HIV-Associated Non-Hodgkin Lymphoma: Title and subTitle BreakIncidence, Presentation, and Prognosis

A 47-year-old man developed painful swelling in his right lower leg. A nonsteroidal anti-inflammatory medication resulted in brief improvement of his symptoms. Four weeks later, he presented to a community hospital emergency department with progressive disabling leg swelling and pain, abdominal distension and pain, fever, and a 20-lb weight loss. He was admitted and a computed tomography (CT) scan of the abdomen showed a large pelvic mass displacing the bowel and bladder with ureteral compression. A needle aspiration of the mass revealed large B-cell non-Hodgkin lymphoma (NHL), and the patient transferred to the National Cancer Institute for further treatment.

On admission to the National Cancer Institute, the patient was febrile (103.1° F [39.5° C]), tachycardic (102/min), slightly tachypneic (18/min), and normotensive (130/75 mm Hg). He appeared acutely ill with dry oral mucosa. His abdomen was distended and tender to deep palpation, chiefly in the right lower quadrant where a large mass extending into the pelvis was evident. There was no guarding or rebound tenderness. There was marked edema of the right lower leg, the right side of the scrotum, and the inferior third of the abdominal wall. His liver and spleen were normal size and he had no palpable lymphadenopathy. The neurologic examination was normal.

The patient's past medical history was remarkable only for a clinical syndrome of hepatitis 23 years prior to presentation, but viral serologies had never been performed. Although there was disclosure of a remote sexual liaison with a female prostitute many years earlier, no other risk factor for human immunodeficiency virus (HIV) infection was identified.

Completion of the staging evaluation revealed a normal complete blood cell count, slightly elevated creatinine (1.7 mg/dL [150.2 µmol/L]) and hepatic transaminase levels. The patient's uric acid (10.4 mg/dL [0.62 mmol/L]; upper normal limit, 8.6 mg/dL [0.51 mmol/L]) and lactate dehydrogenase (398 U/L; upper normal limit, 226 U/L) levels were both elevated. A CT scan of the head and chest were normal, and a repeat abdominal and pelvic CT scan again revealed a large pelvic mass displacing the bowel and bladder to the left, compressing the ureters and right iliac and common femoral veins. Doppler ultrasound of the right lower leg confirmed a right common femoral vein thrombosis. Bilateral bone marrow biopsies and a lumbar puncture were performed and showed no evidence of involvement by lymphoma.

HIV antibody was found to be positive by ELISA with Western blot confirmation. The CD4 cell count was 60 cells/µL and the HIV viral load was 104 722 mRNA copies/mL by reverse transcription polymerase chain reaction (Amplicor, Roche Diagnostics Corp, Indianapolis, Ind).

HIV-associated NHL (HIV-NHL) is a life-threatening complication of HIV infection that has a median survival of 6 to 11 months with standard chemotherapy.^{1 - 2} The median age at presentation is 37 years, so patients are often younger than HIV-negative patients, where the peak incidence is in patients older than 65 years.^{3 - 5} Unlike Kaposi sarcoma, where the risk of developing the disease is greatest among individuals who acquire HIV infection through sexual contact with men who have had homosexual contacts, the risk of developing HIV-NHL is similar among all groups at risk for HIV infection.^{6 - 7} Overall, NHL occurs with a 60-fold increased frequency in HIV-positive compared with HIV-negative populations.⁷ For this reason, in 1985, the US Centers for Disease Control and Prevention included NHL as an acquired immunodeficiency syndrome (AIDS)–defining condition.⁸

Non-Hodgkin lymphoma is the AIDS-defining condition in approximately 3% of HIV-infected persons.^{7 ,9} It is unknown what percentage of HIV-infected patients will develop NHL over their lifetime, although estimates of 5% to 20% have been made.^{10 - 11} As of yet, it is also unclear if the lifetime incidence of HIV-NHL has been changed by the use of highly active antiretroviral therapy (HAART); currently, the opinion is mixed regarding this issue.^{12 - 14} Based on the spectrum of HIV disease in the United States, there does not appear to be a change in the incidence of NHL as an AIDS-defining event over the period 1992-1997.¹⁵ Nevertheless, one might predict an eventual increase in HIV-NHL as patients with AIDS live longer with incomplete immune reconstitution. In contrast, HIV-associated primary central nervous system lymphoma, which comprises about 18% of the cases of HIV-NHL and is virtually always associated with Epstein-Barr virus, appears to have decreased since the advent of HAART.¹⁶

