![[From the Archives of the AFIP main menu]](../../../../../buttons/afipbtt2.gif){kind=small}
![[HOME]](../../../../../buttons/afiphom3.gif){kind=small}
RadioGraphics
The prognosis of patients with primary CNS lymphoma is extremely poor, particularly for immunocompromised patients. The median survival for AIDS patients with CNS lymphoma is only about 45 days from the time of clinical presentation (3). Even for immunocompetent patients, the average survival time is only 3.3 months (6). However, use of radiation therapy or chemotherapy may increase the median survival time, as indicated by reports of patients who survived up to 4 years after clinical presentation and treatment.
The disparity in survival times for the untreated versus treated patients emphasizes the urgency in establishing a definitive diagnosis and is the single most important argument against the use of empiric antitoxoplasmosis therapy to differentiate between CNS lymphoma and toxoplasmosis. The 2-week waiting period to determine the effect of drug therapy may significantly lessen the likelihood of prolonged life for some patients (24). Therefore, prompt evaluation with either PET (if available) or SPECT may promote prolonged survival (3,24,41).
Patients with a more favorable prognosis include those younger than 60 years of age, those with normal or mildly impaired neurologic function (as determined before radiation therapy), and those with the tumor confined to the cerebral hemispheres or cerebellum (86,87).
One curious feature of primary CNS lymphoma is its rapid regression with corticosteroid or radiation therapy, which has led some to dub CNS lymphoma as the "ghost tumor" (88). A few cases of spontaneously remitting CNS lymphoma have been documented in the literature (89,90). Recently, "sentinel" enhancing lesions of primary CNS lymphoma have been described in four immunocompetent patients (91). The lesions all disappeared either after corticosteroid therapy (three cases) or spontaneously (one case). Results of subsequent brain biopsies demonstrated demyelination, nonspecific inflammation, and normal brain tissue. After several months, the patients developed new symptoms, referable to a different location in the brain. Biopsies of the new areas showed B-cell primary CNS lymphoma. The cause for this phenomenon is unknown; explanatory theories include tumor lysis secondary to steroid therapy or cyclic biologic activity (as is seen in systemic lymphoma), host immunity against the tumor, or infarction caused by the tumor occluding its own feeding vessels. Whatever the cause, administration of cortico-steroids before biopsy is contraindicated in suspected cases of CNS lymphoma to reduce the likelihood of false-negative results (2,91).
Surgical resection of these neoplasms, whether they are small or large, has no effect on survival (86). Systemic therapy, which is relatively successful in the treatment of systemic lymphoma (60% of patients survive 5 or more years), is not nearly as effective in halting the progression of primary CNS lymphoma (2,92,93). Radiation therapy involves doses of 4,000-6,500 cGy given over several weeks time and improves patients' chances for prolonged survival (length of median survival increases to 15-20 months) (1,7,94). Chemotherapy in general seems to improve survival, but not all protocols have demonstrated a statistically significant therapeutic benefit (91). Local control of the tumor is improved when chemotherapy is combined with radiation therapy and when higher doses of radiation (>4,000 cGy) are used (86). Consequently, several reports now indicate that survival times for patients with primary CNS lymphoma can be measured in years instead of months (1,94).
Unfortunately, the chances for improved survival that result from undergoing radiation therapy with or without chemotherapy are diminished in the immunocompromised population. Median survival for this group is 4.5-5.5 months (3,95). Patients who undergo radiation therapy more commonly die later and of opportunistic infection, whereas untreated patients more frequently die of CNS lymphoma itself, regardless of histologic characteristics or extent of involvement. This course again emphasizes the need to begin therapy as soon as possible in patients suspected of having primary CNS lymphoma (3).
One of the drawbacks of using intravenous chemotherapy to treat intracranial lymphoma is the large molecular size (200-1,200 d) of chemotherapeutic agents. Although high-grade gliomas and CNS lymphomas cause breakdown of the blood-brain barrier, which allows contrast material to enter the lesion, chemotherapeutic agents with theirlarger molecules do not enter the tumor to the same extent. Consequently, the effectiveness of chemotherapeutic agents is restricted to the extreme margin of a tumor. A recent novel therapy, in which the blood-brain barrier was intentionally disrupted to allow increased passage of chemotherapeutic agents administered through an intraarterial approach, has shown potential in providing more effective chemotherapy (with this approach, the amount of agent delivered was increased by 10-90 times, compared with that delivered by conventional methods) (96). Early results of chemotherapy performed with this method were comparable to those of radiation therapy (97). However, significant side effects, including irreversible ototoxicity, have limited the application of this method (98). If these side effects can be overcome, delivery of chemotherapeutic agents with use of an intraarterial approach and intentional disruption of the blood-brain barrier may be used more frequently in the future.
Because of the dismal prognosis of patients with CNS lymphoma, recurrences have been rare. About half of the cases of recurrent disease occur at the site of the original tumor; other sites of involvement in the neuraxis and systemic sites are not uncommon and take the form of intracranial or spinal cord leptomeningeal spread (1). Rarely, primary CNS lymphoma can hematogenously spread to the dura mater and calvaria, and, as a result of subarachnoid spread, it can infiltrate cranial nerves, the pituitary gland, and the ependyma (29). Extraneural spread of CNS lymphoma occurs most frequently before chemotherapy and involves systemic (nodal) disease as the most common site (10% of cases), followed by sites in the heart, gastrointestinal tract, spinal epidural space, and bone marrow. Institution of chemotherapy appears to reduce the likelihood of extraneural spread: Miller et al (1) reported no cases of extraneural spread after chemotherapy was added to the therapeutic protocol in 1984. However, in a study by Remick et al (8), five of 22 patients developed extraneural disease (testicle, inguinal nodes, bone marrow, shoulder, cervical lymph nodes) despite undergoing chemotherapy.
At the time of recurrence, diffuse involvement of the CNS, including the spinal cord, occurs commonly (60% of cases) and provides an argument for combination craniospinal irradiation or chemotherapy plus irradiation in primary treatment (7). In fact, long-term survival, an unthinkable prospect until recently, is now a possibility for this once uniformly fatal disease (1).
Prognosis and Treatment