached to the vitrectomy cutter hand-piece via a stop-cock. Following infusion, total vitrectomy was performed, and both 1 ml undiluted specimen and the specimen in the vitrector cassette were immediately delivered to the cytopathology laboratory. The specimens were fixed by mixing one part 10% neutral buffered formalin to one part specimen, for approximately 12 hours. A 5-ml fixed specimen was then pipetted into a cytospin chamber and spun at 1000 rpm for 5 minutes, thereby concentrating the cells in the specimen onto glass slides. Cytopathologic analysis was then conducted after air drying and staining of the slides with modified Papanicolaou stain. Flow cytometry analysis and immunomarkers for monoclonality supplemented this light microscopy assay in all cases. A specimen was deemed “negative” if it contained mostly lymphocytes and neutrophils with occasional histiocytes. The cells demonstrated no prominent nucleoli and no irregular outline, and no mitoses were observed. A positive specimen met the cytologic criteria for anaplasia, such as an elevated nucleus-cytoplasm ratio, round or oval nuclei with a coarse chromatin pattern, lobulations of the nuclei, and multiple small nucleoli (Fig 1). Histiocytes and macrophages have been distinguished from the large neoplastic lymphocytes. Histiocytes have large vesicular or watery nuclei with small nucleoli and minimal clumping of the nuclear chromatin. The abundant diaphanous cytoplasm may sometimes be folded. Macrophages have somewhat eccentric nuclei in the cytoplasm and much more opaque cytoplasm and may contain ingested debris including melanin granules. A “suspect” diagnosis included those specimens for which some atypicality existed but not a sufficient amount to diagnose malignancy. The medical records of the patients were reviewed retrospectively. Special attention was paid to patient demographics, visual and neurologic symptoms, initial diagnosis, neuroimaging studies, CSF cytology, and the time interval between the onset of symptoms and definitive diagnosis by biopsy.
Results Thirty-four vitreous cytology specimens harvested from 26 patients were included in the study. All the patients had uveitis
Figure 3. Case 6. Fundus photograph of the right eye shows large clumps of vitreous cells.
Akpek et al 䡠 Intraocular–CNS Lymphoma Table 1. Demographic and Clinical Features of All Patients Who Underwent Diagnostic Vitrectomy Feature
No. of patients Age range (yrs) Mean age (yrs) Female/male Visual symptoms Decreased vision Floaters Decreased vision ⫹ floaters Laterality Bilateral Unilateral Neurologic symptoms Initial diagnosis Idiopathic uveitis Endophthalmitis Amyloidosis Vitreous hemorrhage Acute retinal necrosis Toxoplasmosis Lyme disease Neuroimaging studies Performed Abnormal findings
16 24–96 64 13/3
10 31–82 59 4/6
9 2 5
7 1 2
10 6 3
8 2 5
9 4 1 1 1 — —
6 1 — — 1 1 1
unresponsive to treatment and underwent vitrectomy to rule out a diagnosis of intraocular–CNS lymphoma. The two ocular pathologists concurred in their criteria for interpretation of all specimens. There was a 100% concordance between the initial cytologic reports read independently by the two ocular pathologists and the repeat read-outs done in a masked fashion at the time of the study. Twenty-one vitreous specimens from 16 patients were identified as negative, and 11 specimens from 9 patients were either frankly positive or suspicious of malignancy. The vitreous specimen of one patient (patient 10) was negative initially, but a later second specimen from the fellow eye demonstrated malignant features. Table 1 summarizes the demographics and clinical features of the 26 patients included in the study. Group A consisted of 16 patients with a negative vitreous cytology, and group B consisted of 10 patients with intraocular–CNS lymphoma. Women predominated in group A, with a female-male ratio of 13:3. In group B the number of males was greater than the number of females (4/6). Only one of the patients was black; the rest were white. Patient ages in group A were between 24 and 96 years (mean, 64 years) and in group B were from 31 to 82 (mean, 59 years); three of the patients were younger than age 40. Twenty-four patients presented to the