HIV-NHL frequently presents as advanced stage III or IV disease.¹⁷ The majority of patients will present with either the appearance of a rapidly growing mass lesion or the development of systemic "B" symptoms (ie, unexplained fever, drenching night sweats, or unexplained weight loss in excess of 10% of the normal body weight).¹⁸ Extranodal involvement is common, including the bone marrow (25%-40%), gastrointestinal tract (26%), and the central nervous system (CNS) (17%-32%).^{3 ,19 - 22}

Staging evaluation typically consists of history and physical examination, clinical laboratory assessment of organ function, CD4 cell count, and bilateral bone marrow biopsies. Computed tomography of the chest, abdomen, and pelvis should be performed. Radiologic imaging of the brain and cytological evaluation of the cerebrospinal fluid (CSF) is also needed to assess CNS involvement. Computed tomography of the brain with contrast is adequate to assess parenchymal brain lesions, but magnetic resonance imaging with gadolinium has the potential advantage of producing a better image of the leptomeninges.²³

The CD4 cell count has been the major factor used to establish the prognosis of patients with HIV-NHL. Historically, patients with fewer than 100 CD4 cells/µL had a median survival of 4 months, whereas those with 100 or greater CD4 cells/µL had a median survival of 11 months with standard therapy.^{24 - 25} An even poorer outcome is seen in HIV-associated primary central nervous system lymphoma, which has a median survival of 2 to 4 months, regardless of the CD4 cell count, although most patients have CD4 counts of less than 50 cells/µL. Other factors associated with poor outcome include age older than 40 years, elevated serum lactacte dehydrogenase level, presence of extranodal disease, and a preexisting AIDS diagnosis.^{26 - 27} In non–HIV-associated aggressive NHL, the International Prognostic Index is commonly used to predict prognosis, but it has not been adequately validated in HIV-NHL.²⁸ With the recent advent of HAART, the long-term outlook for patients with HIV-NHL who are cured may significantly improve if they maintain a good virologic and immunologic response.

A needle biopsy specimen of the retroperitoneal mass was submitted from an outside institution for review. It showed a monomorphic proliferation of medium-sized lymphoid cells with round to oval nuclei, 2 to 3 small basophilic nucleoli, and a rim of eosinophilic cytoplasm. Scattered reactive histiocytes containing apoptotic debris were present, and frequent mitotic figures could be identified.

Immunohistochemical studies were performed on paraffin sections. The cells were CD20-positive, CD10-positive, and exhibited a proliferative rate of 100% with MIB-1 (Ki-67). Stains for bcl-2 protein and p53 were negative in the neoplastic cells. In-situ hybridization for EBER RNA of the Epstein-Barr virus could not be performed, due to insufficient material in the sparse needle biopsy specimen. Flow cytometry analysis of the fine needle aspirate showed monoclonal expression of kappa light chains and expression of CD19 and CD20. A diagnosis of high-grade B-cell lymphoma, Burkitt-like, was made.²⁹ While the referring pathologist had favored a diagnosis of diffuse large B-cell lymphoma, the extremely high growth fraction and immunophenotypic features (bcl-2–negative, CD10-positive) favored a Burkitt or Burkitt-like lymphoma.³⁰ In addition, the clinical presentation without a history of clinical AIDS favors Burkitt or Burkitt-like lymphoma over other HIV-associated lymphomas (Table 1).³¹ As is common in the setting of HIV infection, the cytologic appearance was more variable than that of classical Burkitt lymphoma.³² Burkitt and Burkitt-like lymphomas are frequently associated with translocations involving the c-myc oncogene.^{32 - 33}

There are 3 major categories of HIV-associated lymphomas: (1) Burkitt and Burkitt-like lymphomas, (2) B-cell immunoblastic lymphomas, and (3) primary effusion lymphomas. These lymphomas differ not only in their clinical presentations, but also in their underlying pathogenesis. Burkitt and Burkitt-like lymphomas usually present with widespread systemic disease and have a high risk of CSF involvement. However, mass lesions in the brain are less commonly seen. Immunoblastic lymphomas present most commonly in extranodal sites, including the brain. They are positive for Epstein-Barr virus in the vast majority of cases.³⁴ As the name implies, primary effusion lymphomas typically present with pleural and/or peritoneal effusions, but in some cases involvement of soft tissue or gastrointestinal tract may be seen.³³

Other lymphomas commonly seen in the setting of HIV include large B-cell lymphomas,³⁵ classical Hodgkin disease,³⁶ polymorphic lymphoproliferative disorders resembling post-transplant–associated lymphoproliferative disease,³⁷ and lymphomatoid granulomatosis.³⁸

As discussed below, the patient underwent combination antilymphoma chemotherapy while anticoagulation therapy was initiated concomitantly. Antiretroviral therapy was initiated after all the cycles of chemotherapy had been completed.

