Non-Hodgkin’s lymphoma of central nervous system

NHL-CNS, also known as primary CNS lymphoma, arises from the brain, spinal cord, leptomeninges, or the eye, and may then spread throughout the CNS. Systemic spread outside the CNS occurs more rarely and has been reported in fewer than 10% of autopsies. ^, Although NHL-CNS is a rare tumor, its incidence has trebled over the last 10 years, an increase that cannot be fully explained by AIDS or other causes of immunosuppression. The median age for patients with NHL-CNS is between 50 and 60 years; there is a slight male preponderance. ^, Cases in children have been reported, however, ^, and the two youngest patients with reported cases of intraocular lymphoma were 15 and 27 years old. ^, As for NHL-CNS, immunosuppression appears to be a risk factor for the development of ocular lymphoma. Ocular lymphoma has now been associated with AIDS. In addition, ocular lymphoma has been reported in a number of patients undergoing immunosuppression after organ transplantation.

In a large case series of Hochberg and Miller, patients with NHL-CNS showed four distinct clinical profiles: (1) solitary or multiple discrete intracranial nodules, (2) diffuse meningeal or periventricular lesions, (3) subretinal infiltrates or vitritis, and (4) localized intradural spinal masses. Ocular involvement may precede detectable disease in other parts of the CNS, although the percentage of patients in whom this occurs may be overestimated in the ophthalmic literature, in which there is a selection bias for patients who have ocular disease. In a study performed at the National Eye Institute, two-thirds of the patients with ocular disease had undiagnosed CNS disease at the time of their diagnosis. The mean time between the onset of ocular symptoms and the onset of CNS symptoms in one study was 29 months (range, 7–108 months). In this study, systemic spread outside the CNS was noted in 6% of patients and disease was limited to the eye in 22%.

The most common ocular symptoms associated with intraocular lymphoma are blurred vision and floaters; redness and pain are rare. Patients have reduced visual acuity, although the vision is usually better than one would expect from the clinical examination. Slit-lamp biomicroscopic examination often reveals mild anterior segment inflammation, with aqueous cells and flare and keratic precipitates on the corneal endothelium ( Fig. 30-1 ). Vitreous cells, occurring in sheets, are typical of the disease ( Fig. 30-2 ), and examination of the fundus usually shows subretinal yellow infiltrates through a hazy vitreous ( Fig. 30-3 ). Atypical presentations have been reported, including a hemorrhagic retinal vasculitis that resembled a viral retinitis ( Fig. 30-4 ). In some patients the fundus may appear normal, and subretinal lesions may be noted only with fluorescein angiography ( Fig. 30-5 ). Although the disease may start in one eye, bilateral disease is common after several months. ^,

Anterior segment of the eye of a patient with biopsy-proven intraocular large-cell lymphoma demonstrating corneal keratitic precipitates.

Slit-lamp biomicroscopy examination revealing sheets of vitreous cells characteristic of intraocular lymphoma.

Retinal photograph of a patient with biopsy-proven intraocular lymphoma demonstrating multiple yellow lesions deep to retina.

Fundus photograph of patient with intraocular lymphoma showing retinal infiltrates and hemorrhage mimicking cytomegalovirus retinitis. Findings resolved with therapy for lymphoma.

Fluorescein angiogram of a patient with intraocular lymphoma showing multiple circular areas of hyperfluorescence corresponding to tumor infiltration.

Diagnosis

Many patients with intraocular lymphoma have CNS disease at the time of presentation, and a thorough history and neurologic examination are warranted. CNS findings such as headaches, focal weakness, sensory deficits, confusion, and difficulty with gait are common. A history of recent seizure is also a strong indication of CNS involvement. Magnetic resonance imaging (MRI) and a lumbar puncture should be performed on all patients suspected of having an intraocular lymphoma, even if neurologic symptoms are absent ( Fig. 30-6 ). Because lymphoma cells can be fragile, samples should be immediately hand-carried to the cytology laboratory for processing, and at least 5–10 mL of cerebrospinal fluid (CSF) should also be sent for cytologic examination. A repeat lumbar puncture may be necessary for the diagnosis.