clinic with decreased visual acuity, often with floaters, and two patients, one in each group, complained of floaters only, without decrease in their visual acuities. Ten patients (62%) in group A presented with bilateral involvement. In group B, eight patients (80%) had bilateral symptoms. Two who presented with unilateral uveitis went on to develop fellow eye involvement later on. In most cases, the initial diagnosis was idiopathic uveitis. The other diagnoses included endophthalmitis (5 patients), acute retinal necrosis (2 patients), toxoplasmosis (1 patient), Lyme disease (1 patient), vitreous hemorrhage (1 patient), and amyloidosis (1 patient). Characteristic fundus findings of intraocular–CNS lymphomalymphomatous subretinal pigment epithelial (subRPE) infiltrates, retinochoroiditis, and vasculitis (Fig 2) were present in six (60%)
cases, all in group B. The fundus could not be visualized in one patient due to severe vitreous clouding. Three patients (30%) in group B had vitritis only with large clumps or sheets of cells or clumps of cells on vitreous strands and no retinal or vascular involvement (Fig 3). Neurologic symptoms were present in three (18%) of the patients in group A at the initial visit and included headaches (1 patient), history of labyrinthitis (1 patient), and history of epilepsy (1 patient). In group B, five patients reported neurologic symptoms such as diplopia (1 patient), right/left confusion (1 patient), poor memory and imbalance (2 patients), and buzzing sensation in limbs (1 patient). Neuroimaging studies (computed tomography [CT] or magnetic resonance imaging [MRI] scan) were performed on 17 of the patients initially. Five revealed abnormal findings, all in group B. Figure 4 demonstrates the time interval between the onset of symptoms and definitive diagnosis by biopsy. Forty-three percent of patients in group A underwent diagnostic vitrectomy by the end of the first year. Four more (22%) were diagnosed by the end of the second year. In five patients (31%) the time interval between the onset of symptoms and diagnostic vitrectomy was more than 2 years. In group B, eight patients (80%) were diagnosed during the first year of their disease and one (10%) in the second year of the disease course. The interval between the onset of symptoms and the diagnosis was more than 2 years in only one patient (10%). Table 2 summarizes the laboratory investigations and outcomes of ten patients with intraocular–CNS lymphoma. Neuroimaging studies were employed on all patients. Five (50%) exhibited abnormal CT/MRI studies suggestive of CNS lymphoma at initial presentation (Fig 5). Among those five patients, three (patients 1, 4, 5) had fundus findings of lymphoma, in one patient the fundus could not be visualized because of severe vitreous opacification, and in one patient (patient 7) the fundus examination was normal. Three (30%) additional cases (patients 2, 3, 8) went on to develop positive findings on CT/MRI following diagnosis by positive vitreous biopsy. The time interval between the diagnostic vitrectomy and development of positive findings on the neuroimaging studies in these patients was 12 to 20 months (mean, 17 months). All three had fundus findings at the initial presenting visit. Lumbar puncture was performed in seven patients. Three (42%) revealed atypical cells in the CSF. All three patients had classical subRPE infiltrates and abnormal neuroimaging studies suggestive of CNS lymphoma at the time of the CSF examination. In two patients with radiologically documented CNS lymphoma, CSF cytology failed to demonstrate malignant lymphocytes. The follow-up period varied between 1 to 38 months, with a mean of 12 months. Patient 10 was lost to follow-up after the diagnostic vitrectomy. Four of the nine patients received only radiation therapy, and two received both radiation and high-dose intravenous methotrexate therapy. Three patients (patients 4, 6, 9) had not received any form of therapy by the time this study was completed. In two of these patients (patients 4 and 6) with a follow-up of 1 and 2 months, respectively, the treatment was being
Figure 4. Time interval (in weeks) between the onset of symptoms and diagnostic vitrectomy.