After the patient had received 6 cycles of systemic chemotherapy, a CSF sample was submitted for cytologic evaluation and flow cytometry. It showed a marked pleocytosis (423 cells/µL). The cells were large with abundant basophilic cytoplasm and occasional cytoplasmic vacuoles. Flow cytometry demonstrated monoclonal expression of kappa light chain, similar to the primary retroperitoneal tumor.

In non–HIV-infected patients with advanced aggressive lymphomas, doxorubicin-containing regimens produce approximately 45% disease-free survival and 50% overall survival at the 3-year benchmark.³⁹ In contrast, fewer than half of similarly treated patients with HIV-associated lymphomas are alive 1 year after treatment.⁴⁰ The reason for this difference is uncertain, but appears to be related, in part, to increased mortality from opportunistic infection.^{24 ,41} Attempts to balance chemotherapy-induced immune damage, by reducing drug doses, with the need for effective chemotherapy doses, however, have met with unimpressive outcomes.^{2 ,40} A randomized comparison of standard and low-dose treatment with methotrexate, bleomycin, doxorubicin, cyclophosphamide, vincristine, and dexamethasone (m-BACOD) yielded a similar median survival of 31 and 35 weeks, respectively, and a recent study of treatment with cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) showed an equivalently poor outcome.⁴⁰ Clearly, the findings from the former study that among the patients who died after treatment with low-dose m-BACOD, survival was short and 70% of patients had evidence of lymphoma, indicate the need for more effective regimens.

A sobering lesson from the treatment of non–HIV-associated aggressive lymphomas is that the addition of drugs per se to CHOP treatment does not improve effectiveness and suggests that therapeutic advances require a different direction.³⁹ It is important, however, to recognize the incremental benefits that may be accrued by optimizing drug administration. A retrospective analysis of CHOP dose intensity has shown that patients who receive lower doses have a worse clinical outcome and suggests that dose intensity should be maintained whenever possible.⁴² A number of cytotoxic agents show in vitro schedule dependency and raise the question of whether clinical outcome could also be improved by optimizing administration schedules.⁴³ Based on these observations, investigators at the National Cancer Institute developed the EPOCH regimen in which the schedule-dependent agents etoposide, vincristine, and doxorubicin were administered as continuous 96-hour infusions and the non–schedule-dependent agents cyclophosphamide and prednisone were administered on a "bolus" schedule.⁴⁴ A clinical assessment of the EPOCH regimen in previously treated patients with lymphoma showed it to be well tolerated and quite active.⁴⁴ Moreover, the EPOCH regimen showed a promising event-free survival rate of 71% at 2 years in non–HIV-positive patients with previously untreated aggressive NHL, suggesting that EPOCH would be a reasonable regimen to assess in HIV-associated lymphomas.⁴⁵

Based on these results, we began a phase 2 trial of dose-adjusted EPOCH with granulocyte-colony stimulating factor in previously untreated patients with HIV-associated lymphoma.⁴⁶ In this trial, 2 strategies were used to maximize drug dose intensity while minimizing the toxicity in this susceptible population. A "dose-adjusted" strategy was developed in which a reduced dose of cyclophosphamide is administered on the first cycle, based on the patient's CD4 cell count, and thereafter adjusted to the nadir neutrophil counts. Using this approach, the infused agents can be virtually administered at full dose. A second, somewhat controversial aspect of this strategy is the suspension of all antiretroviral agents during the treatment phase. There is legitimate concern that pharmacokinetic interactions and overlapping toxicities between chemotherapy and antiretroviral agents would compromise chemotherapy administration and reduce the chance of eliminating the lymphoma.^{47 - 48} Additionally, the adverse effects of chemotherapy may reduce patient compliance with the antiretroviral regimen, and this could lead to emergence of resistant viral strains.⁴⁹ The possible benefits of suspending antiretroviral therapy must be weighed against the potential for long-term effects of uncontrolled viral replication on immune function,⁵⁰ but we concur with the recommendation of one the authors (A. P.) to suspend antiretroviral therapy during the high-intensity therapeutic period of this EPOCH regimen.