MRI scan showing a mass lesion in the temporal lobe of a patient with intraocular lymphoma.

If the CSF shows no malignant cells, we perform a standard three-port pars plana vitrectomy in an attempt to make the diagnosis of lymphoma. Tissue culture medium (RPMI-1640, Roswell Park Memorial Institute, Mediatech, Inc., Herndon, VA.) enriched with 10% fetal calf serum is added to the collection chamber of the vitrectomy machine to improve cell viability. An initial sample of core vitreous is also obtained at the start of the vitrectomy and taken immediately to the cytology laboratory for processing.

The methods used to process vitreous specimens have been delineated previously. ^, Briefly, the vitreous specimen is centrifuged at 1000 rpm for 8 minutes at 4°C. The cell pellet is resuspended in 4 mL of fresh tissue culture medium and cytocentrifuged for 10 minutes. The machine places centrifuged specimens directly onto gelatinized slides. Giemsa-, Gram-, or Diff Quick-stained slides are prepared, and additional slides are prepared for immunohistochemical staining for B- and T-cell markers and for κ and λ light chains. We prefer a more complete surgical vitrectomy to obtain as much vitreous as possible for analysis, but some investigators have used a vitreous aspiration needle tap in patients with possible intraocular lymphoma.

Even when vitreous specimens are properly processed, diagnosis of an intraocular lymphoma can be difficult. Multiple vitreous specimens may be needed to make the diagnosis. Char and colleagues reported that in three of 14 patients in whom ocular lymphoma was later diagnosed more than one vitreous specimen was required, and we found that multiple samples were needed to make the correct diagnosis in three of 10 patients. In another series of 84 vitrectomy specimens, 14% were diagnosed as definite ‘malignant lymphoma’, 6% as ‘suspicious of neoplastic disease’, 74% as ‘reactive celluar infiltrate’, and 5% ‘insufficient for diagnosis’. Although mishandling of vitreous samples partially accounts for some of the difficulty in making the correct diagnosis, treatment with corticosteroids may obscure the diagnosis of intraocular lymphoma. Corticosteroids can be cytolytic in CNS lymphoma, and this exquisite sensitivity of tumor cells appears to be unique to NHL-CNS. In a number of patients corticosteroid therapy caused CNS lymphoma lesions to reduce in size or disappear, hindering the ability to make the correct diagnosis. Because many patients with intraocular lymphoma are treated with corticosteroids at the time of vitrectomy, the viability of tumor cells in the samples may be diminished, which may contribute to the difficulty in diagnosing this disease. Finally, because the number of malignant cells in a vitreous specimen may be few, it is critical that the cytologist be experienced in the diagnosis of lymphoma.

Recent data suggest that intravitreal cytokine levels can be useful in the diagnosis of intraocular lymphoma. Interleukin (IL)-10 can be produced by lymphoma cells and levels of IL-10 are elevated in both the serum and vitreous of patients with intraocular lymphoma. In contrast, levels of IL-6 are elevated in the vitreous of patients with intraocular inflammation unrelated to a malignant neoplasm. We reported that levels of IL-10 exceeded those of IL-6 in all five patients with intraocular lymphoma examined, but in none of the 13 patients with uveitis. This test, however, is neither 100% sensitive nor 100% specific. We have seen patients with intraocular lymphoma with higher levels of IL-6 than IL-10. However, when IL-10 is elevated in vitreous specimens, the clinician should remain suspicious about an underlying malignancy, and these patients should be closely followed. A study of 51 patients with proven intraocular lymphoma and 108 patients with uveitis of other causes showed that measurement of IL-10 in the aqueous humor can be a good screening test for NHL-CNS. A cutoff of 50 pg/mL in the aqueous humor was associated with a sensitivity of 0.89 and a specificity of 0.93, and the area under the respective ROC curves was only slightly higher for vitreous levels (0.989) than for aqueous levels (0.962).