Ophthalmology Volume 106, Number 9, September 1999 Table 2. Characteristics of Patients Initially Read as Suspect/Diagnostic of Intraocular Lymphoma
⫹ OD Suspect OD Suspect OS Later ⫹ OD ⫹ OS Suspect OS ⫹ OD ⫹ OD Suspect OD Later ⫹ OS Suspect OS Negative OS Later ⫹ OD
3 4 5 6 7 8 9 10
Initially Abnormal CT/MRI Scans
Abnormal CT/MRI Later on (mos)
Follow-up Period (mos)
Rx Rx ⫹ Kx
Yes Yes Yes No No
⫺ ⫹ ⫹ ⫺ ⫹
Atypical Not done Not done Normal Normal
14 1 16 1 2
Rx — Rx ⫹ Kx Rx —
Alive Alive Dead Alive Alive
Yes No No view
⫺ ⫺ ⫹
Atypical Normal Not done
24 4 Lost to follow-up
Rx — —
Dead Alive —
CT ⫽ computed tomography; MRI ⫽ magnetic resonance imaging; OD ⫽ right eye; OS ⫽ left eye; CSF ⫽ cerebrospinal fluid; Rx ⫽ radiation therapy; Kx ⫽ chemotherapy.
planned. In the other patient (patient 9) presenting with vitritis, vitreous biopsy was suspicious for malignancy, and neuroimaging studies demonstrated no abnormal findings. A second biopsy was performed in the fellow eye, disclosing malignant lymphocytes. Bilateral radiation therapy was planned. Two (50%) of the four patients with a follow-up greater than 12 months (14 –36) died due to the consequences of CNS involvement. Patient 5 was discovered to have concurrent CNS involvement at the initial neuroradiologic evaluation, was treated with both ocular and CNS radiation and high-dose intravenous methotrexate, and survived 16 months. Patient 8 had retinal involvement at the initial presenting visit with no abnormal CT/MRI findings. The vitreous biopsy was read as suspicious in the right eye. The patient developed neurologic symptoms, and neuroimaging studies disclosed CNS lymphoma 20 months after the presenting visit. Despite the radiation therapy, the patient survived 4 months after the documented CNS involvement.
biopsies. In only one patient was the initial vitreous cytology negative; a later specimen from the fellow eye demonstrated malignant cells. Lately, the association of elevated vitreous levels of interleukin-10 relative to interleukin-6 in patients with intraocular–CNS lymphoma has been reported.14,15 However, we believe that to make a diagnosis of intraocular–CNS lymphoma, a disease that requires radiotherapy or chemotherapy, demonstration of the malignant cells is essential. Although previous steroid treatment, inappropriate handling of the specimens, occasional multiple specimen requirement, and possibility of incorrect cytologic interpretation may limit the diagnostic yield, the diagnosis of intraocular–CNS lymphoma still depends on vitreous cytology.
Discussion Intraocular large cell lymphoma, previously termed primary reticulum cell sarcoma of the brain or microgliomatosis,9,10 was first described in 1955.11 Intraocular–CNS lymphoma is rare, yet it is important because of the potential for blindness and the almost certain mortality when associated with CNS involvement. The diagnosis still remains difficult and requires a high degree of suspicion. In earlier series reported in the literature, definitive diagnosis of intraocular–CNS lymphoma was based on histopathologic examination of enucleated blind eyes or brain biopsy and studies at autopsy.12 Klingele and Hogan13 reported the first case in which the initial diagnosis was made on cytologic examination of the vitreous specimen. More recently, vitreous biopsy has become a widely performed procedure in the diagnosis of intraocular–CNS lymphoma.6,8,11,12 All ten patients in our series were diagnosed by vitreous biopsy cytology. The cytologic interpretations were reproducible between the two pathologists reviewing the specimens. Three patients required repeat
Figure 5. Case 1. MRI scan shows two contrast-enhancing lesions in the pons (arrow).