Recently, the preliminary results of dose-adjusted EPOCH treatment in 23 patients with previously untreated HIV-associated lymphoma were reported.⁴⁶ The median progression-free and overall survival had not been reached and were 83% and 72%, respectively, at a median follow-up of 23 months. Of significance, 70% of patients achieved complete remission and none had relapsed. The dose-adjusted strategy allowed the administration of over 90% dose intensity of the infused agents with a 12% incidence of hospitalization for fever and neutropenia (based on 115 cycles of therapy administered). The serial assessment of HIV viral loads and CD4 cell counts showed that by 6 to 12 months after chemotherapy, the values had returned to pretreatment levels, suggesting the absence of a significant decrement in immune function or viral control. Based on these promising results, accrual continues to this study.

The role of CNS prophylaxis in HIV-associated lymphomas is an important clinical issue. In non–HIV-associated lymphomas, CNS prophylaxis is usually limited to patients with Burkitt lymphoma and to patients with large B-cell lymphomas who have multiple extranodal disease sites such as the bone marrow.⁵¹ However, quite unlike their non-HIV counterparts, 15% to 20% of HIV-associated lymphomas have involvement of the CNS at presentation, indicating a unique tropism for the CNS.²⁴ If the effectiveness of systemic therapy increases, the CNS is likely to emerge as the limiting factor to cure.²⁴ Indeed, the subject of this case presentation initially presented with a large mass confined to the pelvis and, based on the biology of non–HIV-associated lymphomas, would have been expected to have a remote chance of developing CNS involvement, so CNS prophylaxis was not administered. However, CNS lymphoma emerged late in the treatment course and could potentially have limited his chance of cure, which illustrates the importance of this sanctuary site. Indeed, this patient required triple intrathecal chemotherapy and whole brain radiotherapy to 4000 rad to achieve CNS sterilization. Such intensive therapy increases the risk of late-term CNS toxicity, especially in the HIV-infected patient. Based on our experience with this and other patients, we now recommend that all patients with HIV-associated lymphomas receive CNS prophylaxis with intrathecal chemotherapy.

In the case presented here, the patient was naïve to antiretroviral therapy. Since HAART can result in significant viral suppression, improve immune function, and prolong patient survival,⁵² one would ideally like to initiate HAART with antineoplastic chemotherapy to attain the benefits from both regimens. However, it is essential to recognize that the patient's immediate survival depended on achieving lymphoma remission, and this in part was dependent on optimal chemotherapy administration.⁴² Concomitant therapies raise 3 major concerns: (1) drug-drug interactions, (2) potential additive toxicities, and (3) inability to optimally administer the medications. Concurrent administration of antiretroviral and antineoplastic therapy can be associated with considerable pharmacodynamic and pharmacokinetic interactions leading to changes in the therapeutic index of the various agents administered. The clearance of cyclophosphamide appears to be decreased when administered in the presence of indinavir,⁴⁰ and plasma levels of etoposide are reduced when administered with didanosine.⁵³ Many of the antiretroviral agents, particularly the protease inhibitors and non-nucleoside reverse transcriptase inhibitors (NNRTIs), are metabolized through the cytochrome P450 (CYP450) system.^{54 - 56} All the protease inhibitors and the NNRTI delavirdine are inhibitors of the CYP450 system. Concomitant use of these agents with a drug that is a substrate of this system can lead to accumulation of the substrate and can potentially increase dose-related toxicities. The NNRTIs nevirapine and efavirenz, on the other hand, are inducers of the CYP450 system⁵⁷ and can thus reduce drug concentrations and, potentially, drug efficacy.

Many of the antiretroviral agents and antineoplastic agents have overlapping toxicities including bone marrow suppression, neurotoxicity, and aphthous ulcers. Overlapping toxicities may affect the ability to optimally administer the various drugs, thus potentially leading to inadequate therapy for the malignancy and/or the underlying HIV infection. This could result in decreased potential for lymphoma cure and long-term control of the HIV infection. Chemotherapy-associated nausea and vomiting may lead to decreased antiretroviral drug absorption and regimen nonadherence. Erratic absorption or intermittent dosing may lead to development of drug-resistant HIV, and many authorities emphasize the importance of therapeutic strategies aimed at avoiding resistance.⁵⁸ In this case, there were several potential strategies to consider. As was the case for this patient, one can withhold HAART until completion of chemotherapy. This strategy may avoid development of resistance, and relatively brief uncontrolled HIV viremia is unlikely to have unfavorable long-term immunologic or clinical consequences if appropriate opportunistic infection prophylaxis is used. Resistant HIV infection as a consequence of inconsistent HAART dosing has profound implications for long-term HIV control. This approach also has the advantage of avoiding the pharmacokinetic and toxicity concerns reviewed above.