Typical lymphoma cells are large and pleomorphic with scanty cytoplasm, and hypersegmented nuclei with prominent nucleoli are often diagnostic ( Fig. 30-7 ). However, only a few of these characteristic cells are present in any single specimen, and most vitreous samples contain mainly reactive lymphocytes, histiocytes, necrotic debris, and fibrinous material. The diagnosis of NHL-CNS is usually confirmed by the finding of malignant cells showing B-cell markers with either a κ or a λ light chain monoclonal response. Gross examination of eyes with NHL-CNS shows large gray patches of subretinal and retinal infiltration above a thickened choroid ( Fig. 30-8 ). Collections of lymphoma cells are characteristically found between Bruch’s membrane and the retinal pigment epithelium (RPE) ( Fig. 30-9 ). Reactive T lymphocytes can be seen in the retina and choroid in areas surrounding the tumor cells.

Cells obtained from vitrectomy specimen from a patient with large-cell lymphoma demonstrating large cells with hyperchromatic, pleomorphic nuclei. (Diff Quick stain; Original magnification ×1000.)

(Reproduced with permission from Whitcup et al: Intraocular lymphoma: clinical and histopathologic diagnosis. Ophthalmology 100:1399–1406, 1993.)

Enucleation specimen of patient with intraocular lymphoma shows multiple white areas of retinal and subretinal tumor infiltration with adjacent areas of retinal hemorrhage.

Photomicrograph of enucleated globe shown in Figure 30-8 reveals tumor infiltration between Bruch’s membrane and the retinal pigment epithelium. Reactive lymphocytes and some tumor cells are seen in the overlying vitreous. (Hematoxylin & eosin stain; original magnification ×200.)

The differential diagnosis for patients suspected of having an intraocular lymphoma often includes sarcoidosis. Therefore, evaluation of these patients should include a chest X-ray, angiotensin-converting enzyme (ACE) measurement, serum and urine calcium levels, and pulmonary function testing to rule out this diagnosis. A limited gallium scan of the chest and lacrimal glands may also be useful. If patients have a diminished diffusing capacity on pulmonary function testing, we often obtain a pulmonary consultation and consider a CT scan of the chest and bronchoscopic examination to make the diagnosis of sarcoidosis. Other diseases to consider include syphilis, tuberculosis, acute retinal necrosis, and cytomegalovirus retinitis.

Treatment

The treatment of NHL-CNS is controversial and the ideal therapy is as yet unproven. Radiation therapy was first proposed for the treatment of a chronic uveitis in 1911, and historically, radiotherapy has been recommended as the initial therapy for NHL-CNS. In patients with detectable lymphoma limited to the eyes, controversy centers on whether to limit therapy to the eye or to treat the CNS prophylactically. Char and colleagues reported on two patients with isolated ocular involvement treated with only ocular radiation who had no recurrent disease 24 and 109 months after treatment. NHL-CNS tends to be quite radiosensitive and responds well to 15–45 Gy. However, many researchers recommend cranial irradiation (45 Gy to the whole brain) in addition to orbital radiation in patients with isolated ocular involvement, as many patients have subclinical CNS involvement by the time intraocular lymphoma manifests. ^, In the series of Peterson and colleagues, 10 of 11 patients with lymphoma limited to the eyes at initial presentation later developed CNS disease. In fact, Rouwen and colleagues advise using a combination of chemotherapy and radiotherapy for intraocular lymphoma even if CNS disease cannot be documented on MRI or lumbar puncture. Chemotherapy is recommended to treat possible CNS disease, and radiotherapy is given because penetration of the blood–retinal barrier by chemotherapeutic agents is uncertain. If CNS disease is documented, there is less controversy surrounding the treatment of intraocular lymphoma. In these patients, the combination of radiotherapy and chemotherapy is more frequently recommended, although this therapeutic approach may not be ideal because of its induced CNS toxicity.