Akpek et al 䡠 Intraocular–CNS Lymphoma Intraocular lymphoma usually occurs in middle to late adult life.16 Palexas et al17 previously reported on 42 cases with intraocular–CNS lymphoma and concluded that chronic vitritis in middle-aged or older patients, especially women, should be recognized as characteristics of lymphoma. Among our patients, the mean age was 59, with a range of 31 to 82 years. Importantly, three of the patients were under 40 years of age at the time of the diagnosis. There was a slight male preponderance with a female to male ratio of 4/6. Although an earlier study reported no gender predilection in immunocompetent patients,3 some more recent studies reported a higher incidence in women,4,8,18 whereas one other study reported a higher incidence in men.19 Decreased vision, floaters, or both were the most commonly reported symptoms (90%). However, at early stages of the disease patients may complain of floaters only, without any decrease in their visual acuity. Unilateral involvement is not uncommon at the initial presentation. Among our ten patients, one (10%) reported only floaters, without any decrease in visual acuity. Symptoms were unilateral in two patients (20%) at presentation. In these patients, the fellow eye involvement occurred later in the disease course. The clinical features of intraocular–CNS lymphoma are confined mainly to the posterior segment. The affected eye usually shows no or very mild external signs of inflammation. A mild inflammation may be present in the anterior chamber. The vitreous typically contains large clumps or sheets of cells on vitreous strands. Vitritis with multifocal, large, yellow subRPE infiltrates with overlying solid pigment epithelium detachment have been suggested as the pathognomonic presentation of intraocular–CNS lymphoma.4,5,10,12,20 Whitcup et al8 reported that all 12 patients in their series had retinal findings at the time of diagnosis. Among our ten patients, three (30%) did not have any retinal findings, and in one patient the retina could not be visualized well due to the opacified vitreous. As illustrated in Figure 4, the time interval between the onset of symptoms and definitive diagnosis by vitreous biopsy was frequently about 20 –52 weeks. Eighty percent of the patients were diagnosed during the first year of their disease; in only one patient was diagnosis delayed more than 2 years. In group A, by the end of the first year only 43% of the patients had undergone vitrectomy. This demonstrates that a high index of suspicion based on the clinical findings is important in decision-making for a diagnostic vitrectomy. Previous studies reported an average time interval of 21 months between the onset of ocular symptoms and the definitive diagnosis.8 The radiographic appearance of the CNS lymphoma on CT and MRI studies is distinctive. The tumor is usually located supratentorially and is multicentric in 50% of cases.21 The lesions typically enhance densely and diffusely after administration of contrast material and have distinct borders. Unlike brain metastases or malignant gliomas, ring enhancement is rarely seen. Among the ten patients in group B, five had abnormal CT/MRI findings consistent with CNS lymphoma at the time of diagnosis by vitreous biopsy. An additional three developed positive findings later on. Interestingly, all three patients who developed positive radio-
logic findings (patients 2, 3, 8) had retinal involvement at the presenting visit. The two patients with normal radiologic work-ups (patients 6 and 9) had only vitritis, without any retinal findings. Only one patient (patient 7) presented with no retinal findings but CNS involvement demonstrated by radiologic studies. The CSF cytology did not reveal any malignant cells in this patient. Because the tumor preferentially involves deep brain structures rather than cerebral cortex, seizures and motor symptoms are less common than with other types of brain tumors. It has been reported that because the frontal lobe is the most frequently involved region of the brain, changes in personality and level of alertness are common presenting symptoms in patients with CNS lymphoma.22 Therefore, careful history taking to detect subtle neurologic symptoms must be performed in all patients with clinical findings suggestive of intraocular–CNS lymphoma, as well as in the care of established patients. A careful review of systems revealed that 8 of the 26 patients in our series had neurologic symptoms at the initial presentation. Three of these patients (group A) had nonspecific symptoms, such as migraine, vertigo with past history of labyrinthitis, and a past history of epilepsy. The diagnosis