Another alternative is to continue HAART with close follow-up on drug tolerance and adherence. A third alternative is to modify the HAART regimen to contain agents with the least overlapping toxicities and drug-drug interactions. This may not be feasible in those patients who are fairly antiretroviral therapy experienced. An important consideration is that effective lymphocytotoxic antilymphoma chemotherapy causes CD4 cell depletion.⁵⁹ Since this effect is unlikely to be mitigated by concurrent antiretroviral therapy, the benefits of concurrent antiretroviral therapy may be difficult to assess by surrogate markers (eg, CD4 cell counts).

While lymphoma and HIV were the major problems facing this patient, the clot in the femoral vein potentially threatened his survival. Since the deep vein thrombosis was refractory to routine treatment, we offered this man more aggressive treatment with the thrombolytic agent recombinant tissue type plasminogen activator (rtPA; Activase, Genentech, Inc, South San Francisco, Calif). In contrast to anticoagulants, thrombolytic agents can lyse even large thrombi within several days, thereby restoring flow in the veins and relieving painful pressure. Initially, the marked compression by a tumor of the patient's pelvic veins may have prevented restoration of antegrade flow even if the thrombus were lysed. However, EPOCH treatment resulted in brisk pelvic tumor reduction, thus enhancing the potential that thrombolytic therapy would be successful. We, along with other investigators,⁶⁰ have observed that the risk of a pulmonary embolus is rarely a serious consequence of venous thrombolytic therapy.

Since the collateral circulation tends to divert the drugs around the obstructing thrombus, thrombolytic agents for deep vein thrombosis are given by catheter-directed infusion directly into the clots.⁶⁰ Our patient's symptoms significantly improved with 24 hours of treatment with rtPA. Venography revealed extensive patency of his superficial femoral vein. A second dose of rtPA was given to treat remaining intraluminal filling defects identified by venography. By the next day, our patient's leg was almost asymptomatic and he was ambulatory. Full anticoagulation with heparin was continued throughout the period of thrombolysis and then maintained with warfarin.

With brisk shrinkage of the tumor mass as documented by repeat pelvic CT scans during the first several days of infusional EPOCH treatment, it was deemed potentially advantageous to administer thrombolytic therapy for the venous clot. This strategy resulted in complete resolution of the clot and in secondary alleviation of the primary symptoms that initially led the patient to seek medical attention. EPOCH chemotherapy successfully eradicated his peripheral lymphoma, but at the end of planned chemotherapy, there was late emergence of leptomeningeal lymphoma. This late event highlights the justification for intrathecal prophylaxis against CNS lymphoma in AIDS-NHL even when both cytological examination of the CSF and the bone marrow are negative for malignancy at the time of initial staging. After the emergence of overt CNS disease, patients are at high risk for CNS relapse, thus necessitating a more aggressive and toxic multimodal approach incorporating both chemotherapy and radiation therapy to the CNS, as illustrated here.

This case also demonstrates the feasibility of successfully administering dose-adjusted chemotherapy with resultant high dose intensity while delaying antiretroviral therapy until the completion chemotherapy, even in a high-risk severely immunosuppressed AIDS patient. With a multidisciplinary approach, the patient achieved a complete tumor response to chemotherapy, complete resolution of his thrombosis without evidence of postphlebitic syndrome or symptomatic pulmonary embolism. Suppression of HIV replication below the limits of quantitation of the test (< 200 copies HIV-1 mRNA/mL plasma) was achieved by 12 weeks after antiretroviral therapy was begun. His CD4 cell count subsequently increased to 137 cells/µL within 27 weeks after initiation of antiretroviral therapy. However, 5 months after completion of all cancer therapy (including radiation to the whole brain), spinal magnetic resonance imaging revealed an enhancing lesion in the C2 region. The lesion remained stable, but subsequent progressive neurologic deficits left the patient quadriplegic by 9 months after completion of therapy. Two months later, the patient died of pulmonary failure and pneumonia. An autopsy revealed no evidence of NHL and confirmed the clinical impression that radiation necrosis accounted for the spinal cord defect.

Clearly, the patient's life was most proximally threatened by his malignancy and by the uncontrolled thrombosis. Optimal management of these grave conditions allowed the team subsequently to focus attention chiefly on his underlying chronic condition: HIV infection. This case demonstrates successful interdisciplinary salvage of a patient who presented with poor prognosis due to multiple features of disease. Furthermore, this patient regained his previous quality of life as a result of substantial recent advances in the treatment of lymphoma, thrombosis, and AIDS. However, late radiation-induced spinal cord injury resulted in death, emphasizing the necessity of intrathecal therapy as prophylaxis in this sanctuary site.