A number of chemotherapy regimens have been used for the treatment of intraocular lymphoma. The use of high-dose cytarabine (Ara-C) can lead to therapeutic drug levels in the CSF, and there have been reports of successful treatment of intraocular lymphomas with this therapy alone. However, some studies indicate that CSF levels of systemically administered chemotherapeutic agents may be variable and short-lived. We therefore recommend placement of an Ommaya reservoir for delivery of intrathecal methotrexate to ensure adequate drug levels in the CSF. A number of researchers have used intrathecal methotrexate for the treatment of intraocular lymphoma, but always in combination with radiotherapy. DeAngelis and colleagues showed that the addition of chemotherapy to cranial radiotherapy significantly improved disease-free survival and contributed to overall survival in patients with primary CNS lymphoma.

However, combined radiotherapy and chemotherapy can cause significant toxicity. Cranial radiotherapy produces significant CNS toxicity in long-term survivors. Furthermore, Bleyer and DeAngelis and colleagues emphasize that chemotherapy can exacerbate the toxicity produced by radiotherapy, especially with methotrexate. In contrast to the usual sequence of therapy for intraocular lymphoma, in which radiotherapy is administered before chemotherapy, the use of chemotherapy before radiotherapy may reduce the risk of leukoencephalopathy and protect against late toxicity. ^,

The National Cancer Institute and the National Eye Institute investigated a treatment protocol using systemic and intrathecal chemotherapy for the treatment of NHL-CNS. In this trial, 14 nonimmunocompromised patients received a chemotherapy regimen that incorporated a 24-hour infusion of high-dose methotrexate with leucovorin rescue, thiotepa, vincristine, dexamethasone, and intrathecal Ara-C and methotrexate. The chemotherapy was administered in 21-day cycles and the response rate was 100%. Eleven patients (79%) had a complete response and three (21%) had a partial response. At more than 4.5 years of follow-up the cumulative survival rate was 68.8%, and the progression-free survival rate was 34.3%. The median survival had not been reached at the time of publication, and the progression-free survival was 16.5 months. Toxicity of the treatment regimen included severe drug-related leukoencephalopathy in two patients, grade 3 or 4 neutropenia in 50% of the cycles, ileus in one patient, and seizures in two. Although this was a small trial, we believe that the use of chemotherapy alone as initial therapy should be investigated further. Finally, high-dose thiotepa, busulfan and cyclophosphamide followed by autologous bone marrow transplantation appeared to be effective in some patients with refractory intraocular lymphoma.

Local recurrence of NHL-CNS in the eye can be difficult to diagnose and treat. Definitive diagnosis requires obtaining a vitreous specimen for cytologic examination. One therapeutic option for ocular recurrence is radiation. However, we have seen a number of patients with ocular recurrence who had already received maximal radiation to the eyes. Intravitreal injections of chemotherapeutic agents may be a useful therapeutic approach for these patients. Combinations of intravitreal methotrexate, fluorouracil, and dexamethasone have been studied. The safety of intravitreal injections of retuximab is being assessed.

The presence of intraocular lymphoma is usually associated with concomitant brain disease, and survival is similar to that for other presentations of primary CNS lymphoma. Some data suggest that patients whose ocular disease was identified and treated before CNS progression had a significantly improved survival, thereby emphasizing the importance of early diagnosis and therapy. Although many patients with NHL-CNS have a good initial response to therapy, recurrence is common, and in some studies the 5-year survival is <5%. Hochberg and Miller reported an overall median survival time from diagnosis of 13.5 months in patients with primary CNS lymphoma, and Freeman and colleagues reported a mean survival time of 26 months for patients with intraocular lymphoma. However, some reports suggested that intrathecal chemotherapy combined with radiotherapy of the orbits and whole brain may improve survival. Hopefully, additional research into new therapeutic approaches will lead to improved survival in this disease.

Systemic Non-Hodgkin’s lymphoma metastatic to eye

Systemic NHL can metastasize to the eye and cause an anterior or a posterior uveitis. Unlike the location of NHL-CNS cells between the RPE and Bruch’s membrane, lymphoma cells in patients with systemic NHL characteristically infiltrate the choroid and may be mistaken for an ocular melanoma. Metastatic lymphoma may also cause a hypopyon or hyphema in an otherwise uninflamed eye. Although most patients with NHL have symptoms of lymphadenopathy, fever, and weight loss, ocular involvement may be the initial presentation of the disease.