of intraocular–CNS lymphoma was ruled out in these patients by vitreous cytology and negative neuroimaging studies. The remaining five patients belonged to group B and had suspicious or malignant vitreous cytology. They experienced diplopia, right/left confusion, motor weakness, poor memory, and imbalance. All five patients with neurologic symptoms also had CNS lymphoma revealed by CT/MRI scans. However, in only three of these patients (patients 1, 3, 8) CSF revealed malignant cells. The optimal management of intraocular–CNS lymphoma currently remains undefined and controversial. Radiation therapy has proven to be effective for ocular findings. A dose of 30 Gy is typically given to both eyes because bilateral involvement is the rule.2,18,23 Whole brain radiation therapy with 50 Gy and an additional 10-Gy boost to the tumor site is recommended in cases with documented CNS involvement.18,23 Despite the radiosensitive nature of the malignant cells, the results of the radiation treatment alone have been disappointing because of a high rate of CNS relapse. Margolis et al23 reported that radiation therapy improved the vision but most of their patients died within 1 to 5 years of diagnosis due to CNS involvement. There are no existing data to suggest that prophylactic whole brain radiotherapy reduces the incidence of cerebral lymphoma.6 Chemosensitivity of cerebral lymphoma was initially demonstrated in patients with recurrent or metastatic cerebral lymphoma treated with high-dose methotrexate or cytosine arabinoside as single agents. However, responses are partial and temporary.22 Sustained therapeutic concentrations were reported in the vitreous and CSF with high-dose intravenous cytosine arabinoside or methotrexate.22 DeAngelis et al24 reported a marked improvement in the survival of patients with primary CNS lymphoma when cranial radiation was combined with intrathecal methotrexate and systemic chemotherapy; 12 of 16 patients were alive after a mean follow-up of 22.4 months, in contrast to a median survival of 10 –18 months after radiation therapy alone. Also, intravit-
Ophthalmology Volume 106, Number 9, September 1999 real methotrexate injections have been tried in a small number of patients with intraocular–CNS lymphoma, with promising results.25 The mean follow-up period was 12 months (range, 1–38 months). One of the ten patients in our series was lost to follow-up after the diagnostic vitrectomy. Six patients received radiotherapy, chemotherapy, or both. Of the four patients with a follow-up period greater than 12 months, two died of the consequences of CNS involvement. In conclusion, we believe the vitreous cytology is the best procedure in diagnosing intraocular–CNS lymphoma. Intense intraocular inflammation in the absence of significant pain, photophobia, conjunctival hyperemia, presence of subRPE infiltrates, sheets or clumps of vitreous cells, and steroid resistance should alert the physician. Early diagnostic vitrectomy should be performed in patients with unusual uveitis to avoid delays in diagnosis. Special attention must be paid to the patient’s neurologic status; the presence of subtle signs or symptoms might reflect a possible CNS involvement. Computed tomography and MRI scans are highly specific and sensitive in detecting CNS lymphoma. Therefore, these modalities should be used initially in every patient with a likely diagnosis of intraocular–CNS lymphoma and repeated in the longitudinal care of the patients. Cerebrospinal fluid cytology is unlikely to reveal malignant cells in patients with negative neuroimaging studies. The prognosis in patients with CNS involvement remains poor.
References 1. DeAngelis LM. Current management of primary central nervous system lymphoma. Oncology (Huntingt) 1995;9:63–71; discussion 71, 75– 6, 78. 2. Char DH, Ljung BM, Miller T, Phillips T. Primary intraocular lymphoma (ocular reticulum cell sarcoma) diagnosis and management. Ophthalmology 1988;95:625–30. 3. Rockwood EJ, Zakov ZN, Bay JW. Combined malignant lymphoma of the eye and CNS (reticulum cell sarcoma). Report of three cases. J Neurosurg 1984;61:369 –74. 4. Freeman LN, Schachat AP, Knox DL, et al. Clinical features, laboratory investigations, and survival in ocular reticulum cell sarcoma. Ophthalmology 1987;94:1631–9. 5. Corriveau C, Easterbrook M, Payne D. Lymphoma simulating uveitis (masquerade syndrome). Can J Ophthalmol 1986;21: 144 –9. 6. Ljung BM, Char D, Miller TR, Deschenes J. Intraocular lymphoma. Cytologic diagnosis and role of immunologic markers. Acta Cytol 1988;32:840 –7. 7. Peterson K, Gordon KB, Heinemann MH, DeAngelis LM. The clinical spectrum of ocular lymphoma. Cancer 1993;72:843–9.
8. Whitcup SM, deSmet MD, Rubin BI, et al. Intraocular lymphoma. Clinical and histopathologic diagnosis. Ophthalmology 1993;100:1399 – 406. 9. Yuile CL. Case of primary reticulum cell sarcoma of brain; relationship of microglia cells to histiocytes. Arch Pathol 1938;26:1036 – 44. 10. Henry JM, Heffner RR Jr, Dillard SH, et al. Primary malignant lymphomas of the central nervous system. Cancer 1974;34: 1293–302. 11. Givner I. Malignant lymphoma with ocular involvement. A clinico-pathologic report. Am J Ophthalmol 1955;39:29 – 32. 12. Barr CC, Green WR, Payne JW, et al. Intraocular reticulumcell sarcoma: clinico-pathologic study of four cases and review of the literature. Surv Ophthalmol 1975;19:224 –39. 13. Klingele TG, Hogan MJ. Ocular reticulum cell sarcoma. Am J Ophthalmol 1975;79:39 – 47. 14. Chan CC, Whitcup SM, Solomon D, Nussenblatt RB. Interleukin-10 in the vitreous of patients with primary intraocular lymphoma. Am J Ophthalmol 1995;120:671–3. 15. Whitcup SM, Stark-Vancs V, Wittes RE, et al. Association of interleukin 10 in the vitreous and cerebrospinal fluid and primary central nervous system lymphoma. Arch Ophthalmol 1997;115:1157– 60. 16. Char DH, Margolis L, Newman AB. Ocular reticulum cell sarcoma. Am J Ophthalmol 1981;91:480 –3. 17. Palexas GN, Green WR, Goldberg MF, Ding Y. Diagnostic pars plana vitrectomy report of a 21-year retrospective study. Trans Am Ophthalmol Soc 1995;93:281–308; discussion 308 –14. 18. Siegel MJ, Dalton J, Friedman AH, et al. Ten-year experience with primary ocular ‘reticulum cell sarcoma’ (large cell nonHodgkin’s lymphoma). Br J Ophthalmol 1989;73:342– 6. 19. Hochberg FH, Miller DC. Primary central nervous system lymphoma. J Neurosurg 1988;68:835–53. 20. Gass JD, Sever RJ, Grizzard WS, et al. Multifocal pigment epithelial detachments by reticulum cell sarcoma. A characteristic funduscopic picture. Retina 1984;4:135– 43. 21. Freilich RJ, DeAngelis LM. Primary central nervous system lymphoma. Neurol Clin 1995;13:901–14. 22. Baumann MA, Ritch PS, Hande KR, et al. Treatment of intraocular lymphoma with high-dose Ara-C. Cancer 1986;57: 1273–5. 23. Margolis L, Fraser R, Lichter A, Char DH. The role of radiation therapy in the management of ocular reticulum cell sarcoma. Cancer 1980;45:688 –92. 24. DeAngelis LM, Yahalom J, Heinemann MH, et al. Primary CNS lymphoma: combined treatment with chemotherapy and radiotherapy. Neurology 1990;40:80 – 6. 25. Fishburne BC, Wilson DJ, Rosenbaum JT, Neuwelt EA. Intravitreal methotrexate as an adjunctive treatment of intraocular lymphoma. Arch Ophthalmol 1997;115:1152– 6.